p-

o
0
0
~
ii
AN INTENSIVE COURSE

0
0 BLAST FURNACE .'

o.
0
IRONMAKING
~o
L

0
Volume One
0
fJ PRINCIPLES, ,DESIGN.
0 AND RAW MATERliAlS
0 tF

0
JD
" II ..,
"',: "'

0 McMASTER UNIVERSITY 'l

Ham i Iton,Ontario, Canada f ii
0
JUNE,,1999 II

0
D
~
 AN INTENSIVE COURSE

BLAST FURNACE IRONMAKING

JUNE 7-11, 1999

VOLUME ONE
PRINCIPLES, DESIGN AND RAW MATERIALS

COORDINATING COMMITTEE:

A.J. Fischer, Dofasco Inc. (Chairman)

G.A. Irons, McMaster University (Secretary)
R. Brown, Stelco Inc.
P. Kuuskman, Algoma Steel Inc.
J.J. Poveromo, Quebec Cartier Mining Co.
F.e. Rorick, Bethlehem Steel Corp.
S. Sostar, Lake Erie Steel Co.
Copyright 1999
Department of Materials Science and Engineering
McMaster University
Hamilton, Ontario, Canada
L8S 4L 7

No part of this book may be reproduced in any form, except with the consent of an
individual author concerning his own lecture or with permission from the Department of
Materials Science and Engineering, McMaster University, or the Coordinating Committee
of this Course.

Printed in Canada
at McMaster University
 PREFACE

The efficient operation of the iron blast furnace is essential to the economic well-
being of any integrated steel plant; any improvement in operation usually has a signifcant
impact upon the entire company.

Today's ironmaking technology has evolved over many years through innovations
in raw materials preparation, blast furnace design, refractories improvements, and blast
furnace practice. Much remains to be done; significant gains remain to be realized. Much
is being done.

This course on Blast Furnace Ironmaking was organized in response to a felt need;
the response has been overwhelming. It is an intensive, in-depth course covering every
aspect of blast furnace ironmaking, which should make it useful to many people - managers,
operators, engineers, researchers, and suppliers of equipment, refractories and raw
materials.

The 1999 course was organized by a Coordinating Committee consisting of:

Randy Fischer, Dofasco Inc. (Chairman)

Gord Irons, McMaster University (Secretary)
Rick Brown, Stelco Inc.
Peter Kuuskman, Algoma Steel Inc.
Joe Poveromo, Quebec Cartier Mining Co.
Fred Rorick, Bethlehem Steel Corp.
Steve Sostar, Lake Erie Steel Co.

In developing this course, we adhered to two criteria; the lecturers would be

acknowledged experts in their fields and the contents would be practical, with only
sufficient theory to understand the process.

We, the Committee, hope that this course has satisfied your present needs, that you
wil have made some valuable and lasting "contacts", and that these notes wil continue to
be a valuable reference for you in years to come.

Randy Fischer, Chairman

Coordinating Committee
1999 Blast Furnace Ironmaking Course
 FOREWORD

The first Blast Furnace Ironmaking Course was initiated in 1977 under the
leadership of John Holditch and Don George. The course has been offered 14 times (1977,
1978,1980,1981,1982,1984,1985,1987,1989,1990, 1992, 1994, 1996 and 1998) and owes
its success to the excellent reputations and efforts of the lecturers and of the Coordinating
Committees. This, the 15th course, is being offered at McMaster University in June 1999.

Since 1984 the course has been officially recognized by the American Iron & Steel
Institute, and is jointly supported by the AISI and McMaster University. The overwhelming
response every year to this course has been not only in the number of registrants but also
in their diversifed industrial backgrounds. Another notable fact is that among the
registrants, many are well-known experts in their own right, in certain aspects of
iron making. We would like to take this opportunity to express our sincere appreciation to
all the lecturers who have contributed to this course, and to their employers for allowing
them to take time off from their busy schedules and for defraying their travel expenses.

Gord Irons, Secretary

Coordinating Committee
1999 Blast Furnace Ironmaking Course
 1999 BLAST FURNACE IRONMAKING COURSE

CONTENTS

VOLUME ONE: PRINCIPLES, DESIGN AND RAW MATERIALS

Lecture 1 Historical Development and Principles of the Iron Blast Furnace

J.A. Ricketts, Ispat Inland Inc.

Lecture 2 Blast Furnace Slag

J. L. Blattner, AK Steel Corp.

Lecture 3 Blast Furnace Reactions

A. McLean, University of Toronto

Lecture 4 Blast Furnace Energy Balance and Recovery: Rules of Thumb

and Other Useful Information (Computer Game)
J.W. Busser, Stelco Inc.

Lecture 5 Blast Furnace Design I

J. Carpenter, Paul Wurth Inc.

Lecture 6 Blast Furnace Design II

N. Goodman, Kvaerner Metals

Lecture 7 Blast Furnace Design III

S. Sostar, Lake Erie Steel Co.

Lecture 8 Ironmaking Refractories: Considerations for Creating

Successful Refractory "Systems"
A.J. Dzermejko, Hoogovens Technical Services Inc.

Lecture 9 Iron-Bearing Burden Materials

M.G. Ranade, Ispat Inland Inc.

Lecture 10 Blast Furnace Control- Measurement Data and Strategy

R.J. Donaldson and B. J. Parker, Dofasco Inc.

Lecture 11 Maintenance Relial?i1ty Strategies in an Ironmaking Facilty

G. DeGrow, Dofasco Inc.
 1999 BLAST FURNACE IRONMAKING COURSE

CONTENTS

VOLUME TWO: OPERATIONS

Lecture 12 Coke Production for Blast Furnace Ironmaking

H.S. Valia, Ispat Inland, Inc.

Lecture 13 Day to Day Blast Furnace Operation

A. Cheng, National Steel Corp.

Lecture 14 Challenging Blast Furnace Operations

F.e. Rorick, Bethlehem Steel

Lecture 15 Burden Distribution and Aerodynamics

J.J. Poveromo, Quebec Cartier Mining Co.

Lecture 16 Casthouse Practice and Blast Furnace Casthouse Rebuild

J.B. Hyde, Stelco Inc.

Lecture 17 Environment, Health and Safety Issues in Blast Furnace Ironmaking

E. Cocchiarella and D. Foebel, Dofasco Inc.

Lecture 18 Fuel Injection in the Blast Furnace

F.W. Hyle, USX Corp.

Lecture 19 Ironmaking/Steelmaking Interface

C. Howey and R. Brown, Stelco Inc.

Lecture 20 European Blast Furnace Practice

D. Sert, IRSID

Lecture 21 Japanese Blast Furnace Practice

K. Yoshida, Kawasaki Steel Corp.

Lecture 22 Future Trends in Ironmaking

W-K. Lu, McMaster University
-l

~I LECTURE #1

J
1 HISTORICAL DEVELOPMENT AND PRINCIPLES
OF THE IRON BLAST FURNACE
J
John A. Ricketts
I Manager of Operating Technology, Iron Production
Inland Steel Company

I
I
FOREWORD
I This lecture is essentially a blending of the
material prepared for the previous McMaster Blast Furnace
I Ironmaking Courses, by R. W. Bouman on the Historical
Development of the Blast Furnace and by John F. Elliott
on Principles of the Iron Blast Furnace. A section on
I Modern Aspects of Blast Furnace Theory has been updated
by A. McLean with material drawn from the 1978 Howe
Memorial Lecture by E. T. Turkdogan and also two recent
papers by W-K. Lu which discuss the behavior of silicon
r and alkali metals in the blast furnace. A new section on
iron making 100 years ago has also been added by the
current author.

1-1
 -J

The contents of this lecture have been arranged in J

the following sections:

INTRODUCTION J
EARLY IRONMAKING
J
The First Ironmakers
Ironmaking in the Middle Ages
DEVELOPMENT OF THE BLAST FURNACE J
Pre-Industrial Revolution
Early Industrial Revolution ~
Late Nineteenth Century
Early Twentieth Century
DEVLOPMENT OF BLAST FUACE FUDAMNTALS
I
Early Scientists I
Gas-Solid Contact
Solution Loss
MODERN BLAST FUACES I
Raw Material Preparation
Combined Blast
Large Blast Furnaces
I
Top Pressure
Burden and Gas Distribution
I
MODERN ASPECTS OF BLAST FUACE THEORY
Reduction of Iron Oxides I
Fluxes
Slags
Reactions in the Bosh and Hearth
Energy Considerations
r
í-

CONCLUDING REMA

SOURCES OF ADDITIONAL INFORMTION

1-2
 INTRODUCTION

The ironmaking blast furnace has played an important role in the

development of our industrialized civilization. This furnace has been
a means of producing metallic iron, which has been and continues to be
a major building block of heavy industry. The principal aim of the
iron blast furnace is to smelt iron ores and prepared agglomerates or
iron ore concentrates to produce a liquid crude iron. When liquid,
the crude iron is called hot metal or pig iron, and when solidified,
it usually is termed pig iron. The composition of the product depends
to a considerable degree on the use to be made of the metal. The
principal use is as a raw material for oxygen steelmaking for which a
typical composition is approximately 4.2% carbon, 0.8% manganese, 0.7%
silicon, less than 0.035% sulphur, and from 0.15 to 0.01% phosphorus.
The concentrations of manganese and phosphorus depend primarily on the
composi tions 0 f the iron ores and agglomerates charged to the furnace.

The raw materials consumed in the smelting operation in addition

to the iron-bearing materials, i.e., the ores and agglomerates, are:
coke which is the principal fuel; limestone and dolomite which act to
flux the earthy constituents, or gangue, in the iron-bearing materials
and ash in the coke to form a slag; and hot air and oxygen which are
needed to burn the coke; and minor fuels such as heavy oil, tar and
na tural gas.

The blast furnace produces a slag resulting from the union of

the fluxes with silica (Si02), alumina (A1203) and some of the manganous
oxide (MnO) which are obtained from the coke ash and gangue of the iron-
bearing raw materials. A nominal composition of the slag is 45% CaO,
5% MgO, 35% Si02, 12% A1203, a few percent MnO, and 1 to 2% sulphur.
A large volume of low-grade gas is produced as well. The composition
of this gas varies somewhat with different furnaces and with raw
materials and fuels, but it will be approximately 56% nitrogen, 25% CO,
17% CO2, and 2% H2 on a dry basis. It will also contain some water
vapour. The heating value (low) of the gas is relatively poor, being
in the range of 0.8 to 1-.1 M cal/m3 (90 to 125 BTU/ft3). On leaving
the furnace shaft, these gases will contain considerable quanti ties of
dust, a major portion of which is removed in auxiliary facilities.

The furnace in which the process of smelting occurs is a tall,

refractory-lined steel shell having a circular cross-section. During
operation of the furnace, this shaft is filled with a carefully con-
trolled mixture of the iron-bearing materials, coke and fluxes which are
coarsely granular in form. It is to be noted that in many modern
opera tions some, or in some cases all, of the fluxes are incorporated
in the iron-bearing portion of the charge. Hot air for combustion
of the coke in the èharge is injected into the lower portion of the
furnace through water-cooled nozzles, or tuyeres. The coke and
auxiliary fuels that may be injected into the tuyeres are burned in
the region just in front of the tuyeres to produce a very hot gas that
consists principally of CO and nitrogen. This gas passes up through
the charge in the shaft and heats and alters the charge chemically in
its passage. As a result of burning of the coke at the tuyeres and

1-3
 j
mel ting of the iron and formation of the liquid slag inthe lower
region of the shaft, the solids in the shaft descend slowly and pass J
through the furnace in approximately 8 hours. Accordingly, new charges
of iron-bearing materials, fluxes and coke are added at regular inter-
vals to the top of the furnace, and the liquid slag and hot metal are J
drawn off at the bottom periodically.

The lower end of the shaft below thetuyeres is -a crucible in

which the liquid slag and hot metal is collected. This crucible is J
lined with carbon brick or with high quality refractory brick.

The contour of the shaft is designed very carefully and will vary J
in subtle ways depending on the type of raw materials being smelted,
furnace size, etc. From the top or throat section where the solid
materials are placed on the bed, the shaft widens at a very low angle ~
to allow the bed to expand slightly as it descends. There is a cylin-
drical section, or belt, approximately two-thirds the distance down
the shaft which joins the upper tapered section to the lower tapered
section, or bosh. The bosh is a short, tapered section which restricts I
the cross-section to compensate for the sintering and fusion of the bed
as its temperature rises. The barrel-shaped section below the bosh
contains the tuyeres and the crucible. I
Facilities at the top of the furnace shaft seal it to permit
operation at pressures of 1 to 3 atmospheres, gage. These facilities
provide for collection of the gases after they leave the shaft and for
I
regular and controlled additions of the raw materials and coke. The
furnace is also serviced by facilities for removing the hot iron and
slag. The system for supplying the hot air blast for the tuyeres I
includes very large air compressors, three or four stoves for preheating
the air, and duct-work to distribute the air to the tuyeres. Most fur-
naces also include equipment by which the auxiliary fuels may be
injected into the tuyeres.
I
In the following sections the history of ironmaking is briefly
reviewed. Particular emphasis is given to the major structural and
I
mechanical developments as well as the evolution of blast furnace
theory. The aim of this lecture is to cover the most basic fundamentals
of the ironmaking blast furnace process and show how these fundamentals I
have resulted in furnaces that today are capable of producing over
10,000 tons of pig iron per day. r
EARLY IRONMAING

The First Ironmakers

The first reduction of iron ore to iron probably took place

during the bronze age and was accomplished by using smelting holes of
the type illustrated in Figure 1. By the time of the Romans, iron
smelting was practiced throughout most of the known world. At this
stage the process was a batch operation in which charcoal was ignited
and, when sufficiently hot, produced hot carbon monoxide that ascended

1-4
J
r_2_ -:1
1 _-.r--i=-~~~~,,=-
--41-0 -: -:;-
°Ë-:-::::------~

T-l::::
i L ,.=-/I=;';':~::
'.::.:.:-S. ............ . I'""
- ---~L.L-i~..~
J

~i
i:t::::::~:: .-
I:~:~~:::.. :.:.;..:::.::.:~..:¡¡.'~.:;:..;:.:;..::¡..!.r.r.:;:.i¡¡¡r.¡¡.r.::~~.'~:.:::.'¡:.i.¡¡¡.~.:¡.¡¡:.1¡.:¡..t..¡~.'~:.':~.'¡:..¡¡.1¡.'¡¡l:~:::~:)"

;., i,liiiJII'
i
.:.:.:.:-.:.:.:.:.:.:.:.:.:.:.: ::::::::::::::~::::::::~:::::::~::::::::::::::::::::::::::::::~:::::::::'"

:~:~:~:~:~:~~:Æ~::::.:.. .:::::::::~:::::::::::\~/:é::::::::::::::::::::::::::::::;:Z
;~~~:~:~~~:~~I~~tf~~~::::... .:/~~~ff#II~I~~::~~I:~I~ff~tIII:::

~ lilll~i~¡!l¡MJll,.iii'11l\
I
Figure 1. Early Ironmaking Smelting Hole
I
I
I
I
I
(

Figure 2. Early Bowl or Shaft Furnace for Smelting Iron Ore

1-5
¡

j
to reduce and smelt the ore. Bellows were apparently used qui te early
to provide the air for combustion. These operations were very ineffi- J
cient in the use of both the ore and the reductant. Much of the iron
oxide in the ore was not reduced, and since mel ting temperatures were
not reached, this unreduced iron and impurities such as silica and J
alumina were surrounded by metallic iron at the end of the smelting
operation. The spongy mass, or bloom, was removed from the smelting
hole when the charcoal was spent and formed into tools and weapons.
The forming and shaping operations also served the very important
J
function of removing most 0 f the iron oxides and other impuri ties
trapped in the bloom. Analyses of some of these early iron blooms and
implements indicate that their average composition before surface car- J
burizing was:

Percent ~
Carbon 0.03 - 0.10
Silicon nil- 0.05
a
Manganese nil - 0.15
Sulphur 0.005 - 0.050
Phosphorus 0.05 - 0.50
I
This implies that the iron content of these materials was greater
than 99% and that some of these early irons were relatively pure. These
first attempts at ironmaking produced mostly wrought iron, but some of
the material would today be classified as steel.
I
As the demand for iron increased, ironmakers began looking for
bigger and better methods of producing their blooms. Bowl furnaces I
or short shaft furnaces similar to the one shown in Figure 2 came
into use. The shafts were probably no more than 6 feet in height and
were lined with clays. The advantages of this type of smelter were I
that they could hold a larger charge of ore and charcoal, and eventually
had an opening in the bottom for the removal of the mol ten slag that
formed during the smelting operation. These slags contained the ore
impuri ties such as silica, alumina and lime, and unreduced iron oxide.
I
Air was introduced into these furnaces through one or more openings
located above the slag hole by natural draft and by mechanical blowing
devices. The early shaft smel ters were still batch operations and the I
iron product was still a bloom or spongy mass. After each batch was
processed, the shaft was at least partially dismantled to remove the rr
bloom. Some of these furnaces were constructed or excavated on the
side of a hill and others were free-standing on level ground.

Another type of early iron smelting furnace is shown in Figure 3.

This furnace resembles a beehive coke oven and was constructed with
al terna te layers 0 f charcoal and iron ore. The charcoal and ore mound
was then covered with a thick layer of clay, the bottom charcoal layers
were ignited, and the smelting operation was started. Near the end of
the smelting operation, the clay dome undoubtedly collapsed around the
iron bloom.

1-6
J

The early Japanese smelters produced iron f~oW iron sands and
-I charcoal on an elaborately constructed hearth. This operation,
called the Tatara process, was practiced in Japan as late as the 19th
century. The Tatara furnace was large by early ironmaking standards
J and apparently produced as much as four tons of spongy metal in one
batch. By comparison, it is doubtful that the early ironmaking opera-
tions shown in Figures 2 and 3 produced blooms much larger than 500
J pounds.
The earliest cast of liquid iron was probably produced in China.
There is evidence that cast iron was made in China during the first
J centuries of the Christian era, much before any such activity in Europe.

~ Ironmaking in the Middle Ages

The art of ironmaking spread rather rapidly throughout Europe

I and the Medi terranean area during the Roman era. Roman shaft smelters
similar to that shown in Figure 2 and dating back to the second century
A.D. have been found in Britain. with the decline of the Roman Empire,

I ironmaking seemed to decline in importance. At the beginning of the

14th century, ironmaking was being practiced as i thad been 2000 years
previously . However, the 14th century marks the start of ironmaking
developmen ts that continue today.
I In addition to shaft furnaces, European iron smel ters in the
Middle Ages used hearth furnaces. This type of smelter was eventually

I expanded in size and equipped with a mechanical air blowing device, as

shown in Figure 4. Smelters of this type were used in Spain and
France, and were known as Catalan forges. The air blowing equipment
used with the Catalan forges was a large air aspirator and apparently
I could develop as much as 1.5 to 2 psig of air pressure - considerably
more than could be achieved with the hand or foot powered bellows that
were used during the previous centuries. The Catalan forge did not
I change the basic ironmaking practice that had previously developed but
did significantly increase the size of the blooms produced.

r The most significant ironmaking development of the Middle Ages

was the enlargement of the shaft smelter. A larger shaft smelter,
named the Stückofen, came into use in Germany during the ,early 14th
century. This development is now generally recognized as the earliest
blast furnace. At first the Stückofen was a batch operation and pro-
duced a bloom as in early shaft furnaces. However, the Stückofen was
eventually made taller, probably as a result of the availability of
the higher blast pressures made possible by water-powered bellows.
The Stückofen was constructed as two truncated cones with one on top
of the other as shown in Figure 5, and was made up to 15 feet high and
5 feet in diameter at the widest section.

As a direct result of water-powered bellows to produce higher

blast pressures and the larger Stückofen furnace with reduced heat
losses, mol ten iron started to be produced in Germany during the very
late Middle Ages. The formation of liquid iron in the smelter

1-7
 _1

J
- ~_,~o :'~;";';~:;.~:-.=-~:"7~~:':~"::i""":f';:-C::':_' :.:. ~ :.~.:~:~ : g""
J
-Q..... ....~..Q;L.:Q-CO ...-~ ~ -

~
Figure 3. Beehi ve Furnace for Smelting Iron Ore

~.l
AIR ~i~"
"Ii 11/
I
I

I
I
I
r
L_

Figure 4. Catalan Forge wi th Air Aspirator

1-8
 J
J

J
i
)
~

I
I
.. .-. ~.' .. ... .
'''~;~.';~2~;f~~::r~~7~S:C~~~ii~7l';:;:';? .....
I Figure 5. stückofen or Bloom Furnace

I
I
I
r
,
L-~

Figure 6. Early Charcoal Blast Furnace

1-9
 J

undoubtedly presented problems for the ironmaker. First, be was faced

with a containment problem, and secondly, the liquid product was not J
of the same composition as the previously produced blooms. It appears
that the most common solution to the containment problem was to allow
the mol ten iron to flow from the hearth of the shaft into a forehearth. J
Here the mol ten iron was allowed to solidify and form what is now
called pig iron. The second problem wi thpig iron was its high carbon
content. This problem was solved by the development of a two-stage
process that produced wrought iron. The first stage was the production
J
of pig iron in the Stückofen, and the second stage was the mel ting and
decarburizing of the pig iron in a small hearth furnace, or bloomery.
The two-stage operation then resulted in a product that was similar to J
the blooms that were first produced in shaft furnaces. This two-stage
operation, developed well before the Industrial Revolution, is analogous
to present day steelmaking in blast furnaces and oxygen blown converters. ~
One result of the two-stage process was that the smelting of iron ore
ina blast furnace could be separated from the product-making operation.
This separation of functions eventually played a major role in the
~
enlargement of shaft smelters.

One other notable ironmaking event that took place in the Middle
Ages was the passing from a batch operation to a continuous operation. I
This event has apparently not been noted by historians, but it must be
considered significant in the development of blast furnaces. Continuous
blast furnace operation probably started shortly after liquid iron was
produced in the Stückofen. Once the iron smelters realized they did
I
not have to drag a bloom from the bottom of their shaft, it was a
logical step to continue the charging of raw materials and the casting
of liquid iron. I

DEVELOPMENT OF THE BLAST FURNACE I

Pre-Industrial Revolution

During the 17th century, Britain was beginning to emerge as a

I
leading ironmaking country. Up to this time, other European countries,
notably Germany, France and Sweden, had been the leaders in ironmaking
developments. The ironmaking operations of this era were producing at I
best 1 to 2 tons per day ,and were dependent on the essential raw
materials of iron ore, wood to make charcoal, and water power.
Because of this dependence, ironmaking operations were required to
r
L_
move frequently as the local supplies of wood and ore were exhausted
and new sources were discovered. In Britain, and to a lesser extent,
in other ironmaking countries, the availability of wood became a
problem in the 17th century. The ironmaking operations consumed vast
quanti ties of wood, and concern about the availability of wood for
ironmaking and ship-building was increasing. This supply problem was
recognized by the British iron smelters, and to a lesser extent, in
other ironmaking countries. Attempts to use coal in place of charcoal
were made in the late 17th century. These attempts were largely
unsuccessful due to the high sulphurcontent of the coal and its inability to
support the ore in the blast furnaces without a large pressure drop.

1-10
J
The ironmaker i sunderstanding of his blast furnace increased
1 significantly in the 18th century. In the early 18th century, after
unsuccessful attempts at using coal, a British ironmaker by the name
of Abraham Darby tried to use coke in his blast furnaces. Coke was
J being produced near Darby's ironmaking operations for use in malting
kilns, and after some experimenting with this new ironmaking fuel,
Darby established an ironmaking operation based on coke in 1713. This
I event must be considered one of the most important blast furnace
developments of all time. In view of the serious wood shortage
problems then facing the country, this development was to eventually

i save the British ironmaking industry. In 1740 there were 50 blast

furnaces operating in Britain. The average production of a furnace
was 6 tons per week, and only Darby i s furnace was using coke. By 1790
there were 106 blast furnaces and 81 of these were using coke. The
~ furnaces using coke averaged about 17 tons per week.

Other blast furnace developments that occurred in this Pre-

I Industrial Revolution period were the changing shape of the lower

sections of the shaft and improved methods of blowing. The charcoal
furnace used prior to the use of coke had a small hearth and a flat,
almost horizontal bosh just above the hearth as shown in Figure 6.
I The purpose of the bosh was to support the raw materials in the shaft
above. Because liquid iron and slag dropped to this surface and ran
into the hearth, the bosh eroded rapidly and was probably where these
I early furnaces failed most frequently. With the use of coke instead
of charcoal, the ironmakers soon found the flat bosh was not required
because the coke was much stronger and could support the raw materials

I in the shaft without crushing. Furnacemen also found that with coke
the shafts could be built taller and thus produce more iron.

wi th taller furnaces made possible with the use of coke, air

I blowing requirements increased. At first this was achieved with more
water for the water wheels; horses were also used to produce blast for
the furnaces. However, late in the 18th century, steam engines came
I into use for blowing blast furnaces. At the same time as the intro-
duction of steam engines, piston and cylinder blowing machines began
to replace the bellows that were used with the earlier water wheels.
r These developments significantly increased the blowing and production
capabili ties of exis ting furnaces and, with coke as a fuel, permitted
the furnaces to be increased in size.
In the very early 19th century various grades and quality of iron
had already been established for trade. The ironmaker of this era
had learned how to control the reduction of silica in his furnace and
had apparently long since learned how to make fluid slags with the
addi tion of limestone to the charge. The blast furnaces of this
period were still no more than about 30 feet high and were constructed
entirely of stone and fireclay. The largest of the circular furnaces
(many were rectangular in cross section) were two to three feet in
diameter at the top, up to nine or ten feet in diameter at the top of
the bosh, and had a hearth three to five feet in diameter. The
production from these furnaces was only a few tons per day, and the
coke consumption was, at the very best, two tons per ton of iron. The
furnace tops were open and belched great quantities of fire and smoke.

1-11
 J

Significant developments in methods of refining iron into useful 1

products were also made in this period. The use of cupolas for the
mel ting of pig iron was developed in the 18th century. More importantly,
the puddling furnace was invented by Henry Cort at the start of the
Industrial Revolution. The puddling furnace removed carbon and other J
metalloids from remelted pig iron with an oxidizing flame and the
additions of ore, the result being a spongy mass of wrought iron that
could be formed. This operation was a type of early open hearth furnace J
and further permitted the separation of the ironsmel ting and iron
refining steps.
Thus, at the start of the Industrial Revolution, the ironmakers J
in Britain were in a strong position to provide the building blocks of
heavy industry as a result of the development of coke and steam power
for blowing. The further developments of the two-stage ironmaking ~
process as a result of the puddling furnace invention also opened the
way for the yet-to-come two-stage steelmaking processes.
I
Early Industrial Revolution

During the early part of the Industrial Revolution, the basic I

principles of iron smelting blast furnaces did not change from the
earlier 18th century technology. However, significant mechanical
developments were incorporated into iron blast furnaces in this period.
These mechanical improvements were prompted by the tremendous increase
I
in the demand for iron and iron products. In Britain for instance,
pig iron production increased from about 125,000 tons at the beginning
of the 19th century to about 400,000 tons in 1820 and again to about
I
2.5 million tons by 1850.

The most significant ironmaking development in the first half of I

the 19th century was the invention of preheated blast air in 1828 by
James Neilson, a Scotsman. Up to this time, ironmakers believed that
hot blast would not help their blast furnaces. This belief was based
on their observation that the furnaces seemed to operate more
I
efficiently and produce more iron during the colder winter months.
The early ironmakers did not recognize that this seasonal fluctuation
was due to changes in the moisture content of the air. Neilson
I
apparently made a chance observation that blast furnace air that was
only slightly elevated in temperature made a remarkable improvement in r
the performance of the furnace. He further developed the idea and L_
received a patent for his preheated blast concept. The technique was
quickl y adopted by furnacemen in Scotland and the res t 0 f Bri tain . The
first hot blast systems consisted of an iron pipe enclosed in a refrac-
tory tunnel, with either coal or blast furnace off-gas being burned in
the annular space. These early systems were limited in hot blast tem-
perature; however, the effects on furnace operations were quite
noticeable. The production on the largest furnaces of that day went
from 30 to 40 tons per day. Because of the importance of high hot blast
temperatures in modern blast furnace technology, the development of pre-
heated blast must rank in importance with the use of coke in the histor-
ical development of the blast furnace process.

1-12
 ----_._--- -_. -

By about 1840, blast furnaces were being built up to 60 feet high

1 with an internal diameter of 16 feet at the top of the bosh. The hearths
of these furnaces were up to 8 feet, and the internal reactor volume was
as much as 7,000 cubic feet. One of these furnaces is illustrated in
I Figure 7. It was also apparently in the early 19th century in Scotland
when iron pipes and water-cooled tuyeres were first used to introduce
the air into blast furnaces. Previously, leather and canvas tubing
carried the blast air to the furnaces and clay tuyeres were used to
~I
introduce the blast into the furnace.

i By the middle of the 19th century Britain had become the leading
iron producer in the world and pig iron production by the largest
furnaces was up to 30 tons per day. Coke was the most common fuel and
reductant for blast furnaces in Britain at this time and coke consump-
I tion was abut two tons per ton of pig iron. However, there were at
least two significant ironmaking operations based on the direct use of
coal. Scottish ironmakers were successfully using a hard splint coal
I in their blast furnaces during this period, and American ironmakers
had developed an anthracite blast furnace practice.

I By 1870, blast furnaces were producing up to 60 tons per day. The

incentive to produce more iron and build larger blast furnaces increased
with the development of steelmaking by Bessemer, Siemens and Thomas.
The processes developed by these individuals allowed the conversion of
I pig iron into steel and, as a result, started the modern steelmaking
era. The effect of these developments on iron production in the late
19th century was dramatic. Blast furnace iron production in Britain
I rose from 2.5 million tons in 1850 to 8 million tons in 1895. The
production of steel in Britain rose from about 200,000 tons in 1865 to
3.3 million tons in 1895. However, the growth of the young steel
was most dramatic in the United States. In 1871 blast
I industry

furnaces in the U. S. produced about 1.7 million tons of pig iron per
year, but by 1890 the production of U.S. furnaces was over 9 million
tons per year and greater than that of the British industry. As in
I Britain, the production of blast furnace iron was driven by the increas-
ing demand for steel and steel products, and by 1910 U. s. furnaces were

r producing more than 27 million tons of pig iron per year. As a result,
a new leader in iron producing capability and technology was established.

The American blast furnace in the early 1870 decade was for the
most part still a stone and masonry structure lined with refractory
brick. The furnaces were hand-filled through open tops; however, some
furnaces were using a single bell and hopper arrangement to seal the
furnace between charges. Some furnaces also had facilities for direct-
ing the off-gases to a boiler for steam generation. steam-powered
blowing machines were fairly common, but some furnaces, particularly
charcoal operations, were still blown by water-powered equipment. Hot
blast, when used, was typically produced in iron pipe stoves. A produc-
tion 0 f 30 tons per day was cons idered good in 1870. A production
record of 100 tons per day by the Lucy furnace located near Pittsburgh
in 1874 received world-wide publicity. In 1870 half of the pig iron
produced in the U.S. was made in anthracite furnaces, 30% in furnaces
using coke and 20% in charcoal furnaces.

1-13
 J

I
r
L_

,. .~--..'
:.'l"l ~ _, "' .. ",...J:.
t:il-:' "'''.--
:
--:-&~-
. .:~~.: . .. o ~,_....,.'
,;~"~:- .
.. - .. .... . .-
.;...:.:.:..~...
- ':'..i~:§ " co
. ..
eJ
. c ¿-.:::::
... :"-.l _ .ø - o .. ...-.

Figure 7. Mid-19th Century Blast Furnace

1-14
 I

-I
CHACOAL IRON MAING
1860 TO 1890
l 1870 BLAST FURNACE DESCRIPTION
The typical shape of a blast furnace is a vertical
') shaft formed by two truncated cones joined at their
bases. The upper, taller cone stands upright and is
known as the "STACK". The lower, shorter cone is in-
i verted and is known as the "BOSH". Below the bosh is a
bottom-sealed vessel where liquids accumulate called the
"CRUCIBLE" or "HEATH" (Figure 8).
~ The top of the furnace is open and is called the
"THROAT". The platform on the top of the furnace sup-
ports a short chimney with an opening for raw material
~ charging called the "TUEL HEAD". Gases from the iron
making process are captured at the throat by a ¡'GAS
PORT" and are transported to a boiler or hot blast oven
I by the "DOWN COMER".

I 1'1 LLING HO LI'

GAl PORT
I
I
---
TUN" IL HIAD

I .ID WORIt

r I T AC It

WHI TI THROAT
WORI

aLA S T
aoH11 IPI
l aLAIT
,
CRU~I.LI _
MAl H
TU",..I HOUII

TYM"
HI.RTH I,,OMI

Figure 8 - Charcoal Blast Furnace

1-15
 J

~
Figure 9 - Furnace Stack Overview

I
The massive construction of the tapered rectangular
blast furnace is known collectively as the "STACK PIL-
LA". These stack pillars form a four sided block of I
masonry that is braced with iron tie rods and united by
cylindrical arches on each side which form the "TUYERE
ARCHES". The tuyere arches allow an opening for the I
"BLAST MAIN" to feed hot air through the "BLAST PIPE"
and into the "TUERE" which fits into the furnace (Fig-
ure 9). The inside bosh and stack is lined with a fire
brick called "WHITE WORK". This brick is 15 to 18 I
inches thick and withstands the high iron making tempera-
ture. The outside masonry that supports the fire brick
is ei ther brick or rough stone and is known as "RED I
WORK". A small space, filled with loose sand or slag is
maintained between the white work and red work for expan-
sion as the fire brick heats. I
The crucible or hearth of the furnace has several
parts. The bottom is a solid stone called the "HEATH
STONE". Liquid iron and slag sit on top of this stone.
r
lL

The front of the furnace where the iron and slag is re-
moved is called the "FOREPART". The liquid products
must flow over the "DAM" and under the "TYMP". The
furnace is constantly filled with raw materials through
the tunnel head but is only cast by knocking out a por-
tion of the dam when iron fills the hearth. The slag is
drained continually into "SLAG PITS" , but the iron is
only cast every few hours into a ditch called a
"TROUGH" which leads to small runners called "SOWS"
which have numerous cavities attached called "PIGS".
These iron pigs weigh between 70 and 100 pounds. This
whole process takes place in the "CASTHOUSE".

1-16
 I

-I
other maj or parts of the blast furnace include a
"BOILER" which produces steam for a "BLOWING ENGINE"
J that supplies air for burning the fuel in the furnace. A
"HOT BLAST OVEN" is a rectangular brick structure with
many pipes. Gas collected from the furnace stack is
~I burned in the oven and heats the pipes. As the "COLD
BLAST" from the blowing engine passes through these
heated pipes , it becomes "HOT BLAST" which flows into
J the furnace.

Charcoal which is the fuel in the blast furnace is

produced by partially burning wood in a "CHACOAL
~ KILN". Other raw materials charged into the blast fur-
nace are "IRON ORE" which becomes the pig iron and
"FLUX" which forms the slag. All of these raw materi-
i als are stored in a "STOCKHOUSE" . In the stockhouse,
they are weighed to specific proportions. The raw materi-
al s are then lifted to the furnace top by a "HOIST
i HOUSE" elevator and charged into the furnace (Figure
10) .

I
BLAST FURNACE PLANT LAYOUT
I COLD BLAST

I BOILER AND
BLOWING ENGINE
BOILER AND
BLOWING ENGINE
HOUSE I HOUSE
HOT HOT

I BLAST
OVEN
BLAST
OVEN

I
i R.
...."

HOT BLAST

I
CASTHOUS '
.....CQ.L
ItIL.
MA.Rek
KILN STACK
~
L:
HOT BLAST

STOCKHOUSE

DOCK

Figure 10 - Plant Layout

1-17
 L

J
RAW MATERIALS
Charcoal was the chosen fuel for blast furnace opera-
tion in early industrialized America because there were J
vast forests of hardwood in most unsettled areas. Char-
coal is simply partially burned wood, which is a form of
carbon. Wood normally burns in three stages. First, J
moisture in the wood is driven out as steam. Then the
volatile matter, sap, oils, and pitches, is burned off
which creates gases and smoke. Finally, with only the J
carbon remaining, flames and smoke disappear and charcoal
embers glow releasing great energy in the form of heat.
The production of charcoal for blast furnaces was accom-
plished by allowing only the first two steps of this ~
process which resulted in the final product of high car-
bon charcoal.
~
The preferred wood for charcoal production was hard-
woods, such as maple, oak and birch. The wood was cut
into four feet lengths with a diameter of four to six
inches. I
The average production of a two man crew was to cut
and pile four (4) cords of wood in a ten hour day. The I
wood choppers were paid approximately $0.80 per cord in
the 1860's. The charcoal yield from a cord of hardwood
was about 50 bushels. On the average, one ton of iron I
required 100 bushels of charcoal which is two cords of
wood.
Once the wood was cut, the charcoal could be produced I
by two methods: pi t and Kiln. The pit method could be
used in any open location since it did not require a
permanent structure. The kiln method was performed in I
stationary stone structures that were originally located
in close proximity to the blast furnace. As forests were
cut down and wood supplies were exhausted, the kilns were
buil t farther from the iron plants. A number of blast I
furnaces were permanently shut down due to lack of char-
coal since charcoal transport costs from distant loca- r
tions resulted in iron prices that were too high to re- u_

main competitive. This same issue has resurfaced one

hundred years later because many steel companies cannot
internally support coke requirements and their iron pro-
duction costs increase with the purchasing and shipping
of coke from distant production locations.
The first step in producing charcoal by the pit meth-
od was to clean off a 30 to 40 foot circle of flat,
packed ground. Then 25 to 30 cords of wood were piled to
form a mound. The wood was positioned standing on end
and leaning toward the middle so that the mound looked
like an igloo. Once the cord wood had been put in place,

1-18
 J

1
small dry branches, called lapwood, were placed over the
mound of cord wood. This lapwood was the kindling wood
I for the cord wood. Then a layer of wet leaves was placed
on top of the lapwood and over the entire mound. Final-
ly, a 4 to 6 inch layer of earth covered the mound to
I reduce the amount of oxygen entering into the wood core
(Photo 1).
Once the mound was complete, the pi twas lit and
i allowed to burn for seven to eight days. At no time was
a live fire allowed to burn freely. Remember, only the
moisture and volatile matter were to be removed from the
I wood, so a slow, low heat, smoldering fire was neces-
sary. Slowly the mound decreased to one-third of its
original size as the moisture and volatile matter burned
i off. Finally, the charred wood was carefully raked from
the mound without exposing the remaining wood that was
not fully charcoal. The finished charcoal cooled while
the remainder of the mound was allowed to complete the
I process.

I
I
I
I
Ir
1-,___

Photo 1 - Charcoal Pit

(Courtesy Marquette County Historical Society)
The cooled charcoal was sacked and loaded into wagons
which were drawn by horses or mules. The finished char-
coal was then delivered to the blast furnace plant. The
average pit of 25 to 30 cords of wood would yield 1,000
to 1,500 bushels of charcoal.

Charcoal kilns were hollow, beehive shaped structures

made from local stone or brick (Photo 2).

1-19
 J

J
Wherever possible, the kilns were built along hill-
sides to allow loading the cord wood from the top. If
this hillside location was not available, then a loading
platform was constructed. Each kiln was 14 to 28 feet in J
height. There were two large openings in each kiln; one
at top center and the other on the side at the bottom.
The top hole was 4 to 5 feet in diameter and was the J
charging hole used to stack the cord wood. The bottom
opening was slightly larger, in the shape of a door, and
was used to start the fire and later to remove the char- J
coal. There were also approximately 15 to 30,
four-inch-square openings, called "air vents", located
roughly two feet apart all around the kiln about three
feet from the ground. J

I
I

Photo 2 - Fayette Kiln I

(Photo by Author)
I
The four foot lengths of cord wood were brought in
through the top charging hole. Each piece was piled
parallel to the ground in two concentric circles. The 8
foot diameter center remained vacant and was later filled
I
with dry kindling wood. A small tunnel was made to the
side door to be used for an ignition channel. Anywhere r
from 40 to 75 cords of wood could be placed in a kiln
depending on its size. Once the kiln was filled and
ready, an oil saturated rag was lit and pushed in through
the ignition channel. The kindl ing wood was lit and
allowed to burn until flames were visible through the
charging hole. At this time, the door at the base of the
kiln was sealed and the charging hole diameter reduced by
using stone and plaster. The smouldering fire within the
kiln slowly worked its way from top to bottom. When the
kiln man saw glowing, red coals at the air vents, he
would seal these openings and the remainder of the top
hole. The kiln was now completely sealed and the wood
was allowed to char for eight (8) days.

1-20
J

I
When the charring was complete, the large side door was
opened and the charcoal removed with 15-tine forks and
shoveled into "scuttle-baskets". Each man would carry 2
I to 3 bushels of charcoal in his basket to a wagon or
ri railroad car. Each kiln would produce 2,000 to 3,750
bushels of charcoal which would support 200 to 375 tons
of pig iron production.

The charcoal produced in both the pit and kiln method

i did not have all the volatile matter fully removed. In
some samples gathered around an old furnace, the charcoal
still contained almost 18% volatile matter resulting in a
J 75% fixed carbon. It should also be noted that charcoal
samples had 0.5% K20, an alkal i, which is high compared
to coke and would result in accelerated furnace refrac-
tory lining wear. However, the sulfur content of char-
I coal is very low at approximately 0.05% which yields a
low sulfur, high quality pig iron. A full comparison of
charcoal to coke analysis can be seen below:
I
Parameter Charcoal Coke
I 90.90
Carbon (% ) 75.40
Volatile Matter 17.90 0.90
I Ash
S (% )
(% )
(% )
6.70
0.04
8.20
0.72
CaO (% ) 3.70 0.28
I MgO (% )
Si02 (%)
0.30
1. 50
0.09
4.13
Al203 (%) 0.20 2.24
P (%) 0.03 0.03
I K20 (% ) 0.50 0.16
r Most nineteenth century blast furnaces were built
adj acent to iron ore deposits.
The mines were originally open pits or "cuts". The
ore was mined by blasting solid rock into pieces of ore
that could be lifted by miners onto carts. Once the pits
reached depths of approximately 200 feet, then tunnels
became necessary to follow the veins of rich ore. Iron
ore removal was done by strong men with hand drills,
sledge hammers, pick axes and explosives. Tram cars
carried the ore to the surface. Miners were paid
$2.00jDay for 10 hours of work in 1865.
since the iron ore mined in the late 1800' s was from
new rich deposits, the iron content is better than raw
materials used in today's blast furnaces as seen in the
table below:

1-21
 i

J
Michigan Acid Fluxed
Parameter Ore Pellets Pellets
Fe (% ) 67.80 63.30 59.80
Mn (%) 0.07 0.10 0.06 J
P (%) 0.05 0.02 0.01
CaO ( % ) 0.29 0.20 4.33
MgO (% ) 0.05 0.22 1. 45 J
Si02 (%) 3.40 5.61 5.31
Al203 (%) 0.95 0.33 0.39
Acid pellets used by iron makers today contain only 63% - J
65% iron and fluxed pellets contain 59% - 61% iron. The
silica content of Michigan pellets is 5.5% to 6.0%. It
was the depletion of the high-iron content raw ore that ~
forced the development of concentrating low-iron content
ores with 30% - 35% iron into pellets with the 60% plus
iron content. ~
Another raw material required in ironmaking is lime-
stone. High calcium and dolomitic limestone are both I
suitable as fluxes for the blast furnace. Fluxes are
used in the ironmaking process to form slag of a proper
chemistry to remove sul fur from the iron. sul fur causes
cast iron to be brittle and break easier, therefore, the I
highest quality and highest priced iron has the lowest
sulfur. Most blast furnaces were built in the immediate
vicini ty of limestone deposits. Enough flux should be I
charged to remove sulfur from the iron, but too much flux
can result in a thick, gummy slag that will not run out
of the blast furnace. Therefore, iron masters moni tared
flux additions, slag properties and iron chemistry to get I
the right balance.

A good blast furnace flux should have large percents I

of calcia (CaO) and magnesia (MgO) since they remove the
sulfur and low quantities of silica (Si02) and alumina
(Al203) since they do not remove sulfur but increase I
the quantity of slag produced.

BLAST FURNACE OPERATION r

RAW MATERIAL CHARGING w_

Once all of the _ raw materials had arrived at the

blast furnace plant, they were usually stored in a build-
ing or at least under a roof to keep them dry. This
storage area was known as the stockhouse. The stockhouse
not only contained the various ore types, charcoal and
flux but also included a crusher and a scale. The crush-
er was driven by a steam engine and was used to crush ore
and flux to a smaller, nugget sized material to improve
furnace permeability and efficiency. The scale was used
to weigh the ore, charcoal and flux to the right propor-
tions to make the desired iron and slag qual i ty.
1-22
 I

J The charging process began by hand loading wheelbar-

rows with each type of material. These wheelbarrows had
l two side mounted wheels, sturdy legs and good balance for
easy dumping. The capacity of these barrows ranged from
500 to 1,500 pounds. Once the wheelbarrows were full,
they were rolled onto the scale and weighed. All weights
~I were recorded in a charging log. The charcoal furnaces
in the Upper Peninsula used 30 bushels of charcoal as the
standard fuel charge. This charcoal would be balanced
i with SOo to 1,000 pounds of ore and 40 to 60 pounds of
flux. This complete set of materials was called a
"charge". The charcoal would be kept in separate wheel-
barrows, but the ores and flux could be mixed into one
J barrow.
Once the materials had been weighed, they were taken
i to the top of the furnace. If the furnace was built at
the base of a bluff, a platform called the "stock bridge"
connected the flat top of the bluff where the stockhouse
i was located with the furnace top platform. If the fur-
nace was not built at the base of a bluff, an elevator
was constructed (Figure 11). These elevators were called
"hoist houses" and consisted of a hollow, roofed tower
I with two adjacent lift platforms. The tower also con-
tained a flight of stairs to the furnace top in case the
elevator malfunctioned. The elevator platforms were
I hoisted by small stearn engines.

I
I
r

Figure 11 - Hoist House

After the wheelbarrows reached the furnace top, they
were dumped into a charging hole by pushing the wheels
against a charging ring and lifting the back handles _ of
the wheelbarrow. Charcoal and ore/flux were dumped in
al ternating layers.
1-23
 J

It was this concern and the inability to increase J

production with this method of charging that resulted in
the installation of the first inclined skip hoist on a
Pennsylvania blast furnace in 1883. J
Originally, raw materials were dumped into an open
mouthed stack through a tunnel head. This could be dan- j
gerous to the individual charging the furnace since he
could fall into the furnace. Early blast furnace opera-
tors and technicians realized that an open top furnace
had two disadvantages. First, the flammable gas from the )
stack could not be captured to fire boilers or heat hot
blast ovens and second, that the distribution of the raw
materials was causing furnace inefficiency. When raw J
materials are dumped directly in the center of the fur-
nace, they form a conical heap. The fine material stays
at the center of the heap while the coarse particles roll
down and deposit at the furnace wall. This resulted in I
the wall area having higher permeability, and, therefore,
most of the gas and heat ran up the furnace walls. This
was detrimental to the furnace operation since the materi- I
al at the center of the furnace arrived unprepared for
melting in the bosh area and the excessive gas flow at
the wall would wear out the lining. I
The first device placed into the open mouth of the
furnace to allow the capture of all the gas and in an I
attempt to help the distribution of raw materials was the
"cup and cone". It consists of an inverted conical
cast- iron funnel fixed to the top of the furnace. This
"cup" was approximately one-half the diameter of the I
throat. Inside of this cup would sit a cast-iron cone
which was suspended from a fulcrum beam with a counter-
weight. In first design, the cone sat on the top of the
cup, and was raised by a hand winch (Figure 12). This I
system worked in closing the top to capture gas, but all
the raw materials still were piled at the center so the
burden distribution problem was not resol ved. In the I
second design, the cone sat below the cup suspended by a
chain to a counterweight fulcrum that pulled the cone up
against the rim of the cup (Figure 13). r

Figure 12 - Cup and Cone Charg ing

1-24
J
I

J
)
J
~

I
I Figure 13 - Bell Type charging
When the charge was placed on the cone, the cone would be
I lowered and the raw materials would slide off the cone
towards the wall of the furnace. In this case, the mate-
rial peaked in the form of a ring at the furnace walls.
The fine material stayed at the wall and the coarse mate-
I rial rolled toward the center. This resulted in a change
of gas flow patterns from the heavy wall flow with the
open top or top opening cone system to a heavy central
I flow wi th the bottom opening cone system. This change in
gas flow patterns resulted in less wall wear and improved
stability of the operation and a more consistent iron
I quality. It is interesting that one hundred years later,
devices are still being developed to measure and control
gas flow in the blast furnace.
If'
Once the cup and cone was installed, all the gas,
except that released when charging, was captured. The
gas was collected below the cup and in openings at the
side of the furnace that led to a large cast iron pipe
called a downcomer. This large pipe left the top of the
furnace and was then split into smaller pipes. Some of
the gas was diverted to boilers which provided steam to
the blowing engine, hoist engine or crusher engine and
the some of the gas was diverted to the hot blast oven
and burned to heat the cold blast.
This description of furnace charging gives an idea
of the numerous steps and equipment, but it also indi-
cates that blast furnace iron masters understood what was
happening inside their furnaces. Many of their improve-
ments were the basis of our current day equipment on a-
typical North American furnace.

1-25
 J

J
BLAST FURNACE STACK
The four main parts of the furnace stack from top to J
bottom were the throat, stack, bosh and hearth. The
charcoal furnace stood 40 to 50 feet high. The throat
was a vertical cylinder that was 4 to 5 feet in diame-
ter. Therefore, the top of the stack region was the same J
diameter, but it tapered outward as it descended to a
diameter of 8 . 0 to 9 ~ 5 feet. At this point, the stack
region met the bosh. The top of the bosh was its widest J
point and its diameter decreased as it descended. The
bosh bottom diameter was usually 3.5 to 4.0 feet. The
diameter of the lower stack and upper bosh was determined
by the type of fuel, type of ore and quantity of air ~
blown into the bottom of the furnace. If the bosh was
too narrow, the passage of hot gases moved more quickly
and smelting occurred higher in the stack. If the bosh I
was too wide, the hot gases moved more slowly and smel t-
ing occurred lower in the furnace. The optimum bosh
diameter would yield the best balance between the chemi-
cal reactions occurring between gases and the lumpy parti- I
cles in the stack and the physical reactions when liquid
slag and iron forms. This balance was critical to maxi-
mize production and minimize fuel requirements. I
The hearth of the furnace may be a vertical cylinder
or slightly tapered truncated cone toward the furnace
bottom. The hearth diameter was 3. 5 to 4. a feet. The I
hearth was smaller than the bosh because it held only
liquid which was denser than the raw materials, there- I
fore, required less space.

Approximately 30 to 40 inches above the hearth bot-

tom was where the tuyeres were located. The tuyeres were I
the openings where the blast was introduced to the fur-
nace. Charcoal furnaces normally have two (2) tuyeres
each at 900 from the front of the furnace or three (3) I
tuyeres with two (2) at 900 and one (1) at 1800 from the
front of the furnace (Photo 3). The inside of the
throat, stack, bosh, and hearth was lined with fire If
brick. This brick was wedge shaped to give a tight fit L

at the desired diameter.

There were usually two to three rows of brick making

the lining 16 to 24 inches thick. One main wear mecha-
nism of this brick was the high alkali content of char-
coal which dissolved the refractory. Immediately behind
the fire brick was a layer of sand or crushed brick.
This material was loosely packed and granular to allow
the fire brick to expand as the furnace was heated.
Behind the loose layer was the outer layer of the fur-
nace. This was usually made of large stones held togeth- -
er by mortar. Horizontal iron tie rods were placed

1-26
l

I
along the outer stack on all four sides to add support
(Photo 4). These tie rods were usually wrought iron bars
1 from 1-1/2 to 2 inches thick. At each end of the bar was
a large cast iron washer and a nut.

J
J
I
~

I
Photo 3 - Tuyere Arch - Fayette, Michigan
I (Photo by Author)

I
I
I
r

Photo 4 - Tie Rod Support - Fayette, Michigan

(Photo by Author)
At the front of the hearth were two sections re-
quired for iron and slag removal: the dam and the tymp.
The dam rose from the hearth bottom to a height of 15 to
25 inches. The dam held the liquid iron and slag in the
hearth. The tymp hung down from the upper hearth and
directly over the dam. A small gap was left between the
tymp and dam for liquid slag to run out of the furnace.
A hole was made in the dam for removing the iron from the
furnace. The complete casting operation will be dis-
cussed in a future section.
1~7
 J

J
The tympjdam and tuyeres were located under arches
that were built into the outer stone stack column. These
arch roofs were formed by keyed brick and tapered toward
the furnace. These arches could be 8 to 16 feet wide and J
8 to 12 feet high at the outside. The arch met the inner
lining of the furnace where the lining was supported by a
horizontal cast iron beam. The arches over the tuyeres J
were called "tuyere arches" and the arch over the
tympjdam was called the "casting arch". In many cases,
the casting arch was bigger to allow better access to the
dam. J
The total inner volume of the furnace was 1400 to
1500 cubic feet. As time went on into the 1890' sand ~
1900' s, the blast furnaces were constructed to have big-
ger volumes, more tuyeres, and, therefore, higher produc-
tion rates. A typical blast furnace of today has a a
53,000 cubic feet working volume.

Furnace Dimensions
Typical i
Charcoal Furnace
Furnace Toda V
I
Volume 1464 Ft. 3 53,000 Ft.3
Tuyeres 2 20
Hearth Diameter 3 Ft. 8 In. 28 Ft. I
Bosh Diameter 9 Ft. 4 In. 30 Ft.
Throat Diameter 4 Ft. 6 In. 22 Ft.
Total Height 45 Ft. 100 Ft.
Tuyere Height 32 In. 12.5 Ft. I
from Hearth

BOILERS AND BLOWING ENGINES I

The key difference between an ordinary furnace and a
blast furnace is the blast of air forced into the furnace
through the tuyeres. Originally, blast furnaces were fed
I
air by a water wheel connected to an eccentric shaft that
pumped a leather bellow. The blast furnaces in the Lake r
Superior region of 1870 used steam engines to deliver air G

or "wind" to the furnace.

The first step in running a steam engine is generat-

ing the steam within a boiler. Blast furnaces were built
near water sources not only for shipping purposes but
also for water supplies to generate steam. The boilers
were usually located in a stone building adj acent to the
furnace stack. The boilers were fired by the gas collect-
ed at the top of the blast furnace stack. These boilers,
approximately 30 feet long, were positioned vertically
above a boiler chamber. There were no internal flues-
inside the boilers. Gas was burned in the boiler chamber

1-28
I
I
and flames directly contacted the bottom of the boiler to
heat the water. Each boiler chamber had its own stack to
I draw the flames across the boiler which formed a draft.
The waste gases were then exhausted from this stack. The
steam generated in the boiler was then piped to the
blowing engine.
J
BL(IING ENGINE

il

J FLYWHEEL

COlD
BLT

I BLOWING CYllHDER

COlD
BLAT

I
Figure 14 - Blowing Engine
I The blowing engine was mounted horizontally on a
timber frame. It consisted of a single steam cylinder
that was 18 to 72 inches in diameter with a stroke rang-
I ing from 28 inches to 48 inches. The steam cylinder
piston rod was connected via a crank to a heavy, large
diameter flywheel. This flywheel was then connected to
I either two cylindrical blowing tubes that compressed air
in one direction or one blowing cylinder that compressed
air in both directions (Figure 14). These blowing cylin-
ders were 4 a to 50 inches in diameter and had a stroke of
I 3 to 5 feet. The cold blast pipes were connected to the
end of these blowing cylinders and wind called "cold
r blast" was then piped to the hot blast ovens. The blast
pressure was 2 to 3 psi, but the volume of cold blast was
not measured. This system remained in use on all blast
furnaces until 1910 when the first turbo blower was used
at the Empire steel Company in Oxford, New Jersey. Cur-
rently, turbo blowers deliver 80,000 to 120,000 Cu. Ft.
per minute of air to the blast furnace at 25 to 30 psi
pressure.
HOT BLAST OVENS AND WIND DELIVERY TO FURNACE

The first use of hot blast was in 1829 in Glasgow,

Scotland. In 1831, a New York blast furnace engineer
applied for a patent on "heated air blast". The idea did
not become a reality until 1836 and its success was mini-
mal. Other operators tried various methods to heat the
cold blast, but the first success came in 1840 at Dan--
ville, Pennsylvania.

1-29
 1
J

J
Throughout this developmental period, a controversy
still raged over which is better; hot blast or cold
blast. Theoretically, hot blast should deliver a higher J
energy blast to the furnace with a heat value that should
offset some of the charcoal consumption, but actual blast
furnace operation showed that cold blast furnaces operat- J
ed in the winter with a lower blast temperature used less
charcoal than the same furnace run in the summer with a
higher blast temperature. Therefore, the colder the J
blast, the better the fuel rate. What the operators did
not know, and was explained later, was that the moisture
content of air is lower in winter than in summer, and
that it was the high humidity in summer blast that caused ~
fuel rates to increase not the temperature difference.
Once this was explained to the blast furnace operators,
further development of the hot blast oven and hot blast ~
stove continued.

The hot blast oven used on most charcoal blast fur- I

naces was a simple heat exchanger. The oven was a rectan-
gular brick structure (Photo 5). The cold blast pipe was
fed into one end of the oven. The pipe was than connect-
ed to several rows of hair pin shaped pipes that stood I
upright in the oven similar to radiator coils. These
hair pin shaped pipes reached to the top of the oven and
were connected in series. In the bottom of the oven was
a combustion chamber. The gas collected from the blast
I
furnace stack was brought through the downcomer into a
gas flue in the combustion chamber. This gas flue con-
tained numerous sl its where the gas was burned. The I
burning gas heated the inside of the oven and all of the
cold blast hair pin pipes. The exhaust gas was sucked
out of the opposite side of the combustion chamber by a I
draft created by an external ~tack.

I
r

Photo 5 - Hot Blast Oven - Fayette Circa 1880

(Courtesy Marquette County Historical Society)

1-30
J
I As the cold blast passed through the numerous pipes, it
became progressively hotter (Figures 15 & 16).
J The blast left the oven as "hot blast" and continued
underground through the hot blast main. The hot blast
main went to the base of the furnace where it was split
J to feed the two or three tuyeres (Figure 17). The pipes
then turned upward out of the ground, came up to tuyere
level and turned 900 toward the furnace. The right angle
pipe is currently called the bootleg. The blast then
J continued through the blast pipe, now known as the blow
pipe, and into the tuyere. Since the volume of air being
generated by the steam blower engine could not be easily
~ controlled, a valve was placed on each blast pipe to
control the quanti ty of hot blast going into the fur-
nace. This valve allowed wind volume adjustments during
I start-ups, shutdowns or during cast. This valve was
usually a slide type orifice (Figure 18).

I
I .... .... ... ...
....
I

... ... I

.... .
I
. ..
l
I
... ..
r.. 'lr . ~"'Air

I
Figure 15 - Hot Blast Oven Plan View

I / " ...- L_::

"

r
L_
¡'
6.
Ii
l. '\ \
i ¡
i ,

i'
! ¡
~
~
:¡ ¡
¡I i ~
I II
ii ¡
1'1
IL
.i
¡ ~.~_. .

Figure 16 - Hot Blast Oven End View

1-31
 L
J

Figure 17 - Hot Blast Main to Tuyere

~

~\
.
\. I

--
~\-~
-"
~.~.',,--rr4l \'
,- -;

--' ~ .::-:~ \
;-'
I

~~ I

I
Figure 18 - Hot Blast Slide Valve

The tuyere design was different for a cold blast or hot I

blast furnace. A cold blast furnace used a sol id conical
copper nozzle for a tuyere and the blast pipe was connect-
ed to the bootleg with a flexible leather tube. A hot I
blast furnace required a water cooled tuyere with a solid
ball-and-socket joint at the blast pipe. These water
cooled tuyeres were double walled, liD" shaped tubes that r
were tapered into the furnace. Clay was used to help L

seal connections between the tuyere and blast pipe.

steam dr i ven pumps forced water through the tuyeres. At
the bend of the bootleg was a small hole with a shutter
containing a piece of glass or mica. This allowed the
operator to look at the heat intensity inside the furnace
in front of the tuyere. This opening was also used to
introduce various fusible metal samples to determine the
temperature of the blast when they fused at their known
melting point. Originally, cast or wrought iron tuyeres
were used followed by bronze tuyeres. These metals would
not conduct the heat away from the tuyeres so they
burned. Copper, an excellent conductor, carried the heat
1-32
J
I
quickly to the water and the standard tuyere became cop-
per as it is today. Leaky tuyeres were a constant
J problem causing a great waste of fuel since water cooled
the furnace. It also effected iron quality because more
whi te rather than gray cast iron was made due to the
cooling effect of the water in the furnace. In the ex-
1 treme case of water leaking tuyeres, explosions and break-
outs were caused in the hearth. The normal tuyere diame-
ter was 3 to 5 inches, and this depended on the number of
i tuyeres and the volume of blast delivered to the furnace.

The first regenerative type of stove in the united

i States, similar to those used on today' s blast furnaces,
were erected at the Cedar Point Iron Company, Port Henry,
New York and at the Rising Fawn Furnace in Dade County,

i Georgia in 1875. Hot blast temperatures now average from

17000 F to 19000 F on a typical blast furnace of today.

I
CASTHOUSE AND CASTING OPERATION
I The casthouse was the very heart of the furnace
operation. The building extended from the front of the
I furnace and was approximately 30-40 feet wide and 50-70
feet long. The roof was slightly raised above the walls
to allow smoke and fumes to escape. There were numerous
I doors to allow sand to be brought in and pig iron and
slag to be taken out (Photo 6).

The casthouse contains areas for iron casting and

I slag casting (Figure 19). The side for iron removal
consisted of a large ditch called a trough that sloped
from the front of the furnace to the casthouse floor. It
rL then split into two runner systems. A main runner on
each system ran parallel with the length of the cast-
house. As this runner sloped down hill, a series of dams
were made at regular intervals. At a right angle before
each dam, a smaller runner called a sow was produced.
Then off of this sow were numerous cavities called pigs.
This system looked like a series of piglets suckling
their mother. There were several parallel rows of sows.
These were produced by pushing "D" shaped wooden forms
into moist beach sand on the casthouse floor. During the
cast, as each sow and its pigs were filled, the sand dam
on the main runner was knocked out with a bar and the
iron ran downhill to the next sow and pig bed. There
were two such complete systems so that as one side had
its pigs removed and beds reformed, the other side could
be cast. This allowed an uninterrupted furnace opera--
tion.
1-33
 J

Photo 6 - Casthouse - Fayette, Michigan

(Photo by Author) I

~ SLAG
I

~
FURNACE RUNNERS
SLAG J
PIT I
WALL

DAM I
I
crh I
r
L_

~. JjN RU~R ~ l!
Figure 19 - Casthouse Layout
The other side of the casthouse was used for slag re-
moval. Slag was constantly running over the front of the
dam down a runner toward a pit. The dam was divided into
two halves, each one feeding a separate slag runner and
slag pit. The slag pit was a large depression in the sand
with sand ridges. These ridges would act as cracking

1-34
J

I
points when it was time to remove the slag. In some cast-
houses, a jib type wood crane is used to remove thick
I pieces of slag. If the casthouse man saw the slag layer
getting too thick, he would place a bar in the center of
the liquid slag. Then when the slag froze around the
J bar, a rope or chain could be wrapped around the bar and
hoisted by the crane. Once again, there were two complete
slag systems so that as one was being used, the other
could be cleaned and made ready.
J
What is the origin of the word "casting"? It proba-
bly was originated when the first furnace men saw the
J iron being "cast" or thrown from the furnace. The cast-
ing operation was in two parts. As mentioned earlier,
while liquid slag was formed and its level reached the
dam, it flowed between the dam and tymp, down the runner
~ and into the pit. The other part of the casting opera-
tion was the removal of the iron. First, the blast was
shut off the furnace using the valves at each tuyere.
I This was done for the casthouse man's safety. Then the
taphole in the middle of the iron side of the dam was
opened. The taphole was opened as one man held a wrought
I iron bar in the taphole and another man drove the bar
through the dam with a sledge hammer. The iron ran down
the trough, into one of the runner systems and into the
I sow/pigs closest to the furnace. When this pig bed was
filled, a dam in the main runner was knocked down and
iron ran into the next pig bed (Photo 7). This filling
of pig beds continued until iron stopped running from the
I taphole. The furnace men then replaced the taphole with
a moist mixture of sand-and-fire clay or sand-and-coal.
The blast valves were then reopened and wind put back
I into the furnace.

Photo 7 - Casthouse During Cast

(Courtesy Marquette County Historical Society)

1-35
 J

J
When the cast was complete, the men removed the iron
from the sow and pigs. This was done while the iron was
still hot since the pigs broke loose more easily when the
iron was red and slightly mushy. This was accomplished J
wi th sledge hammers and pry bars. Once the pigs had been
loosened from the bed, they were allowed to cool. The
cooled pigs were then loaded on to carts or railroad J
cars. This was a hard job since each pig weighed between
70 and 100 pounds. The casthouse men also wore wooden
clogs on their shoes to protect their feet from the heat J
of the pig beds.

The blast furnace was cast approximately six times a

day and produced 4 to 6 tons per cast. The iron produced ~
was classified into No.1, No. 2 or No. 3 grade. Exact
specifications for each were not discovered but the best
iron was a gray cast, low sulfur, low phosphorous iron ~
mainly used for railroad car wheels and rails. Charcoal
furnaces could produce this low sulfur iron due to the
small quantity of sulfur in wood versus the large quanti-
ties of sul fur carried by anthracite coal or coke in I
other blast furnace operations. The iron also had only
3 . 5% carbon versus 4. 3 % carbon in the iron of today.
This was due to the fact that iron making temperature was I
much lower in these old furnaces, and, therefore, did not
have as much carbon in solution. It should be noted that
these furnaces used an acid slag practice since the sul- I
fur was so low and the slag volume was 400-500 Lbs./Ton
of iron.

Some samples of iron and slag found around a char- I

coal blast furnace were analyzed and the chemical analy-
sis is presented here:
I
Iron Slaq I
C
si
3.54% CaO ll.16% r
3.37% MgO 5.94% L_

S 0.013% Si02 53.87%

Mn 0.27% Al203 10.42%
p 0.14 % S 0.014%
Ti 0.06% K20 5.08%
Cr 0.02% FeO 5.2%
Temp. 2300°F B/S 0.32
B/A 0.27

The slag samples found around the furnace are blue,

purple, green, black and white. (Note: B/S = CaO +
MgO/Si02 & B/A = CaO + MgO/Si02 + Al203) .

1-36
 J

J Today a much lower silicon iron is required in the

steel making process and iron sulfur is controlled with a
higher basicity slag. Typical iron and slag chemistries
J of today are listed below:

~I Iron Slaq
C 4.30% CaO 40.25%
i si
s
0.60%
0.040%
MgO
Si02
11. 45%
37.50%
Mn 0.50% Al203 7.72%
P 0.04% S 1.37%
~ Ti 0.03% K20 0.47%
Cr 0.01% FeO 0.26%
Temp. 2670°F B/S 1.38%
i B/A 1. 13%

I BLAST FURNACE OPERATING RESULTS

The blast furnace iron master both of 100 years ago
and of today looks at several key indicators to gauge his
I success. These indicators are production, fuel rate,
yield and cost. The charcoal blast furnaces of the Upper
Peninsula of Michigan showed progress in production and
I fuel rate from 1865 to 1890, but the cost and market
value of their iron finally shut them down. In the
1860's, the furnaces produced 15 to 20 tons per day which
I was increased to 40 to 50 tons per day by the mid
1880' s. Today, the typical furnace makes 3,000 NT/Day.
The fuel rate was approximately 115 bushels (2,000 Lbs.)
per ton of iron in the 1860' s, which was decreased to 96
I bushels (1,920 Lbs.) per ton by the mid 1880's. Today

typical fuel rate is 1,000 Lbs./NT iron. The productivi-

a

r ty, measured by net tons of iron per 100 cubic feet of

L _
furnace working volume, ranged from 2. 1 in the 186 a's to
3.5 in the 1880' s. Today productivity ranges from 6.5 to
8.5 NT/lOa Cu. Ft.. The yield for these furnaces which
is the ratio of the quantity of metallic iron put in the
furnace top to the quantity of iron sold was 90%. This
is lower than modern day standards of 97% since a large
amount of iron went into the slag 100 years ago.
These results show maj or improvements not only in the
20 years from 1865 to 1885 but from 1885 to today. There
are some world class blast furnaces today with 4 a tuyeres
and 4 tapholes that are producing 9600 NT/Day at 910 Lbs.
Fuel/NT of iron.

The table below compares an average operation in the

life of a charcoal furnace in 1880 to a typical blast-
furnace of today.

1-37
 J

J
Furnace
1880 Toda v
Production (Ton/Day) 33 3000 J
Productivity (Ton/100 Cu. Ft. Vol.) 2.25 6.5
Blast Pressure (PSI) 2.63 25
Blast Temperature (0 F) 750 1750 J
Charges/Day 107 150
Fuel Rate (Lbs ./Ton Iron), 2280 1000
Flux Rate (Lbs./Ton Iron 130 250 J
Ore Rate (Lbs . /Ton Iron) 3430 3000
Pig Yield (%) 90 97
SAFETY ISSUES ON THE CHARCOAL BLAST FURNACES J
The charcoal blast furnace operation was very hazard-
ous. One of the main concerns was fire. Many blast ~
furnace plants had fires caused by charcoal kilns or
furnace breakouts.
The men who filled the blast furnace were also in I
constant jeopardy. One furnace man fell into the furnace
during charging and another man, who became fatigued
while pushing his wheelbarrow, lost his balance and died I
when he fell off the furnace top. The worst incident
occurred when a man charg ing the furnace opened the cone
to dump material just as the stock in the furnace
"slipped". This slip sent flames shooting out of the
I
furnace top and engulfed the man. He died within a week
from severe burns.
I
The other maj or hazardous area was the casthouse.
Mol ten iron and slag shooting out of the furnace were a
constant threat. Here is the account of a furnace fore- I
man's near miss, as told in the March 13, 1874, Mining
Journal of Marquette, Michigan.

The furnace was acting like an animal does after

I
taking a large dose of caster oil and was pretty soft
inside, but rather hard at the forebed: and as he was
trying to ease her of her burden, she flew at him r
like a fiend. He succeeded in getting a hole in her
and after pulling the bar out, the cinders flew like
water from a hose, striking him on the shoulder, back
and legs, burning his pantaloons badly. But he was
quick in getting them off, he escaped with little or
no inj ury .
The casthouse also contained other hazards such as
wet sand. When the casthouse men would prepare the sow
and pig beds for casting, the sand had to be moist enough
to retain the molded shape. If the sand was too wet,-
molten iron would trap the water and quickly convert it
1-38
J

1
to super heated steam. This would cause a severe erup-
tion sending liquid iron high into the air and spraying
the whole casthouse.
I

J NEW TECHNOLOGY
Cornelius Donkersley, the Iron Master, of the Morgan
blast furnace in the Upper Peninsula of Michigan was
~ truly ahead of his time. There are two very interesting
stories of his ingenuity that are forerunners to modern
operating practices.
~
First, in May of 1871, the furnace took a big "slip"
and cold raw material fell into the hearth, causing it to
I chill. In fact, all the iron froze, filled both tuyeres,
resul ting in a 45 inch thick mass known as a "sala-
mander". The Mining Journal of May 31, 1871 explains
what was done in this seemingly hopeless situation.
I
Mr. Donkersley not being inclined to give it up and
always fertile in expedients, went deliberately to
I work to save the stack if possible. By his direction
the arch was broken through, a large tuyere inserted
above the chilled mass and coal oil was then forced
I into the stack through a pipe leading from the
top-house to the tuyere. The effect of this experi-
ment was most satisfactory, for in a short time af-
ter, iron and cinder were running out above the dam.
I The oil has been used steadily ever since and has
gradually cut the iron away until the present writ-
r ing. The mass has been reduced to wi thin eight inch-
es of the top of the hearth; and the prospects are
that the furnace will soon be making iron as usual.
Seventeen days later, the Mining Journal printed this
follow-up.
The salamander has been entirely removed with but
trifling damage to the hearth structure. The furnace
is now running as smoothly and as successfully as if
no accident had occurred.

This whole salamander removal process required six

days and seven barrels of oil. This is the first record-
ed case of using oil as a inj ectant to heat a blast fur-
nace. oil injection is now commonplace as an auxiliary
fuel and as an operating variable to control furnace heat-
levels, but it was not fully developed until the 1960' s.

1-39
 J

c1
The second bi t of technology used by Mr. Donkersley
to improve the furnace operation was also amaz ing for
that era. The fuel in the blast furnace was charcoal.
However, this iron master would also charge raw wood. 1

The Mining Journal of August 24, 1972 gives Mr. Donker-

sley's reasonings:
J
Charging white pine in connection with charcoal for
fuel, in a small percent, we find the furnace works
admirably by being supplied in this way with hydrogen J
which serves as a lubricant for the stock, giving a
tougher fibrin to the iron and effecting a saving of
over ten percent in fuel.
~
It is a well known fact that today's blast furnace
operators add hydrogen in the form of moisture or fuel
inj ection and get the same results. The increase of ~
hydrogen gas allows smoother furnace raw material de-
scent, reacts wi th iron ore to remove the oxides and
results in a more efficient operation. Maybe Mr. Donker-
sley didn' t exactly understand the mechanism was for this I
phenomenon, but it is amazing that he used this technique
over 100 years ago.
I
The charcoal ironmaking era laid the foundation for
new theory and practice to be further developed in the
twentieth century. I
I
I
I
r

1-40
 I

Early Twentieth Century

1
By 1910 radical changes had occurred to blast furnace equipment.
The new furnaces of this era would be recognized by a modern ironmaker,
I whereas many of the 1870 operations might have been mistaken for a
large house. The furnace lines had changed from the low, flat boshes
that came from the previous charcoal and anthracite eras to steeply
J sloping boshesas shown in Figure 20. Furnace hearths had

in diameter to 17 feet, and the height had reached 100 feet. This
been increased
height was of considerable concern to furnace-men; some in fact felt
i that 100 feet was too high for a blast furnace and 75 to 90 feet was
much more reasonable. The internal volume of these furnaces was about
25,000 cubic feet.

J Furnace construction had improved with ,the use of steel plates and
beams, and water cooling plates had been introduced to protect the steel
structure and extend the furnace campaign life. Prior to 1890, a typical
I furnace campaign was two years, but with the introduction of steel struc-
tures and water cooling, furnace campaigns were increased to over eight
years. A cross-sectional view of a furnace representing the latest 1910

I technology is shown in Figure 21. While the furnace shown in this figure
is much smaller in diameter than the newer furnaces in use today, it has
lines and external features similar to those of modern furnaces. The
the 1910 furnace was
I distance from the tuyeres to the stock line in

about 70 feet, whereas the largest furnaces in operation today have a

tuyere to stock line distance of about 85 feet.

I Regenerative hot blast stoves had replaced the iron pipe stoves
and blowing equipment had become much more powerful by the early twen-
tieth century. The numer of tuyeres in the furnaces had increased

I from two or three to as many as twenty, but more typically eight to

twelve. Internal gas combustion engines using blast furnace off-gas
were introduced for blowing in 1902, and turbo blowers were first being
considered. Since the furnace off-gas was commonly being used in boilers
I or combustion engines, dust catchers and gas cleaning devices were being
developed and used.
ir The type of raw materials used in U.S. blast furnaces also changed
L__
markedly from the late 19th century to the early 20th century. Coke
made from bituminous coal had become the standard blast furnace fuel
and reductant. The rich Mesabi iron ores had been discovered and were
being used in the blast furnaces located in Pittsburgh, Chicago, and
Cleveland. The beneficial effect of sized and washed ores was appre-
ciated, .and certain operations were practicing some form of raw material
preparation. The equipment for delivering raw materials to the furnaces
is probably the area that improved the most in this era. Most notably,
skip hoists replaced wheelbarrows as the most common method of charging
the furnaces. Also, large bulk carriers of all types had appreciably
al tered the way in which ores and coals were mined and transported to
steel plants.

Blast furnaces in the early 20th century were equipped with a

variety of fairly sophisticated top charging and raw material distribu-~

1-41
 '1

",' 8/eeer
J

I
I
I ~
I
I --Down comer
I
i _---
I

ôfock---1-
line i i
~
I
I
I
I
I
I

13
I
~
II
I
I
I
I
I
i
I
I
I
i
I
I
I
i
I
I
i
i
I
I
I
i
1
I
I
I
I
Melfing
zone --
I
.iII I
~i
I
I
I
I
I
Tuyeres --t------
r
1 _ _--~
Slaq - - +----
norch

Figure 20. Late 19th Century Blast Furnace

1-42
'I

i
I Stock
Line,

I
I
I
I ,'.. //.." ,.',
.,i--£;;sfèr--.,
Oar i
\ '" \. )
"ÍÍ ---"d ,-'
!r ----i
TJIi
T T
l.i 1
I
r

Figure 21. Early 20th Century Blast Furnace

1-43
 tion devices. The double bell and hopper arrangement that later became
the industry standard had been introduced but had not yet become the ~j

most popular top charging and sealing device. A particularly interest-

ing historical note for the modern ironmaker is that blast furnace in-
strumentation had begun by 1903. The temperature and pressure of gases J
entering and leaving the newest furnaces of this era were being moni-
tored, and J.E. Johnson, Jr. had developed .an automatic stockline
recorder that greatly improved the furnaceman' s knowledge of his opera-
tion. These simple measuring devices were the start of bIas t furnace
J
control developments that have continued through to today.

The production of the largest blast furnaces in the early 20th J

century had increased from 60 tons per day in 1870 to as much as 500
tons per day. The coke required to produce one ton of iron ranged from
1750 to 2100 pounds and, as today, depended a great deal on the type of ~
ore and the blast temperature.

The blast furnace developments described to this point have

centered on the furnace structure and the amount of iron produced. ~
The next section will show how the ironmaker' s understanding of the
physical and chemical phenomena occurring in ironmaking smel ters has
evolved. This evolution started with the first ironmaking smelting I
hole and continues today. But, because of its importance in modern
blast furnace operations, the portion of this evolution that has
occurred in the past 100 years will be emphasized. I
DEVELOPMENT OF BLAST FURNACE FUNDAMNTALS
I
Early Scientists
One of the earliest researchers of the chemical and physical I
phenomena occurring in a blast furnace was Charles Schinz. Schinz
studied "the art of measuring heat and applying it rationally in the
various branches of industry" in his native Germany, and was struck with
the lack of understanding of blast furnace phenomena that existed in the
I
mid-19th century. Because his early studies of "heat" convinced him
that many accepted theories of blast furnace operations were incorrect,
he embarked on an extensive study of the blast furnace. The results of I
his work were compiled in a book that was published in 1868. Schinz
attempted to make quantitative mass and energy blances of blast furnace If
operations but was severely limited by the lack of accurate thermodynamic
data. He conducted laboratory experiments to determine heat capacity and
heats of formation and apparently was the first to determine the reduci-
bili ty of iron ore in the laboratory. More importantly, Schinz defined
different zones of the blast furnace and major chemical reactions taking
place in each zone. In comparison to the present understanding of blast
furnace phenomena, Schinz' s theories were incomplete and in some cases
inaccurate. However, he was one of the first to attempt to change the
art of ironmaking into the science of ironmaking.

Many of the principles recognized today by ironmakers were first

postulated by Sir Lothian Bell, a well educated scientist and an

1-44
I

ironmaker entrepreneur who worked diiring the midd:L~ and late 19th
I century in England. His book, "Chemical Phenomena of Iron Smelting",
published in 1872, is recognized as the first text on blast furnace
ironmaking. While Bell had many ironmaking "firsts" during his career,
I only a few of the more important ones will be mentioned here. In 1884,
he was apparently the first to document the function of different con-
sti tuentsinblast furnace slags and note that the melting temperature
a range of
I of the slags was important. He also observed that there was

slag compositions which resulted in good fluid properties and good de-
sulphurizing capability and that blast furnace slags were complex
structures.
iI
Probably the most important of Bell's many contributions was his
understanding of the chemical reactions in the blast furnace. He deter-
I mined that certain concentrations of CO and CO2 could be oxidizing or
reducing to iron or iron oxide depending on the temperature of the
system. He made these observations under carefully controlled condi-
I tions in the laboratory and was therefore apparently the first to start
defining equilibrium in the Fe-O-C system. Bell also recognized that:
"a considerable excess of carbonic oxide (CO) is indispensable for the
reduction of the oxides of iron", and felt that, ideally, the best that
I could be achieved at the top of a blast furnace was a CO :C02 ratio of 2.
But, in addition to being an experimentalist, Bell was a practicing
ironmaker and felt that blast furnaces would not be able to achieve the
I "minimum" CO:C02 ratio because of the need for a "little margin" to
allow for upsets in the furnace. It will be shown later that Co: C02
ratios of 1 and lower are achieved frequently in modern blast furnace
operations.
I
Bell also recognized that the stack ofa blast furnace was impor-
tant for the preheating and pre-reduction of ores prior to entry into
I the higher temperature zones of the furnace. He had observed black
slags and off-grade iron quality as a result of poorly prepared iron
oxides dropping into the bosh and hearth of his furnaces, and he
I related these experiences to his laboratory work. These observations
and experiences led Bell to a concern for the influence of furnace
height on iron production and fuel requirements as shown by the consid-

r erable space devoted to this subject in his second book. It was Bell
who first stated that there is an optimum height for each furnace: a
shorter furnace would not properly pre-reduce and prepare the ore, and
a taller furnace would be a waste of capi tal resources .

Sir Lothian Bell made carbon, oxygen and nitrogen balances of his
blast furnace operations and showed that some of the charged carbon
was consumed in the stack by carbon dioxide. A legacy left by Bell to
modern ironmakers is the use of the term "solution loss" to designate
the carbon consumed by carbon dioxide in the blast furnace stack. Bell
and other earlier ironmaking theoreticians did not fully understand
the role of this blast furnace reaction and felt that they could achieve
the ideal blast furnace operation only when this reaction was eliminated.

Another well-known late 19th century scientist-ironmaker was

M.L. Gruner, a professor of metallurgy in France. Gruner expanded

1-45
 J

Bell's method of determining blast furnace heat balances by comparing

many different furnace operations. Gruner observed large differences 1

in heat requirements among furnace operations and related these differ-

énces to furnace volume and height. The most quoted statement by
Gruner concerns the "ideal working furnace" and what eventually became ,1
known as Gruner i s theorem. The theorem states that the "ideally perfect
working" of a blast furnace will be achieved when "the reduction of iron
ore is made as far as possible by the transformation of CO into CO2,
that is, without any consumption of solid carbon". Gruner, like Bell, J
believed that the minimum fuel rate for bIas t furnaces would be reached
when the "solution loss" was eliminated. Gruner and Bell believed this
because they felt that the solution loss reduced the total amount of J
heat produced in the combustion zone. As will be discussed later this
misconception remained a part of ironmaking art until the middle of
the 20th century.
~
One of the blast furnace mysteries of the Bell and Gruner era was
why hot blast had such a large and dramatic effect on furnace production
and fuel rates. Bell incorrectly explained the effect of hot blast as a
the result of increased residence time of both solids and gases in the
furnace. In making this explanation, he failed to recognize that the
availability of energy above certain temperatures, that is the Second I
Law of Thermodynamics, was important in the process. The Second Law of
Thermodynamics had been stated in 1850 and ironmakers were probably
quite familiar with steam engines and the important role of "steam
quality". However, at this point in time the implications of the
I
Second Law with respect to ironmaking were not yet understood.

The first to apply the Second Law to the blast furnace process was I
J.E. Johnson, Jr., an American ironmaker in the late 19th and 20th
centuries and the author of two well-known and often quoted books on
blast furnaces. As related by Johnson in his second book, he was often
bothered by the explanation offered by Bell for the effect of blast
I
tempera ture on blast furnace production and fuel rates. As a resul t,
Johnson postulated that the fuel rate of blast furnaces was determined
by two thermal equations, these being the First and Second Laws of
I
Thermodynamics. With these principles Johnson was able to explain the
effect of blast temperature on furnace performance, and in so doing he
made a major breakthrough in the Understanding of blast furnace opera- I
tions. This line of reasoning eventually lead him to postulate that
there is a critical furnace temperature above which a minimum amount of
heat is required. This minimum amount of heat he called "hearth heat". r
lL
He used this principle to explain the high fuel rates experienced with
the production of ferromanganese in a blast furnace and to explain the
effect of dry blast, as proposed and practiced by Gayley. Possibly
more important than the specific explanations provided by Johnson's
thermal equations is the fact that the application of his critical
temperature and hearth heat concepts further convinced furnacemen that
their process was rational and, as a result, predictable. The thermal
equations were not Johnson i s only contribution to blast furnace iron-
making. He was a very active engineer and responsible for many equip-
ment innovations made in blast furnace plants during his lifetime.

1-46
I

Gas-Solid Contact
J
During the period 1920 to 1930, the flow of solids and gases in
blast furnaces was studied extensively by a group of workers at the
l u.s. Bureau of Mines. This group, composed of P.H. Royster, S.P. Kinney,
C.C. Furnas and T.L. Joseph, was interested in the physical and chemical
phenomena occurring in blast furnaces, and in,.order to understand these
phenomena they felt it was necessary to sample and probe operating
I furnaces. Their work started with a small experimental furnace at
Minneapolis, and eventually lead to studies in commercial furnaces and
to studies in laboratory cold models. The initial work of this group
J on an experimental blast furnace showed that the flow of gases and
solids was not uniform across any horizontal plane in a blast furnace.
This observation was confirmed in a large commercial furnace in a classic
I work by S.P. Kinney. The most important result of this work was the
group realization that the efficiency of the ironmaking blast furnace
could be significantly increased by improving gas-solid contact in the

I stack of the furnace. Before this time, furnacemen apparently thought

that the ultimate in blast furnace efficiency was represented by a top
gas CO:C02 ratio of 2 as stated by Bell. However, Kinney's work showed
that much lower ratios were reached in certain areas of operating fur-
I naces. This finding was particularly significant in 1929 because
equilibrium in the Fe-O-C system was not well understood until the work
of Darken and Gurry in 1945-46. This observation by Kinney lead to an
I intense interest in raw material and gas distribution in blast furnaces
that has occupied the time of many ironmaking investigators for the past
50 years.
I As a result of their belief in the prospect of improved blast
furnace efficiency, Furnas and Joseph conducted a series of blast
furnace cold model tests in an effort to determine methods of improving
I gas-solid contact. This work resulted in a reasonable understanding of
the furnace charging parameters that affected raw material distribution
in the furnace stack. But, more importantly, Furnas and Joseph realized
I from this work that raw material size was a critical parameter in deter-
mining both raw material and gas distribution in the furnace stack and
that raw material size was therefore important in determining furnace

r efficiency. They observed that large pieces of raw material rolled to

the centre of the furnace after charging and provided a minimum path of
resistance for the gases. Such segregation allowed relatively unused
hot reducing gases to leave the furnace and thus decrease the efficiency
of the operation. They also observed that very small pieces of raw
material restrict gas flow and caused channeling of gases. These obser-
vations led Furnas and Joseph to make thé important suggestions that
iron ore be crushed to a maximum size of two inches and that undersized
materials be agglomerated.

Furnas and Joseph were not the first to recognize the benefits of
closely sized raw materials and furnace performance. Frank Firmstone,
an operator of anthracite blast furnaces during the late 19th century,
reported the benefits of sized ore used in his furnaces during the
period 1882-1886. Firmstone also noted that others before him had
recognized the possibilities of improving furnace performance with

1-47
 J

sized raw materials. 'However, Furnas and Joseph started the era of
prepared blast furnace raw materials when they clearly demonstrated the J
effects of raw material sizing. Fortunately, the technique of iron ore
sin tering had been developed before they started their work and was
available for the agglomeration of undersized ore. J
Along with their concern for the effect of raw material size on
gas-solid contact, Furnas and Joseph were concerned about the effect of
iron ore reducibility on furnace efficiency and about the effect of ore J
size on pressure drop and permeability in the furnace stack. The latter
effect is illustrated in Figure 22, where the resistance to gas flow in
a packed bed is shown as a function of the size of particles in the bed. J
This finding led Furnas and Joseph to speculate that the optimum size
of iron ore in blast furnaces would be a compromise between permeability
and reducibility considerations. They were apparently the first to ~
state this basic conflict in blast furnace technology.

Another major contribution to the understanding of the interaction

of gases and solids in ironmaking blast furnaces is the ability to pre- a
dict the pressure drop in the stack region of the furnace. A quanti ta-
tive expression of pressure loss in a blast furnace is difficult to
derive, because first, the stack is a non-homogeneous packed colum and I
then, lower in the furnace, the flow phenomena are complicated by the
mel ting and trickling of iron and slag. The most accurate expression
for quantifying pressure loss in the stack region of a blast furnace is
an equation developed by Sabri Ergun. This equation was developed in
I
1952 and has been widely used by blast furnace engineers ever since.
Based on his study of fluid dynamics, Ergun speculated that production
of current blast furnaces might increase four-fold with proper sizing of I
raw materials and the use of furnace top pressure. This speculation has
proven to be remarkably accurate, because the largest furnaces in 1952
were producing about 1000 tons of iron per day whereas these same I
furnaces are now capable of producing three to four times as much.

Solution Loss
I
A review of blast furnace fundamentals would not be complete
without putting the "solution loss" reaction and Gruner's theorem in I
proper perspective. As stated earlier, Bell and Gruner believed that
the ideal working blast furnace would be achieved if the solution loss,
that is the oxidation of coke by carbon dioxide: r
LL

CO 2 + C -+ 2 CO

could be eliminated from the furnace. The elimination of solution

loss was a goal of many ironmaking researchers and furnacemen into the
1950s. The realization that solution loss played a beneficial rather
than a detrimental role in the blast furnace was apparently first
recognized in the late 50s. This recognition was made possible by the
complete definition of equilibrium in the Fe-O-C system and the detailed
mass and energy calculations that were being made for the first time in
this period. The simplest statement of the role of solution loss was

1-48
 I

J t
REISTANCE
i TO
GA FLOW
J
I
114 112 3/4 1 1Y4
I PARTICLE SIZE, inches

I Figure 22. Relationship Between Particle Size

and Resistance to Gas Flow

I
I
I
f

,r t
CARBON
RATE

100"0 1004f
INDIRECT SOLUTION LOSS'" DIRECT
REDUCTION REDUCTION
Figure 23. Relationship Among Indirect Reduction, Direct Reduction,
Solution Loss and Carbon Rate in a Blast Furnace

1-49
 J
made by R.L. Stephenson in 1962. Stephenson pointed out that iron
oxide reduction in a blast furnace is a combination of the following J
reactions:
FeO + CO 7 Fe + CO2, i. e. Indirect Reduction J
and FeO + C 7 Fe + CO, i. e. Direct Reduction
The point to note is that indirect reduction followed by solution loss J
is direct reduction. Stephenson pointed out that the two iron oxide
reduction routes shown above have quite di fferent chemical and thermal
requiremen ts. Indirect reduction requires about three moles of CO for J
each mole of FeO reduced because of equilibrium considerations, but
direct reduction requires only one mole of carbon to reduce a mole of
FeO. On the other hand, direct reduction is highly endothermic whereas ~
indirect reduction is only slightly endothermic. Using these considera-
tions to determine carbon rates for all combinations of these two reduc-
tion routes as a function of solution loss results in the plot shown in
Figure 23. This plot first of all shows that the total carbon required a
in a blast furnace is determined by either chemical or thermal require-
ments, whichever is greater. It also shows that some amount of solution
loss actually reduces total carbon requirements for reduction. The I
solution lO$s reaction is thus seen to be a critical balancing reaction
that regenerates reducing gas and cools the hot gases rising from the
combustion zone. Since the reaction is very temperature dependent, it
regenerates reducing gas only at high temperatures and has for the most
I
part stopped by the time temperatures are near l600oF. The chemical and
thermal requirement lines shown in Figure 11 are different for different
blast furnaces and are dependent on blast temperature and iron ore I
reducibili ty, among other variables.

Although it is not known with certainty when this situation was I

completely understood or who first explained it, Stephenson was one of
the earliest and he explained it very well.
I
* * * *

I
To sum up the development of blast furnaces to this point, at the
beginning of the 1960 decade the important principles governing furnace
operations had been discovered and stated. From the time of the late
Middle Ages to the early 20th century, ironmakers had learned mostly
r
by trial and error how to build large blast furnaces and what combina-
tion of operating variaples seemed to maximize performance. The effect
of preheated blast was demonstrated by British ironmakers and explained
by Johnson. The importance of fluid dynamics was shown by the U.s.
Bureau of Mines Group, and as a result, furnacemen knew that closely
sized raw materials dramatically improved furnace performance. As will
be discussed next, many of the furnaces operating both here and abroad
in the early 60s took advantage of these principles. However , it has
been the aggressive Japanese steel industry that has taken full advantage
of this technology in the past 10 years.

1-50
 J.
MODERN BLAST YURNAÇES
I
The best blast furnaces operating in the early 20th century were
producing up to 500 tons per day with a coke consumption of about one
I ton per ton of product. Furnacemen in the U. S. did not believe this
was the ultimate and began designing a blast furnace capable of produc-
ing 1000 tons per day. A special committee of the Blast Furnace and

J Coke Association in the Chicago district was formed to design such a

furnace, and their report was presented in April 1930. The general
arrangement of this furnace is shown in Figure 24. This furnace has
a working volume (stockline to tuyeres) of about 35,000 cubic feet and
J a hearth diameter of 25.3 feet. Many of these furnaces were constructed
in the U.S. and at the start of World War II this type of furnace was
the most modern being used in the world.
I There were many mechanical and structural improvements in the 1000
ton furnaces in comparison wi th the earlier 500 ton furnaces, but as a

I chemical reactor, the larger furnaces were a direct scale-up of the

smaller. Ma terials handling equipment improved markedly in this period,
and furnace construction was much more substantial by 1940. Improved
hot blast stoves had been developed, and large turbo blowers were being
I used with the newest furnaces. Furnace tops had been improved for raw
material charging and for containment of the furnace off-gas. The
McKee double bell hopper arrangemetn was accepted as the best furnace
I gas sealing and raw material charging device by 1940. However, the
blast temperature and types of raw materials used in the 1910 and 1940
models were about the same, and the increased production with the 1940

I version was obtained by blowing twice as much air into a furnace with
about twice the volume.

The next step in blast furnace development was the design and
I construction of the classic 28-foot hearth diameter furnace. Many of
these furnaces were built after World War II in the U. S. and later in
Europe and Japan. These furnaces were originally designed to produce
I 1200 to 1500 tons per day and were an extension of know-how developed
wi th the 1000 ton per day furnaces. The basic 28-foot furnaces have
been modified many times, and today the f~rnaces range from 28 to 31
r
i'
II
feet in hearth diameter with working volume of 50,000 to 55,000 cubic
feet. The soundness of this furnace design is indicated by the fact
that after 30 years they are still the workhorse of the U. S. steel
industry. The producti vi ty of these furnaces has improved three-fold
in this 30 year period of time due mostly to improved iron bearing raw
materials. This raw material improvement is discussed in the next
section.

Raw Material Preparation

The most important development in blast furnace technology and

practice in the' past 25 years has been the use of beneficiated and
sized raw materials. This breakthrough started with the development of
the sintering process and was given a solid basis with the u.s. Bureau
of Mines work in the 20s and 30s. However, more significant for the use_of

1-51
 '1
J

OOJ!!!ER J

a
199'-0.

I
PYROMETER
PLA 'FORM

I
I
r
,;

=-;

Figure 24 . Blast Furnace Designed for 1000 Tons/Day

1-52
J
beneficiated ores in the u. S. was the depletion. of _thE!direct shipping
1 Mesabi ores after World War II. This led to the development and use of
the pelletizing process, which in turn marked the beginning of dramatic
improvements in blast furnace productivity and efficiency. These im-
J provements are illustrated in Figure 25 , where changes in the average
U.S. blast furnace coke rate and the use of sinter and pellets are
shown for the period 1957-1966. The improvements in this period were
due to closer sizing of agglomerate materials versus direct ores, lower
~J gangue content and thus lower slag production with beneficiated ores,
and a moderate increase in blast preheat temperature. Another step in

i the development of high productivity blast furnace operations was the

realization of the beneficial effects of very close sizing of sinter,
pellets, and coke on furnace performance. It was also found that there
were important relationships between the size of coke and the size of
ferrous raw materials, and that these relationships were important in
~ the optimization of blast furnace operations. A relatively small
increase in coke size can improve the permeability of the furnace

l burden and thus permt a higher wind rate and production.

While the above information confirmed on an industrial scale the
principles formulated by the Bureau of Mines Group earlier in the
I century, the most dramatic demonstration of these principles was made
in Japan. New steel plants were being constructed in Japan during the
post-war period and these plants were designed to take full advantage
I of raw material preparation and sizing. Large blending and bedding
facili ties for coals and ores were built to produce homogeneous raw
materials for coke plants and sintering operations. Because Japan
does not have indigenous supplies of iron ore and metallurgical coal,
I these new steel plants were constructed for high efficiency and the
capability of handling a wide variety of imported raw materials.
Close sizing of all raw materials charged to the blast furnace was a
I primary objective of these facilities. The results of applying the
latest raw material and ironmaking technology in the new Japanese
steelplants started to be realized in 1965. An example is shown below

I for the Sakai No.1 furnace of Nippon Steel Corporation.

Furnace Sakai No. 1

r Date
Furnace ,Size
July 1966

Hearth Diameter, ft. 32.8

Working Volume, CF 63,100
Production, NT/day 4427
Coke Rate, lb/NT 1022
Blast Temperature, 0 F 1890
Ore Burden, %
sinter 65
Pellets 14
Sized Ore 21

1-53
 ~i

.J

J
80 1800
~

60 1600
a
PECENT AVERAGE u.S.
PELLETS 40
AND
14 COKE RATE,
Ibsl NT
i
SINTER

20 1200 I

I
o 1957 58 59 60 61 62 63 64 65 66 100
YEAR I
Figure 25. Trend of Coke, sinter and Pellet Use in U.s.
Blast Furnaces During 1957-66 Period I

I
r
!~
~-

1-54
 J

This operation was achieved with the following raw materials size
I ranges:
Size Range, inches
J Coke 0.4 - 3.0
sinter O. 4 - 3.0

J Pellets 0.2 - 0.8

Ore 0.4 - 1.0

i Current material preparation practices in some operations in North

America and Japan have reduced the size range of ferrous materials even
further than that shown above. It is standard practice in some plants
to size sinter and ore to the range of 0.25 - 1 ~O inches and screen
I pellets to 0.25 - 0.60 inches. In many operations coke is screened
into two size fractions and charged separately to the furnaces in an
effort to minimize pressure drop and improve efficiency.
~ Another improvement in blast furnace raw material preparation has
been the production of fluxed and superfluxed sinter. Fluxed sinter
not only removes the thermal load of limestone calcination from the
I furnace but also produces a smaller, stronger and narrower size range
of raw material as compared with acid or unfluxed sinter. Thus, the
production and use of fluxed sinter has affected furnace performance
I for thermal, chemical and physical reasons.

Thus far, the discussion of improvements in blast furnace perfor-

I mance during the post war era has been concerned with raw material prep-
aration and gas-solid contact. Another development of great significance
in modern blast furnace operations is the use of blast additives and very
high blast temperatures. Blast addi ti ves include steam, hydro-carbon
I fuels and oxygen. The use of these with high blast temperatures is
called combined blast. The historical development of these aspects is
outlined below. A more detailed discussion of their influence on opera-
I tions is given in Lecture 14 by R. W. Bouman.

fì Combined Blast
li~
It was probably some time in the early 20th century when blast
furnace operators first noticed that their furnaces would not respond
to higher and higher blast temperatures. This phenomenon has been
described in many ways, but it is typically referred to as a "tight",
or hanging, furnace. In the period between 1910 and 1950, most
furnacemen believed blast temperatures higher than l200-1400°F could
not be used, particularly with the Mesabi or "lake" ores. In the
1950s it was found that steam additions to the blast relieved a tight
furnace, thus permitting the use of higher blast temperatures and
higher wind rates.
In 1957 it occurred to a new group at the u.s. Bureau of Mines
that hydrocarbon fuels injected in the blast might improve furnace
operations even better than steam. This group was composed of

1-55
 "i
N . B . Melcher, J.P. Morris, E. J . Ostrowski andP. L . Woolf. They were i

interested in hydrocarbon injection not only as a method of improving J

the flow of gases and solids in the furnace, but also as a method of
subs ti tuting low cost hydrocarbons for expensive metallurgical coke.
J
Soon after initial experiments with natural gas, industrial adoption
of hydrocarbon injection followed quickly in the u.s. By the end o£ 1963,
67 of the 134 blast furnaces operating in the U.S. were equipped for
hydrocarbon inj ection. In addi tion to natural gas and coke oven gas, J
some operations were using fuel oil, and trials with powdered coal were
being conducted. Fuel oil was tested in a low-shaft furnace in Belgi ur
as early as 1958, and European operations quickly adopted this form of J
hydrocarbon inj ection. More recently, Japanese and North American blast
furnace operations have also made extensive use of fuel oil injection.
A recent development that increases the maximum amount of oil that can ~
be injected is the use of oil-water emulsions. The emulsion helps
atomize the oil stream and thus improves burning characteristics.

At the present time, fuel oil is the most common hydrocarbon used ~
for injection into blast furnaces. Except in Russia, natural gas is
not commonly used because of the availability problem and its poorer
coke replacement characteristics compared to other hydrocarbons. Coal I
tar is used in many North American and Japanese operations with results
similar to those obtained with fuel oil. However, coal tar is usually
more di fficul t to handle, and as a result coal tar inj ection rates have I
been lower than oil injection rates.

Oxygen enrichment of blast air has been used in many furnaces

throughout the world. The first industrial trial of oxygen enriched I
blast was carried out by National Steel in 1951. The benefits of
oxygen enrichment are increased furnace production due to increased
fuel burning capability and an ability to use more hydrocarbon tuyere I
inj ectants. There are heat transfer and heat capacity limits to the
amount of oxygen enrichment that can be used in an ironmaking blast
furnace, and these limits would be typically reached in North American
operations with air enriched to about 25% oxygen. However, this limit
I
is not normally reached in practice because of the high cost of oxygen.
The justification for oxygen enrichment is in the need for iron produc-
tion that could not otherwise be obtained, or in the replacement of very I
expensive coke by fuel injection.

Overall, the use of high blast temperature with various types of r

L.
tuyere additives has played an important role in the development of
modern blast furnaces. Combined blast has provided the furnaceman with
a process tool that permits much flexibility in the establishment of a
good operation. This tool and the raw material preparation techniques
discussed previously have been combined by the Japanese steel industry
to produce blast furnace operations that, with a few exceptions, are
unmatched in the world today. As of the middle 1960s, Japan became the
new leader in blast furnace technology.

1-56
I

Large Blast Furnaces

J
Around 1951 Japan literally made the expansion of their steel
industry a national goal and today Japan is clearly the leader in the
I efficiency and size of primary steelplant operations. Japan's success
in blast furnace operations is due to the application of ironmaking
principles established earlier in North America and Europe and to the
J implementation of their own technical and practice developments
resul t of the Japanese developments are operating furnaces with working
. The

volumes 2.5 times larger than the nominal 28 foot hearth diameter

i furnaces that were being constructed in the late 1940s. In addition,

these furnaces are capable of producing more with a unit of working
volume due to an intensification of the process. Improvements in furnace
performance have been achieved by:
~
o Increased use of agglomerated and closely sized raw
materials.
~ o Higher blast temperatures with oil injection and oxygen
enrichment.
I o The application of top pressure.

o Control of burden and gas distribution in the furnace stack.

I The importance of raw material preparation and combined blast have been
discussed previously. The last two items listed above along with the

I construction of large blast furnaces have been more recent developments

and will be discussed below.

I Top Pressure

The pressurization of a chemical reactor is usually an advantage

I because it intensifies the process and reduces the critical size of the
vessel required for a specified output. This is true in the case of the
ironmaking blast furnace because increased pressure increases the resi-

r dence time of gases in the furnace and, as a result, increases gas-solid

contact. In addi tion, from a fluid flow standpoint, increased pressure
will decrease the pressure drop experienced in a packed bed reactor at a
constant mass rate of gas. This is illustrated in Figure 26 from
Furnas' work and can also be demons tra ted with the Ergun Equation. The
use of top pressure in blast furnace operations began in Russia and the
u. s. at about the same time during the 1940s. The initial efforts in
the u. s. were limi ted to 5 - 10 psig by the double bell and hopper charg-
ing equipment. Later charging and sealing equipment developments in
Japan and Europe have led to furnace top pressures as high as 2.5 atmos-
pheres (gage). These equipment developments have included the use of
3 or 4 bell and hopper arrangements, the use of sealing valves with the
normal double bell and hopper, and the use of sealing valves with a
rotating shute inside the furnace for raw materials distribution. The
last mentioned arrangement is called the "bell-less" top and has been
one of the most revolutionary blast furnace developments in modern times.

1-57
 J

a
f SYSTEM PRESSURE. Atm.
PRESSURE 1.0
DROP 1.5
2.0
i
I

I
GAS FLOW, SCFM --
I
Figure 26 . Rela tionship Among Gas Flow, System Pressure I
and Pressure Loss in a Packed Bed

I
r
Li

,-

1-58
I

This developmènt was made in Luxembourg and first used on a commercial

I furnace in Germany. Not only is the bell-less top an important innova-
tion for the use of high furnace pressure, it also has significantly
increased the flexibility of raw materials charging and distribution.
_I

Burden and Gas Distribution

I The great importance of raw materials and gas distribution in the
blast furnace has been appreciated by furnacemen for about 100 years.
Many different types of top charging and materials distribution devices
J were designed and used in the early 20th century, but the usefulness
of these devices for con trolled distribution was limited. The use of
sized raw materials improved the distribution of solids and gases in the
J furnace as discussed earlier. However, as larger blast furnaces were
built, it became apparent that additional measures were needed to control
the movement of solids and gases. The basic problem in the distribution
of raw materials in a blast furnace is the large difference in density
~ and angle of repose between iron ores and coke as shown below:

Bulk Densi ty, lb/CF Angle of Repose, degrees

I Sinter 120-140 32-36
Pellets 130-150 28-32
I Coke 24-28 36-44

These differences cause the ferrous materials and coke to radially

I distribute qui te differently, and since coke provides the least resis-
tance to gas flow, the furnace gas will preferentially flow up through
the thickest part of the coke layers. This phenomenon is accentuated as
the furnace diameter increases, and since increased furnace capacity
I has been achieved mostly by increasing furnace diameter, burden and gas
distribution has received more attention in the last 20 years.

I The first attempts to mechanically alter the distribution of raw

materials inside the furnace were made in Germany in the late 1960s.
This was accomplished by installing movable panels at the throat of the
furnace that could be set at different angles for ores and coke. This
r basic technique with several mechanical variations is being used on most
of the very large furnaces in operation today. The control of burden
and gas distribution has received a large industrial and research effort
in the past few years and has resulted in significant improvements in
furnace performance. The more recent bell-less top has been used on
several 50,000 - 55,000 cubic foot furnaces and is now being installed
on some of the larger furnaces. The result of using a bell-less top on
a large furnace is of considerable interest to furnacemen.

Another consideration that has a very important effect on gas flow

in blast furnaces is coke quality, in particular, coke strength. Fur-
nacemen have said, probably from the time of Abraham Darby, that coke
quality is crucial in the operation of a blast furnace. Sometimes these
statements have been true and sometimes they have been a convenient
alibi for some difficul t-to-explain event. However, it has become clear

1-59
 J
wi th the construction and operation of large blast furnaces that coke
strength requirements increase as the size of the furnace incrèases. A J
comprehensive investigation by Sumi tomo Metal Industries on the effect
of coke strength was recently reported. Interest in coke quality con-
siderations will be increasing in the future because of the limited
supply of good metallurgical coking coals. J
To sum up, the use of prepared burdens, combined blast, top
pressure, and raw materials distribution techniques have all had an impor- ~
tant role in the development of large blast furnaces. These have had the
effect of intensifying the process, that is, producing more iron for a
uni t of internal reactor volume. A commonly used method of expressing
this productivity is to compute the tons of iron produced per day per
I
100 cubic feet of working volume, or NT/day/100 CFWV. When this produc-
tivi ty factor is plotted chronologically for some of the monthly record
blast furnace performances in the past 17 years, a curve as shown in I
Figure 15 is obtained. This shows that in the early 1960s the classic
28- foot furnaces were producing at the rate of 5.5 - 6.5 NT/day /100 CFW
and, more recently, that the high producti vi ty furnaces in Japan have I
reached 8.7 - 8.8 NT/day /100 CFWV. These figures represent a remarkable
improvement.
A simple method of sumarizing the development of the modern blast
I
"i furnace is to review how 20th century furnace performance has evolved.
Furnace operations for different periods during the past 70 years
present just such a sumary: r
1910 1940 1963 1974

Production, NT/day 500 1000 3000 10000

Fuel Rate, lb/NT 2000 1800 1200 950
Blast Temperature,oF 1000 1200 1400 2100
Furnace Size
Hearth Diameter, ft 17 26 29 45
Working Vol lie, CF 25,000 47 , 000 50,000 130,000
Productivi ty
NT/day/lOO CFW 2.0 2.1 6.0 7.8
These are not record performances as shown in Figure 27 but rather
represent what the typically good operations were doing in each period.
The increase in furnace size and the improvement in productivity shown
above are the result of monumental efforts by blast furnace designers
and operators. The results of their efforts must rank with the best of
modern engineering achievements.

1-60
I
1

J
i
J 9 ø-
KIMlTSU 3
I FUKUYAMA 2

8
I PRODUCTIVITY,
NT / DAY /100 CF
WORKING 7 ø SAI 1
I VOLUME

I 6 EACH POINT IS A
ø PORT KEMBLA 4 ONE MONTH
OPERATION
I MIDDLETOWN 3

;961 62 63 64 65 66 õT 68 69 70 71 72 7! 74
r YEAR

Figure 27. Change in Record BIas t Furnace Producti vi ty

Performance During 1961-74 Period

1-61
 MODERN ASPECTS OF BLAST FURNACE THEORY
J
Reduction of Iron Oxides

with few exceptions, the iron-bearing components in the charge to

the furnace are the simple oxides of iron, Fe203 and Fe304. The natural
J
ores usually are hematites (Fe203) or magnetites (Fe304). Pellets are
principally Fe203. Sintered ores can range in composition from Fe203
and Fe304 to fused mixtures containing magnetite, fayalite, 2FeO.Si02' J
and dicalci um ferrite. The reduction of iron oxides generally takes
place in steps. The reactions with carbon monoxide (CO) are:
J
3Fe203(s) + CO(g) 2Fe304(s) + C02(g); ßH -11,537 cal ( 1)

Fe304 (s) + CO(g) 3FeO(s) + C02 (g); Lm -5,200 cal (2) I

FeO ( s) + CO ( g) Fe(s) + C02 (g) ; Lm -2,620 cal ( 3)

These reactions are accomplished at successively higher temperatures, I

and farther down the furnace. As shown in Figure 16, successively
higher percentages of carbon monoxide are required to complete reactions
(1), (2) and (3) by the rising gases. It is to be recognized that it is
not possible for all of the CO in the gases to be converted to C02 for
I
each reaction. For example, there is an equilibrium ratio as given by
the constant for Equation (3) and from Figure 16:
I
K3 = P Co2/P CO
I
at each temperature. At 800°C, the equilibrium gas mixture contains
about 65% CO and 35% CO2. If the C02 con tent exceeds this value in the
gases in contact with FeO and solid iron at this temperature, iron I
present will tend to be oxidized back to FeO. Accordingly, to force
these reactions to occur, there must be a considerable concentration of
CO in the gases at each step as indicated in Figure 28 , and it is not
possible to convert CO completely to CO2 by the reactions.
I
ir
Because of hydrogen in the auxiliary fuels and moisture from the
fuels and the air blast, the gases leaving the tuyeres may also contain
up to 2 or 3% hydrogen. Steam may be added to the blast as an aid in
controlling the furnace. The reduction of steam by carbon in the
coke and fuels proceeds by the overall reaction:

H20(g) + C(s) = CO(g) + H2(g); åH = 31,380 cal ( 4)

This reaction is endothermic whereas the oxidation of carbon by oxygen

in the blast to form carbon monoxide is exothermic:

C(s) + 1/2 0 (g) = CO(g); åH = -26,420 cal (5 )

The reduction of iron oxides by hydrogen also proceeds by steps:

3Fe203(s) + H2(g) = 2Fe304(s) + H20(g); åH = -1,698 cal ( 6.)

1-62
 I

I
I
,)

i
S
i
i
I Sl
~
0 ~
0

I 0u
¡
fj

I sa

I o
20 40 EO eo io 12
Temperature .C
I
Figure 28 . The Fe-C-O System Showing the Fields of
l1 Stabili ty of Iron and Various Iron Oxides

1-63
 Fe304(s) + H2( g) = 3FeO ( s) + H20(g); L1H 15,040 cal ( 7)
,1

FeO(s) + H2 (g) Fe (s) + H20(g); Ll 7,220 cal ( 8)

The effect of temperature on the equilibria of these reactions is shown J

in Figure 29 .
The water gas shift reaction can take place among the various J
species in the gas phase to redistribute the oxygen and bring the
hydrogen-bearing and carbon-bearing gas species into equilibrium:

CO2 (g) + H2 (g) = H20(g) + CO(g); L1H = 9,840 cal ( 9)

J
This reaction requires very little heat and the equilibrium constant
~
(PH 0 . Pc ) / (p PCO ),
2 0 H2 2
is unity at 8250C. I
The gases in the shaft will react with the carbon of the coke as
well as with the oxides of iron in the charge (Eqs. 1, 2 and 3). The
overall reaction of carbon monoxide and carbon dioxide wi th carbon as I
graphite is the "solution loss" or Boudouard reaction:

C02 (g) + C(s) = 2CO(g); L1H 41,200 cal (10) I

The equilibrium of the reaction is shifted strongly to the right at
temperatures above 750oC. Below 6000C the equilibrium is strongly to
the left, resulting in the deposition of carbon as soot in the furnace
I
burden:
2CO(g) =C(s) +C02(g); L1H=-4l,200cal (lOa) I
The "s" shaped curve leading from the lower left to the top center of
Figure 16 represents the equilibrium of Eqs. (10) and (lOa). A gas I
whose temperature and composition place it above the line will tend to
deposit carbon by reaction (lOa), and one whose composition and tempera-
ture place it below the line will oxidize carbon in accordance with r
U
reaction (10). The principal effects of the carbon solution reaction at
high temperatures are a relative reduction of heat generated at the
tuyeres where it is needed and an increase in the concentration of co
in the gases at regions of the furnace above 700oC. This latter condi-
tion is particularly desirable as it increases the volume of the gases
and aids in heat transfer, a point that will be treated in greater detail
later. It is to be noted that the combination of Eq. (10) with Eq. (3)
corresponds to the "direct" reduction of FeO by carbon:

FeO(s) + C(s) = Fe(s) + CO(g); L1H = 31,380 cal (11)

It will be evident from Figure 16 that the gases passing up the

stack cannot generally be in equilibrium with carbon in the coke and
the iron oxides in the descending burden. Measurements of the tempera-
tures and compositions of gases in operating furnaces show that they do

1-64
 J

J
J
i
8 0
~
0 Iron 0
Q) N

I
f6 ~
~ ~
0

I 0
:i
..
0~ £
0..
:i

I 0N 0
Q)

I o
20 40 ti BO 100 1200 1400
._~
Temperature °C
I
Figure 29. The Fe-H-O System Showing the Fields of
Stabili ty of Iron and Various Iron Oxides
lJ

1-65
not follow either set of equilibrium curves, those for the iron oxides ~
or that for carbon. They tend to fall between the CO/C02-C line and
the wustitela-iron line above 8000C, touch the wustite/a-iron line at
between 6500 to 8000C, and then remain at or just above the Fe304/a-iron
line as shown in Figure 30 : J

The actual relationship between gas Gomposition and temperature in

the blast furnace stack will depend to a great extent on the actual J
practice employed. This aspect is illustrated in Figure 31 , which is
taken from E. T. Turkdogan i s Howe Memorial Lecture. The lower curve in
Figure 31 is for regular blast furnaces operating with acid sinter or J
pellet and lump ore, and a high coke rate of about 800 kg/tonne of hot
metal. The upper curve is for high-pressure furnaces operating with
basic sinter, oxygen-enriched high-temperature airblast, and a low coke
rate of less than 400 kg/tonne of hot metal. In older type blast furnace ~
operations, the gas composition is reducing to wusti te at all levels in
the stack. In modern blast furnace operations however6 the gas composi-
tion is oxidizing with respect to iron below about 950 C. I
Fluxes I
Limestone charged to the furnace will calcine by the following
reaction at approximately 8000C (14720F):
I
CaC03(s) = CaO(s) + C02 (g) ; ~H = 42,500 cal ( 12)

Magnesi ur carbonate in dolomitic limestone in the charge calcines I

by a similar reaction at 500 to 1000C (900 to l800F) lower temperatures:

MgC03(s) = MgO(s) + C02(g); ~H = 40,000 cal (13) I

These reactions result in several undesirable conditions in the
furnaces. The first is that they require considerable heat and the
second is that CO2 is released in the furnace. The additional C02 I
raises the oxygen potential of the gases which inhibits the final step
in the reduction of the iron ore, i. e., FeO to Fe. It also favours
"solution" of carbon from the coke by Eq. (10). A significant improve-
ment in furnace operations is obtained when "self-fluxing" agglomerates
r
of iron-ore concentrates are the principal iron-bearing charge to the
furnace. Limestone and dolomite may be added to the feed of sintering
machines and pelletizing furnaces. When the sinter is fired and the
pellets are indurated, the fluxes are calcined and reacted with iron
oxides to form calciur-ferri tes and other more complex compounds. The
CaO and MgO carried into the blast furnace by these agglomerates are
then free of CO2.

Slags
The oxide system that forms the basis for blast furnace slags is
the lime-silica-alurina (CaO-Si02-A1203) system as shown in Figure 32 .
Slags with compositions in the region of 40% Si02, 48% CaO and 12% A1203

1-66 I
J

I 100
dP \ \ Japanese
80 \( German
') N
0u 60
ll
,+ FeO
J 0u 40 ~- i .. ..
ll
'-0 20 Fe 3°4 / J
J. .. .-
-- ---
~ /
ll
U ." ,-
0
1000 1200
I 200 400
Temperature,
600 800
°C

I Figure 30. CO Content og Gas Samples from Operating Furnaces

I
I
I
r N
o
(.
+
ôo
(. U
+ +
00
.. N
x: x:
+
N
x:

0.1

80 80 100 1100 120 130

, ~EMPERATUAE, .c '

Figure 31. Gas Compositions in Blast Furnaces

Wi th Different Operating Practices

1-67
 J

Figure 32. The CaO-Si02-A1203 Phase Diagram

I

1-68
 I

have low melting points, i.e., l3000C (23750F) ,and are appropriate for
1 control of sulphur and silicon in the metal. Often 6 to 10% MgO is used
in place of an equivalent amount of CaO to lower the viscosity of the
slag. Small amounts of MnO, FeO, Na20, K20 etc. help to lower the melt-
J ing point of the slag.
Essentially there are two slags in the furnace. The first is the
"early" slag that is formed principally from the gangue constituents in
') the ores and agglomerates and CaO and MgO from the calcined fluxes, or
the self-fluxing portions of the agglomerates. This slag is relatively
basic compared to the final slag and would contain some iron oxide.
J The "final" slag is formed by the union of the early slag with consti tu-
ents of the coke ash that are freed from the coke when it is burned
before the tuyeres. This final slag continues to have its composition
~ modified as it passes down into the hearth and mingles with liquid iron
that also is flowing down into the crucible. There is an adjustment in
the silica con tent of the slag, iron oxide may be reduced from it and it
I may absorb sulphur from the coke and liquid iron.

I Reactions in the Bosh and Hearth

Sulphur is a troublesome element in blast furnace operations

because hot metal for steelmaking must be Hlowinsulphur ¡ levels of
I 0.035 to 0.02% are usual. The reaction by which sulphur is removed
from liquid iron into the slag is often represented by the reaction:

I S + (CaO) + C = (CaS) + CO(g) ( 14)

where sulphur and carbon in the metal react with lime dissolved in the
slag to form calcium sulphide in the slag and CO gas. The distribution
I of sulphur between slag and metal, (S) /~, is strongly
numer of factors:
influenced by a

~ (a) Increasing the basicity of the slag (lime/silica ratio) tends to

raise the thermodynamic activity of lime in the slag which pushes
reaction (14) to the right.
(b) An increased oxygen potential in the system pushes the reaction to
the left. This is shown by rewriting the reaction as follows:

S + (CaO) (CaS) + 1/2 02(g) (15)

The effect is very strong, and the presence of a small concentration

of FeO in the slag will seriously limit the sulphur ratio, (S) /~.
(c) Fortunately both silicon and carbon in hot metal raise the thermo-
dynamic activity of sulphur in the metal at a given concentration
level. Accordingly, sulphur at 0.02 to 0.035% in ordinary hot
metal for steelmaking is 5 to 7 times easier to remove than it would
be in liquid steel that contains relatively little carbon and
silicon.

1-69
 j
The sulphur distribution ratios found in the blast furnace gener-
ally vary between 20 and 120. On the other hand experiments have shown J
that when metal and slag samples from the blast furnace are remelted in
graphite crucibles at 1 atm CO, the distribution ratio increases to
between 120 and 220, depending on the slag basicity. This suggests that J
the oxygen potential of the system is higher than might be expected for
C-CO equilibrium in the furnace hearth. Thus while thermodynamic condi-
tions favour sulphur removal from hot metal wi thin the blast furnace,
kinetic considerations imply that the reaction can be more readily J
accomplished outside the furnace by external desulphurization. The
implications of this approach are discussed by A .M. Smillie in the
lecture on External Treatment of Hot Metal (Lecture 18). J
For many years it was considered that silica and manganese oxide
were reduced directly from the slag by reaction with carbon in iron ~
according to the reactions:
Si02 (slag) + 2C si + 2CO (g) (16) I
MnO (slag) + C Mn + CO (g) (17)
It was thought that mol ten iron droplets picked up silicon as they I
passed through the slag phase and on into the hearth. Research during
the last decade however, has shed new light on these reactions and also
those involving sulphur. Several laboratory studies together with plant I
data from Japan have shown that at the temperature of the combustion
zone, about 20000C, silicon monoxide gas is produced during the combus-
tion of coke by the reaction:
I
Si02 (coke ash) + CO + SiO(gas) + CO2 (18)
Combining Eq. (18) with the coke oxidation reaction: I
CO2 + C (coke) + 2CO

yields the overall reaction:

I
Si02 (coke ash) + C (coke) + SiO (gas) + CO ( 19)
r
While the presence of FeO in slag is likely to make SiO formation from
slag very difficult, an additional source of silica would be reduced
silica-rich slag adhering to coke particles. Following these reactions,
silicon is transferred to iron droplets by reaction with silicon monoxide
in the gas phase:

SiO(gas) + C + Si + CO ( 20)

As iron droplets containing silicon pass through the slag layer, some of
the silicon is oxidized by iron oxide and manganese oxide, and taken up
by the slag:
2 (FeO) 1 + Si
s ag (Si02) slag + 2Fe ( 21-)

1-70
¡
!

I
2(MnO) 1 + Si = (Si02'..) 1 . + 2Mn
s ag s ag (22)
1
Assuming sulphur in coke ash is present as CaS, the following
reaction can occur with SiO in the combustion zone to form volatile sis:
J CaSco(ke) ash
+ SiO ( )-+
gas CaO + SiS ( )
gas (23)

1 To a lesser extent, some CS gas may form by the reaction:

J
co e as gas
CaS (k h) + CO -+ CaO + CS ( ) (24 )

Sulphur transfer from these volatile species to molten iron droplets

then takes place with the bosh zone. Turkdogan has shown that when iron
droplets containing Si and S are allowed to fall through mol ten slag, in
~ the absence of MnO, the Si content of the metal actually increases, and
there is no transfer of S. In the presence of MnO, Si is removed from

I the metal by reaction( 22) and Mn transfers from slag to metal together
with S transfer from metal to slag. Based on the various results avail-
able, Turkdogan suggests the following sequence of reactions in the bosh
and hearth:
I 1. The formation of sio and SiS in the combustion zone.
2. The transfer of silicon and sulphur to metal and slag
I droplets in the bosh.
3. The oxidation of silicon by FeO and MnO in the slag as the

I iron droplets pass through the slag layer.

4. The desulphurization of metal droplets as they pass through
the slag layer.
I other reactions involving the formation of volatile species are
those associated with so-called rogue elements such as sodium, potassium
and zinc. These elements have adverse effects on furnace operation due
r to refractory attack, generation of fines, accretion formation and
decreased burden permeability. Problems of this type were accentuated
during the 60s as more furnaces began to operate with higher driving
rates, increased flame temperatures, lower slag volumes and relatively
high basicities. During the 70s, our understanding of these phenomena
has been greatly enhanced both by laboratory studies and results from
plant operations. A leading contributor to this field has been W-K. Lu
and co-workers.
The alkali metals, sodium and potassium, generally enter the
furnace with the raw materials in the form of very stable alumino-
silicates. Certain amounts of sodium and potassium leave the furnace
with the top gas and in the slag phase, while the remainder accumulates
in the furnace in the form of cyanides, carbonates and intercalation
compounds in coke, e. g., C6K and C8K. These compounds decompose in the
higher temperature regions of the blast furnace to form metallic and
cyanide vapours, e. g. :
K2Si03 + C -+ 2K(gas) + Si02 + co (25j

1-71
 Li
~I

These vapoursare carried by massive gas flow to lower temperature

regions where condensation reactions occur and compounds are reformed J
to be transported back down the s tack wi th the burden materials, e. g. :
2K(gas) + CO + K20 + C ( 26) J
Decomposition reactions of the type indicated bYEq. (25) are strongly
endothermic and bring cooling to the hearth zone. During condensation
reactions in the cooler regions of the furnace, heat is released. J
Since the alkali metals form basic oxides , they are readily neutral-
ized by the use of acidic slags. Their removal with the slag is there- J
fore enhanced when the slag has a low basicity (Figure 33) and tempera-
tures in the hearth are relatively low. These conditions do not favour
the production of low pulphur hot metal and the behaviour of alkalies
consti tutes another reason, why hot metal should be desulphurized ~
outside the furnace. The reactions of sodium are similar to those of
potassium except that sodium is more difficult to gasify and thus it
can be more easily removed in the slag phase. I
The main source of zinc in the blast furnace is from sinter which
contains dust from steel furnaces in which a high proportion of galvan-
ized scrap has been melted. Following reduction of ZnO, zinc vapour
I
will recirculate through the furnace with the subsequent formation of
ZnO and ZnC03 in the regions of lower temperature and higher oxygen
potential. Unlike the alkali elements, zinc does not form stable I
silicates and cannot be removed in the slag phase.

I
Energy Considerations

The counter-flow of gases and solids in the shaft of the blast

furnace provides for highly efficient use of the heat and reducing
I
power of the gases leaving the tuyere region. Heat transfer from gas
to solids is accompanied by oxygen transfer from solids to gas. It will
be recalled that the strongest reducing gases are employed first in I
carrying out the most difficult reduction step, that of converting FeO
to iron. Similarly, the gases when hottest complete the highest temper-
ature work, that of melting and superheating the slag and metal, and
providing the heat required for reactions in the bosh and hearth zones.
r
There is a gradual transfer of heat from the gases ascending the furnace ~~

to the solids that are descending. At steady state operations, the tem-
perature profiles of gases and solids do not alter their positions in
the furnace. The temperature of the gas, T, is always higher than that
of the solids, 8, and the difference (T-8) is the driving force for the
transfer of heat from the gases to the solids.

Application of the concept of thermal flows in heat and mass

exchange in a counter-current, gas-solids reactor to the operation of a
blast furnace can be useful and illuminating. The thermal flows of
gases and solids may be represented in terms of the products of their
respect thermal capacities (G and S) and velocities (U and V). Thermal
flow for the gas phase is given by the product I UG I while that for the

1-72
 J

J
~I

i
i 3 A

i
2 D

I SLAG
K10
..
I 0 c
c Cc

D D I If c

I D 0
11'1 i
c
cc c
c ::
c:
0.80
.90 1.0 1.1 1.2

I BASICITY

,- Figure 33. Relatiqnship Between K20 Content of Blast Furnace Slag

and the Slag Basicity Ratio (CaO + MgO/Si02 + A1203)

1-73
 J

o J

J
l 400 r
Thermal capaity
;" of or, coe ash
ë.. and ftuaCi 'J
~
t' 800

J
1200
~
o SOO lO 15 200
Thrmal capacity keal/hr .C
I
Figure 34. Contributions to the Effective Thermal Capacity of
Solids in the Stack of a Blast Furnace
I
I
Stock Il£V£1
i I
I
i I I
¡ :£
i :
!i I
I ¡ ii
¡; -ï.- ,- ,- ,
oi
"ë;
.c
!I i
I
CD
I
uei .. I
o
..c
X
iI I r
~ ¡i
u.

,: I \ I i I
I ! r
r

---'--f--- ,-- ,-_.. --

I 1'£ ¡ ',I I
Tuyër~ ,¡ : "' ì i I~,
ICVI£Î: i ! I ""i
o 400 800 1200 160 200
Tcmp~raturc, °C

Figure -J§. Temperature Profiles in a Blast Furnace Stack

1 - Temperature of Solids 2 - Temperature of Gas
A - Direct Reduction B - Indirect Reduction

1-74 I
 l
solids is Ivs I. The products may be expressed in terms of J/S. m2 .0F.
1
The heating of the charge in the blast furnace can be expected to
be most uniform if IUGI ~ Ivs I. However, both UG and VS will change as
~J the streams pass through the shaft. Generally the heat capacities of
many substances, particularly coke and gas species increase with increas-
ing temperature (Figure 34,). Accordingly,_both G and S tend to increase
J with temperature. Loss of combined water, drying of solids, and calcin-
ation of limestone and dolomite will all increase UG and decrease VS.
These reactions also absorb heat which in effect results in an increase

i in the value of VS.

The net effect of the changing thermal flows of the gases and
solids on the temperature profile in a furnace is shown in Figure 35 .
~ In the region Hi, I UG I ~ I vs I and the temperature profile is pushed
high in the shaft. The region H2 is present in most furnaces where
I UG I ~ I vs I. This region is often termed the thermal reserve zone or
I thermal pinch point, because there is little heat transfer and the tem-
peratures of the two streams change very little. Depending on the type
of burden and the blast furnace practice, the temperature level for the
thermal reserve zone varies from about 8500 to 10500C, and the length
I of this zone varies from about 1 to 4 m. with burdens containing
hydra ted ore and carbonates, addi tional thermal pinch points may occur
at lower temperatures where the hydrates and carbonates dissociate.
I The region H3 is in the bosh just above the tuyere level and the thermal
flow of the gases tends to be lower than that of the solids principally
because of the heat required to melt the slag and iron and because of
I reduction reactions in this location.

If operating conditions are established to give a good balance of

thermal flows, the descent of the solids is uniform and the permeability
I of the solids is also uniform, it is to be expected that the solids will
be heated reI a ti vely uni formly as they pass down the shaft. If, however,
the gas channels locally through the bed because of non-uniform packing,
r there will tend to be very hot colums of gas reaching very high in the
furnace. On the other hand, portions of the bed will be starved of gas
which will result in cold, partially reduced material arriving in the
tuyere region. Such a condition leads to serious operating problems.

It will also be evident that it is not possible to replace all of

the air in the blast with oxygen. with the substitution, the amount of
ni trogen in the gas stream decreases and the thermal flow of the gas
stream decreases dramatically. As a consequence, the.heating of the
materials in the stack is impaired. steps have been taken to reduce
the thermal flow of solids per unit of production of pig iron. By
increasing the blast temperature, it is necessary to burn less coke.
Thus there is less coke in the solids and the value of VS is decreased
so that the value of UG may be decreased somewhat by a replacement of a
small amount of the air blast by oxygen. The use of self-fluxing
sintering also allows a reduction in UG because of the elimination of
the need for the supply of heat in the furnace for calcination of the
fluxes.

1-75
 This concept of thermal flows in heat and mass exchange helps to
illustrate why a blast furnace operates so successfully on raw materials J
that are uniform in size and composition. As mentioned in a previous
section, because of the differences in the physical characteristics of
pellets, sinter and coke, the placing of materials in the furnace to
obtain a uniformly permeable bed is extremely important. Similarly, it
J
is essential that the materials in the burden retain their size and
shape and do not degenerate into fines as they pass through the furnace.
J
In concluding this discussion on energy aspects, it it worth noting,
as pointed out by W-K. Lu, that the major difference between Japanese
and North American blast furnace practice, may be characterized in terms J
of the temperature and silicon content of the hot metal. In North
America an increase in hot metal temperature is usually associated with
an increase in silicon level and vice versa. At the present time hot
metal silicon contents are about 0.8% or higher. In Japanese plants g
however, silicon concentrations are 0.4% or less. In spite of these low
silicon levels, the hot metal temperature is about 500 to 1000C higher
than in North America. From the standpoint of melting scrap in the BOF, I
Lu has indicated that the beneficial effect of 0.1% silicon in hot metal
is equivalent to raising the hot metal temperature by 12. 30C. In the
process of using the chemical heat provided by silicon oxidation, oxygen
must be supplied and a basic slag formed in the converter. These
I
requirements are decreased when low-silicon, high-temperature hot metal
is used. From an overall energy conservation standpoint, it would appear
likely that increasing use will be made of this approach in the 80s. To I
a~complish this objective, however, further improvements will be required
in the quantity, reducibility and high temperature characteristics of
the burden materials. I
CONCLUDING REMARKS I
The modern blast furnace operating with a low coke rate is an
efficient processing unit primarily because of the intrinsic character-
istics of a counter-current gas-solids reactor. A successful use of I
this concept requires that each of the materials charged to the furnace
be of uniform physical character, and have a uniform composition. In r
addi tion, each material must retain this good physical character as it
passes down through the furnace to where melting occurs. It is important
to note that much of the improvement in furnace operations that has been
achieved in recent years has resulted from improvements in the physical
and chemical characteristics of the materials charged to the furnace and
in procedures for distributing the charge wi thin the furnace. Other
crucial developments have been the use of high-temperature blast, tuyere
injection processes, high-top pressure and external desulphurization of
hot metal.

During the 60s and 70s substantial progress has been made in our
understanding of the physical and chemical aspects of blast furnace
ironmaking. This has been accomplished by an appropriate blending of
laboratory experiments, plant trials and production experience. Signif-
icant advances have been achieved in the areas of burden reducibility,

1~6
 J

1 fluxed charge materials, coke properties, slag-metal reactions, alkali

behaviour, and heat and mass transfer aspects. A schematic representa-
tion of current thinking on the behaviour of materials within the blast
furnace is shown in Figure 36. A detailed discussion of the reactions
:1 which take place wi thin the various zones indicated on this diagram, is
given in the lecture on Blast Furnace Reactions (Lecture #3) by
C.M. Sciulli.
cl

i
~
LUMPY ZONE

I
I SOFTENING
FROT

MELTING
FRONT

I · ~~:tETS
o SLAG
DRPLETS

I COKE
SLIT

I ¡SOFTENING/
MELTING ZONE

RACEWAY

Figure 3.6. Schematic Representation of Reaction

Zones in a Modern Blast Furnace

1-77
 J

SOURCES OF ADDITIONAL INFORMTION

J
1. Aitchinson, L., A History of Metals, Vol. 1, Interscience
Publishers, Inc., 1960.
J
2. Tylecote, R.F., "Roman Shaft Furnaces in Norfold", JISI, Vol. 200,
January 1962, p.19.
J
3. Maddin, R., "Early Iron Metallurgy in the Near East", Transactions
ISIJ, Vol. 15, 1975, p.59.

4. Matsushita, Y., "Restoration of the Tatara Ironmaking Process, an ~

Ancient Ironmaking Process of Japan", Supplemental Transactions
ISIJ, Vol. 11, 1971, p..2l2.
I
5. Morton, G.R. and W.A. Smith, "The Bradley Ironworks of John
Wilkinson", JISI, Vol. 204, July 1966, p.66l.

6. Morton, G. R., "The Furnace at Duddon Bridge", JISI, Vol. 200,

I
June 1962, p. 444.
7. Murray, D., "Who Invented the Hot Blast?", Steel Times, April 1965, I
p.597.
8. Firms tone , F., "Development in the Size and Shape of Blast Furnaces I
in the Lehigh Valley, as Shown by the Glendon Iron Works",
Transactions AlME, Vol. XL, 1909, p.459.

9. Gramer, F. L., "A Decade in American Blast-Furnace Practice",

I
Transactions AlME, Vol. XXXV, 1905, p.124.

10. Birkinbine, J., "The United States Iron Industry from 1871 to 1910", I
Transactions AlME, Vol. XLII, 1912, p.222.

11. Johnson, J.E., Jr., "An Automatic Stock-Line Recorder for Iron
Blast-Furnaces", Transactions AlME, Vol. XXXVI, 1906, p.79.
r
12. Bell, I.L., Principles of the Manufacture of Iron and Steel,
George Routledge & Sons, 1884.

13. Howe, H.M., "Biographical Notice of Sir Lothian Bell, Baronet",

Transactions AlME, Vol. XXXVI, 1906, p. 412.

14. Bell, I. L., Chemical Phenomena of Iron Smelting, Iron and Steel
Institute, 1872.

1-78
J

15. Gruner, M.L., Studies of Blast Furnace Phenomena, translated by

1 L.D.B. Gordon, Henry Carey Baird, Publisher, 1874.

16. Johnson, J.E., Jr., Blast-Furnace Construction in America,

i McGraw-Hill Book Company, 1917.
17. Johnson, J.E., Jr., The Principles, Operation and Products of the
i Blast Furnace, McGraw-Hill Book Company, 1918.

18. Gayley, J., "The Application of Dry-Air Blast to the Manufacture

i of Iron", Transactions AlME, Vol. XXXV, 1905, p.746.

19. Royster, P.H., T.L. Joseph and S.P. Kinney, "(a) Reduction of
Iron Ore in the Blast Furnace; (b) Significance of Hearth Tempera-
I tures; (c) Heat Balance of the Bureau of Mines Experimental Furnace;
(d) Time Element in Iron Ore Reduction, (e) Influence of Ore Size
on Reduction", Blast Furnace and Steel Plant, VoL. 12, 1924,

I pp. 35-38; 97-101; 200-204; 246-250; 274-280.

20. Kinney, S.P., "The Blast-Furnace Stock Column", u.S. Bureau of

Mines Technical Paper 442, 1929.
I 21. Furnas, C.C. and T.L. Joseph, "Stock Distribution and Gas-Solid
Contact in the Blast Furnace", U.S. Bureau of Mines Technical
I Paper 476, 1930.

22. Darken, L.S. and R. Gurry, "The System Iron-Oxygen", Journal

I American Chemical Society, Vol. 67, 1945, p.1398 and Vol. 68,1946,
p.798.
23. Ergun, S., "Pressure Drop in Blast Furnace and in Cupola",
I Industrial and Engineering Chemistry, VoL. 45, No.2, 1953, p.477.

24. Stephenson, R.L., "Improved Productivity and Fuel Economy Through

r Analysis of Blast-Furnace Process", Iron and Steel Engineer, 1962,
p. 601.
25. Sweetser, R. H., Blast Furnace Practice, McGraw-Hill Book Company,
1938, p.ll.

26. Strassburger, J.H. (Editor), Blast Furnace Theory and Practice,

Vol. 1, Gordon and Breach Science Publishers, 1969.

27. Melcher, N.B. et aI., "Use of Natural Gas in an Experimental Blast

Furnace", U.S. Bureau of Mines Report of Investigation 5261, 1960.

28. British Iron and Steel Institute Special Report 72, Part I -
Injection Processes, 1962, pp.1-7l.

29. Ashton, J.D. and J.E.R. Holditch, "Homogenized Oil Injection at

DOFASCO", AlME Ironmaking Proceedings, VoL. 34, 1975, p.261.

1-79
l

30. Strassburger, J.E., et al., "Solid Fuel Injection of the Hanna

Furnace Corporation", AlME Blast Furnace, Coke Oven and Raw Materials J
Conference Proceedings, 1962, p. 157.
31. Bell, S.A., J.L. Pugh and B.J. Snyder, "Coal Injection - Bellefonte
Furnace", AlME Ironmaking Proceedings, VoL. 26, 1967, p. 180. J

32. Strassburger, J .H., "Blast Furnace Oxygen Operations", AISI

Yearbook, 1956. J
33. Higuchi, M., et al., "High Top Pressure Operation of Blast Furnaces
at Nippon Kokan K.K.", Journal of the Iron and Steel Institute, J
September, 1973, p.605.
34. Furnas, C.C., "Flow of Gases Through Beds of Solids", u.S. Bureau
~
of Mines Bulletin 307, 1929.

35. Legille, E. and K.H. Peters, "Operation of a Blast Furnace

Incorporating a Paul Wurth Bell-Less Top Charging System and its I
Application to Large Blast Furnaces", AlME Ironmaking proceedings,
Vol. 32, 1973, p.144.

36. Hatano, M. and M. Fukuda, "The Effect of Coke Properties on the

I
Blast Furnace Operation", AlME Ironmaking Proceedings, VoL. 35,
1976, p.2.
I
37. Elliott, J.F., M. Gleiser and V. Ramakrishna, Thermochemistry for
Steelmaking, Vol. II, Addison-Wesley Press, N. Reading, Mass., 1963.

38. Levin, E.M., C.R. Robbins and H.F. McMurdie, Phase Diagrams for I
Ceramists, The American Ceramics Society, Inc., Columus, Ohio,
1964, p.2l9.
I
39. Turkdogan, E.T., "Blast Furnace Reactions", Met Trans. B, AlME,
Vol. 9B, No.2, 1978, p.163.

40. Lu, W-K., and J.E. Holditch, "Alkali Control in the Blast Furnace: I
Theory and Practice", Blast Furnace Conference Proceedings, ArIes,
France, June, 1980.
r
41. Kitaev, B.I., Yu.G. Yaroshenko and V.D. Suchkovi, Heat Exchange in
Shaft Furnaces, Pergamon Press, London, 1967.

42. Elliott, J.F., and J.C. Humert, "Heat Transfer from a Gas Stream to
Granular Solids - An Idealized Analysis", Proceedings, Blast Furnace,
Coke Oven and Raw Materials Committee, AlME, Vol. 20, 1961, p.130.

43. Elliott, J.F., "Some Problems in Macroscopic Transport", Trans.

Met. Soc. AlME, Vol. 227,1963, p.802.

44. Elliott, J.F., R.A. Buchanan and J.B. Wagstaff, "physical Conditions
in the Combustion and Smelting Zones of a Blast Furnace, Trans.
AlME, Vol. 194, 1952, p. 1168.
45. Lu, W-K., "Silicon in the Blast Furnace and Basic Oxygen Furnace", -
Iron and Steelmaker, VoL. 6, No. 12, 1979, P .19.
1-80
J
C)

BIBLIOGRAPHY

1. Bartholomew, Craig L. and Metz, Lance E., The Anthra-

J cite Iron Industry of Lehiqh Valley, Center for Canal
History and Technology, Easton, PA., 1988.
J 2. Beard, Directory & History of Marquette Country -
Early History of Lake Superior - Mines and Furnaces,
Detroi t, MI., 1873.
J 3. Benison, Saul, Railroads. Land and Iron: A Phase in
the Career of
Lewis Henry Morgan, University Microfilms Internation-
~ al, Ann Harbor, MI., 1954.

4. Blast Furnaces, Article from Marquette County Histori-

I cal Society.

5. Boyer, Kenyon, Historical Highliqhts, Radio Manu-

I script from Marquette Country Historical society,
Marquette, Michigan, 1958.
6. Claney, Thomas, "Charcoal Humor". Michiqan Historical
I Maqazine, Volume 5, 1921, Page 410.

7. The Daily Mininq Journal, Marquette, Michigan,

I 1890-1900.
8. Data on Furnace of Jackson Iron Company, Article from
I Marquette County Historical Society.
9. Directory to the Iron and Steel Works of the united
States, The American Iron and Steel Association,
r Philadelphia, PA., 1884.

10. The Iron and Steel Works of the united States, Ameri-
can Iron and Steel Institute, 1880.

11. Johnson, J . E., The Principles. Operation and Products

of the Blast Furnaces, McGraw-Hill Book Company, New
York, 1918.

12. King, C.D., Seventy-Five Years of Proqress in Iron

and Steel, AIME, New York, 1948.
13. Lake Superior Journal, Marquette, Michigan,
1855-1865.
14. Lake Superior Mining Institute, Published by the
Institute, Printed by
D. Thorp, Lansing, Michigan.

1-81
 ~
J
15. The Lake Superior Mininq and Manufacturinq News,
Negaunee, Michigan, 1867-1868.
16. Letter to William Mather from CVR Townsend, Marquette J
County Historical Society.
17. The Mininq Journal, 1869-1891. J
18. Rist, Donald E. , Iron Furnaces of the Hanqinq Rock
Iron Reqion, Hanging Rock Press, Ashland, Kentucky, J
1974.
19. Schallenberg, Richard H., Innovation in the American
Charcoal Industry 1830-1930, 1970. ~

20. Strassburger, Julius, H., Blast Furnace - Theory and

Practice, Gordon and
York, 1969.
Breach Science Publishers, New I
21. Swank, James M., The Manufacture of Iron in All Aqes,
The American Iron and Steel Association, Philadel- I
phia, PA., 1892.

22. Weale's Rudimentary Series, Metallurqy of Iron, Cros- I

by Lockwood & Company, London.

r
~-

1-82
I

l LECTURE #2

J Blast Furnace Slag

J J. L. Blattner
Principal Research Engineer
Primar Process Research
~ AK Steel Corporation
Middletown, Ohio

I
I
INTRODUCTION
I The importance of blast furnace slag in achieving good furnace operation is illustrated
by the old saying that goes something like "If you take care of the slag, the furnace will
I take care of the rest". There has been a tremendous amount of work done studying the
properties, formation mechansms, and impacts on furnace operations of blast fuace
slag. The purose of this paper is to sumarize the concepts of this previous work to
I answer the following questions on blast furnace slag:
1. What is it;
2. Why do I care;
r 3. How do I manage it; and
4. What do I do with it when I'm done with it.

The fudamentals ofblast furnace slag are complex. At approximately 40 weight

percent, oxygen is the largest single element in slag. Slag is, therefore, a oxide system
and ionic in nature. Due to the_nature of the blast furnace process, slag formation is a
multi-step process involving signficant changes in composition and temperature.
Slag's four primar components form numerous compounds which result in a wide
range of chemical and physical properties. The lesser components of slag are of
particular interest with respect for hot metal chemistr and fuace control, and add to
the complexity of the physicochemical properties of slag.

Fortunately, there are general relationships which provide a more practical view of the
nature of slags which can be used on a daily basis. It is important, however, to have a
basic understanding of the fundamental nature of blast fuace slag to understand these
general relationships.

2-1
i'

.J

'1

SLAG FUNDAMENTALS J
The following is a brief discussion of some fundamental issues of the blast fuace
process and blast furnace slag. The issues include the slag formation, flow in the
J
hearh, the molecular strcture of slag and how the strctue relates the chemical indices
known as basicity, slag solidification, and the impact of changes of the thermal state of
the fuace on slag composition.
J

Slag Formation J
The iron blast fuace is a pressurized, counter-current heat exchanging, refluxing, gas-
solid-liquid, packed bed reactor. The iron blast furnace has 3 primar fuctions:
Fe oxides to metallic Fe;
1. Reduction of ~
the metallic Fe and oxides; which provides for the
2. Fusion of

the impurities ofthe burden and fuel from the molten Fe.
3. Separation of

These characteristics of the process lead to the division of the furnace into 3 vertical I
zones with respect to slags; Granular, Slag-Formation, and Hearh Zones. These zones
and some specific reactions for each zone are given in Figures 1 and 2.
I
The granular zone is located in the upper part of the fuace where all charged
components are in solid phases. The granular zone is bounded by the stockline on the
top and by the start of the formation of liquid phases, the cohesive zone, on the bottom. I
As the burden descends through the granular zone it is heated by gases from the lower
the iron oxides is performed. The
part of the furnace and a portion ofthe reduction of

amount of reduction that occurs in the granular zone is a fuction of the nature of the I
iron bearing materials, burden distrbution, and the gas composition and flow patterns.

The slag-formation zone begins at the cohesive zone, where softening of burden begins,
I
and continues down to below the tuyere elevation. The slag-formation zone thus
includes the cohesive zone, active coke zone, deadman, and raceway. The slag formed
in the upper par of the slag-formation zone is call the 'Bosh' or 'Primar' slag, and the
I
slag leaving the zone at the bottom is the 'Hearth' slag. The Primary slag is generally
assumed to be made up of all burden slag components including the iron oxides not r
reduced in the granular zone, but does not include the ash from the coke or inj ected
coaL. The slag composition changes as it descends in the furnace due to the absorption r~-

ofthe coke and coal ash, sulfur and silicon from the gas, and the reduction of the iron
oxide. The temperature of the slag increases of the order of500 °C (1,000 OF) as it
descends to the tuyere elevation. These changes in composition and temperature can
signficantly impact the physical properties of the slag, specifically the liquidus
temperature and the viscosity.

The third zone is the slag layer in the hearh of the fuace. The slag produced in slag-
formation zone collects in the slag layer, filling the voids in the hearh coke and
'floating' on the hot metallayer. The hot metal passes through the slag layer to reach
the hot metal layer. The high surface area between the hot metal and slag as the hot
metal passes through the slag layer enhances the kinetics of the chemical reactions. -
These reactions result in signficant changes in the hot metal chemistry. In paricular

2-2
I

I the (Si) and (S) contents prior to entering the slag layer are much higher than the those
in the hot metal layer.

I The formation of slags in the slag-formation zone is very furnace specific due to the
impact of burden properties and fuace operation, and is not discussed further in this
the hearth
) paper. The remainder of

slag.
this paper is directed primarly at the properties of

J Slag Flow In The Hearth

The control of the slag level in the hearth is important for maintaining stable fuace
operation, especially as the hot metal production rates have been increased. High slag
~ levels result in increasing blast pressure and bosh wall working, and disrupting the
uniform descent of the burden.
I One of the issues in controlling slag level is slag flow in the hearh during casting. In
the hearh, slag flow to the taphole is more diffcult than the flow of hot metal to the
I taphole. Hot metal flow has a larger driving force due the higher density of

compared to slag. The hot metal flow path is thought to be primarily through 'coke
hot metal

free' regions below and/or around the deadman coke. The slag flow path to the taphole
I is through deadman coke.

Figure 3 is an illustration of the hearth and a possible sequence of

the configuration of

I stages of the hearth durng casting that lead to a false dr-hearth condition at the end of
the cast. The surface of the hot metal is thought to remain relatively flat across the
hot metal and the 'coke
I
entire hearh area throughout the cast due to the high density of

free' path to the taphole. The slag surface maybe signficantly lower in the region about
the taphole than at other regions ofthe hearth. When the slag cast rate is greater than
the slag flow rate across the hearth to the taphole region, a depletion of slag occurs in
r the taphole region and the slag surface begins to curve down towards the taphole, Step 4
of Figure 3. The slag depletion continues until there is no slag at the taph6le and the
furnace appears to be dry when there is still signficant slag remaining in the hearth,
Step 5 of Figure 3.

Minimizing the resistance to slag flow in the hearth minimizes the slag remaining in the
hearth at the end of a cast. Resistance to slag flow in the hearth is reduced as the
porosity of the hearth coke bed is increased and the slag viscosity is reduced.

Slag Structure
The conceptualization of slag structure is based upon the structure formed by silica,
Si021. On the molecular level, the silicon atom is located in the center of a tetrahedron
surounded by 4 oxygen atoms, one oxygen atom at each comer of the tetrahedron as
illustrated in Figue 4. Each oxygen atom is bonded to two silicon atoms, thus each
oxygen is a comer of two tetrahedrons. The sharing of oxygen atoms results in a
polymer or network in three dimensions in the crystalline state where all comers are

2-3
 J

shared, Figure 5. As silica is heated, some of the comer bonds are broken but the J
polymer nature of the structue is maintained even when molten as illustrated in Figue
6.
J
The addition of metallic oxides, such as CaO and MgO breaks down the polymer
strcture. These oxides act as oxygen donors, replacing an oxygen atom in one comer
of a tetrahedron and breakng the tetrahedron-to-tetrahedron comer bond, Figure 7. The J
breakdown of the polymer structure continues with the addition of more metal oxides
until the molar ratio of metal oxides to silica equals two, at which point all tetrahedron-
to-tetrahedron comer bonds are broken, Figure 8. The molar ratio of2 is the J
orthosilicate composition, 2CaO-SiOi, 2MgO-SiOi, and CaO-MgO-SiOi. Ah03 acts
in a similar fashion as SiOi in forming polymers and accepting oxygen atoms from
basic oxides. ~

Oxides that accept oxygen, SiOi and Ah03, are termed acid oxides. Oxides that donate
oxygen, CaO and MgO are termed basic oxides. I

Slag Basicity I
It is very useful when relating the properties of a multi-component system to its
composition to develop an index based upon the composition. The problem in I
developing an index is how to reflect the signficance of each component of the system
in the index.
I
The different natue of the acid and basic oxides has been used in the development of
slag composition indices, generally termed basicities. Examples of basicity indices that
have been developed are given below in equations 1 to 4. I
Excess Bases = r (CaO)+ (MgO) ) - r (SiOi) + (Ah03) ) (1)
I
Basicity = r (CaO)+ (MgO) ) / r (SiOi) + (Ah03) ) (2)

Bell's Ratioi, = r (CaO) + 0.7*(MgO) ) / r 0.94*(SiOi)+ 0.18*(Ah03) ) (3) r

Optical Basicitl = (CaO) + 1.11 *(MgO) + 0.915*(Si02) + 1. 03
*
(Alz.;Ù (4)
(CaO) + 1.42*(MgO) + 1.91 *(SiOi) + 1.69*(Ah03)

Basicity indices can be grouped into general catagories:

a) Differences between the amount of bases and acids, equation 1;
b) Bases to acids ratios based upon the weight percentages, equation 2;
c) Bases to acids ratios based upon the molar concentrations, equation 3; and
d) Sum of the basicity of each component and its molar concentration, equation 4.

As would be expected based on the previous description of slag structure, those indices
which reflect the molecular natue of the slag composition, equations 3 and 4, tend to b_e
better predictors of slag properties. However, as the index defined by Equation 2 is '

2-4
l

I probably the most commonly used definition, it is used throughout the remainder of this
paper as B/ A.

J Temperature Impact -ISil, Basicity, and Slag Volume

The (Si) increases with increasing hot metal temperatue for all blast fuaces as
illustrated in Figure 9. The amount of (Si) increase for a given temperature increase
~i
varies from furnace to furnace, but the trend is the same for all furnaces. As the (Si)
increases, the (SiOi) decreases and therefore the basicity increases and the slag volume
i decreases. The amount of increase in the basicity for a specific increase in (Si) is a
fuction of the slag volume.

~ Shown on Figure 9 is the change in BfA for initial slag volumes of 200 and 300 kg /
THM and for the (Si) andhot metal temperature relationship given on the figure. The
general trend demonstrated here is that the larger the slag volume the smaller the change
i in BfA for the same change in (Si) or hot metal temperature.

I Slag Solidification
The common definition of melting temperatue only applies to a single component
system such as water, where only liquid water exists above the melting temperature and
I only solid water exists below the melting temperature. Slags are a multi-component
system and, therefore, do not have the common definition of melting temperature except
at specific compositions. Most slag compositions have both solid and liquid phases
I present over a range of temperatures. The lowest temperature at which only the liquid
phase exists for a specific composition is called the liquidus temperature.

I The solidification path of a slag is ilustrated on the simplified phase diagram shown in
Figure 10. Star with slag of composition Cstart at temperatues where only liquid slag
exists. As the slag cools, moving down vertically on the diagram, the composition of
~ the liquid slag does not change until the intersection with the Liquidus Line. The
intersection with the Liquidus Line is the liquidus temperature for the composition Cstart.
A very small amount of the solid compound on the left forms at the liquidus
temperature. Three changes continue as the temperature is further reduced below the
liquidus temperatue:
a) More of the solid compound is formed;
b) The amount of liquid slag decreases; and
c) The composition of the liquid slag changes, moving towards the right along the
Liquidus Line.
In the example, where the compound formed is 2CaO.SiOi, the basicity of the liquid
slag decreases as the slag is cooled because 2CaO.SiOicontains approximately twice as
much CaO as SiOi.

The solidification path illustrates how a compound can be formed even when the liquid
slag composition is significantly different than the composition of the compound. The
weight ratio of CaO to SiOi = 1.86 for the compound dicalcium silicate, 2CaO.SiOi. -
Whle no blast fuace has ever been successfully operated using slags with a CaO to

2-5
 .J

Si02 approaching 1.86, significant amounts of dicalcium silicate can be formed in the J
slags of operating blast furnaces. The formation of suffcient dicalcium silicate results
in a solid slag that breaks down into dust upon cooling, know as a 'Falling' or 'Dusting'
slag. The breakdown is caused by the 10% volume expansion of dicalcium silicate as it J
goes through a phase change at 675°C. The following guideline for avoiding a fallng
slag has been reported4:
J
(CaO) Less Than 0.9 * (Si02) + 0.6 * (Ah03) + 1.75 * (S) (5)

It is important to remember that phase diagrams are based upon equilibrium conditions. J
Equilibrium conditions imply that the cooling rate is slow relative to the rate of the
reactions, such as the formation of dicalcium silicate. The solidification path described
above is 'bypassed' if the cooling rate is very high as in slag granulation and, to a lesser ~
extent, slag pelletization. The rapid cooling locks the composition in a solid glass
phase, where the kinetics of the reactions are too slow for the compounds to form.
I
SLAG PROPERTIES I
The physical and chemical properties of slags are primarly a fuction of the slag
composition and temperature. The following describes these relationships for the
purose of developing general trends. I
Liquidus Temperatures
The definition of liquidus temperatue was described previously in the section on I
solidification. The relationships of liquidus temperature and composition for the four
primary components of slag are represented on a quaternary phase diagram. Figures 11
and 12 were generated from ternar planes of the quaternar phase diagram. Note that I
Figures 11 and 12 are not phase diagrams.

There are two general trends derived from these figures. First, the liquidus I
temperatures increase with increases in BfA and (Ah03)' Second, (MgO) in the range
of 8 to 14% tend to minimize the increase in liquidus caused by the increase in either
BfA or (Ah03).
r
r-
Viscosity
Viscosity is a measure of the amount of force required to change the form of a material
and is reported in units called Poise. The higher the viscosity, the more force required
to cause a liquid to flow. For comparison purposes consider that at 20°C the viscosity
of water is 0.01002 poise, while a typical acceptable slag viscosity is 2 to 5 poise, and
the viscosity of molten Si02 is ofthe order of 100,000 poise.

The high viscosity of liquid Si02 is caused by the polymer strcture discussed
previously. The breakdown ofthe polymer strctue by the basic oxides, lowers the
viscosity. The decrease in the viscosity of all slags with increasing the BfA is
liquid

shown Figure 13.

2-6
 I

I
In general the viscosity ofany liquid/solid mixtue increases as the amount of
suspended solids increases. The impact of temperature on slag viscosity is significantly
I greater at temperatures below the liquidus temperature than above the liquidus
temperatue, Figure 14.

:1 There are two general trends that for viscosity. The viscosity of liquid slags, above the
liquidus temperature, decreases with increasing B/ A and temperature. At temperatures
below the liquidus temperatue, the viscosity decreases with decreasing B/ A and
J increasing temperature.

J Sulfur Partition Ratio

The iron blast furnace is a very good desulfurizing process compared to the steelmaking
process because of the difference in the oxygen potential of the slags of the processes.
I The effect of the oxygen potential on desulfurization can be illustrated using Equation
6, where the oxygen potential is indicated by the (FeO). The higher the (FeO) the more
the reaction is driven to the left and the higher the (S). Steelmaking slags with (FeO) of
I 15 to 25 % are, therefore, weaker desulfurizing slags than the blast fuace hearth slags
with (FeO) of less than 1 %.

I (CaO) + (S) = (CaS) + (FeO) (6)

Essentially all the sulfur into the blast furnace leaves the fuace in the hot metal and
I slag. A relationship for the prediction of (S) can be developed based upon a mass
balance of sulfur for one ton of hot metal, Equation 7, and the defined term sulfur
partition, Equation 8. The prediction of (S), Equation 9, is derived by substitution of
I (S) from Equation 7 into Equation 8 and then solving for (S).

ST = (S) / 100 * 1,010 + (S) 1100 * SVol (7)

n where
1,010 is the kg of hot metal including a 1 % yield loss.
hot metal in a ton of

The remaining terms are defined in the Nomenclature section.

SP = (S)/(S) (8)

(S) = ST*100 / ( SP * SVol + 1,010 1 (9)

The slag SP can be predicted based upon Equations 10 and 11. Note that the
coefficients in Equation 10 were developed from regression analysis of a specific
furnace.

SP = 147.7 * BB + 37.7 * (Si) - 190 (10)

BB5 = ( (CaO) + 0.7*(MgO) 11 ( 0.94*(SiOz)+ 0.18*(Ali03) 1 (11)

2-7
 Cj

Equations 10 and 11 were used to construct Figure 15, and Equations 9, 10, and 11 were J
used in the constrction of Figue 16.

The general trends that can be derived from the above equations and figures are: ,j
a) (S) decreases with decreasing ST and increasing SP and SVoi;
b) SP generally increases with B/ A; however
c) CaO is a better desulfurizer than MgO; and J
d) Ah03 has a smaller effect on SP than SiOz.

Alkali Capacity J
A 'refluxing' or 'recycling' phenomena occurs in the furnace due to the counter-curent
flow of gases versus solids/liquids, paricularly for sulfu, zinc, and alkalis. The
~
recycling of the alkali potassium, K, is illustrated on Figure 17. The recycling'
phenomena is when an element travels down the fuace in a solid or liquid phase,
reacts to form gas species in the higher temperatue regions of the fuace, then travels
back up the furnace as gases, where it reacts and is absorbed by the solid/liquid phases
I
in the lower temperature region of the furnace. The recycling results in much higher
internal concentrations of the recycled element than the concentration going in or out of I
the fuace. For example the internal loading ofK may be 10 kg / THM when the
materials being charged contain only 2 kg / THM.
I
Alkalis have no beneficial, but many deleterious effects on the blast furnace. Alkalis
are absorbed by refractories, coke, and ore causing degradation of the refractories and
coke, and ore swelling. Alkalis can also form scabs which can peal off upsetting the I
thermal condition of the fuace, or build up and constrict burden and gas flow.

Alkalis cannot be avoided as they are contained in all coals, cokes, and to a lesser extent I
ores. The alkali loading should be minimized whenever possible.

A portion of the alkalis leave the fuace in the top gas, the amount being a function of I
the top temperature profile. The remaining alkalis must be removed in the slag. The
ability of slag to remove alkalis from the furnace is referred to as the alkali capacity of
the slag. The relationships of alkali capacity to slag composition and temperature are r
shown in Figure 186. In general the alkali capacity increases with lower B/ A and
temperature.

Silica Activity
The (Si) produced is dependent upon the burden materials, furnace operation, and slag
chemistry. The impact of the slag chemistry is shown in Equation 16. Equation 16 is
developed from the equilibrium constant, Equation 13, for the reaction given in
Equation 12, the definitions of the activities of(SiOz) and (Si), Equations 14 and 15,
and assuming that the activity of the carbon in the hearh equals one.

2-8
J

I (Si02) + 2 C = (SiJ + 2 COgas (12)

Keq = t ASi * p2 co ) / t Asi02 * Ac ) (13)

,i
ASi02 = (Si02) * YSi02 (14)

,i ASi = (SiJ * YSi (15)

i (SiJ = (Si02) * YSi02 / YSi * Keq / p2 co (16)

The reader is referred to the work by Chaubal and Ricketts9 for details of the above
I equations. The trend implied by Equation 16 is that the (SiJ decreases as the (Si02)
decreases.

I SLAG DESIGN FACTORS

In North America, a typical slag composition that would be formed from the gangue in
the ore and ash from the coke is 9% CaO, 5% MgO, 75% Si02, and 10% Ah03. A slag
I of this composition would have a liquidus temperature of

OF ) and would not flow well even above it's liquidus temperature. CaO and MgO are
the order of 1,600 °C (2,900

I added to the burden to 'flux' the gangue and ash resulting in acceptable liquidus
temperatures and flow characteristics.

I Basic slag design is the selection of fluxes to be used with a

the tyes and amounts of

burden and coke to produce a slag of acceptable properties. Burden and coke selections
are largely drven by economic issues such as local verses foreign sources and degree of
I beneficiation. These economic driving forces have resulted in a wide range of slag
compositions throughout the world, Table 1.

r The following are the general factors to be considered in designng a slag for normal
operation:
1. Liquidus Temperature - the slag must be completely liquid in the hearth and
casthouse;
2. Viscosity - the slag must have a low viscosity, high fluidity, so as to drain from the
hearh and down the casthouse runners;
3. Sulfu Capacity - the SP must be suffcient to produce hot metal with sulfu
contents within specifications;
4. Alkali Capacity - the slag alkali capacity must be suffcient to prevent alkali build
up in the furnace;
5. Hot Metal Silicon Control- the effect of the slag chemistry on the (SiJ must be
considered;
6. Slag Volume - the slag volume should be high enough to contrbute to the stability
of the slag properties and hot metal quality, but not so large as to require excessive
fuel or contribute to furnace instability;
7. Robust Properties - the slag properties should be as insensitive to variations in
normal variations in fuace operation as possible, specifically hot metal
temperatue; and

2-9
 -1
I
8. End Use - the requirements of the end use of the slag must be considered. J

Slag design must recognze that the above factors are not independent and that the
design always involves a balancing of the above factors to resolve the conflicting J
trends, Table 2. Two examples of the slag design are given below.

In the first example, the problem was to increase the alkali removal without increasing J
the (S). The resolution ofthe problem was to increase the slag volume through the use
of additional SiOz in the burden, while decreasing the slag basicity.
J
The problem in the second example was to lower the (Si) without negatively impacting
the other properties of slag and furnace operation. This problem was resolved
decreasing the (SiOz) by increasing the (Ah03) using diaspore, a high (Ah03) burden ~
material, while holding the (CaO) and (MgO) constant. Note that the change in slag
chemistry resulted in a decrease of both (Si) and (S).
I
SLAG AFTER THE BLAST FURNACE I
The use ofblast furnace slag is driven by the economics of processing and market
demand. In the past, when the processing and marketing was performed by the
company producing the slag, the markets tended to be local in nature with minimal I
processing. The trend to use independent companies that take ownership of the liquid
slag at the end of the slag runner, has lead to wider markets with more extensive
processing. The product slag can be classified by the rate of cooling. I
Air-cooled slags are those produced with low cooling rates. These are slags that are
solidified in pits and frequently cooled with water sprays. The largest uses for air- I
cooled slag are in road construction, railroad ballast, and aggregate. Air-cooled slag has
also been used in the production of cement, mineral wool insulation, roofing, and glass.
I
Pelletized and granulated slags are those produced with high cooling rates. Pelletized
slag is produced by pouring liquid slag onto a rotating drm, sometimes with water.
Granulated slags are produced by either pourng the liquid slag directly into a large pit r
of water or through the use of high pressure water sprays which breaks the slag up into
droplets. Rapidly cooled slags have been used for the same applications as air-cooled
slags. The high glass content of rapidly cooled slags makes it particularly sui tab Ie for
portland cement production.

ACKNOWLEDGEMENTS
The objective of this review is to summarize the work done by others. Due to the
magnitude of the work that has been done, it is difficult to give the personal recognition
due. The author would like to recognze the previous authors of this section,
paricularly R.L. Shultz, who provided the foundation of the strcture and contents of -=
this paper.

2-10
 I

I Table 1 - Examples of Typical Blast Furnace Slags?

North
J America Japan Europe India Australia
__ÇQ.IlQositiQn__ --------------- ~-------------------- ------------ ------------- --------------------
%SiOi 37-41
-------------------- 34
------------- 36
------------- 33 35-38
--------------------- -----------------
,) ------%-AI;Õ;---- 7-10
--------------- ------------ 13-15
------------ ----------
11-13 21- 25
------------ 15-17
------------------
%CaO
--------------- ------- --------
37-41 41 37-43--------- 33 37-42
------------------ --------------
i %MgO
V olume*
10-12
175 - 280
7
310-320
6-11
300-320
7-10
500-600
3-7
300-420
(350-560) ( 620-640) (600-640) (1000-1200) (600-840)
~ Alkali 2-4 2-3 2.5-5 7-10** 2.5-3.5
Loading* (4-8) ( 4-6) (5-10) (14-20) (5-7)
hot metal)
i * Units are kg/metric ton of

* * estimated
hot metal (lb/short ton of

Table 2 - General Conflcting Trends

I
Basicity (Ah03)
Lower Liquidus Temperature Lower Lower
I Lower Viscosity Higher
Higher K Removal Lower Lower
I Lower (S)
Lower (Si)
Higher
Higher
Higher
Higher

I Table 3 - Example of Designng Slag for Increased KiO Remova18

Slag K20 S
~ Basicity Volume (K20) Removed (S) Removed
0 (kg/THM) (wt% ) (kg/THM) (wt% ) (kglTHM)
r 1.10 225 0.47 1.30 1.82 5
1.05 282 0.55 1.55 1.77 5
1.00 290 0.63 1.85 1.72 5
0.95 298 0.71 2.10 1.68 5

Table 4 Example of Designng Slags for Lower (Si)9

Period Base No 1 No2 No3

Basicity 1.12 1.13 1.13 1.12
(MgO) 11.8 11.5 11.7 11.5
(Ah03) 7.8 10.2 10.3 11.7
(Si) 0.76 0.53 0.54 0.49
(S) 0.043 0.031 0.029 0.026

2-11
 Nomenclature J

(X) = weight percent of X in the slag

(XJ = weight percent of X in the hot metal
SP = Sulfu Partition Ratio = (S) / (SJ J
ST = Sulfur Loading = total weight of sulfur per unit weight of hot metal. Units
are kg per metric ton or pounds per short ton of hot metal
Alkali Loading = total weight of alkali per unit weight of hot metal. Units are kg J
per metric ton or pounds per short ton of hot metal.
SV 01 = Slag Volume = weight of slag per unit weight of hot metal. Units are kg
per metrc ton or pounds per short ton of hot metal. J
B/ A = basicity as defined by Equation 2
BB = basicity as defined by Equation 3
Kea = Equilibrium constant
a
ASi = Activity of Si in hot metal
ASiOZ = activity of SiOz in slag
Ae = Activity of carbon in the hearth coke I
YSiOZ = Activity coeffcient of SiOz in slag
YSi = Activity coeffcient of Si in hot metal
Pea = Partial pressure of CO in the hearh
I
References I
1 Richardson, F.D., Physical Chemistry of
Steelmaking, Edt. J.F. Elliott, MIT, Mass.,
1958, pp. 55-62.
z Kalyanram, M.R, Macfarlane, T.g. and Bell, H.B., " The acitivity of Calcium Oxide I
in Slags in the Systems CaO-MgO-SiOz, CaO-Ah03-SiOz and CaO-MgO-Ah03-SiOz
at 1500 °C," Joural of the Iron and Steel Institute, 1960, pp 58-64.
3 Sommervile, LD., and Sosinsky, D.J., "The application of
I
the Optical Basicity
Concept to metallurgical Slags," Second International Symposium on Metallurgy Slags
and fluxes, edited Fine,H.A., and Gaskell, D.R, published by the Metallurgical society I
of AIME, 1984, pp. 1015-1026.
4 BisWas, A.K. Principles of
Blast Furnace Ironmaking, Cootha Publishing House,
Australia, 1981, pp. 347.
5 Kalyanam, M.R, Macfarlane, T.G. and Bell, H.B., "The acitivity of
r
Calcium Oxide
rc ~
in Slags in the Systems CaO-MgO-SiOz, CaO-Ah03-SiOz and CaO-MgO-Ah03-SiOz
at 1500 °C," Journal ofthe Iron and Steel Institute, 1960, pp 58-64.
6 Poos, a., and Vidal, R, "Slag Volume and Composition for Optimal Blast fuace
Operation," 12th McMaster Symposium on burden Design for the Blast Furace, Ed. W-
K Lu, May 1984, pp 67-89.
7 Shultz, RL., "Blast Furnace Slag," Blast Furnace Ironmaking, published by McMaster
University, 1990.
8 Sciulli, C. M., and Ravasio, D., "Alkalies in Raw Materials and Their Effect on the
Blast Furnace," 51st Anual Meeting, Minnesota Section AIME, and 39th Anual
Mining Symposium, Duluth, Minnesota, Januar 1978.
9 Chaubal, P.C. and Ricketts, J.A., "Slag Properties Optimization Program at Inland's _
Eight Meter Blast Furnaces," Ironmaking Conference Proceedings, 1991, pp 445-455..-

2-12
 I

J Figure 1 - Blast Furnace Zones Figure 2 - Blast Furnace Slag Zones - Reactions

Fe,O, =0. FeD

~I Granular Zone ¥" FeO =0. Fe

Granular Zone
FeO + Gangue + Fluxes
=0. Bosh Slag

i Cohesive Zone J
Active Coke Zone
and Deadman
Slag
Formation
Zone
Slag
Formation
Zone
(FeD) =o.IFel
SiO." ~o.ISil or (SiO,)
SiOiCoke =;: SiOgas
Raceway J AshCoke =:; Slag

~ Hearth eartb .- (SiO" MoO, S) = ISi, Mn, Si

I
I Figure 3 - Ilustrtion of Slag Flow in Heart Figure 4 - Silica Atomic Strcture

.. Coke above Slag Layer

~t
Atomic structure of (SiO,)-4 is Tetrahedral structure witb:

I Taphole
. = Si atom in the center

. = Oxygen atoms at the corners.

I
fJ
~
Figure 5 - Crystallne Silica Strcture Figure 6 - Molten Silica Strcture

2-13
 J

~J

Figure 7 - Addition of Bases to Molten Silica Figure 8 - Orthosilicate Strcture - 2MOoSiOi

J
.. :: Base Oxide

~ ~ ~'WÄ
-,l-v-.l--v- ,l-~-~-V- J
-Ä.Ý..Á: Ý-Ä-Y-Ä-Ý-
-~i-v-~;-v-~;.v-A;.v- J

........ ..
.._..,~, .. ,.,.
:.YI- :.'i- ,~,.*l(*
A- ~~- ~A- V;. A- ~ g

Figure 9 - Temperature & Slag Volume Impact Figure 10 - Ilustration of Slag Solidification I
..Ø/A. Slag Vol- 200 -oB/A - SlagVol- 300 (Sll Start with only Liquid Slag, Composition = Cstar

1.20 1.00% Liquidus

I
Tem, -
1.5 0.95%
'"
..
~
el 1.10 0.90%
~
H
i: Liquidus
I
1.05 0.85% e'" K- Line
¡.
1.00 0.80%
1,350 1,400 1,450 1,500
Chqud - I
HM Temperature Compound

(ex. 2CaO' 8i02)

Figure 11 - Liquidus Temperature ~ BfA = 1.0

r
Figure 12 - Liquidus Temperature (f 10% (Ali03)
rr
2,00 2,(0
1,9 i,
~ 1,l
~
!Ap,
10 8
'"
l, BlA
1.3
e 1,70
=
.. 1,70
is 3.. 1.2
f 1,(ß
'" 5-10 ..
'" 1,60 1.0
i: i:
~ 1;0
¡.
e'"
¡.
1,
1,40 1,40
1,3 1,
o 10 ~ 20 30 0 10 () 20 30

2-14
 I
UJ

J Figure 13 - Viscosity Verses BfA Figure 14 - Viscosity Vs Temperature

35
7
A B
At i,500C 30 c
6
~J 25
Llauidus. BfA
~ A) 1,250 C - Low
~ 'õ
'õ 5 .e 20 B) 1,:345 C . Mlddl~

i
S- q 1.390 C - High
~0 15
f
;;
4
;i
;; 10
3

~ 2
0.7 0.8 0.9 1.0 1.1 1.
0
1,3 1,350 1,4 1,45 1,50
Tempratu (C)
BIA

I
I Figure 15 - Sulfur Partition Figure 16 - HM Sulfur Prediction

Where (CaO) I (MgO) = 4; ISil = 0.8 % Where (CaO) f (MgO) = 4; fSi) = 0,8 %; ST ~ 3 kgfHM; SlagVol ~ 200 kgflHM

I 60

50
0.08

0.0
BfA
0.95

BfA

I ~
re 30
40 1.10

1.05
f2
0.06

0,05
1.00

0.04 1.05
1.00
20 1.10

I 10
3.0 3.5 4.0 4.5
0.95

5.0
0.03

0.02
3.0 3.5 4.0 4.5 5.0

(SiO,) f (A1,0,) (SiO,lf(AI,O,l

Figure 17 - Alkali Recycling Figure 18 - Alkali Capacity

In As Out in
Solid Gas 4.0
3.5

K Condensation
;t 3.0
~.. + Si0i,llId + COi
K~O Reduction =;: KSi0.i,s,lId + CO I 2.5
.ia 2.0
-¡
K¡OlllId + C =)0
2~..+CO ~ 1.5

lJ 1.0
(KiO,)+ (CaO) + C ='" ¡; 0.5
~.. + (CaOSiO,) + CO
0.0
0.85 0.90 0.95 1.00 1.05 1.10 1.5

BIA

2-15
J

I
I

I
I

r
n
 l

J LECTURE #3

I
~I BLAST FURNACE REACTIONS

i Alex McLean

Deparent of Metallurgy and Materials Science

I University of Toronto
Toronto, Ontao M5S 3E4
Canada
I
I
I Abstract :- Durng the latter half of this centuy, in parallel with developments
pertaining to greater productivity, there have been increasing demands for improved hot
meta quality. To a large extent, these demands have been met by advances in our
I knowledge of the chemical, physical and thermal interactions between gas, solid and
liquid phases that tae place within the fuace and during external treatment of hot
metal. In this lectue, the reactions discussed include those involving carbon and

I oxygen, the reduction of iron and other oxides, the behaviour of alkalies and sulphur,
and interactions with slag. The concept of optical basicity is described and examples are
presented of how it can be used to design slags with appropriate characteristics for
~ specific operations.

1-0 INTRODUCTION
Over one hundred years ago in 1890, Hemy Maron Howe, a distinguished steelmaker,
an eminent professor of metallurgy and President of AIME, published his classic text
entitled "The Metallurgy of Steel"(l). In spite of the time difference, much of the
material contaned in this volume is stil worthy of study today. Howe had the great gift
of being able to express in vivid terms, some of the basic truths of iron and steelmakng
technology. An example of this may be found in his use of the term "The Treachery of
Steel", which he employed in a very graphic maner to discuss what we today would
call, "The Management of Quality":

"Owing to the very nature of the processes by which steel and wrought-iron are
made, carelessness and ignorance, whether in selecting materials, in conducting the
processes, or examining the product, is likely to lead to the making and sellng of
treacherous steel, treacherous simply because it is unsuited to the purpose_ f~r
which it is sold."

3-1
 J

J
Major technological changes have transformed the iron and steel industry during the
past centu. These changes have had a profound effect on process intensification, J
energy Utilization, metal yield and product quality. They include: high productivity blast
fuaces, external treatment of hot metal, oxygen steelmakng, alternate ironmaking
processes, ultra high power arc fuaces, ladle metallurgy, vacuum processing, J
continuous casting, thermo-mechanical processing and novel coating technologies.

Coupled with the implementation of advanced production technologies there have been J
ever increasing demands for improved steel performance which in tur have strongly
infuenced changes in steel chemistry and steel quality. For example, in 1911 at the time
of the launch of the Titanic, the steel plates used for construction of the hull met all of ~
the required stadards. The ship was built by Harland and Wolff at their Belfast
shipyard in Northern Ireland, Figue 1. The steel was manufactued at the Motherwell
works of David Colvile & Company in Scotland. This is the same company which I
twenty five years later provided steel for the construction of the world's largest
passenger liners, the Queen Mar and the Queen Elizabeth.
I
On Sunday April 14th, 1912 at 11.40 PM, the Titanic struck an iceberg and san two
hours and forty minutes later, with a loss of over fifteen hundred men, women and
children, Figure 2. Six years ago, metallurgists at the Metals Technology Laboratories, I
CANMET, in Ottwa, Canada, published a report on their investigation of a number of
cast iron, wrought iron and steel samples recovered from the Titanic wreck site, on the
bed of the Atlantic, over 4,000 meters below the surace of the ocean. (2) Chemical I
analysis of a section of hull plate, which was approximately 25 mi thick, Figure 3,
indicated that the steel contained 0.2%C, 0.52%Mn, 0.025%Si, 0.065%S, O.OLO%P,
o:O.005%Al and 0.004%N. From ths analysis, it is evident that the hull was constrcted I
from low carbon, semi-killed steel, produced by the open hear process. The high
sulphur content is of paricular significance. The micro-strctue shown in Figure 4,
indicates extensive carbon banding, typical of hot rolled O.2%C steel and more I
importtly, elongated in the rolling direction, long MnS stringers, some of which
exceed 25 mi in length. In Figure 5, the results of Chary tests performed on samples
taken in the longitudinal direction are compared with data for a semi-killed steel of
r
similar composition but with considerably lower sulphur content which would be typical
of steels produced in the early fifties. With a seawater temperature of approximately
zero degrees Celsius, the hull plates had essentially no ductility.

A major thrst in curent steelmakng technology is the production of steel with lower
residual concentrations of sulfu. This element has a profound infuence on the quality
of the final steel product because of the effects on mechanical properties. Today, high-
quaity steels are produced for demanding applications such as Arctic pipelines, offshore
platforms, ice-breaker vessels and ships for the transporttion of liquid natual gas.
These steels are produced with extremely low inclusion contents and the residual
sulphur levels can be less than 10 ppm.

3-2
 .1

I THE RELATIVE STABILITY OF OXIDES

As iron oxide, coke and slag-makng materials move down through the stack of the
I fuace, several important exchange processes take place. Heat is removed from the
ascending fuace gases that consist mainly of carbon monoxide, carbon dioxide and
nitrogen and transferred to the descending burden materials. Oxygen is removed from
'I the descending iron oxides and transferred to the ascending reducing gases. Thus within
this very efficient counter-curent reactor, chemical reactions take place as the charge

i descends, the temperatue of the 'burden materials increases, fusion of the reduced iron,
iron oxide and slag-makng materials begins and finally liquid metal and slag collect in
the hear of the fuace. Much of the coke charged to the fuace is bured with

l oxygen in the hot air blast at the tuyeres to provide both heat and the reducing agent
carbon monoxide.

The relative stabilty of varous oxides is plotted against temperature in Figure 6 which
I is adapted from Gaske1i3. This is known as an Ellngham Diagram and is extremely
useful for understanding the behaviour of oxides in the blast furnace. The relative
I stability is measured in terms of the free energy of formation of the oxides. The greater
the negative free energy of the oxide, the greater is the oxide stability. This
formation of

means that oxides that are located in the upper part of the diagram have a relatively low
stabilty, while oxides located in the lower portion of the diagram have a high stabilty.
I Oxides located in the center of the diagram have a moderate stability.

I . Oxides with a relatively low stability include potassium oxide, sodium oxide,
phosphorus oxide and iron oxide.

I . Oxides with a moderate stability include manganese oxide, chromium oxide,

silica and titanum oxide.

r . Oxides with a high stabilty include, alumina, magnesia and lime.

r-oo It is also useful to consider this diagram in terms of the affinity of an element for
oxygen. For example, elements that are located at the top of the diagram have a low
affnity for oxygen, while elements located towards the bottom of the diagram, have a
high affinity for oxygen. Ths means that oxides at the top are relatively easy to reduce,
while those at the bottom, are diffcult to reduce.

Ths is ilustrated by the line for the formation of phosphorus oxide which lies above the
line for formation of iron oxide at temperatues corresponding to those found in the
hearh of the blast fuace. This implies that phosphorus oxide has a lower stability than
iron oxide and consequently, since reducing conditions in the fuace are sufficient to
reduce iron oxide, essentially all of the phosphorus entering the fuace wil end up in
the hot metal.

On the other hand, stable oxides such as alumina, magnesia and lime are not reduced
under blast fuace conditions, and end up in the slag phase. Oxides with a moderate

3-3
 stability such as manganese oxide, chromium oxide, silca and titaum oxide are J
parially reduced to give some manganese, chromium, silcon and titaum dissolved in
the hot metal, while the remaining uneduced oxide constitutes par of the slag.
J
The Ellngham Diagram is constructed on the basis that a pure element at unit activity
reacts with one of mole of oxygen gas to form pure oxide at unt activity. The
thermodynamic term "activity" is a paricularly useful concept for discussing the J
behavior of elements dissolved in molten iron, or oxides dissolved in molten slag. For
example, when small concentrations of elements such as oxygen or sulphur are
dissolved in molten steel, their activity can frequently be taken as equal to their J
concentration in weight percent. However, in the presence of high concentrations of
other elements, for example, carbon in hot metal, the activity of sulphur is greater than
the concentration, while the activity of oxygen is less than the concentration. In such g
cases it is importt to distinguish between activity and concentration.

· The concentration of a component in solution is a measure of how much of the ~

component is present.

· The activity of a component in solution is a measure of how the component I

actually behaves.

All the lines on the Ellingham Diagram except those involving carbon, have a positive
I
slope, indicating that the oxide stability decreases with increasing temperatue. The lines
for the oxides of potasium oxide, sodium oxide, magnesia and lime, each show a shar
increase in slope at the temperatues corresponding to the boiling points of the
I
respective metals.

The line for the formation of carbon dioxide from carbon and oxygen has almost zero
I
slope indicating little change in stability with increasing temperatue, while that for
carbon monoxide has a strong negative slope which means that the stability of carbon I
monoxide actually increases as the temperatue increases. The lines for the two oxides
of carbon cross at about 700 C. Above this temperatue, carbon monoxide is more stable
than carbon dioxide while at lower temperatues, carbon dioxide is more stable than r
carbon monoxide.

CARON-OXYGEN REACTIONS

The pre-heated air blast injected through the tuyeres at a temperatue of about 1000 C
and two to three atmosphere's pressure, produces a pear shaped reaction zone in front of
each tuyere. The temperatue in this region is about 2000 C and rapid reaction first
occurs between excess oxygen and coke to give carbon dioxide. This is an exothermic
reaction.

C + O2 = CO2 (1)

3-4
 I

I Immediately outside this zone, there is no longer free oxygen available and the carbon
dioxide reacts with excess coke to give carbon monoxide. This is known as the
Boudouard reaction and is endothermic.
J
CO2 + C = 2CO (2)

cl Combining reactions 1 and 2 gives the reaction for partial combustion of carbon with
oxygen to provide carbon monoxide.

J 2C + O2 = 2Ca (3)

The heat evolved in the formation of one mole of carbon dioxide is about three and one
I half times that for the formation of one mole of carbon monoxide and one measure of
the efficiency of the blast fuace is the degree of conversion of carbon in the coke to
carbon dioxide.
I
Below 700 C, carbon dioxide is more stable than carbon monoxide and reaction 2
I proceeds to the left:

2CO = C + CO2 (4)

I Ths reaction is often referred to as the carbon deposition reaction and wil be mentioned
again later.
I Above 700 C, carbon monoxide is more stable than carbon dioxide and reaction 2
proceeds to the right. Ths is sometimes called the carbon solution loss reaction and in
I this sense implies a negative behavior. On the other hand the reaction represents a
regeneration of reducing gas within regions of the fuace above 700 C. This is one of
the important fuctions of coke within the blast fuace and is paricularly desirable as
~ it increases the volume of the gases and helps in heat transfer. However this reaction is
endothermic and when it occurs within the tuyere zone it creates a cooling effect within
F-' a location where high temperatues are important.

The effect of temperatue on the equilibrium reaction between coke and a gas mixtue
containing carbon monoxide and carbon dioxide at one atmosphere pressure and also
three atmospheres pressure, which is more typical of modem blast fuace practice, is
shown in Figure 7. To the right of the graph, carbon monoxide is more stable than
carbon dioxide, while at lower temperatures, to the left of the graph, carbon dioxide is
more stable than carbon monoxide. From this figure it is evident that above 1000 C, the
percentage of carbon dioxide in equilibrium with coke is essentially zero . On the other
hand, at temperatures below 400 C, the concentration of carbon monoxide is smalL.
Thus as the temperatue decreases between 1000 and 400 C, the stability of carbon
monoxide decreases while the stability of carbon dioxide increases and the partial
pressure of both gases in equilibrium with coke is significant.

3-5
 J

The gases leavíng the top of the furnace are usually about 200 C and íf equílíbríum was L
obtaíned wíth coke, the ratío of carbon monoxíde to carbon díoxíde would be about 10-5.
In fact, the ratío ís usually between 1 and 3, Le. the gas ís very much more reducíng
than that predícted from equílbríum consíderatíons and full use ís not beíng made of the J
reducíng potentíal of the gas. Tils ímplíes that the coke rate ís ín excess of theoretícal
requírements.
J
Thís lack of equílbríum between the gases and coke can be attríbuted maínly to the ilgh
gas velocíty ín the stack. The gas retentíon tíme ín the fuace ís only about 10 seconds,
and extremely hígh velocítíes can occur, parícularly ín loosely packed, coke rích J
regíons. Another factor ís that the gas temperatue drops by about 1800 C as ít ríses
through the fuace and so there ís líttle opportíty for equílbríum to be maíntaíned.
~
THE CARON DEPOSITION REACTION

Sínce the carbon monoxíde content of the gas wíthín the stack of the blast fuace at ~
temperatues below 1000 C ís consíderably ilgher than ít should be, there exísts a
drvíng force for the carbon deposítíon, or sootíng, reactíon to proceed. Thís dríving
force ís partícularly strong between 500 and 700 C. A gas with a temperature and I
composítíon above the líne ín Fígure 7 wíll tend to deposit carbon by reactíon 4, and one
wíth a composítíon and temperatue below the líne wíll oxídíze carbon ín accordance
wíth reaction 2. Fortately reactíon 4 ís sluggísh and equílbríum ís never attíned,
I
otherwíse seríous cloggíng of the spaces wíthín the burden at the top of stack could
occur. Ths ín tu could lead to írregular flow of the reducíng gases and uneven
descent of the burden. Even for paríal reaction, a suítable catalytíc surface ís required,
I
upon whích the carbon can nucleate and grow. Iron parícles, partíal reduced íron ore
and íron carbíde have all been suggested as possíble catalysts. The reactíon appears to
be enhanced by hydrogen and water vapor whíle nítrogen and sulphur compounds, for
I
example, amorua, hydrogen sulphíde and carbon dísulphíde act as ínhbítors. Zínc
oxíde and alkalíne compounds oppose the ínhíbítíng effect of sulphur, and although the
concentratíon of these compounds ín the furnace ís generally small, they volatílze at
I
hígh temperatues ín the hearh and condense agaín ín the cooler regíons of the stack.
The cumulatíve effect ís that such compounds can offset the ínfluence of sulphur. r
The carbon deposíted by the reactíon ís ín a very finely dívíded form and some may be ,-
accommodated wíthín the pores of the íron ore parícles and cared back down the stack
agaín. Tils can affect the reductíon process ín several ways.

. Because of the actíve natue of the carbon and íts close assocíatíon wíth the ore,
reductíon by solíd carbon can take place at lower temperatue than that requíred for
reductíon by coke, parícularly sínce coke canot penetrate the pores and reductíon
can only take place at poínts of contact between the solíd parícles. The rate of such
reductíon wíl depend upon the rate of díffusíon of oxygen from the ínteríor of of the
parícle to the poínt of contact. In the upper par of the fuace, the reductíon by
coke is neglígíble, compared wíth gaseous reductíon. It becomes sígníficant only
above about 1000 degrees C. when the gaseous reactíons are ímpeded by slag

3-6
"¡i.

I formation. In contrast, reduction by precipitated carbon may occur at temperatues

as low as 800 C.

J . The formation of carbon monoxide during reaction within the pores tends to open up
deep fissures within the paricle, thus increasing the gas-solid contact area, and
increasing the efficiency of gaseous reduction.
1
. When carbon dioxide is produced within the pores of a paricle by the gaseous
reduction reaction, it can be rapidly regenerated to carbon monoxide by reaction
) with the carbon in the pores, thus allowing the reaction to continue.

Unfortately, the carbon deposition reaction can also have certain adverse effects.
~
. The reaction can cause splitting of refractories by deposition on active iron spots, in
I regions where the temperatue is about 500-550 C., for example in the outer shells at
lower levels in the stack, or within the inner shells at the upper levels.

I . If excessive, carbon deposition can cause ore pellets or sinter to cruble into powder
and this can cause irregular gas flow and uneven descent of the burden.

I . Since the reaction is exothermic, the temperature of the exit gases is increased.

Although the overall effect of the carbon deposition reaction may be debatable, certain
I facts remain.

the exit gases.

I
. The reaction does decrease the CO/C02 ratio of

. The reaction recirculates a certain amount of carbon, which otherwise would be

cared out of the fuace, thus increasing the time available for reaction with carbon
L and increasing the chemical effciency of the reduction process.

REDUCTION OF IRON OXIDES

The reduction of iron oxides by carbon monoxide can be represented by the following
reactions:

3Fe203 + CO = 2Fe304 + C02 (5)

Fe304 + CO = 3FeO + CO2 (6)

FeO + CO = Fe + CO2 (7)

These reactions are accomplished at increasingly higher temperatues and as shown in

Figue 8, with increasingly greater percentages of carbon monoxide. Ths means th~t
reactions 5 and 6, which are relatively easy to achieve, can take place within the upper

3-7
 regions of the fuace. Reaction 7 which entails the removal of the last amount of
oxygen from the iron, is in fact the most diffcult to achieve and therefore takes place
fuher down the fuace where the temperatues are higher and the carbon monoxide L
i
content of the reducing gases is greater. Below 570 C, the non-stoichiometric wustite J
phase (FexO) is unstable and it is possible to reduce magnetite directly to iron.

At any paricular temperature, there is a minimum carbon monoxide content in the gas J
mixtue required for reduction of a specific oxide. This means that it is not possible for
all the carbon monoxide in the gases to be converted to carbon dioxide if the reduction
reactions are to continue. For example, at 800 C the equilibrium gas mixtue in contact J
with FeO and solid iron contains about 65%CO and 35%C02. If the
the CO2 content of

gases exceeds this value at this temperature, iron wil tend to be oxidized back to FeO.
Accordingly, for these reactions to occur, there must be a minimum concentration of ~
carbon monoxide in the gases at each step as indicated in Figure 8, and it is not possible
to convert CO completely to CO2 by these reactions. Fortately at these temperatures
the carbon dioxide produced by the reduction reactions is unstable in the presence of I
coke and carbon monoxide is regenerated based on reaction 2 so that the reduction
reactions can continue.
I
It is wort noting that the combination of reaction 2 with reaction 7 corresponds to the
"direct" reduction ofFeO by carbon and this is a strongly endothermic reaction:
I
FeO + C = Fe + CO (8)

The reduction of iron oxides may also take place by hydrogen which is generated by I
parial combustion of auxiliar fuels injected through the tuyeres to produce two
reducing gases, carbon monoxide and hydrogen. Hydrogen is also produced when
steam is added to the blast as an aid in controllng the fuace. Excellent discussions on
I
tuyere additives and their effects on blast furnace operation are presented in the two
chapters" Blast Furace Energy Balance and Recovery", and "Fuel Injection in the Blast I
Furace" .

Whle the oxidation of carbon by oxygen in the air-blast to form carbon monoxide is r
exothermic, the reduction of moistue by coke to form carbon monoxide and hydrogen is
strongly endothermic:

H20 + C = CO + H2 (9)

The reduction of iron oxides by hydrogen again proceeds in a sequential maner:

3Fe203 + H2 = 2Fe304 + H20 (10)

Fe304 + H2 = 3FeO + H20 (11)

FeO + H2 = Fe + H20 (12)

3-8
 J

I The effect of temperatue on these reaction equilbria is shown in Figue 9. Whle

reaction 10 is slightly exothermic reactions 11 and 12 are endothermic. The presence of
hydrogen, which because of its small size has a high diffsivity, markedly reduces the
J density and viscosity of the blast fuace gases and, paricularly at high temperatures,
enhances the reduction of low reducibility raw materials.

J The water gas shift reaction can take place between the different components in the gas
phase to bring the hydrogen-bearing and carbon-bearing gases into equilibrium:

) C02 + H2 = H20 + CO (13)

It wil be evident from Figure 8, that the gases passing up the furnace canot be in
~ equilibrium with carbon in the coke and at the same time in equilbrium with iron oxides
in the descending burden. Above about 800 C the reaction of the gases with carbon is
more rapid than with oxides and the equilibrium between coke and the gas phase is
I probably approached fairly closely. As shown in Figue 10, measurements of the
temperatues and compositions of gases in operating fuaces indicate that they tend to
I fall between the CO/C02-C line and the FeO/Fe line above 800 C, cut the FeO/Fe line
between 600 and 800 C and then remain at or just above the Fe30JFe line. At
temperatues below 600 C, the very rapid gas flow allows little time for reaction with
solids and the CO content of the gas is far in excess of that which would be in
I equilibrium with coke.

I If the iron oxide is chemically associated with other oxides, its activity in the blast
fuace will be decreased. This means the iron oxide will be more difficult to reduce
and the CO/C02 ratios required will be greater than those considered here. For example
I with ferrous silicate, the minimum CO/C02 ratio required for reduction at 700 C would
be increased from about 1.5 to about 22, i.e. from about 60%CO to almost 96%CO on a
carbonaceous gas basis. Since combined oxides are more diffcult to reduce, higher
temperatues are required for reduction and thus the amount of reduction obtained with
IJ
CO before slag formation occurs will be decreased. This implies an increase in coke
1- rate since the amount of reduction required in the lower par of the fuace wil be
increased.

REACTIONS IN THE BOSH AND HEARTH

Reduction of Other Oxides

The reduction of oxides more stable than iron oxide such as manganese oxide and silica
would not take place in the blast fuace if the products were pure metals since the
reaction:
MnO + CO = Mn + CO2 (14)

would have, at equilibrium, a percentage of CO very close to 100 percent. That is, the
effciency of reduction is extremely low and enormous quantities of gas would be
required for very small amounts of manganese reduced. The situation with silca is even

3-9
 more extreme since it is a very stable oxide. However, by dissolving the manganese and
silicon in iron, the reactions:

MnO + co = Mn (dissolved in iron) + CO2 (15) '1

,-

and
J
Si02 + 2CO = Si (dissolved in iron) + 2C02 (16)

are moved somewhat to the right so that there is a distribution of manganese and silicon J
between metal and slag which is a fuction of the slag composition and of the
temperatue. Since the reduction of both of these elements is endothermic, the amount
of each in the hot metal increases with temperatue and the extent of the reactions wil to ~
some degree be controlled by controlling the temperature in the hearh of the fuace.
Of greater importance is the fact that the CO2 produced by these reactions will react by
the Boudouard reaction and wil cause an increase in the coke consumption. i
The amount of manganese reduced clearly also depends on the amount in the charged
ore. Ores such as Wabush from the Labrador trough with up to 2 percent manganese I
give much higher than normal manganese contents in hot metal with consequent higher
coke rates per tonne of iron produced. Silicon "swings" caused by erratic burdening of
the fuace or by temperatue variations can also have another serious effect: as the I
silicon is reduced into the hot metal it is depleted from the slag, increasing the basicity
the slag sometimes dramatically.
ratio and changing the melting point and fluidity of
I
Effects of Silcon Monoxide Formation

For many years it was considered that silca and manganese oxide were reduced directly I
from the slag by reaction with carbon in iron according to the reactions:

Si02(slag) + 2C = Si + 2CO(g) (17)

I
MnO(slag) + C = Mn + CO(g) (18) r
It was thought that molten iron droplets picked up silicon as they passed through the
slag phase and on into the hearh. Research however, has shed new light on these
reactions and also those involving sulphur..(4) Several laboratory studies together with
plant data from Japan have shown that at the temperature of the combustion zone, about
2000C, silicon monoxide gas is produced during the combustion of coke by the reaction:

Si02(coke ash) + CO = SiO(gas) + C02 (19)

Combining Eq. (19) with the reaction for coke oxidation:

cO2 + C(coke) = 2CO (2)

3-10
 J

I yields the overall reaction:

Si02(coke ash) + C(coke) = SiD(gas) + CO (20)

J
While the presence of FeD in slag is likely to make SiO formation from slag very
diffcult, an additional source of silica would be reduced silica-rich slag adhering to
LI coke paricles. Following these reactions, silcon is transferred to iron droplets by
reaction with silicon monoxide in the gas phase:

J SiO(gas) + C = Si + CO (21)

As iron droplets containing silicon pass through the slag layer, some of the silicon is
~ oxidized by iron oxide and manganese oxide, and taken up by the slag:

I 2(FeO) slag + Si = (Si02) slag + 2Fe (22)

2 (MnO) slag + Si = (Si02) slag + 2Mn (23)

I Behaviour of Sulphur

I Sulphur is a troublesome element in blast fuace operations because hot metal for
steelmakng must be low in sulphur; levels of 0.035 to 0.02% are usuaL. The reaction
by which sulphur is removed from liquid iron into the slag is often represented by the
I reaction:

~ + (CaO) + C = (CaS) + CO(g) (24)

I Where sulphur and carbon in the metal react with lime dissolved in the slag to form
calcium sulphide in the slag and CO gas. The distribution of sulphur between slag and
~ metal, (S) /~, is strongly infuenced by a number of factors:

r . Increasing the basicity of the slag (lime/silica ratio) tends to raise the
thermodynamic activity of lime in the slag which pushes reaction (24) to the right.

. An increased oxygen potential in the system pushes the reaction to the left. This is
shown by rewriting the reaction as follows:

~ + (CaO) = (CaS) + 0 (25)

This effect is very strong, and the presence of even small concentrations of FeO In
the slag wil seriously limit the sulphur ratio. (S) /~.

. Fortately both silicon and carbon raise the thermodynamic activity of sulphur in
hot metal by 5 to 7 times. Accordingly, sulphur in hot metal is 5 to 7 times easier to
remove than it would be from liquid steel that contains relatively little carbon and
silicon.

3-11
 J

Assuming sulphur in coke ash is present as CaS, the following reaction can occur with :J
SiO in the combustion zone to form volatile SiS:

CaS (coke ash) + SiO (gas) = CaO + SiS (gas) (26) J

To a lesser extent, some CS gas may form by the reaction:
J
CaS (coke ash) + CO = CaO + CS (gas) (27)

J
Sulphur transfer from these volatile species to molten iron droplets then takes place
within the bosh zone. Turkdogan has shown that when iron droplets containing silicon
and sulphur are allowed to fall through molten slag, in the absence of MoO, the silicon ~
content of the metal actually increases, and there is no transfer of sulphur. (5) In the
presence of MoO, silicon is removed from the metal by reaction (23) and manganese
transfers from slag to metal together with sulphur transfer from metal to slag. Based on I
the various results available, Turkdogan suggests the following sequence of reactions in
the bosh and hearth:
I
. The formation of SiO and SiS in the combustion zone.
. The transfer of silcon and sulphur to metal and slag droplets in the bosh.
. The oxidation of silcon by FeO and MoO in the slag as the iron droplets pass
I
though the slag layer.
. The desulphurzation of metal droplets as they pass through the slag layer. I
The sulphur distribution ratios found in the blast fuace generally var between 20 and
120. On the other hand experiments have shown that when metal and slag samples from I
the blast fuace are remelted in graphite crucibles at 1 atm CO, the distribution ratio
increases to between 120 and 220, depending on the slag basicity. This suggests that the
oxygen potential of the system is higher than might be expected for C-CO equilibrium in I
the fuace hearh. Thus while thermodynamic conditions favour sulphur removal from
hot metal within the blast fuace, kinetic considerations imply that the reaction can be
more readily accomplished outside the furnace by external desulphurzation. r
Alkalies and Zinc

Sodium, potassium and zinc, often called the "rogue elements", can cause serious
operating problems in the blast fuace and must be monitored and carefully controlled
if stable conditions are to be maintained. The alkali metals enter the blast fuace as
constituents of the gangue in the ore and also as a part of the coke ash, generally as
silicates. In the stack ofthe fuace, the silicates react by the formulas:

KzSi03 + CO = 2K + SiOz + COz (28)

NazSi03 + CO = 2Na + SiOz + COz (29)

3-12
 I

I In the blast fuace, the potassium reaction can take place above 500 C. While the
sodium reaction occurs at about 600 C. At temperatues of about 900 C, the alkali
.I metals are above their boilng point so they join the gas phase. However, as these gases
sta to rise up the fuace, the metal becomes unstable with respect to other compounds
that can form and cyandes, oxides and carbonates all sta to precipitate from the gas
,i phase as very fine fues or mists, since the cyanides are liquid over a wide temperatue
range. These fine paricles of solid and liquid can deposit on the iron ore paricles, the
coke, and the fuace wall, with some, of course, being swept out with the fuace gas
J and being captued in the dust catching system. Paricularly the liquid alkali compounds
can penetrate the brick lining of the fuace and cause serious deterioration and spalling.
As well, these compounds can build up on the wall and cause scaffolding, hanging and
J slipping.

I The alkalies which land on the iron and coke are cared to the lower par of the furnace.
There, they are again reduced to the metal which rises up the stack as a gas, forms the
same alkali compounds, and repeats the cycle, joining new material in the process. The
reduction requires carbon, increasing the coke rate and cooling the fuace, and the
I recycling material can build up to the point where it degrades the coke in the fuace,
causing it to break into small pieces and increasing the reactivity of the coke to C02.
I This increased reactivity can again reduce the temperatue of the furnace and decrease
the heat efficiency of the whole system. The high concentration of alkalies in the
fuace also effects the strength and reduction characteristics of the iron bearng
I materials, causing dramatic swellng and catalyzing carbon deposition on the pellets.
These deleterious reactions with both the coke and the ore can have serious impacts on
the gas permeabilty in the fuace and on the stability of the blast fuace operation.
I Fortately, the alkali oxides are very basic oxides and can be fluxed with SiOi in acid
slags and removed from the furnace. Generally, decreasing the slag basicity can car
increasing amounts of alkali away in the slag. This is in direct contrast to sulphur
n removal, where increasing the slag basicity increases the sulphur removaL. When most
ï de sulphurizing took place in the blast fuace, there was a confict between the
attinment of low sulphur and removal of alkalies and the basicity of the fuace was
carefully controlled to balance both problems. With external desulphurization, this is no
longer a problem and the fuace can generally be burdened to minimize alkali attck.

Zinc normally originates in steelmakng off-gas dust from furnaces using galvanzed
scrap which in some fashion has been recycled to the blast fuace. Occasionally, the
zinc content of iron ores or coal ash may also be a signficant source. Behaving not
unlike sodium, zinc is reduced from the oxide or ferrte at about 600 C, forms a vapour
that subsequently forms oxides or carbonates that can react with the sidewalls or he
caried down the furnace on coke or ore to be reduced and fuher cycled, consuming
coke at each tu. Zinc that escapes as a fue in the gas stream enters the blast fuace
filter cake, makng it unsuitable to recycle if present in a high enough percentage.
Unlike the alkalies, zinc is not captued to any extent in the slag and can only effectively

3-13
 J

be removed by decreasing the input and allowing the recycling vapour to slowly leave J
via the gas phase.

Clearly, the best protection against alkali metals and zinc is to ensure that the absolute J
minimum are par of the blast fuace feed. Because of the tendencies of these elements
to circulate in the fuace, they are unseen and unown consumers of coke and cause
refractory, ore and coke problems. Unfortately, the symptoms of the problem are not J
always evident until the problem is of fairly major proportions and then requires fairly
draconian measures, such as eliminating certin feed materials, to effect a solution.
)
Titanium and Lead

Lead is seldom a problem in blast fuaces but occasionally enough can enter a blast ~
fuace through the ore or sinter to cause a problem. Lead is very easily reduced in the
iron blast fuace and falls to the bottom of the hearh which normally has a chilled hot
meta11ayer which protects the hearh refractories. Lead has virtally no solubility in the ~
hot metal so it forms a low melting point liquid pool on which the hot metal floats, and
thus promotes more rapid hearh attack. In certin fuaces where this problem is
known to occur, a second tap-hole, deeper than the iron notch, can be used to I
periodically tap the lead.

Titanum is an even more stable oxide than silica but in the blast fuace it can form I
extremely stable carbides and nitrides. These titanium compounds, if present in small
quantities can be effective in forming a light protective layer on the hearh surfaces and
prolonging hearh life. For this reason, especially in Japan, titaiferrous ores are added I
judiciously to sinter mixes. However, at high concentrations, these same compounds
can stiffen the slag while building up a heavy hearh layer, reducing the hearh capacity
of the fuace. As with zinc, the best solution is to reduce the input and slowly I
eliminate the titaum from the fuace.

CORRLATION OF SULPHIDE AND ALKALI CAPACITIES WITH OPTICAL

I
BASICITY OF BLAST FURNACE SLAGS

Optical basicity is a relatively new concept which provides a good foundation for a
r
better understading of the behaviour of molten slags than the conventional basicity ~-
ratios. The simple (CaO/SiOi) ratio ignores the effects of other oxides and, as indicated
in the chapter devoted to blast fuace slags, the relationship (CaO + MgO/ Ah03 +
SiOi) implies that lime and magnesia behave as equivalent basic oxides and that
alumina and silica have the same degree of acidity, neither of which is the case.

The concept of optical basicity was developed by glass scientists(6) and introduced to the
metallurgical community by Sommervile and co-workers in the late seventies. This
approach has proved to be a valuable tool for designing slags or fluxes which wil have
the required characteristics in terms of, for example, sulphide capacity, phosphorus
capacity, magnesia capacity and even viscosity. (7-1i). Details of the method used to
calculate the optical basicity of molten slags are provided in the appendix.

3-14
 l

I
The relationships between the composition of lime-silica and lime-alumina slags with
respect to optical basicity are shown in Figure 11. From this diagram it is clear that
i lime-alumina slags have a greater basicity than lime-silica slags and therefore would be
expected to have a higher sulphide capacity. As can be seen from Figue 12, where
sulphide capacity data for several slags systems are plotted against the mole percent of
I basic oxide, this is indeed the case. This type of information is often plotted against
the lime-silca ratio. With either method of plotting, there is a separate line for each slag
system.
J
On the other hand, if the sulphide capacity of slags is plotted against the optical basicity,
the behaviour can be represented by a single line, Figure 13. This is because the optical
J basicity parameter is a more fudamental measure of the slag behavior.

Figue i 4, shows lines of iso-sulphide capacity for lime-alumina-silca slags at 1600 C

I i.e. 2910 F. From this diagram it can be deduced that a binary slag consisting of 50
percent lime and 50 percent silica has a sulphide capacity of approximately 5xl0-4, On
the other hand, a binar slag consisting of 50 percent lime and 50 percent alumina, has a
I sulphide capacity of approximately 8xlO-3, which is sixteen times greater than the
sulphide capacity of the equivalent lime-silica slag. Thus, a lime-alumina slag is a more
effective desulphurzing agent than a lime-silica slag. However, to enhance alkali
I absorption, a lime-silica slag would be more effective than a lime-alumina slag. Ths
aspect is discussed in more detail below.
I As discussed previously, the formation of volatile species associated with sodium and
potassium have adverse effects on the fuace operation due to refractory attack,
I generation of fines, accretion formation and decreased burden permeability. Problems
of this type are accentuated when fuaces operate with higher driving rates, increased
flame temperatues, lower slag volumes and relatively high basicities. Our
understanding of these phenonmena has been greatly enhanced both by laboratory
~ studies and results from plant operations. Major contributions to this field have been
r- made by W-K. Lu and his co-workers.(13,14)

Figure 15 and Figue 16 show the relation between the KiO solubility in slags of blast
fuace composition and slag basicity defined in terms of the (CaO/SiOi) ratio and
optical basicity, respectively. The data were derived from equilibrium experiments on
the KiO solubility in slags of blast fuace composition caried out at 1500° C by
KarsrudYS), In these Figures, there are two data points representing slags with a high
basicity containing 50% CaO, 49% Ah03 and 0.35-0.40% SiOi. With slags of these
compositions it is clearly not appropriate to use the ratio of CaO to SiOi to characterize
the basicity. Using the ratio of CaO to Ah03 to express basicity is possible, however it
is really not accurate to assume that Ah03 is equivalent to SiOi in terms of acidic
behaviour. Comparing Figure 15 with Figure 16, it is evident that the optical basicity
approach provides a more reasonable expression of slag basicity than the simple
CaO/SiOi ratio.

3-15
 From the experiments conducted by Karsrud. (IS) it is possible to calculate the K20
capacity of the slag, defined as follows:

1 J
2K(g)+-02 (g) = KiO(slag ) (30)
2

a
J
K I -(p~.p~:)
K,O (31)

J
CKiO
_(Wt%K20)_~
(i 05) + (32)
PK'PO~ j KiO ~

I
As shown in Figure 17, the alkali capacity decreases, with increasing optical basicity.
This behaviour can be represented by the following equation which is valid for a slag
temperatue of 1500 C:
I
Log CK20 = -11.57 A + 13.43 (33) I
Included in Figure 17, is a line showing the dependence of sulphide capacity on slag
optical basicity obtained from the work of Sosinsky and Sommervile (7). In contrast with I
the behaviour of alkalies, the sulphide capacity increases, with increasing optical
basicity:
I
Log Cs2- = 12.60 A - 12.30 (34)

r
It will be evident from this discussion, that with the optical basicity model, an optimum
slag composition can be designed in order to meet paricular operating requirements in r
terms of alkali removal and/or sulphur removaL. It should also be noted, that since the
stability of oxides increase with decreasing temperatue, operating at a lower
temperatue rather than at a higher temperatue, will improve the recovery of alkali
oxides within the slag phase.

CONCLUDING COMMENT

In his 1987 Extractive Metallurgy Lectue, Professor Julian Szekely reviewed the state
of extractive metallurgy and its important place within the national economy.(16) In this
excellent paper, he emphasized the fact:

3-16
 I

I "Both process optimization and process control require a quantitative

representation of the process."

I He also stressed the concept:

"Calculations and measurements are not alternatives, but most often must be
:~I pursued in a complementary fashion. "

Professor Szekely went on to say:

J
"The main barrier to the implementation of these concepts tends to be the
nonavailabilty of suitably trained personneL."
~
The continued existence of this Blast Furace Course at McMaster University

I represents a major contribution to the training of individuals equipped with the

knowledge and understanding of the importance of measurements and process models
both of which are essential for the control and optimization of ironmaking operations.
I ACKNOWLEDGEMENTS

I Acknowledgements are due to Professor T.R Meadowcroft, Deparment of Metals and

Materials Engineering, University of British Columbia who presented the lecture on
Blast Furnace Reactions at the 1998 Ironmakng Course. In the present lectue, the
I sections: "Reduction of Other Oxides", "Alkalies and Zinc", as well as "Titaum and
Lead" have been reproduced in their entirety from Professor Meadowcroft's 1998
lectue. In addition, some material originally prepared by the late Professor J.F. Elliott
I for the lecture on "Principles of the Iron Blast Furace" when ths course was first
offered has been incorporated within the text of this chapter.

r REFERENCES

I (1) H.M.Howe, "The Metallurgy of Steel," The Scientific Publishing Company, New
York, 1890.
(2) RJ.Brigham and Y.A.Lafreniere, "Titaic Specimens," Metals Technology Labor-
atories, CANMET, Ottwa, Canada, Report No.92-32 (TR), 14 pages.
(3) D.RGaskell, "Introduction to Metallurgical Thermodynamics," 2nd ed.,
Hemisphere Publishing, New York, 1981.
(4) W-K.Lu, "Silicon in the Blast Furace and Basic Oxygen Furace," Iron and
Steelmaker, VoL. 6, No. 12, 1979, p.19.
(5) E.T.Turkdogan, "Blast Furace Reactions," Met. Trans B, Vol.9B, 1978, p.163.
(6) J.A.Duffy and M.D.Ingram: "Establishment of
an Optical Scale for Lew's Basicity
in Inorganc Oxyacids, Molten Salts and Glass," J. American Chemical Society,
December 1971, pp. 6448-6454.
(7) D.J.Sosinsky and LD.Sommervile, "The Composition and Temperatue
Dependence of the Sulphide Capacity of Metallurgical Slags," Met. Trans. B,
Vol.17B, 1986, pp.331-337.

3-17
 J

(8) D.J.Sosinsky, I.n.Sommervile and A.McLean, "Sulphide, Phosphate, Carbonate :J

and Water Capacities of Metallurgical Slags," Fifth International Iron and Steel
Congress. Process Technology Proc., ISS-AIME, Vol.6, 1986, pp.697-703.
(9) A. Bergman, "Some Aspects on MgO Solubility in Complex Slags," Steel J
Research, Vol.60, No.5, 1989, pp. 191-195.
(10) I.D.Sommervile, "Optical Basicity as a Control Parameter for Metalurgical
Slags," Advanced Materials-Application of Mineral and Metallurgical Processing J
Principles, SME-AIME, 1990, pp.147-159.
(11) R.W.Young, J.A.Duffy, G.J.Hassall and Z.Xu, "Use of Optical Basicity Concept
for Determining Phosphorus and Sulphur Slag-Metal Paritions," Ironmaking & J
Steelmaking, VoL. 19, No.3, 1992, pp. 201-219.
(12) Y.Yang, A.R.McKague, I.D.Sommervile and A.McLean, "Phosphate and
Sulphide Capacities of CaO-CaCh-CaF2 Slags," Canadian Metallurgical ~
Quarerly, VoL. 36, No.5, 1997, pp. 347-354.
(13) W-K.Lu, "Fundamentals of Alkali-Containing Compounds," Proceedings of
Symposium on Alkalies in the Blast Furnace, McMaster University, 1973, pp. 2-1 a
to 2-18.
(14) W-K.Lu, and J.E.Holditch, "Alkali Control in the Blast Furace: Theory and
Practice," Blast Furace Conference Proceedings, ArIes, France, June, 1980. i
(15) K.Karsrud, "Alkali Capacities of Synthetic Blast Furace Slags at 1500°C,"
Metallurgy, VoU3, 1984, pp. 98-106.
Scandinavian Joural of

(16) J.Szekely, "The Mathematical Modeling Revolution in Extractive Metallurgy,"

I
Met. Trans B, VoL. 1 9B, 1988, pp.525-540.

I
ADDITIONAL SOURCES OF INFORMTION

Sul!l!ested bv Professor T.R. Meadowcroft

I
1. Ellott, J.F., Gleiser, M., Ramakshna, V., "Thermochemistry for Steelmakng,"
Volumes 1 and 11, Addison-Wesley Publishing Co, Reading, Mass. U. S. A.,
I
1963, now out of print but many copies in various steel companes. Stil an
excellent source of data in very comprehensible form. r
2. Thompson, W.T.*, Pelton, A.D.o, Bale, C. W.o, "Facility for the Analysis of r--

Chemical Thermodynamics," (The FACT System), Interactive Computer

Softare available through the authors at *Royal Milita College, Kingston,
Ont., and ° Ecole Polytechnque, Montreal, Quebec.

3. "HSC Chemistry," vers 1. 10, Outokumpu research Oy, Pori, Finland, a very
easy to use softare package for PC's with an excellent data base for iron and
steelmaking.

4. Stadish, N., Lu, W-K., "Alkalies in the Blast Furace," Proceedings of the 1973
McMaster Symposium, stil the best collection of aricles on this subject.

3-18
 J

I APPENDIX

CALCULATION OF OPTICAL BASICITY OF BLAST FURNACE SLAG

J
i Optical basicity of the molten slag is calculated using the following equation:

i L. i I
A = " A ,N,
i=1
(1)

~ N,i= nXinOi (2)

¿XinOi
i=1

i Here A: Optical basicity of the slag

Ai: Optical basicity value of component "i "
I N¡: Compositional fraction
Xi: Mole fraction of component "i " in the slag

I lli: Number of oxygen atoms in component "i"

Optical basicity of the slag is calculated by the following procedure:

I 1) Select the optical basicity value Ai for each component of the slag. The optical
basicity values for several oxides are given in Table 1.

I
Table 1. Optical Basicity ValDes of Varioos Oxides (6)
~ Oxides K20 Na20 CaO FeO MgO MnO Ah03 Si02 Ti02
l' Ai 1.40 1.15 1.00 0.51 0.78 0.69 0.61 0.48 0.61

2) Calculate compositional fraction N¡ using equation (2).

3) Calculate the compositional optical basicity value for each component of the slag
using the relation: Ai Ni.

4) Calculate the sum of the compositional optical basicity values using equation (1).

An example of the calculation process is outlined in Table 2.

3-19
 J

Table 2. Calculation of Slag Optical Basicity J

Items CaO MgO Ali03 Si02 Sum
Optical basicity of oxide, Ai 1.00 0.78 0.61 0.48 - J
Number of oxygen atoms, 1 1 3 2 -
Ili
Molecular wt. of oxide, M¡ 56 40 102 60 - J
Composition, wt% i 47.52 2.90 12.07 37.51 100
Number of moles, 0.8486 0.0720 0.1183 0.6252 1.6641 J
n¡ = (wt% i)/Mi
Mole fraction, Xi=n¡ /¿ni 0.5099 0.0432 0.0711 0.3757 1.0000
g
Compositional parameter, 0.5099 0.0432 0.2133 0.7514 1.5178
X¡.Ili
Compositional fraction, 0.3359 0.0285 0.1405 0.4951 1.0000 I
N¡= X¡Il¡/¿X¡noi
Compositional optical 0.3359 0.0222 0.0857 0.2376 0.6814
basicity, Ai N¡ I
Example: l1aO = (wt% CaO)/Mcao= 47.52/56 = 0.8486
I
XCaO = l1aol(l1ao+nMgo+nA103+ns¡02) = 0.8486/1.664 = 0.5099

XcaO.Il in CaO = CaO mole fraction. the number of oxygen atoms in CaO I
= 0.5099xl = 0.5099
I
XS¡02.nOinSi02 = 0.3757x2 = 0.7514

NCao= 0.5099/1.5178 = 0.3359 I

NSi02 = 0.7514/1.5178 = 0.4951
r
Acao .Ncao =lx0.3359 = 0.3359 1-

ASi02 NSi02 = 0.48x0.4951 = 0.2376

A = Ncao. ACao + NMgO.AMgO + NAl203.AAl203 + NSi02. ASio2

= 0.3359xl + 0.0285xO.78 + 0.1405xO.61 + 0.4951x0.48

= 0.3359+0.0222+0.0857+0.2376

= 0.6814

3-20
 ')

J
:i

J
~

I
I
Fig. 1 The Titanic prior to departure on her maiden voyage.
I
I
I
I
¡i

Fig. 2 The sinking of the Titanic in the North Atlantic off

Newfoundland as depicted by artist W. Stoewer.

3-21
 J

J
Fig. 3
Sample of Titanic hull plate
showing location and orient- )
ation of test specimens. (2)

I
Fig. 4
Microstructure of section from the
hull plate showing carbon banding
I
and MnS stringers elongated in the
rollng direction. (2) I

I
80,
'ã
(l
70 r Mild steel / Acier doux I
j - 0,18% C
3, i
o 60 ~
- 0,54% Mn
!: 50~
(! ¡
a: :
- 0,07% Si r
Z
W I'
W 40'"
("-
.. 30r
)- !
g 20~
W'
~ 10r ~itaniC
0'
-40 -20 o 20 40 60
TEMPERATURE ( C)

Fig. 5 A comparison of the results of Charpy tests on

specimens of plate steel from the Titanic with
a similar steel produced in the early 1950'sY)

3-22
 I

I
Hz/HzOratia
10-5
CO/COz ralio - .
o
J 0 10' 10' 10" 10" 10-4

!/ 7 7
J
-10 0 /
~ Ll ..riV =:.
I
~ v- --.
..vc../ "' i.O"'~
~~.'! !__
,"
-20 0
J
~
-30 . :: l.~ ~ "",.
~.0.,"
fJ --- /1/ v~ 01.'~~"°.-
-f- CIl"01.S
O~~i'k"~"'!~~ 0"
~ ./ O'! ""Vi; I~~ ~
,,~
o_.si.eaO
~..01. ~~~.=~
Mm
~
I 0"

I 0"

4:
1- -,0."
¿ -/I~.... v 0 C+Oz'
ife" ifeO COz
I H
-400

-50
~.. ~--~?~~V
__ __ ;r.'!04.. 7
V .o~. ~c ' V m
o \f.:.~ Ai.. /v 7..¡-,¿.. ..
I

1
10310
Z

103
..
.. ~ ~~ ~~ /~V m~~~~~ 104

I C
S
;;0
~rui ß.; ..~c~~",I\o __,," ..~ J::
/"..~ / ..V
i

:o -600 104
Ñ 10'

I
rP
.!
l-
e: -700
./ ~~ illI\ · ...-
.... "/
~~~ ,I ..\o~..
.~c¡\i..c¡'~
~ ",
.. ~ ') .- / ..:; ;. ~
/
/
I

/
"

.. .. ..
i
16

I (:
-800
,r~
.. V V
.~ V // /.- .. 7 i O.
106

I -900 . p.\ .. 0;;

..~
-;
p.\,,O'!

/
/ .. / i

~ -1000
..
/ V
.-

~ .. -'
/.. .. ~V
¡: b

,
I

- ° ?
I

r
-1100
..
"" it 01. 'lCO V
'l"'~"'V
m-

'l"'~ 0;: em melting point of metal I

i. 'lC~
VV
l! b boiling poini of metol
.M melting poinl of oxide
-1200 I

I I I 1 1
o 200 400 Temperoture,.C
Paz (atm) 10.100

CO/COz rolio
Hz 1HzO rolio

Fig.6 Ellngham Diagram (adapted from Gaskeli3) for the effect of temperature
on the standard free energy of formation of oxides, including the
nomographic scales produced by Richardson. The diagram has been
modified to include the behaviour of phosphorus, potassium and sodium. ~

3-23
 J

l J

~ J
a''
=
ou J
~
~
....
~
~
2
~
~
100 20 300 .c, 50 800 700 10 10 1000 1100 12001:, 140
Teil"re (deg C~)
I
Fig. 7 The effect of temperature on the CO content of a
CO/C02 gas mixture in equilbrium with carbon
for total pressures of 1 and 3 atmospheres. I

B
I

i I
II
:2
s I
:I
o ~
u g
~'
r
2 T"-

(:
fife
20 40 &0 80 io 12 14
T emprëlure .C

Fig. 8 The effect of temperature on the CO and CO2

contents of gas mixtures in equilbrium with
carbon and various iron oxides.

3-24
 J

~ o
J
i Iron

J ~
~..
:r a~
) ¡ '~ 'X ..

~ oCD
~
~ 40 eo eo 100 IZO 1400
I lemperature .C

I Fig. 9 The effect of temperature on the H2 and H20

contents of gas mixtures in equilbrium
with various iron oxides.

I
I
100
I tI \ \ Japanese
80 y\ German
N
0u
~ ø.
60
+ FeO
r 0u 40 ~- - JJ ....-
..-0 ---
ø. .. ...-.. ....

20 Fei04 /
u
ø. ,,//
.. ,
0
200 400600 800 1000 1200
Temperature, °C

Fig.l0 Actual gas compositions at various temperatures

based on samples taken from operating furnaces,
in comparison with the conditions for equilbrium
either with carbon or with different iron oxides.

3-25
 j

I
I
I

I
I
I
1'-

Fig. 11 Optical basicity values for lime-silca and lime-alumina slag systems.

3-26
 J

J
1

i
~

i
I
I
I
I
~
1-

Fig. 12 The effect of basic oxide content on the sulphide capacity of different ,
slags at 1500C.

3-27
 1

-l

J
i.o . CaO-AI203
. CaO-Si02
o CaO-Ali03 -SiOi
 CaO-MgO-Ali03 J
Å CaO-SiOi-8iÜJ
o CaO-MgO-SiOi
o CaO-MgO-AliÜJ -SiDi ~
10
o I
UVI
-ot:
I
I
4.0
I

5D I
T :15O.C

m
.$ .60 ;6 .70 .15 .eo .85
~n
" . Oøtical Baicity

Fig. 13 The relationship between the sulphide capacities of

different slag systems and optical basicity at 1500C

3-28
 I

J
1

)
~

I
I
I
I
I
I
, CoO AllOi

Fig. 14 Iso-sulphide capacity lines derived from optical

, basicity calculations for liquid slags in the
lime-alumina-silca system at 1600C.

3-29
 i

_J

8
Equilbrium J
7
- 6
~
T= 1500 °c
J
-~
?F

..~
r:
5 (ò
..r:
Q) 4 J
0()
3
0N ~
~ 2
1 a

0
0
o 1 2 3 4120 130 140 150 i
CaO/Si02
Fig.15 Dependence of K20 solubility in SF slag on CaO/Si02 ratio at 1500°C. I

8 I
Equilibrium
- 7
T = 1500 °c
I
-~~
?F 6
5
~8
(( I
r:
..
Q)
r: 4
0
() r
0N 3
~ 2
F'-

o
0
0.64 0.66 0.68 0.70 0.72 0.74 0.76 0.78
Optical basicity
Fig.16, Dependence of K20 solubility in SF slag on optical basicity at 1500oC.

3-30
I
J

,)

J
J
I
I
I
I
I
~

Fig. 17 The alkali and sulphide capacities of slags

at 1500C as a function of optical basicity.

3-31
J

r
f"
 l

I LECTUR #4

I
BLAST FURACE ENERGY
J BALCE AN RECOVERY
RULES OF THUMB AN OTHER
i USEFUL INFORMTION

J John W. Busser
Supervisor
Ironmking System
Steleo Ine.
~ Haml ton, Ontario

I Abstract: Simplified mass and energy balances are outlined

for the purpose of optimising blast furnace operations. A
I sumary of useful blast furnace related data from numerous
sources is presented. Tuyere zone, stack and general blast
furnace reactions are reviewed from an energy standpoint. The
impact of variability in blast furnace input parameters is
I discussed. 'Rules of Thum' relating furnace raw material and
practice changes to energy consumption are reviewed. These
principles are demonstrated through a computer simulation
I model "The Blast Furnace Game" that uses mass, energy,
chemical and cost balances to assess means of improving the
blast furnace process.
I
INTRODUCTION
~;

In order to make changes to the blast furnace process that

will meet their intended goals, the Ironmaker must have an
understanding of his process, his facilities, and the costs
associated with each of the process changes. Although rules
of thum can be applied, each situation is somewhat different
due to the variety of constraints and physical limitations
that apply to individual furnaces. The purpose of this paper
is to promote an understanding of the process. In this
regard, many of the concepts presented have been simplified,
and some liberties have been taken in estimating data.

Models, in general i do not have to be exact in order to be

useful. What is most important is that the direction and the
magnitude of change can be predicted. This, in essence, iB
why rules of thum have been developed.

4-1
 EXHIBIT 1 BLAST FURNACE ENERGY BALANCE J
(MMBTU PER NET TON OF HOT METAL)
:J
ENERGY RECYCLED ENERGY
INPUTS ENERGY OUTPUTS
BFG is USED FOR STEAM BFG EXPORTED 1.0 ~J
FURACE COKE 12.7 TO DRIVE BLOWING ENGINS IRON PRODUCED 7.5
INJECTED FUEL 1. AN PUMPS FOR COOLING TOTAL LOSSES 5.9
--- WATER AN ALSO TO HEAT --- J
TOTAL INPUT 14.4 HOT BLAST STOVES TOTAL OUTPUT 14.4

EFFICIENCIES LOSSES CONSUMD ~

FUACE FUL 63% TOP TEMP LOSS 0.3 TOTAL OUTUT 14.4
STOVE HEATING 71% DUST BTU LOSS 0.1 LESS BFG
BLOWER STEAM 9% TOP BFG LOSS 0.1 EXPORTED 1.0 ~
FE YllLD 96% STOVE ENERGY 2.1 ---
STEAM ENERGY 2.1 CONSUMED 13.4
PROCESS 56% FUACE COOLING 0.3 I
SLAG BTU LOSS 0.3
CALCINATION 0.3
OTHR LOSSES 0.3
I
---
TOTAL LOSSES 5.9
I
INPUTS OUTPUTS
TOP TEMP 0.3 I
DUST LOSS 0.1 EXCESS
BFG LOSS 0.1 BFG TO
PLA NT 1.0 I
STOVES 2.1
STEAM 2.1
FE BURDEN
CONVERSION TOTAL ENERGY
I
COKE 12.7 LOSSSES AT FOR CONVERSION
FL UXES ir
STOVES 0.6 TOP LOSSES;
BLOWERS 1.9 COOLING LOSS;
CALCINA nON;
TUYERE HEA T FROM SENSIBLE HEA T
INJECTED 1.7 jlh .. STOVES 1.5 IN SLAG;
¡ , BLOWERS 0.2 OTHER LOSSES;
FUEL TOT AL 5.9
COOLING 0.3 IRON 7.5
SLAG TEMP 0.3
CALCINE 0.3
TOT AL 14.4 OTHER 0.3 TOTAL 14.4

4-2
 I

I An Energy Balance

The blast furnace process is a significant consumer of

I energy, consuming about two-thirds of the total energy
required for an integrated steel plant. The blast furnace
typically consumes all the coke produced by the coke ovens,
as well as some additional inj ected fuels, in the production
~I of hot metal for steelmaking.

As shown in Exhibit 1 about 14.4 million BTU of energy is

J required to make a ton of hot metal. This energy is provided
mainly by coke and supplemented by inj ected fuels such as
natural gas, oil, tar or pulverised coal. All of these fuels
~ are burned in the raceway of the furnace with a limited
amount of air to provide the reducing gases for the iron ore
smel ting process. Since these fuels are not burned to
i completion, significant amounts of by-product top gas or
Blast Furnace Gas (BFG) is produced.

Of the total top gas energy produced of 5.3 million BTU/NTHM,

i about 2.1 million BTU/NTHM or 40 percent is used in the blast
furnace stoves to preheat air for the blast furnace. The
stoves are fairly efficient recycling more than 70 percent of
I the energy or about 1.5 million BTU/NTHM to the blast furnace
process.
Another 40 percent or 2.1 million BTU/NTHM is used to make
I steam to drive blast furnace blowing engines and cooling
water pumps. Of this amount, only about 10 percent or 0.2
million BTU/NTHM is recovered in the heat of compression of
I the cold blast air that is returned to the blast furnace.

Only about 20 percent of the total top gas is available for

export to the rest of the steelworks. This represents less
I than ten percent of the energy that was provided to the
~ì
process.
There is some fuel energy loss from the process, including
the fuel energy in the dust that is carried out of the
furnace, and the BFG that escapes during the raw material
charging opera tion .
Finally, there is also a significant (about 4 percent) loss
of iron from the process. Iron is lost through poor iron/slag
separation, through runner and other scrap losses, and in the
form of iron bearing dust exiting the top of the furnace.
Iron yield loss has an impact on furnace energy performance,
increasing the energy required per net ton of hot metal
produced.
Of the total energy input of 14.4 million BTU/NTHM that is
provided to the blast furnace process, only the BFG that is
exported to the plant (1.0 million BTU/NTHM) is for non blast

4-3
 i

furnace use. The remaining 13.4 million BTU/NTHM is all i

directed in some manner toward the heating and reduction of
iron. Since the actual iron reduction process requires only
7.5 million BTU/NTHM of energy i the difference of 5.9 million
J
BTU/NTHM is lost in the conversion process. Since the energy
required to reduce iron of consistent chemistry is constant,
this loss becomes directly related to fuel rate. As a
process i the blast furnace shown in the model is only about J
56 per cent energy efficient. (i. e. 7.5 million BTU /NTHM
divided by 13.4 million BTU/NTHM)
J
Due to the tremendous amount of energy required to make hot
metali much attention has been given to reducing blast
furnace fuel rates i reducing heat losses, and to recovering
energy from the blast furnace process. Furnace fuel rates ~
are the result of the sum total of the energy demands of the
process, and can be viewed from several perspectives.
~

The Reduction Process

The purpose of the blast furnace is to reduce oxides of iron
I
and to melt them for subsequent refining in steelmaking. The
main reactions are reviewed here from an energy perspective.
The energy requirements have been developed from Standard I
Heat of Formation data.
HEMTITE Fe203 (s) + 294 BTU /LB Fe -- Fe304 (s) I
MAGNETITE Fe304 (s) + 821 BTU /LB Fe -- FeO (s)
I
WUSTITE Fe ° (s) + 2056 BTU/LB Fe -- Fe (s)
IRON Fe ( s) + 600 BTU/LB Fe -- Fe (l) I
TOTAL Fe203 (s) + 3771 BTU/LB Fe -- Fe (l)
I
In total about 7.5 million BTU/ton Fe is required to convert ~.
hemati te to liquid iron. The actual amount of energy required
to make hot metal will differ somewhat based on incoming raw
material and resultant hot metal chemistries. A typical hot
metal chemistry is shown below.

Composition of Hot Metal

Iron 93. ° %
Carbon 3.9 g,o
Manganese 2. ° %
Silicon 1. ° %
Phosphorus o. i %
Sulphur 0.04 %

4-4
 I

1 The main blast furnace reduction reactions (plus the carbon

reaction) are shown below. The carbon content of hot metal is
determined by solubility of carbon in hot metal and is
I consistent at a given temperature. The reduction energy
requirement for manganese is similar to iron but far more
energy is required to reduce silica than iron ore.
cl IRON Fe203 + 317l BTU/LB Fe -7 Fe
CARON + 14093 BTU/LB C -7 C
J CO2

MAGANSE Mn°2 + 4077 BTU/LB Mn -7 Mn

~ SILICON Si02 + 13490 BTU / LB Si -7 Si

l PHOSPHOROUS P20S + l0452 BTU/LB P -7 P

SULPHUR S02 + 3991 BTU /LB S -7 S

I Energy Inputs

I The energy inputs to the process are twofold. The first is

the chemical energy content of the fuels and the second is
the sensible heat of the hot blast. Energy contents of
I various fuels are shown below:

Coal l3,600 BTU /LB

Coke BTU / LB
I Tar
Oil
12,8 ° °
16,800
18,700
BTU / LB
BTU /LB
Natural Gas 23,800 BTU /LB
I The sensible heat of the hot blast air is provided by stoves,
i ¡r' which are about 70 percent efficient in converting top gas
fuel energy into hot blast energy. Rough estimates of
specific heats and densities for solid and gaseous materials
are shown in the following tables. 1,2

SOLID Specific Heat Density

MATERIALS BTU/LB/oF LB/ SCF

Wa ter 1. ° 62
Sinter 0.2 100
Pellets 0.2 145
Iron 0.2 424
Slag 0.3 206
Silicon 0.2 l45
Oil 0.4 60
Tar 0.4 75
Coal 0.4 65
Coke 0.4 35

4-5
GASEOUS Specific Heat Density j

MATERIALS BTU/LB¡OF LB / SCF

Air 0.26 0.076 J

Ni trogen 0.26 0.074
Oxygen 0.26 0.085
J
Carbon Monoxide 0.25 0.074
Carbon Dioxide 0.23 0.117
J
Tuyere Reducing Gas 0.27 0.069
Blast Fce Gas (Dry) 0.26 0.081
~
Hydrogen 3.50 0.005
Natural Gas 1. 2 0.042
Stearn o . 6 * for comparison 0.044 * ~
** g 212 degrees F 0.037 **

I
Energy au t:Du t s
Fuel energy inputs to the blast furnace not consumed in the I
process exit in the form of top gas. The BTU heating value
and volume of the top gas decrease as the furnace becomes
more fuel-efficient. The volume of top gas produced can be
calculated from the specific wind rate (SCF wind /NTHM) by I
using a nitrogen balance. Nitrogen is inert in the process;
hence the volume of nitrogen remains constant. A typical
BFG/Wind ratio is calculated as follows: I
(79% Nz in Air / 55% Nz in BFG) = 1.44
I
Energy Losses

Losses from the process result primarily from losses in the I

conversion of energy i. e. wind generation, cooling water
pumping i and hot blast stove heating. Il

Blast Furnace Gas generated by the blast furnace process is

typically used to make stearn to drive turbines for blowing
engines and water cooling pumps. Most of the energy, however,
is lost along with the condensate from these turbines. The
only heat recovered from these two processes is in the form
of the sensible heat of compression of the hot blast air i
which is in the order of 300 degrees F depending on the blast
pressure. This represents only about 10 percent of the fuel
energy input to the boilers i an area of significant energy
loss.

4-6
 J

I Blast Furnace Gas is typically enriched with Coke Oven Gas or

Natural Gas to fire the blast furnace stoves. The stoves are
used to further heat the compressed blast air to the normal
J hot blast temperature of around 1900 degrees F. Since the
stoves are reasonably combustion efficient i the stack and
cooling losses are limited to about 30 percent. The major
loss of energy is due to the stack loss associated wi th was te
,J gas exiting the stoves at about 600 degrees F.

Cooling losses result mainly from water cooling members in

J the hot blast stream in the walls of the furnace. Cooling
members in the hot blast stream are devices such as hot blast
valves and tuyeres. Cooling members in the furnace walls can
I be either stack plates or staves, or external water sprays,
depending on the design of the cooling system. All of these
devices can remove a great deal of heat from the process.
~ The next maj or group of losses are in the form of the
sensible heat lost in the slag, top gas, and radiant heat
from the casting process etc..
I Other losses include the heat of reaction required for
calcination, the chemical energy lost in flue dust and filter
I cake, and top gas losses from top filling equipment.

It should be noted that Higher Heating Values are used in

calculations to avoid double counting moisture losses.
I (The HH measures the total heat release cooling the products
of combustion to 60 of that includes the latent hea t 0 f
vaporization of water.)
I
Sumry of Losses MMTU/NTHM
I Top Temperature Loss 0.3

r Dust BTU loss O. 1

Top BFG loss 0.1

Stove Fuels 2.1
Blowing /Pumping Fuels 2.1
Furnace Cooling Loss 0.3
Slag BTU Loss 0.3
Calcination Reaction Loss 0.3

Total Losses 5.9

4-7
 J

Hot Blast Temperature J

Prior to the 19th century, all blast furnaces operated with

air at ambient temperature with the tuyere zone of the
furnace serving a dual purpose. Firstly, it provided the J
reducing gases for the reduction process, and secondly, it
provided the heat to drive the reduction process and melt the
hot metal and slag. J
In 1828, James Nielsen found that an elevated blast
temperature had a remarkable effect on furnace performance. J
Energy introduced to the process via the hot blast was energy
that did not have to be provided by coke. As the fuel rates
were reduced, the volume of hot blast required was also
reduced, generating diminishing returns, but quite favourable ~
in any event.

As better designed stoves were developed to produce even ~

higher hot blast temperatures i blast furnaces began to
operate erratically due to too much heat being generated at
the tuyeres and further increases were achieved only after
this problem was resolved with the injection of steam into
I
the hot blast. It was later found that hydrocarbons inj ected
through the tuyeres had the same stabilising effect and
seemed to save even more coke than expected. I
The mechanism explaining these developments can be understood
by reviewing the reactions at the tuyeres. I

Tuyere Zone Material Heat Content

I
All materials entering the tuyere zone, both from inside and
outside the furnace must be heated up to the flame I
temperature as part of the combustion process at the tuyeres.

Generally burden materials are preheated by the ascending

gases and do not play a maj or part in determining flame
I
temperature. Hot Blast air and steam inj ected prior to the
stoves temperature must be increased from that provided by If
the stoves up to flame temperature. Inj ected fuels must be
heated from the injection temperature, usually ambient
temperature, up to the flame temperature.

Exhibit 2 displays the heat content of hearth zone materials

and Exhibit 3 displays the heat content of tuyere zone
materials.

4-8
 I

I
EX 2 HE ZC ~ HF a:
140
J
120
LL
,.
J ,. 1(0
W
~
om-lO
i c: m
i- ::
Z~ 1m
W r-
i- m
J Z_
o 40
()
i-
i ~ 20
i:

I
I
o'
o
~IN~~
50 1(0 150 :i 2S :D

I EX 3 'l 2' MM HE C'

12000

I 10000
H2

I 8000
r' Q
~
\ 6000
=
!- MET HAE
Q
4000 PLUS ENERGY
DE~pæES C+2H2
01 ::C+2H2
2000
CH4 H20
C
o
AIR
o 500 1000 1500 2000 2500 3000

TEMPERATURE IN DEGREES F

4-9
 j

Tuyere Zone Reactions J

Energy is introduced into the blast furnace process via two
maj or elements by the following general reactions:
J

Carbon C + O2 -- CO2 + 14093 BTU/LB C

J
C + Y: 02 -- CO + 3935 BTU/LB C
Hydrogen H2 + Y: O2 -- H20 + 61095 BTU/LB H2 J
However, in the tuyere zone, other reactions occur as wind
and inj ected fuels are blown into the furnace. Initially J
there are a numer of step reactions, which absorb heat prior
to the above reactions to produce CO and H2 reducing gas. The
energy absorbed by these reactions serves to moderate the ~
heat (i.e. flame temperature) generated by the combustion of
coke with hot blast oxygen. The energy shown is the heat
required to break up the molecules into reducing gases and
does not include the energy required to heat the injected I
materials up to flame temperatures.
F1ame Temperature Moderating Reactions: I

Steam H20 + 5800 BTU/LB -- H2 + l/ O2 I

Methane CH4 + 2010 BTU /LB -- C + 2 H2
Ethane C2 H6 + l206 BTU/LB -- 2C + 3H2
Propane C3H8 + 1028 BTU/LB -- 3C + 4H2 I
Butane C 4 Hio + 936 BTU /LB -- 4C + 5H2

I
F1ame Temerature
The temperature of the gas leaving the tuyere zone can be
calculated by assuming that the chemical energy of whatever
I
reactions occur in the raceway (complete or incomplete) is
totally converted to heat energy. Thus, the temperature of ~-
the flame in the raceway assuming no heat losses (adiabatic L_

condi tions) is called the Raceway Adiabatic Flame

Temperature.

RAT = Heatinq Value

Sum of combustion product weights x mean specific heats

The American Iron and Steel Institute Technical Committee on

Blast Furnace Practice has adopted the following formula
developed by Naren Sheth of Bethlehem Steel Corporation.

4-10
 J

I
RAT = 2686 + 0.82 (BT) - 23.5 (BM) + 95 (OE)
J -124 (Oil) - 80.7 (Tar) - 8.7 (HW) - 7.0 (AS)

-29.9 (Coal) - (68 + 0.034 (GHV)) (NG)

~I

The definitions of terms are:

i RAT Raceway Adiabatic Flame Temperature, F
BT Blast Temperature, F
BM Blast Moisture, grains/SCF dry blast
~ OE Oxygen Enrichment, (% O2 in blast - 21)
Oil Dry Oil Injection Rate, lb/1000 SCF dry blast
Tar Dry Tar Injection Rate, lb/1000 SCF dry blast
~ Coal Dry Coal Injection Rate, lb/1000 SCF dry blast
HW Homogenizing Water, lb/100 lb dry oil or tar
AS Atomizing Steam, lb/100 lb dry oil or dry tar
I NG
GHV
Natural Gas Injection, SCF/lOO SCF dry blast
Gross Heating Value of Natural Gas, Btu/SCF

I The equation shows that increased blast temperature or oxygen

enrichment will increase the RAFT while greater use of any
hydrocarbon or water will decrease the RAFT. The above
I equation can be simplified by assuming a constant wind rate,
e. g. 42,000 SCF /NTHM. (Note: 7000 grains = 1 pound and the
GHV of Natural Gas is lOOO BTU/SCF). The resultant formula
I can be broken into components as follows:

RAT (OF) = 2686 Constant

I +0.82 * Blast Temp (OF)
in +95 * Oxygen Enr i chmen t (% )

-3.9 * LB/NTHM Steam/Water

-5.8 * LB/NTHM Natural Gas
-2.9 * LB/NTHM Oil
-1. 9 * LB/NTHM Tar
-0.7 * LB/NTHM Coal
Reducing Gas Volume and RAT
Natural Gas has a greater moderating effect on flame
temperature than steam because it is inj ected at ambient
temperatures and requires in the order of 4000 BTU/lb to be

4-11
heated to flame temperature. Natural Gas also completely j
dissociates in the raceway requiring another 2010 BTU/lb.
The Reducing Gas to Wind Ratio exiting the tuyere zones will
be lower than the BFG/Wind Ratio because the Carbon, Oxygen, J
Ni trogen, and Hydrogen in the burden has not been liberated
to form CO, CO2, or H2 gases. If only H2, N2, and CO gases are
present at the tuyeres, the Reducing Gas to Wind Ratio is J
about 1.33. The reducing gas has a specific heat of about
0.27 and a density of about 0.68 lb/SCF.
J
Considering a wind rate of 42,000 scf/NTHM and the reducing
gas parameters above, about 1,050 BTU is required to change
the RAFT of the reducing gas by one degree F. Since Natural
Gas, for example, requires about 6,000 BTU/lb for heating and g
dissociation, it will reduce the flame temperature by just
less than 6 degrees.
~
It should be noted that the AISI coefficient for coal appears
to be low relative to the other hydrocarbon based injected
fuels. The overall flame temperature effect should be the
sum of the effect of dissociating the hydrocarbon and then
I
heating the hydrogen and carbon components. To have a flame
temperature effect less than the heating of pure carbon would
appear to be incorrect. As shown previously, the moderating I
effect of hydrocarbons are related to the size of molecule
involved (i. e. hydrogen to carbon ratio) in terms of the
energy required for dissociation.1 (See Exhibit 4) I
Hydrocarbon H2 (Wt%) C (Wt%) N¡ (Wt% ) lli- / N¡ I
Natural Gas 22.5 69.4 8.1 0.32
Bunker "C" Oil 9.3 88.6 0.3 0.10 I
Tar 7.1 91. 4 1. 1 0.08
Bi twninous Coal 5.0 80.1 0.0 0.06
I
IT
Anthracite 2.8 80.6 0.0 0.03

4-12
I

:1
EXHIBIT 4 FLAME TEMPERATURE VS H2 TO C RATIO
r- 1400
Z
~ c:
l- EFFECT Cf PURE HYDRCGN
~ 1200
I ..Z
~ 1000 -
I o
o
,.
(J
((
~
-i
0:
a: 800
I
CC
W ::
I-
c. ((
W -i ;;
en 600" UJ -:
l- ()
I c:
w
a:
t)
-:
u
f,
::
o 400
CC
:: o
I- ;;
W ;; ,~ -i
C -: :: Õ
~-
c. m
:æ 200
W EFFECT Cf PURE CARBQ\
l-
AISI EFFECT Cf DISSO:IATlO\
o
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
HYDROGEN TO CARBON RATIO

DISCUSSION OF TUYRE ZONE RECTIONS AN RAT

The coefficient for Hot Blast Temperature is 0.82 because the

sensible heat in the DRY hot blast excludes 18 percent of
other materials, (including natural humidity and steam) i
injected at the tuyeres.

The explanation for the benefit associated with hydrocarbon

fuel inj ection versus steam can be found by comparing the
overall reactions occurring at the tuyeres.

When steam is inj ected to moderate the flame temperature it

unavoidably reacts with coke absorbing a great deal of energy
at tuyere level. It also generates a significant coke penalty
because i t involves the carbon in coke in the reaction.

4-13
il

The reaction of steam with the carbon in coke is: J

H20 + C + 3138 BTU/LB Hp -7 CO + Hi I

Wi th hydrocarbon fuels less energy is consumed when compared
with water and the reaction does not involve carbon in coke. I

CH4 + 1¡ 02 + 959 BTU/LB CH4 -7 CO + 2 Hi I

Furthermore, the energy required to dissociate the
hydrocarbon is not a penalty because it is already considered
I
in calculating the Higher Heating Value of the fuel. For
example, the reactions considering one pound of methane are:
I
CH4 + 2010 BTU -7 C + 2Hi
r
L

C + O2 -7 CO2 + 10570 BTU

2Hi + °i -7 2HiO + 15270 BTU

CH4 + 202 -7 CO2 + 2HiO + 23830 BTU

Using hydrocarbons to control flame temperature then has a

double barrel effect. Firstly, it replaces water in the
tuyere zone and eliminates the associated coke penalty.

Secondly, it acts as a replacement for coke because it

provides addi tional reducing gas to the process.

If hydrocarbons are injected at a level greater than required

to control flame temperature then they act only as a coke
replacement.

4-14
 EXHIBIT 5 ici FURNACE ENERGY CONSUMPTION
18

16

:æ 14
J:
l-
-
Z
-
~
l- 12
II
:æ

~
:æ
, ¡ 10
. .
8

6
1977 1978 1979 1980 1981 1982 1983

-- Net Energy -- Coke Energy -. Total Energy

Exhibi t 5 shows the changes that occur wi th varying fuel

injection practices on 'C' Furnace at Hilton Works. The
higher quanti ties of fuel inj ection replace coke on an energy
basis and also increase the top gas recoveries. interestingly
the net energy required to make a ton of hot metal of
consistent analysis remains about the same.

Blast Furnace Efficiency

Since the blast furnace is a reduction process, the objective

at tuyer~ level is to generate reducing gas (i. e. CO and H2 ).
From a combustion standpoint, the objective at the top of the
furnace is for the exiting gases to contain only the products
of complete combustion (i. e. CO2 and H20). Although these
conflicting obj ecti ves make it impossible for the blast
furnace to be efficient from a combustion standpoint, steps
can be taken to make it more efficient. Since the majority
of the fuel is carbon based and hydrogen efficiency parallels
carbon efficiency, only the carbon-based reactions will be
discussed.
The common measure of furnace efficiency is the percentage of
carbon burned to completion in the furnace top gas.

% CO2 x 100 = % Efficiency

% CO- + % CO2

4-15
The main reactions being dealt with, considering one pound of
carbon being burned to completion in step reactions are:

C + Vi02 -- CO + 3,935 BTU/LB C

CO+ Vi02 -- CO2 + 10,158 BTU /LB C

C + O2 -- CO2 + 14,093 BTU/LB C

Considering a furnace that is 40 percent efficient at

converting carbon to CO2, this means that for each pound of
carbon in the process 40 percent burns to CO2 and 60 percent
burns to CO. (It is important that the CO2 produced by the
calcination of limestone/dolomite be deducted prior to these
calculations) .
The energy available to the process is:

o . 40 LB C x 14, 093 BTU = 5, 637

0.60 LB C x 3,935 BTU = 2,361
Total = 7,998 BTU/LB C

This constitutes about 56.8 percent of the energy into the

furnace if burned to completion. Hence, the furnace is 56.8
percent fuel-efficient. A one percent improvement in gas
efficiency would provide more energy to the process:

o . 41 LB C x 14, 093 BTU = 5 , 778

0.59 LB C x 3,935 BTU = 2,322
Total = 8, LOO BTU/LB C

This constitutes about 57.5 percent of the energy into the

furnace, an improvement of about 1.28 percent. Considering
an 1100 lb/NTHM fuel rate this is equivalent to 14 lb/NTHM.

4-16
Variation In Furnace Efficiency

There are a number of factors, which affect furnace

efficiency, and consequently the effective fuel rate of the
blast furnace. Factors in this category include burden
distribution, coke stability and raw material fines. For
ease of analysis the effects of variations in these factors
will be discussed in total.
Variations in furnace efficiency are exhibited as variations
in top gas BTU content specifically the CO/C02 ratio. A
blast furnace that becomes inefficient will generate more CO
in the top gas, robbing the reduction process of energy,
lowering the hot metal silicon.
The "Rule of Thumb" for these changes in efficiency is an
increase of 14 lb coke rate for each 1 per cent decrease in
gas utilisation.

Effect Of Efficiency Variation On Top Gas Energy Content

Top gas BTU content can be calculated from the analysis of

blast furnace gas. The main gases with a fuel value are CO
and H2. (C02 and N2 are inert).

BTU/SCF = 3.25 BTU/% ~ + 3.23 BTU/% CO

In general, H2 and C fuel efficiencies tend to follow one

another (i. e. a more efficient furnace will burn more of both
of these fuels). It should be noted that H2 does not suffer
incomplete combustion. It either burns to form H20 vapour or
exi ts the furnace as H2 in the top gas. Hydrogen forms only
a small part of the fuel to the furnace and its efficiency is
affected by many of the same factors that affect carbon
efficiency.

Tyical Top Gas Analysis

Heating Value
Component Percentage BTU / SCF

CO 22.7% 73.32
CO2 18.5% 0.00
H2 3.8% 12.35
N2 55.0% 0.00
Total 100.0% 85.67

4-17
A 1% increase in furnace efficiency will change both the
percent CO and CO2 in the top gas by 1 percent. Assuming
that the total volume of CO + CO2 remains about the same at
41% of the top gas i the percent CO2 in the top gas will
increase by 0.41%.

Component Before After

CO 22.70 22.29
CO2 18.50 18.91
CO + CO2 41.20 41.20
CO2 / (CO + CO2) 0.449 0.459
BTU Contribution 73.32 72.00
BTU % Change 1. 8%

Assuming that hydrogen efficiency changes at the same rate as

carbon efficiency a one percent increase in CO2 utilisation
will reduce the top gas BTU content by 1.8% x 85 BTU/SCF or
about i. 5 BTU/SCF.

4-18
Effect Of Efficiency Variation On Hot Metal Silicon

The objective of applying the "Rules of Thum" has

tradi tionally been to obtain answers, which are expressed in
terms of "coke" or "carbon in coke". It can, however, also
be used to determine the change in hot metal silicon, given a
constant ore to coke ratio, for a step change in raw
materials or operating practice. For example, the reduction
in hot metal silicon associated with a loss of hot blast
temperature can be both calculated and readily observed.

Using the coke rate formula in this manner, the step cast-to-
cast changes in raw materials and operating practice can be
associated with step cast-to-cast changes in hot metal
chemistry. Furthermore, the effect of blast furnace input
variability on hot metal chemistry can be statistically
quantified. Two measures of variability used in this
discussion are the standard deviation (S. D.) and the variance
(VAR) where VAR = (S.D. )2.
Since the fuel rate change due to a one percent change in
efficiency is about the same as the fuel rate change for a
0.1% change in hot metal silicon, a "rule of thum" that can
be used is that a loss of 1% in furnace efficiency will lower
hot metal silicon by 0.1% (Si J and will raise the top gas BTU
value by 1.5 BTU/SCF.

A review of blast furnace operating data at Stelco indicates

that the BTU value of individual top gas samples is typically
83 :! 4.3 BTU/SCF. A review of continuous top gas analysis
data indicates that a standard deviation of :! 2.5 BTU/SCF
reasonably represents cast-to-cast variations due to changes
in furnace efficiency only. (Both the mean and standard
deviation of BTU content can change as the result of varying
fuel inj ection practices, blast moisture, etc.)
The variation in furnace efficiency can be calculated
directly or indirectly as follows:

S. D. (Efficiency) = Top Gas BTU Std Dev.

1.5 BTU/% Efficiency
= :t 2.5/1.5
= :! 1. 67 % (1 S.D.)

This variation in Gas utilisation (Efficiency) as shown above

causes a corresponding variation in hot metal silicon content
of:!O.167% (SiJ (1S.D.).
The relative importance of efficiency variation on hot metal
silicon can be calculated using the coefficient of
determination.

4-19
 2
COEFFICIENT OF DETERMINATION = S.D.ll (EFFICIENCY)

S. D. (SiJ (TOTAL) 2

= (0.167)2
(0.245)2
r2 = 0 . 4 6
Consequently, about half of the variability in the process
results from factors such as coke stability and burden
distribution which affect furnace gas utilisation or fuel
efficiency.
Discussion of Efficiency Variation

Due to the relative importance of efficiency variation in the

operation of a stable blast furnace, the factors that affect
furnace efficiency will be significant with regard to their
effect on hot metal quality.

The primary factors affecting furnace efficiency are burden

distribution, raw material size distribution and coke
stabili ty. These are the same factors that have allowed the
tremendous improvements in furnace fuel rates.

Other factors such as alkalis, high temperature ma terial

breakdown, scaffolding, etc. which also affect gas / solid
contact are much harder to quantify.

Discussion of Other Variation

All factors that have an impact on the charged fuel rate of
the furnace have an effect on cast-to-cast hot metal
chemistry. Cast-to-cast variation in ore weights, coke
weights, coke moisture, hot blast temperature, hot blast
moisture, natural humidity, have a cast-to-cast impact on
furnace operation.
Operating practices and irregularities can also have a cast-
to-cast effect on hot metal chemistry. One example is casting
practice, particularly on one-taphole furnaces. Variation in
casting time and residual hearth hot metal will change
parameters such as the burden descent rate. Late casting for
example will fill the hearth with hot metal and float more
coke into the tuyere zone generating more heat on late casts.

4-20
 OTHER ARAS TO BE CONSIDERED

Raw Materials Preparation

The quality of the ferrous material charged to the blast

furnace improved dramatically in the mid nineteen sixties
with the introduction of pelletized ore of an optimum size to
promote good gas/solid contact.

Improving the gas/solid contact decreases the amount of

carbon required to reduce the ore since the reaction to CO2
reduces twice as much ore as the reactions to CO. This
improves the furnace fuel efficiency, which in turn reduces
the furnace coke rate.

Raw Flux

The calcination of raw flux (dolomite or limestone) in the

blast furnace consumes energy and releases CO2 which serves
to dilute the top gas produced by the furnace. The reaction
proceeds as follows:)
Ca CO) + 766 BTU/LB = CaO + CO2 (g)
The energy requirements to make a ton of hot metal can be
reduced if this reaction does not take place in the furnace.
(i. e., if previously calcined flux can be charged to the
furnace. ) The most commonly used method of providing a
calcined flux to the blast furnace is by using fluxed sinter
or pellets, or by charging BOF slag.

Coke Properties

Coke is the main reductant for the blast furnace process

since it serves to support the burden and provide a means for
gas to flow through the burden. Increasing the stability of
coke improves the gas/solid contact and makes the furnace
more fuel-efficient since more carbon can be converted to
carbon dioxide.
Increasing coke ash has two effects. Firstly, the carbon
content of the coke is reduced. Secondly, the ash must be
mel ted and drained from the furnace as slag, robbing the
process of energy.

4-21
Hot Metal Chemistry

The two main specifications for hot metal chemistry concern

silicon content and sulphur content.

The silicon content is determined by how much silica is

reduced in the process. Since it takes more energy to reduce
silica than iron oxide, there is an energy penalty associated
wi th increasing the silicon content of hot metal.
The sulphur content of iron is determined mainly by slag
basici ty. Higher basicities are usually achieved by adding
flux, which requires a calcination reaction and contributes
to slag volume. There is an associated energy penalty for
each of these factors.
The effect of other elements making up the hot metal
chemistry can be determined in a similar fashion.

Scrap
The benefit of charging scrap to the blast furnace can be
found by reviewing the energy required to reduce Fe203 to hot
metal versus the energy required to only mel t iron. The
melting component constitutes only about 16 percent of the
total energy requirement. The average coke rate for the
furnace can be significantly reduced with only a small
addi tion of scrap to the burden.

Any partial reduction to Fe304 or FeO will also reduce the

total furnace energy requirements mainly through increased
top gas recoveries. This is because the reducing gas in the
upper stack can exit the furnace without having to perform
these steps in the reduction process and will consequently
have a higher BTU value.

4-22
Producti vi ty

The productivity of the blast furnace is proportional to two

basic factors. The first is the amount of wind (oxygen)
blown and the second the furnace fuel rate.

In the past, significant increases in productivity were made

with the conversion from raw to pelletized ore, mainly due to
the higher wind rates allowed by the more permeable burden
and the lower slag volumes allowed by the lower gangue
content.
For a given bed of materials, there is a maximum wind rate
for the furnace beyond which the furnace becomes unstable.
At this point, the producti vi ty of the furnace depends mainly
on the furnace fuel rate since the volume of wind to the
furnace is fixed. In this situation the relationship that
applies is:

New Production = Old Production x Old Fuel Rate

New Fuel Rate

OxYgen Enrichment

In cases where the volume of wind to the furnace is a

limiting factor, further increases in producti vi ty can be
made by increasing the oxygen content of the hot blast above
21 percent.

Oxygen enrichment reduces the amount of inert nitrogen in the

system, thereby concentrating the process. With oxygen
enrichment there is more oxygen available to form reducing
gases at tuyere level. In addition, more injected fuel can
be introduced at tuyere level to replace coke while
maintaining the same flame temperature. These gains,
however, are partially offset by the reduced sensible heat of
hot blast per ton of hot metal.

Some discussion of the use of Oxygen enrichment is in order

since this the use of oxygen enrichment can have a profound
effect on the Blast Furnace process.

4-23
OxYgen Enrichment (continued)

The amount of oxygen required to make a ton of hot metal is

determined by combustion requirements in the raceway.

Generally the amount of wind required to make a ton of hot

metal is calculated as the Specific Wind Rate. For example
wi th a specific wind rate of 42,000 SCF/NTHM with no
enrichment, the amount of oxygen required to make a ton of
hot metal is
42,000 SCF * 21 percent Oxygen in air = 8,820 SCF Oxygen/NTHM

Oxygen is typically added to the cold blast after the blowers

and the total amount of oxygen added can be calculated

Percent Enrichment * Wind Rate / 0.79 = Oxygen SCFM

Percent Oxygen in Wind * Wind Rate = Oxygen SCFM

Considering that the total amount of oxygen required to make

a ton of hot metal will remain the same for a constant
furnace fuel rate, the specific wind rate will decrease with
the use of oxygen enrichment.

Oxygen Specific
Percent wind Rate (SCF/NTHM)

21 % 42,000
22 % 40,090
23 % 38,347
24 % 36,750
25 % 35,280
26 % 33,923

The reduction in specific wind rate for the furnace will

significantly reduce the amount of wind required from the
blowers and will also significantly reduce the amount of fuel
required for heating that wind in the stoves. Appliances
using BFG such as the stoves and Boilers will also become
more efficient due to the reduced N2 content.

4-24
Since the use of oxygen enrichment has the effect of
concentrating the process by eliminating nitrogen, the top
gas BTU value will rise about 2.8 BTU per percent enrichment.
This value will decrease slightly as the level of oxygen
enrichment is increased.

Tyical Top Gas Analysis

BFG with NO enricluent BFG with 1 % enricluent
Gas Percent Volume MMTU Percent Volume BTU

CO 22.7 13 f 701 4.43 23.4 13 f 701 4.43

CO2 18.5 11, 116 0.00 19.1 II f 116 0.00
H2 3.8 2 f 294 0.74 3.9 2 f 294 0.74
N2 55.0 33,180 0.00 53.5 31, 270 0.00
Total 60,358 5. l7 58,431 5.17
BTU / SCF 85.7 88.5

wind SCF 42 f 000 40,090

02 % 21 % 22 %
02 SCF 8 f 820 8 f 820
N2 SCF 33 f 180 31,270

Rules Of Thwr
The Rules of Thumb for blast furnace operations are based on
the work of many people. Special mention should be given to
R. V. Flint and his Flint Carbon Rate Formula. (4).

The Rules of Thum are the result of multiple regression

analysis of blast furnace operating data. Most steel
companies to suit their individual operating conditions have
established similar data.
Table 1 presents the most common or accepted "Rules Of Thumb"
and describes their application to blast furnace practice.

4-25
 Overview Of Rules Of Thum

The data in Table 1A outlines Hilton Works Blast Furnace

Operations since 1964. Although a strict comparison cannot be
made due to the magni tude and numer of changes, several
observations of step-wise improvements are worthy of note.

A) The introduction of sized low gangue pelletized

ore in 1964 significantly improved the efficiency
of i A i Fce. The reduction in coke rate caused a
corresponding reduction in slag volume, which
provided further energy savings. The introduction
of pellets provided a more stable operation, which
allowed the hot metal silicon content to be
lowered.
B) Higher hot blast temperatures on i B i furnace
significantly lowered coke rates in spite of the
amount of steam injected. The addition of fluxed
sinter to the burden (not shown) combined with the
lower coke rates allowed a large reduction in raw
flux used and the slag volumes generated.

C) The introduction of Burden Distribution on i D i

Furnace significantly increased top gas efficiency
by about 3.5 percent. Combined with a further
increase in hot blast temperature, minimum raw
flux charge, and minimum hot blast moisture, fuel
rates were cut by one third over this period.
(1964 to 1988)

D) The net energy required to make a ton of hot metal

of similar chemistry has remained about the same.
Note that no corrections have been made for
variations in hot metal chemistry.

E) The majority of energy savings over this 25 year

period were achieved through reduced conversion
costs, mainly in wind and stove heating.

4-26
 A Furnace A Furnace B Furnace D Furnace
Parameter 1964 1964 1970 1988
Raw Ore % 75 40 0 0

Flux lb/NTHM 671 544 104 17

Hot Metal (Si J % 2.14 1.20 1. 03 1.1l

Slag Volume 680 630 406 388

H.B.T. of 1450 1500 1750 1900

Moisture gr / SCF lO 9 19 8

Coke lb /NTHM 1520 1270 1084 890

N.G. lb/NTHM 69 45 75 68

Total Fuel Rate l599 1315 1159 958

Wind SCF /NTHM l20,OOO 80,000 65,000 44,000
RAFT of 3500 3570 3394 3680
Top Gas

% CO 22 21 24 22

% CO2 14 l7 17 20

% H2 5 5 5 4

CO2/ (CO+C02) % 39 45 42 48
Energy Balance MMTU /NTHM

Hot Blast In 4 3 3 2

Energy In 21 17 16 13
Out 13 9 8 5

Net 12 11 11 10

4-27
Discussion Of "Rules Of Thum"
Some of the "Rules of Thumb" can be verified by reviewing the
amount of energy involved for the variable concerned and
determining the energy input required considering the fuel
efficiency of the furnace.

For example, the addition of 1 lb/NTHM of raw flux requires

766 BTU for the calcination reaction. If the furnace is 55
percent fuel efficient, this means that about 1392 BTU will
have to be added. Since 1 lb of coke contains 12,800 BTU,
about 0.1 lbs of coke will be required. This happens to
correspond exactly with the empirical "Rule of Thumb".

New "Rules of Thum" can be estimated on the same basis. For

example, if a more efficient cooling system removes more heat
from the process that heat can be quantified in BTU/NTHM.
Again, considering a furnace fuel efficiency of 55 percent,
the amount of additional energy required for the process can
be calculated.
100, 000 BTU/NTHM / (0.55 * 12,800 BTU/LB Coke ) = 14 LB/NTHM

Other "Rules of Thum" can not be so easily developed due to

the number of variables involved. For example, an increase
in hot metal silicon is usually accompanied by other changes
such as a change in hot metal temperature. Increasing the
silicon content of hot metal requires more energy not only
for the reduction of silicon, but also for the increased
heating of the hot metal and slag. Accompanied by minor
changes in furnace efficiency, the effect of a change in hot
metal silicon content is difficult to estimate and is best
evaluated from experimental data and the empirical "Rule of
Thumb" .
COMPARING TWO PERIODS OF OPERATION USING THE RULES OF THU
An example of how the Rules of Thum are used to compare two
periods of blast furnace operation is shown in Table 2. The
Table is constructed by listing all the parameters that have
changed that will affect the furnace fuel rate. The
corresponding fuel rate correction can be calculated for each
variable. The total amount of these adjustments will
generally explain the change in fuel rate between the two
periods of operation.

4-28
 TABLE 2 COKE RATE ASSESSMENT
PARTER BASE CURRNT COKE RATE ADJUSTMNTS
Blast Moisture 8. a 11.1 +3.1 * 4 = + 12.4
( grains/scf )
Blast Temp 1875 1864 -11 * 22/100 = + 2.4
( degrees F )
Hot Metal %Si 1.00 0.8 - 0 . 2 * 13/0.1 = 26.0
Hot Metal 'YoM 1.00 0.9 -0.1 * a = 0.0
ASTM Coke 57.0 54.1 -2.9 * -16 = + 46.4
Stabili ty
Coke % Ash 7.5 7.8 +0.3 * 30 = + 9.0
Slag Volume 350 382 +32 * .25 = + 8.0
( lb/NTHM )

Flux Rate 40 24 -16 * .15 = 24.0

( lb/NTHM )

Nat Gas Rate 29 28 -1 * -2 = + 2.0

( lb/NTHM )
Tar Rate 85 84 -1 * -1 = + 1.0
( lb/NTHM )

Total Adjustments ( lb/NTHM ) = + 52.8

BASE CURRNT
Actual Fuel Rate 985 1034 ( lb/NTHM )

Dry Coke Rate 871 922 lb/NTHM )

Coke Rate Adjustment 0 53 ( lb/NTHM )

Adjusted Coke Rate 871 869 ( lb/NTHM )

4-29
Model Blast Furnace Examle
To further illustrate the comparison between periods as shown
above, the effect of step wise practice changes on a model
blast furnace are shown in Table 3. The model Blast Furnace
shown in Case 1 is operating on a burden of lump ore with low
blast temperatures and no injected fuels. The furnace is
inefficient due to poor gas/solids contact and has a tendency
to slip, limiting the wind rate.

Swi tching to a low gangue pellet burden with no other changes

would result in excessively high slag basicity. Accordingly,
the raw flux consumption must be cut in half as in Case 2. A
tremendous producti vi ty gain is made due to the increased
wind rate allowed by the improved burden materials. Making
these changes serves to reduce coke rate in three ways, flux
rate, slag volume, and furnace efficiency changes.
Case 3 shows further coke savings can be made by swi tching
part of the burden to sinter and eliminating the raw flux.

Increasing the blast temperature alone results in excessive

RAFT, hence, the blast moisture must be increased as well as
shown in Case 4. The BTU value of the top gas increases due
to the higher Hydrogen content. Using oil instead of moisture
to control RAFT has a tremendous effect on coke rate as shown
in Case 5. Lowering the hot metal Silicon content as in Case
6 lowers the coke rate and has a secondary effect on slag
basici ty since more silica from ore is diverted to the slag.

Case 7 outlines the effect of burden distribution equipment

on the furnace, generating a large reduction in coke rate and
also the amount of top gas produced.

The use of Oxygen enrichment to increase productivity

combined with an increased oil inj ection rate to maintain
RAFT is shown in Case 8. Note the higher BTU value of the BFG
due to the lower ni trogen content.
Finally, the addition of scrap to the burden displacing
pellets is shown in Case 9. The significant reduction in fuel
rate has a significant effect on producti vi ty.
OveralL, the producti vi ty of this model Blast Furnace has
more than doubled and there is still an opportunity for
improvemen t .

4-30
 Case Case Case Case Case Case Case Case Case
TABLE 3 1 2 3 4 5 6 7 8 9
NTHM per day 2477 4351 4464 4525 4723 5976 5241 5688 6100

Wind Rate MSCFM 100.0 160.0 160.0 160.0 160.0 160.0 160.0 160.0 160.0
Wind MSCFM/NTHM

Scrap lb/NTHM
Lump Ore lb/NTHM
58.1

3280
0
53.0

0
51. 6

0
0
50.9
0
0
48.5
0
0
46.3

0
0
44.0 40.5

0
0
0
0
~
~
37.8

0
Pellets lb/NTHM
Sinter lb/NTHM
Raw Flux lb/NTHM 700
0
~
0 1300~11 300~1
0
1960
1300
0
1960
1300
0
1960
1300
0
1974
1300
0
1974
1300
0
1974
1300

~ 0
1300
0

Blast 02 %
Blast Temp. F
Moist gr/scf
RAT deg F
1500

3775
21

6
21
1500

3775
6
21
1500

3775
6
118001~
4021
17
21

3763
~ 21
1800

3767
1800

3753
21

6
21
1800

3739
6
1800

3752
6
23
1800

3719
6

Coke lb/NTHM
Oil lb/NTHM
Total lb/NTHM
1395

1395
0
1287

1287
0
1254

1254
0
1238

1238

1. 00
0 ~1086

1186

1. 00
1025
100
1125

0.501
969
100
1069

0.50 0.50
~918

1068
846
150
996

0.50
Hot Metal (Si) % 1. 00 1. 00 1. 00
Hot Metal (S) % 0.029 0.029 0.026 0.026 0.026 0.03 0.028 0.028 0.018
Slag lb/NTHM 742 432 511 509 500 519 515 512 488
Slag B/A 1.14 2.02 1.13 1.13 1.13 1.17 1. 08 1.10 1.11 1. 20

BFG BTU/ scf 83.7 83.1 84.4 87.5 89.0 85.5 79.3 87.1 88.1
BFG MMTU /NTHM

Efficiency %
6.84
38.0
6.38

42.0
6.19
42.0
6.44

42.0
6.34

42.0
5.59
42.0 ~
4.86 5.13
46.0
4.89
46.0

The Blast Furnace Game

The blast furnace computer model used to demonstrate the

effects of practice changes was originally constructed with
the assistance of Mr. Duncan Ma of McMaster Uni versi ty and
has been revised several times since then.
This model incorporates nearly all of the principles involved
in the use of the "Rules of Thum". The model involves the
simul taneous solution of mass i energy, chemical and cost
balances and reasonably reflects changes in operating
practice and changes to the process via equipment
modifications.
The challenge that is presented in the Blast Furnace Game is
to optimise this model blast furnace by judiciously
purchasing equipment and improving on the furnace operating
practice. The obj ecti ve is to ...... in the words of Bill
i Taylor, a retired Blast Furnace Operator,
i

.. I

"Keep the wheels turning and the costs down".

4-31
 REFERENCES

1. "North American Combustion Handbook" , North

American Manufacturing Company, Second Edition,
1978, Page 356.

2. "Perry i s Chemical Engineering Handbook", McGraw

Hill Chemical Engineering Series Fourth Edition,
Page 9-43.

3. Strassburger et al., "Blast Furnace - Theory and

Practice" . Gordon & Breach Science Publishers
1969, Page 697.

4. R. V. Flint, Blast Furnace and Steel Plant, 50, 1,

1962.

5. I.N. Gibra "Probability and Statistical Inference

for Scientists and Engineers" Prentice Hall Inc.,
Englewood Cliffs, N. J. 1973 Page 110.

6. A. J. Duncan "Quality Control and Industrial

Statistics" Richard D. Irwin Inc. , Homewood,
Illinois 60430,1974, Page 767,768.

7. M. R. Spiegel "Theory and Problems of Statistics"

Schaum Publishing Company, New York 1961, Page
253.

8. J. B. Hyde and J. W. Busser "Use of a Charge

Control and Coke Moisture Gauge System at Stelco 's
'E' Blast Furnace" 46th Ironmaking Conference,
pittsburgh, P.A., 1987.

4-32
 LECTU #5

BLAST FUACE DESIGN i

John A. Carenter

Paul Wur Inc.

600 Nort Bell Avenue
Buidig 1, Suite 230
Caregie, Pennylvana i 5 i 06

Abstract: - 1bs paper is of a general natue, coverig the blast fuace proper and
those ancilar components imedately upstream and downstream of the fuace. It
will focus on the stockhouse. the chaging equipment, the fuace top, the fuace
proper, the cooling system, and the casouse area of
typical blast fuaces.

Blast fuace ironmakg is a system comprised of many components

fuctionig in hanony. Proper application and operation of these components is
necessar to support the ironmakg process. Selection of specific components is
dependent upon such factors as existing conditions, physical constraits, production
requiements, cost, schedule. reliabilty, and maitaability. Interdependence of
components is as importt to the system's operation as their individua capability.

1bs paper will illuste the major requirements and "usua" practices for each
area or component. It \\ ill also explore some alternative technologies which are
commercially available. The inerent advantages and disadvantaes of those
alternatives will be discussed.

In the overal ironmakg course, other disserttions providing more detal on

specific components and the process will be presented. 1bs paper is complementa to
those more specific presentations.

5-1
 INODUCTION
Ths paper is organed in the followig maner:

(1) The Introduction pro\ides a genera description of blast fuace ironmakg.

(2) There are eight sections which describe in more detal a blast fuace's
components and equipment.
(3) A short design exercise which is provided to demonstrate component sizig,
i equipment selection, and the interaction between equipment and process.
,i

Thè bl~ fuace (Figu 1), converts iron bearg ores and revert into molten
meta. Associated with blas fuces are coke plants which convert coal into coke and
pellet plants, which prepare iron ore for the blast fuace. The blast fuace convert
these prepared raw materials into a product of greater value. Iron from some blast
fuace operations will be made dictly into saleable cast iron products in a foundr.
Other operations produce a lower in silicon, "hot meta", which is converted into steel.
Blast fuace by-products are slag, off gas, flue dust, and fiter cake. These by-products
may have either positive or negative economic impact, depending on the local
possibilities for utiliztion.

Blast fuace ironmakg is a four hundred year old technology. Even so, the
int~grated mill using blas fuce hot meta is still the most common method used for
the production of steel. Today' s integrated steel plant process relies upon the blast
fuace to provide on schedule. predictable quatities of molten iron of consistent
quaity. Varation in any of the ascts of the supply of molten meta has a serious
impact on the rest of the steel production processes. Therefore, the blast fuce is a
key component in the modem integrted steel mill.

There are some who say tht the blast fuace is at the end of its useful life.
Ths is not so. Consider that twenty thee years of operating data from a tyical pair of
medium size blast fuaces shows an average increase in productivity of thee percent
per year (Figue 47). At the sae time, the average reduction in fuel rate was one
percent per year (Figue 48)18. Also, the productive time between relines, "the
campaign", has been extended though improvements in equipment, materials, and
designl8. As a result, the overal cost of makg iron, corrected for ination, ha
improved even more than the opetig data indicates. The Blast Furace is not dead; it
is a four hundred year old technology which is stil progressing in every area at a
signficant rate. The Blas Fure is a dynamc science supported by constatly
improving technologies.

Some of the many aspects afecting every consideration for blast fuace design
or re-design are profit, employee health, safety, environmenta protection, governenta
regulation, market requirements, downstream processing, available workforce,_

5-2
constrction, maitenance resoures, changing technologies, equipment obsolescence,
raw materials, utilities, and so on. Any serious constrait in one of these factors could
jeopardize the viability of an exig uit (or even a steel plant) or preclude or
necessitate the consction of a nev..' blast fuace.

Blas fuaces are generly grouped by size. "Small" - under five thousand net
tons of hot meta per day (Nflday), "Medium" - six thousand to eight thousand
NTIday, and "Large" - nie thousd to twelve thousand NTHday. A given
integrted steel plant will operate the number and size of blast fuaces requied to
provide the hot meta needed. \-iulti-fuace mills are less afected by individua
fuace repai relines or control prblems. Small fuaces have shorter relines than
large fuaces and are considere to be easier to operate. However, the hot meta from
small fuaces is higher in cost. An individua mill will operate the mium number
of cost effective fuaces. In some cases, upgrades are made in order to reduce the
number of fuaces in operation.

Blast fuces are "relined" periodically. In the past, ths involved the
replacement of the internal brick lig of the mai vesseL. In recent times, extensive
component rebuilding, replacement, and general maitenance was performed at the
same time. With ths practice. the more effcient plant with fewer "large" blast fuaces
will lose a higher percent of it' 5 production durng a reline than the plant with more
small fuaces. In order to have both the low operating cost and the minium
interference from relines, the indus has worked to maxe the blast fuace
campaign (time between relies) and to reduce the reline duration.. The clear trend
today is for mils to operate tèwer large fuaces and to utilize technques and design
which will extend their campaign indefitely.

At the same tie, Ùle reduction in product varability has become more
importt so investments have ben made which improve monitorig and control of the
process. Blast fuace operators, researchers, maitenance personnel, and designers
have applied modern technology and analytical methods to the process in order to better
monitor and control the process. As a consequence, the stadard deviation of the hot
meta quaities have been reduced. Improved data collection systems also provide more
inormation for suppliers and manufacruers. Tbs improves the materials selection and
the design of fuaces and equipment. Campaign lengts have increased from thee or
four years to more than eight year.

BLAST Fù~ACE PLAN LAYOUT

The layout of a blas fuace plant is essentially an exercise in integrating the

equipment requied to handle the varous materials required to make iron and the
resulting product and by-products. The most effcient design will properly
accommodate the process and will be judged effective from the stadpoint of both_

5-3
 intial capita investment and ongoing operating costs. Plant layout is dependent upon
many factors such as terr raw marial delivery method, in-plant raw material
processing systems, climatic conditions. downeam processing systems and locations,
quatity/flow requirements for "hot meta", hot meta delivery "fleet" size, and so on.

A blast fuace plant (Figu 2) typically comprises the followig elements:

(a) Raw Material Storae, Handlg and Reclai

(b) Stockhouse
(c) Charging System
(d)' Furce Proper
(e) Casthouse
(f) Slag Handling
(g) Hot Meta Handling
(h) Stoves and Hot Blas System
(i) Gas Plant
(j) Utilities
(k) Control Systems
(1) Maitenance Facilities
(m)Personnel Support Facilities

RAW MATERI STORAGE AN HALING

Bulk materials such as ore, pellets, fluxes, and coal are normally delivered by
bulk carers (ship or barge) or by ra car. Coke could be delivered in the same fashion
or be produced in-plant by coke ovens. Sinter can be delivered to the plant or can be
produced in-plant from ore and in-plant generated materials (mill scale, B.O.F. scrap,
pellet fines, coke breeze, etc.). These raw materials, whether purchaed or produced in-
plant, requie sufcient controlled storae to support the blast fuace plant operations.

Storage capacity is required in the event of predictable delivery disruptions (i.e.

normal cessation of seaway shipping due to witer ice conditions) or unpredictable
disruptions (such as possible late delivery due to ship mechancal problems).

Additional storage capacity can be required due to possible changes in the

source of cert raw materials. Separ storage locations are requied due to different
physical or chemical characteristics in simar materials. Mixig of "simlar" materials
could cause process control/metaurgical problems.

The storage piles must be searted to prevent intermxig of dissimilar

materials. The piles must be placed on prepared beds to enable the raw material reclai
equipment operators to distinguish bet\\"een prie and tramp material. Piles are laid out

5-4
to mie material degrtion and to prevent wid pick-up of fies. Water sprays
and cocoonig agents may be used to mi dust pick-up/car-offby wids.

Many different technques ar avaiable for raw material laydown and retreval.
Laydown: self-unoadig ships, ore bridges, stackig conveyors, scrapers, etc.
Retreval: bucket wheel reclaiers, frnt-end loaders, scrapers, diectly from bin or
pile bottoms, etc.

Obviously, the lay down and retreval systems must be sized to ensure the
thoughput requied for the blas :fe plant.
,i. .

STOCKOUSE

The stockhouse is the blas :fe operator's storage unt for direct feed of the
burden to the fuace. Storage bin ar provided for each of the burden materials for the
blast fuace. Individua bin are provided for simlar materials (i.e., pellets) having
different metalurgical propertes.

The stockhouse provides adequate capacity for the varous burden materials in
the event of short term disruption of supply from the raw material storage areas.
Typical stockhouse bin thoughput caacities, in the event of loss of raw material feed,
are:

(a) Coke 2 to 8 Hours

(b) Pellets and Sinter 4 to 16 Hours
(c) (c) Miscellaneous Materials 8 to 24 Hours (fluxes, scrap, etc.)

These capacities are basd on rated fuce production and var depending
upon the reliability and the access tie for their replacement from inventory or from a
supplier.

Burden materials tend to degre due to climatic conditions and repeated

handling. The greater the number of ties that the material is handled (stockpiling,
reclaig, dumping, conveyor chutes, ore bridge buckets, etc.), the greater the percent
of fines in the burden. The blas fue process requies controlled permeability and
hence controlled burden. The chagig of excessive fines, either generally thoughout
the charge or concentrted over spifc short charging periods, can be disruptive to the
process and daaging to the fue equipment. The stockhouse provides the last
reasonable opportty for removal of fies prior to charging into the fuace. Where
possible, vibrating screens are ined afer the coke, sinter, and pellet storage bin to
elimate the major portion of the fies. The removed fies are collected for
reprocessing or sale. Some blas fu operators charge fies to specific areas in the
fuace to adjust local fuace permeailty and control heat loads on the fuace walls. _

5-5
 Moistue gauges are often provided in the stockhouse to monitor the actu
water quatities charged to the fwe. Ths inormation permts adjustments to the
charging quatities to compens for varing ambient conditions (i.e. higher coke
moistue due to rai fall).

Since different tys and varing amounts of burden material are requied to
support the continuous operation of the blast fuace, the burden materials must be
provided in a specific sequence (which itself can be changed frequently to support
varing fuace operatig parete). Hence the stockhouse must be provided with
reliable equipment for extctig and feeding accurte quatities of specific burden
materials tò meet a specific schedule.

The most common type of stockhouse has been the highine tye (Figue 3).
Ths type of stockhouse is located diectly adjacent to the fuace. Ral cars or bridge
craes feed the storage bin: the storae bins feed directly to a trveling scale car. A
scale car operator manualy controls the bin discharge gate to feed specific amounts of
material into the scale-equipped hopper located in the scale car. Afer collecting the
proper tyes and amounts of maerial, he moves the scale car to a position above the
"skip pit" and dumps the burden. \ ia a chute, into a waiting skip car. The skip car will
then be hoisted to the fuace top.

Placement of the stockhouse adjacent to the fuace often results in layout

congestion and restrcts flexibilty for futue modifications.

Many stockhouse have been modified to accommodate automatic coke

hadling. Coke is often fed to the storage bins via conveyor. Upon demand, the coke is
discharged from the storae bin, over vibrating screens (for fies removal), and
diected into weigh hoppers in the skip pit. When the fuace charging sequence
dictates, the coke is discharged into the skip car.

Improvements to the highe tye of stockhouse have primarly focused upon

automation of the bin gates (prmision ofpnewnatic or hydraulic actutors on each gate)
and the scale car. A trckig system is usually provided to ensure that the automatic
scale car is selecting material and amount from the correct bin.

The highline tye of stockhouse, in conjunction with a scale car, has presented
few options for the provision of ferous charge (pellets and sinter) screenig.

As knowledge of the blas fuace process has increased, more strgent

requiements for the burden have developed. The concept of "engineered burden" is
well recogned in the indus. It is generally accepted that there are limts to the
flexibilty and adaptability of the highine stockhouse to support ths requiement.
Where circumstaces have permtted (i.e. major fuds available for rebuilding or for_

5-6
new intalations), automatconveyorized stockhouses have been implemented
(Figue 4).

Provision of an automated stockhouse can provide more effcient feed of raw

material to the stockhous and more effcient selection, screenig, weighg, and
delivery of the burden to the fue.

The automated stackhous can be located directly adjacent to the fuace

feeding skip car (i.e. conversion of an existig highline type stockhouse), or can be
located re~ote to the fue for chagig via a conveyor belt.
.~. p

HOISTING SYSTEM

Modem blast fues ar chaged with skip cars or by conveyor belt.

Skip Car Hoisting

The use of skip car (Figue 3) for blast fuaces evolved from the mig
industr .

Blast fuace skip c.arare sized to suit the fuace thoughput (small
fuace/low thoughput/smal skips; large fuacelhgh thoughput/large skips).
Obviously, many factors such as hoist capacity, skip bridge design, and so on, have
their own inuences or consts upon the skip size.

Generally, two skips operate in opposing fashion (to reduce hoisting power
requiements) on a common hoist. Skips travel on rails on a skip bridge, usualy
intaled at an approxiate inclie between 60° and 80° to horionta. The ful skip
accelerates slowly as it leaves the skip pit, accelerates as quickly as possible reachig
and traveling at maxum sp for most of the lift. The hoist slows the skip down as
it approaches the top of the skip bridge. Dumping and horn rals gude the wheels of the
skip as it is overted into the fuace top charging eqrupment. As the hoisting skip
reaches and stops at the fi dumping position, the empty skip (descending at the same
speeds) is just reachig the bottom of its travel into the skip pit, awaiting filling.

The skip charging system is a very reliable, effective technque for deliverig
the burden to the fuace top. However, it lacks flexibility for the operator in tht the
skips can only hold a specifc amount of material (overloading results in overflling or
excessive hoist loads) or beomes ineffcient if small volumes of specific burden are
required.

5-7
 For a 5000 THday fuce, a tyical skip hoisting system would comprise:

(a) Two skips, each with 375 cu.ft active capacity.

(b) One hoist 50,000 lb capacity, two drves at 400 hp, 600 fpm maxum rope
speed.
(c) Vertcal hoisting height, 200 ft.

Furce Charging Conveyor

With the conveyorition of the stockhouse has come the conveyorization of the
hoisting system?. It is now common tor stockhouses to be located remote from the
fuce and one large conveyor belt (Figu 5) will car the burden to the fuace top.

If the fuace top is about 200 feet high, then a conveyor belt (inled at 80)
inclination to horionta will position a stackhouse at least 1,135 feet from the fuace.
Steeper belt inclinations are usualy a\"oided to mize pellet roll back. It is common
to charge miscellaneous materials ditly over and afer the end of a pellet charge on
the conveyor belt in order to hold the pellets in place until they reach the fuace top.

For a 5000 THday fuace, a typical feed conveyor would comprise:

(a) Two drves (including one stdby) 500 hp each

(b) Belt speed 350 fpm
(c) Belt width 54 il.
(d) Belt lengt -2700 ft. tota

CHAGING SYSTEM / FURACE TOP

The fuce proper is operated with some amount of positive (gauge) top
pressure. Blast fuace gas consistg priarly of carbon monoxide, carbon dioxide
and rutrogen is generated by the fuace process along with large amounts of entrained
dust. The blast fuce operator wants to maita the top pressure due to process
benefits and to conta the gases and dus (bth for fuel value and environmenta control
puroses). However, he mus reguarly place burden material inide the top of the
fuace in order to replerush the internal process, without losing the fuace top
pressure.

Bell Type Top

For many years, the most common tye of fuace top has been the two-bell top
(Figure 6). As the burden reaches the fuace top (by skip or conveyor), it falls into a
receiving hopper and into the smal bell hopper. The small bell (corucal shaped steel
casting about 8Y: feet in diameter and 412 feet high for a 5,000 THday fuace)~

5-8
 lowers and permts the burden to fal into the large bell hopper. The small bell is lifted
and seals agai a fixed seat on the smal bell hopper. Depending on the volume of the
large bell hopper, additiona loads of burden are sequenced into the large bell hopper by
the small bell. Thoughout ths process the large bell has remaied closed, sealing the
fuace. When the correct number of load of burden have been collected, the large bell
(conical shaped steel casg about 18~-S feet in diameter and 11 Yi feet high for a 5,000
THday fuace) lowers and alO\\'s the burden to slide down the bell into the top of
the fuace proper. Afer the burden discharges the large bell is raised and seals agait
the underside of the large bell hopper.

Ooviousy, the burden distbuton control with the fuace for ths style of top
is limted by how evenly the burden is placed on the large bell (skip dumping results in
uneven placement of burden into th style of top) and the falling cures of the specific
burden materials (i.e. coke or pellets) as they slide and falloff the large bell.

The two-bell top is susceptible to loss of sealing of the large and small bells and
of the packig between the large bell rod and small bell tube. Bell leakge results from
abrasion by the burden material sliding oyer the bell sealing suraces. The rod packig
leakage is a result of abrasion from fies either from with the fuace or from
collecting on the large bell rod afer burden is dinped in the receiving hopper.

In an effort to mie wear of the large bell sealing surace, blast fuace gas
taen from with the fuace is introduced between the bells to equaize the space
(reducing the pressure differential across the large bell sealing suiace). 1bs gas is
relieved to atmosphere prior to openig the small bell to permt introduction of more
burden.

The followig are some options ayailable to improve the limtations of the two-
bell type top system.

The McKEE Distrbutor

The McKEE Distrbutor (Figue 7) for many years was the mai burden
distrbution improvement available for the two-bell tye top. It is stil in operation on
some fuaces today. However, it is quickly being replaced by other technologies.

1bs design incorporates the abilty to rotate the small bell and small bell hopper
together while the skip car is dischagig. Burden is evenly distrbuted into the small
bell hopper, thus improving the even placement of burden onto the large bell.

1bs style of top is stil prone to small and large bell wear and subsequent loss
of sealing effect.

5-9
 The CRM Universal Rota Distbutor Top

The CRM (Centr Recherhes Metalurgiques - Belgium) Universal Rota

Distrbutor (Figue 8) was develope to elimte the loss of small bell sealing effect.
Two bells (a sealing bell and a marial bell) are intaled in place of the normal small
bell. A revolving burden hoppe is mounted on the material bell. The sealing bell is
located beneath the material bell and seals agait a fixed seat. Durg skip dischage,
the burden hopper and the close material bell are rotated to evenly fill the hopper.
When filling is complete, the hoppe rotation stops. When it is time to dump onto the
large bell, ,the revolving hoppe, material bell and sealing bell lower. The sealing bell
lowers below the- fied seat. Pan way though the lowerig process, the hopper descent
is stopped and the material bell and sealing bell contiue to descend until they reach
their stop position. As the gap opens between the material bell and the hopper, the
burden dischages evenly into the large bell hopper. As the burden leaves the gap, it
does not come into contact v.ith the sealing valve seating sUDace, thus maitag the
top sealing capability. TIs style of top is capable of maitag two atmospheres of
intern pressure.

The CRM Top! improves me sealing capabilty and longevity of the two-bell
top. It does not provide howewr a dratic fuace burden distrbution improvement
over the McKEE Distrbutor and does not eliminate the vulnerabilty of the large bell
sealing sUDace.

The GI- Lockhopper Top

The "Lockhopper Top" (Figu 9) is marketed by MA-GI- (Germanyl TIs

modification to the two-bell top reuces dependency on the large bell to maita a gas
seal. The addition of lockhoppers \\ith separate seal valves for each skip dump location
provides an additional capacity for sealing the top. The large bell can be operated with
no differential pressure across its sealg surace (i.e., fuace top pressure equas large
bell hopper pressure).

The operation is as follows:

(a) A skip dumps the load of burden into the lockhopper via a receiving hopper
and open seal valve. The burden is placed on the rotating small bell and
unfonny fills the rotatig distbutor hopper above the small belL. A seal
between the lockhopper and the rotating small bell hopper is open while the
rotation is underway.
(b) When the burden dischage from the skip is fished, the seal valve and the
seal between the lockhopper and small bell hopper are closed. Equaizig
gas is introduced and the lockhopper is pressurzed to fuce top pressure.

5-10
 (c) The small bell is then lowered to introduce the burden into the large bell
hopper.
(d) The small bell closes and the pressure from the lockhopper is relieved to
atmosphere.
(e) The seal valve on the opposite side (i.e. at the other skip dumping position)
is opened.
(1) The seal between the lockhopper and the small bell hopper is opened.
(g) Rotation of the smal bell and hopper commences.
(h) The top is now able to accept burden from the other skip.
-
Ths style" of top improves the sealing capability and the longevity of the two-
bell top. The "Lockhopper Top" however, does not provide a dratic fuace burden
distbution improvement over either the McKEE or CRM Top. Although the large bell
no longer is requied to perform a sealing fiction, the small bell sealing effect
longevity is stil critical.

Movable Arour

The major step taen to improve the burden distrbution of the bell tye top was
the development of movable anour (Figure 10). Adjustable deflectors are installed in
the thoat area of the fuace to deflect the burden after it slides off the large belL.

The movable anour is adjused depending upon the specific burden material
being discharged and where the operator wants to place that burden with the fuace.

Several manufactuers pro\ide alternate styles of movable anour,3,4. Individua

arour segments can be moved unformy (simultaeously and equaly) inside the
fuce to place the burden in an anular pattern.

Other styles of movable anour are available to provide individua control of

the anour plates in order to achieve non-circular distrbution pattern.

Some disadvantaes assoiated with movable anour are:

(a) Most mechancal and wear components lie with the harsh envionment of
the fuace top cone.
(b) Some loss of internal workig volume is required to provide clearance
between the movable anour and the design stockline level (albeit ths area
of the fuace canot be classified as a rugh productivity zone for fuace
workig volume consideration).

5-11
 (c) Limted capability to deflect burden to the very center of the fuace,
paricularly when the stockline level is aleady high. The rolling
chacteristc of pellets often negates the limted displacement of the
movable anour.

The PAUL WUTH Bell-Less Top

In the early 1970,s PA.ll WUTI S.A. of Luxembourg developed the Bell-
Less Top Charging System (Figu 11). lbs style oftop5,6 is a radical depare from
the bell ~e top.
.t, .

Burden can be placed \\ iùi the fuace in any pattern requied by the fuace
operator. Anular rigs, spir, segment and point placement are common pattern
achievable by synchroni or independent tilting and rotation of a burden distrbution
chute located with the top cone of the fuace.

Furace top sealing is maitaed thoughout the campaign of the fuace.

Maitenance activities are simple and of short duration.

Generally, the bell-less top consists of a receiving chute or hopper (receiving

burden from the skips or from a conveyor belt), a lockhopper with upper and lower seal
valves, a material flow control gate, a mai chute drve gearbox (a water or gas-cooled
unt used for chute rotation and titig), and the burden distrbution chute.

There are three mai styles (Figue 12) of bell-less tops available today, namely:

(a) Parallel Hopper

(b) Central Feed
(c) Compact Style

Typically, the parel style incorporates two lockhoppers (thee hoppers have
been intaled on some fuces for thoughput and backup puroses; a one "eccentrc"
hopper style has been ined for an application with restrcted clearce). Since the
early 1980's, many fuces haw selected the "central feed" single lockhopper style for
its improvements in burden segregation and burden distrbution control resulting in
enhced fuace operation.

A "compact" stle of bell-less top has been developed for small to mid-sized
fuaces to permt the introduction of the bell-less top (and its advantages) to fuaces
where the other larger types of bell-less tops canot be used due to cost or physical
constrts.

5-12
 Bell-less top operation for a central feed tye (Figue 13) is as follows:

(a) Burden is dischaged frm a skip or conveyor belt though a receiving chute
or hopper pas an open se valve into the lockhopper.
(b) Afer the burden is reeived in the lockhopper, the upper seal valve is closed
and equaizg gas is introduced to pressurze the lockhopper to fue
pressure.
(c) The lower seal valve opens.
(d) The burden discharges from the lockhopper. The material gate ha been set
to the preselected openig to suit the specific burden material to be
dischaged.
(e) The burden drops vercaly though the feeder spout with the mai
trmission geabox and falls onto the burden distrbution chute.
(f) The burden distbuton chute directs the burden to the requied point(s)
with the fuce (Figu 14).
(g) When the lockhopper is fuly discharged (monitored by load cells and/or
acoustic monitorig), the lower seal valve is closed.
(h) A relief valve is opened to exhaust the lockhopper to atmosphere (or
though an energy reovery unt).
(i) The upper seal val\'e opens and the sequence can repeat.

Users regularly repon bell-less top advantages over other top charging systems,
such as:

(a) Higher top pressur capability (i.e. 2.5 atmospheres).

(b) Furace fuel sa'\ ings.
(c) Increased fuace production.
(d) More stable operation.
(e) Reduced maitenace in terms of cost and tie.
(f) Increased fuace campaign life.
(g) Improved fuce opeonal control when employing high coal injection
rates at the tuyeres.

Fl"RACE PROPER

The fuace proper is the mai reactor vessel for the blast fuace ironmakg
process.

Its internal lines ar designed to support the internal process. Its external lines
are designed to provide the necessa systems to conta, mainta, monitor, support
and adjus the internal process.

5-13
 The blast fuce is a counterfow process:

(a) Burden at ambient conditions is placed in the fuace top onto the colum
of burden with the fue.
(b) As the burden descends \lith the burden colum, it is heated, chemically
modified and fily melted.
( c) Furer chemical modicatons occur with the molten material.
(d) The molten products ar extted near the bottom.
(e) Melting of the burden materal and extaction result in the descent of the
_ burden colum and the nee for replenishment of the burden at the top.
(£) Hot blat ai is introduced thugh tuyeres near the bottom.
(g) Blas fuace gases ar genered in front of the tuyeres and ascend though
the burden. They chemicaly modify the descendig burden and they
themselves are chemicaly modified and cooled.

(h) Blast fuace gas (and dus) is extcted near the top of the fuace.
(i) Heat is extacted from the vessel in all directions (priarly though the
ling cooling system) and along with the blast fuace gas, molten iron and
molten slag.

Furace Stvle

Furaces are constrcted to be mantle supported or free stading (Figure 15).

Mantle type fuaces (most ~ort American fuaces) characteristically have a

rig girder (mantle) located at the bottom of the lower stack of the fuace. The mantle
is supported in tu by colum \vhich re on the mai fuace foundation. The hear,
tuyere breast and bosh are also supported by the foundation. Furaces with mantle
support colum tend to have restcted access and reduced flexibilty for improvements
in the mantle, bosh and tuyere breas aras.

Since thermal expanion is a major consideration in fuace shell design, the

mantle style of fuace provides an interestig design consideration. The mantle
support colum are relatively cooL. The mantle tends to maita a constat height,
thoughout the fuace campaign \\ith respect to the fuce foundation. Thermal
expanion of the stack due to process heat is considered to be based at the "fixed"
mantle (i.e. the top of the fuace rases with respect to the mantle). The effective
height of the bosh, tuyere breas and hear wall shells (supported on the fuace
foundation) increases due to the theral expanion of the shell caused by the process
heat. The lower portion of the fuace li upwards towards the fixed mantle; therefore
the provision of an expanion joint of some type is required at the bosh/mantle
connection or somewhere appropriately located in the lower portion of the fuace.

5-14
 Free stadig fues wer developed to elimte the colum and permt the
inlation of major equipment and fuace cooling improvements. Ths fuace style
has a thcker shell for stctu support. Instalation and maitenance of a reliable
cooling and ling system is esstial in order to susta the strctu longevity of the
shell.
Two varations of the fr stdig fuace have been employed. One stle
provides for a separte stctu support tower to car the fuace off-gas system and
chagigloistig system load. The other style (while it does employ a separate
support tower for shell replacement puroses durg relines) uses the fuace proper to
support th: off-gas system and chagloistig system loads.
.~. R

Special consideration to the fuace shell design mus be made regardless of the
fuace stle. The vessel is subjected to internal pressures from the blast and gas,
burden, molten iron and slag. De and live load durg all operating, maitenance
and reline states mus be consider as weif.

Furace Zones

The major fuce prope zones (Figue 16) are as follows:

(a) Top Cone

(b) Thoat
(c) Stack
(d) Mantle/Belly
(e) Bosh
(f) Tuyere Breas
(g) Hear Walls
(h) Hear Bottom
(i) F oiidation

Top Cone

The top cone or dome is the uppermost par of the fuce proper. It support
the fuace top charging equipment, and the off-gas collection system (tyically in
Nort America). Stock rods (stockle recorders or gauges) are usualy placed here to
monitor the upper level of the buren in the fuace. These devices are the unts which
provide the permssive or indication signals to charge the next scheduled burden input
to the fuace.

Typically, they are weights lowered by special wiches, or microwave unts.

Some fuaces incorporate raioactive isotope emitters and detectors moiited in the
fuace thoat to monitor the burden leveL. Infared camera can be instaled in the top

5-15
 cone to monitor the off-gas tempetue distrbution as it escapes the fuace burden
stockline.

The top cone is the coolest zone of the fuace proper but can be exposed to
extemely high temperatues if burden "slips" (rapid, uncontrolled burden descent afer
a period of unusua lack of descent). The newly charged burden falls though ths zone;
off-gas is cared away from th setion.

Thoat

Stèel wea plates or arour are instled in ths zone. Here, abrasion of the
fuace ling from the charged burden is the prie cause of deterioration. Furace
operators work to maita the upper level of the burden (the stockline) in ths region.
As noted earlier, movable anour can be intaled in ths area in order to deflect the
burden falling from a large bell.

With the advent of the PAlL WUTH bell-less top, wear of the stockline area
can be greatly dimshed. Some users have elected to elimate the anour plates and
use an abraion resistat refrctory ling intead.

Stack

The stack (someties caled the "in-wall") is the zone between the mantle (or
belly on a free stading fuce) and the stockline area. Smooth, unform lines (the
process "workig surace") of the stck are essential for unform and predictable burden
descent, blast fuace gas ascent and stable process control thoughout the fuce
campaign. Process considerations dictate a larger diameter at the base of the stack than
at the top. Typical stack angles ar -850 from the horizontal.

Mantle/Bellv

The mantle or belly (free stding fuace) area provides the tranition between
the expanded stack and bosh setions. Maitenance of the effectiveness of the
coolingling system is parcularly importt for the mantle tye fuace in order to
protect the mantle strctue. Ob\ iously thermal protection is importt for the free
stading fuace stle as well; however, the free stading design is less complicated and
more accessible in ths area

Bosh

The bosh area lies between the tuyere breast and the mantle/ belly of the
fuace. The bosh diameter incres from bottom to top. The inclination of the bosh
pennts the effcient ascent of the process gases and has been found to be essential in_

5-16
 order to provide the necessar zone seice life (the process gases are extemely hot and
internal chemical attck conditions ar severe). Typical bosh angles are ~80° from the
horizonta.

Boshes are constrcted in two baic styles, banded and sealed (Figue 17). They
can be cooled by varous technques.

Banded boshes are found in older mantle supported fuaces (they canot be
applied to free stding fuces). A number of steel bands are placed in incrementaly
the bosh; largest at the top) and are tied
increasing diameters (smallest at the bottom of

together With cenecting strps. Gas between the bands pemit the introduction of
copper cooling plates. Ceramc brick lig must be used as ai inltration would result
in oxidation of carbon-based lings. Ga leakage though the banded bosh can be high.
Ths style is not suitable for fues with high blast pressurelhgh top pressure
requiements. Banded boshes pro\ ide sucient flexibility to elimate the requiement
for a shell expanion joint in the lower porton of the fuace.

Sealed boshes, using contiuous steel shell plate instead of separate bands, are
employed to pemit the use of improved cooling/ing systems, elevated fuace
operating pressures, and the free stdig fuace style. Sealed boshes retain valuable
gases with the fuace, thus imprO\ ing the metalurgical process. As well, the seal
bosh, since it precludes ai entry into the linng, supports the use of carbon based
refractories.

Tuvere Breast

Hot blas ai is introduced to the fuace though tuyeres (water-cooled copper

unts) located with the tuyere brea The munber of tuyeres requied depends upon
the size (production capacity) of the fuce.
The tuyere breast diameter, tuyere spacing and number of tuyeres are inuenced
by the expected raceway zone sizg in front of each tuyere.

Tuyere stocks (Figure 18) convey the hot blast ai from the bustle pipe to the
tuyeres. The tuyeres are supported by tuyere coolers (water-cooled copper unts) which
are in tu supported by steel tuyere cooler holders (either welded or bolted to the
fuace shell). Special consideration mus be made in the tuyere breast shell and ling
design in order to maita effective sealing of the varous components in order to
prevent escape and loss of the fue gases8.

5-17
 Hear
The hear (Figue 19) is the crucible of the fuce. Here, iron and slag are
collected and held unti the fuce is tapped. The hear wall is penetrated by tap holes
(often called iron notches) for the removal of the collected iron and slag. The number of
tap holes is dependent upon the size of the fuce, hot meta and slag handling
requirements, physical and capita constrts, etc.
Many fuaces are equipped with a slag or cinder notch (usualy one per
fuce, although some fuaces could have two). The slag notch openig elevation is
usualy seyeral feet higher th the iron notch elevation. In earlier days, when slag
volumes were high the slag was flushed from the slag notch periodically. Ths
simplified the iron/slag separtion process in the casthouse. More commonly now,
however, the slag notch is retaed solely for intial fuace st-up procedures or for
emergency use in case of irn notch or other fuace operating problems.

Hear Bottom

The hear bottom support the hear walls and is flooded by the iron with
the fuace. As the campaign progresses, the hear bottom ling wears away to a
fixed (hopefully) equilibrium point.

The remaig refrctory contas the process and with sufcient cooling or
inerent insulation value protects the fuace pad and foundation.

Cooling System

Little reference ha ben made in ths paper so far in the provision of specific
ling technologies to the varous zones and areas of the fuace proper. Specific ling
technologies will be covered in grater depth by other authors in the course.

The application of spifc cooling technques (if at all) to individua fuace

zones is dependent upon many factors such as campaign life expectacy, fuace
operation philosophy, burden types, refrctories, cost constraits, physical constraints,
available cooling media, preference, etc. Different cooling technques can be provided
for different zones to assist the lig to resist the specific zone deterioration factors.

Generally speakg, the provision of adequate cooling capacity is essential in

each of the applicable fuce zones if the linig system located there is to surive.
Where the thermal, chemical and to some extent the abrasive conditions of the process
are exteme, suffcient cooling mus be provided to maitan the necessar unform
interior lines of the fuace and to protect the fuace shelL.

5-18
 Typically, the top cone and droat areas of the fuace are tucooled. The hear
bottom can be "actively" cooled by underhear cooling (ai, water or oil media) or
"passively" cooled by heat conduction though the hear bottom ling to the hear
walL.

The basic cooling options for the balance of the fuace are:

(a) No Cooling (tyicaly the upper portion of the stack is uncooled in many
fuaces)
(bl Shower or Spray Coolig
(c) Jaclæt or Chanel Coolig
(d) Plate Cooling
(e) Stave Cooling

Shower Cooling

Water is directed by sprays or by overfow troughs and descends in a film over

the shell plate. Effective spray nozze design, numbers and positionig are importt
for proper coverage and to mie rebound. Proper deflector plate design is essential
to ensure effcient cooling water distbution and to mize splashig. Shower
coolig is often employed in the bosh and hear wall areas. Spray cooling is
commonly applied for emergency or back-up cooling, primarly in the stack area.
Exterior shell plate corrosion or organc fouling are common problems which can
disrupt water flow or ÌnlÙate the shell from the cooling effect of the surace applied
cooling. Water treatment is an importt consideration to reta effective cooling.

Jacket or Chanel Cooling:

Fabricated cooling chabers or indeed strctual steel chanels or angles are

welded directly to the outside of the shell plate. Water flows at low velocity though the
cooling elements in order to cool the shell and ling. Jacket or chanel cooling is often
applied to the hear walls, niyere breas and bosh areas. Scale build-up on the fuace
shell and debris collection in the bottoms of the extema cooling elements can
compromise their effectiveness. Hence periodic cleanng of the cooling elements is
essential.

The critical area of concern in the cooling schemes mentioned so far is the
necessity for the shell plate to act as a cooling element. If exteme heat loads are acting
, ,
I upon the inide face of the shelL there will be an extemely high thermal gradient across
the shelL. Ths effect reslÙts in high thermally induced shell stresses and eventul
crackig. The cracks \\ill st from the inide of the fuace and propagate to the
outside. The cracks will remai invisible (other than a "hot spot") until they can fully
penetrate the shell plate. Thoug crackig of the shell plate results in blast fuace gas ~

5-19
 leake, exposed shell carburon and disruption of the cooling effect (parcularly
spray or shower cooling). Shell crackig into a sealed cooling jacket or chanel is
diffcult to locate and can resut in long fuace outae time for repai. Entr of water
into the fuace (often when the fuace is off-line and intern fuace gas pressure
canot prevent entr of coolig \'\-ater though shell cracks) can have detrenta effect
upon the fuace ling. Water in the fuace could be potentially dangerous due to
explosion risk (steam or hydrogen).

Shower and jacket coolig rely on the shell plate to conduct the process heat to
the cooling media; plate and stve cooling are confgued to isolate the shell from
process. ~. .
Plate Cooling

Instalation of coolig elements though the shell of the fuace (Figue 20) has
been a major fuace design improvement resulting in effective cooling of the fuace
ling and protection of the shell plate. Cooling is provided along the lengt of the
cooling element penetration into the linig. The inserted elements provide positive
mechancal support for the refrctory ling.

Typical cooling plate manufactue is cast high conductivity copper. Single or

multiple passes of cooling water can be incorporated.

Cooling boxes (Figue: 1 J \\ith larger vertical section have been produced from
cast steel, iron or copper.

Cigar tye (cylindrcal) coolers of steel and/or copper have also been
successfully employed5.9.

The philosophy of dens plate cooling (i.e. vertical spacing of 14" to 16" center-
to-center, and horizonta spacing of 24" center-to-center) has enhanced the cooling
effect and increased ling lie5.

Copper cooling plates haw traditionally been anchored in the shell plate with
retainer bars or bolted connections to permt ready replacement if plate leakage occurs.
More recently, plates have ben designed with steel sections at the rear of the plate for
welding directly to the steel shell. Whle sometimes tag longer to replace, ths stle
provides a positive seal agait blast fuace gas leakage.

Plate coolers are tyicaly intaled in areas above where the molten iron
collects in the fuace. Hence the mid point of the tuyere breast, right up to the
underside of the thoat anour is the rage of application.

5-20
 Stave Cooling

Cast iron cooling elements (Shaon plates or staves) have been used for many
years in the bosh and hear wal ar. These castings have cored cooling passages of
large cross-section. Whle their service lie was not remarkable in the bosh, multiple
campaign were common for the hea wcù1.

These staves often sufered frm low flow rates of marginal quaity cooling
water (scaling and debris depositioIlbuid-up) and someties casting porosity. Water
leak into !he hear wall can be a signcant problem.
~, .

In the 1950's, the USSR develope a new style of stave cooler (Figue 22) and
"natual evaporative stave cooling"lo. For ths design castings were of gry cast iron
contag steel pipes for water pases. The pipes were coated prior to casting to
prevent carburzation of the cooling pipe and metalurgical contact with the stave body
material. The staves are instaled in horionta rows with the fuace and the cooling
pipes project though the shell. Vertcal colum of staves are formed by the
interconnection of the projecting pipes from one stave up to the corresponding stave in
the next row.

Staves can be applied to al the wals in the zones below the anour (Figure 23).
Staves in the hear wall and tuyere bre are supplied with smooth faces. Staves in the
bosh, mantlelbelly and stack usualy haw rib recesses for the instalation of refrctory.

Evolution of the stave cooler design has been dramatic. Staves in the higher
heat load areas are now typicaly cas from ductile iron for improved thermal
conductivity and crack resistace.

Whle early stave design usd castable refractory (intaled afer stave
inlation with the fuace), ribs now normally incorporate refractory bricks, either
cast in place (with the stave body at the foundr) or slid and morted in place prior to
inlation in the fuace.

Staves are normally expected to reta a refractory ling in front for some tie.
Afer loss (expected) of the ling the st\"es are designed to resist the abrasive effects of
descending burden and ascending dirty gas. As well, they must absorb the expected
process heat load and resist thermal load cyclig and shock.

Four generations of stves (Figue 24) are commonly recognzed In the

1i .
industr

5-21
 First Generation (no longer commonly used):

(a) Four cooling body ciruits (with long radius bends which did not effectively
cool the stave comer).
(b) Gry iron casgs.
(c) Castable rib refrctory.

Second Generation:

(a), Four cooling body ciruits with short radius bends for improved comer
coolhi.
(b) Ductile iron casgs.
(c) Cast-in or glued-in rib bricks.

Thd Generation:

(a) Two-layer body coolig incorporating four or six cooling body circuits
(stave hot face) and one or two serpentine cold face circuits (stave cold face)
for additional or back-up cooling in the event of hot face circuit loss.
(b) Additional edge coolig (top and bottom).
(c) More frequent use of cooled ledges to support a refractory ling.
(d) Cast-in or morted-in rib bricks.

Four Generation:

(a) Two-layer cooling (simar to thrd generation).

(b) Cooled ledges.
(c) Cast-in wall brick lig elimatig the need for a manualy placed interior
brick ling.

Staves incorporating hot tàce ledges are more effective in retag a brick
ling than the smoother rib faced bricks. However, once the brick ling disappears,
the ledges are very exposed \\-ith the fuace. The ledges disrupt burden descent and
gas ascent. Exposed ledges tend to fail quickly. They are often servced by cooling
water separate from the mai stye cooling circuit(s). In ths way leakg ledge circuits
can be more easily located or isolated. Some stave suppliers are now providing separate
ledge castings so that ledge crakig and loss will not damage the parent staves. As
well, there is some curent change in philosophy to abandon the application of ledges
entirely.

Varations of the basic stve generation styles are common5. For example,
staves of four generation style utg a refractory castable for the wall ling have

5-22
 been employed successfuy. Alternely, brick lings have been anchored to the stave
bodies. Such approaches can be used to substitute for brick support ledges.

A "fift" generation of stves design ha been the developed. It is the copper

stve (Figue 42). Rolled copper plates are drlled to form cooling passages. Rib
recesses are machied to permt the ination of low conductivity refrctory bricks.
Test intalations of ths stve ty have ben successfui12.13.

Natual EVaporative stve Coolig (NVC) (Figue 25) is a technque where

boiler quaity water is introduced into the bottom row of staves and flows by natual
mean up the vètcal cooling circuits. As the process heat conducts though the stave
and cooling pipe into the water, the water in tu heats up. As the water wans, it
expands. Since cooler water is being intruced below, the wan water tends to move
upwards. At some point in the vertcal cooling circuit, the water will be at the boiling
point. As the water changes phae to steam, due to the latent heat of vaporization,
additional heat is absorbed (drving the phase change). Afer boiling begin, two-phase
flow (water and steam mie) ascends the cooling pipes to the top of the fuace.
Usualy located on the fuace top platíòrm are steam separator drs used to extact
and vent the steam to atmosphere. Make-up water is introduced to the dr (to replace
the discharged steam). The water is piped back by gravity to the fuace bottom and is
fed once more to the staves.

Ths cooling technque is very effcient and has low operating costs; there is no
pumping equipment.

More recent improvements ha'-e been to boost the flow of the cooling water
with recirculating pumps (Forced EVaporative Cooling -FEVC) in order to ensure
unform cooling water flow and to cool the recirculating water (Forced Cold Water
Cooling - FCWC)14. Both of these approaches have resulted in improved stave and
ling life.

Russian stves and natu evaprative cooling were first used in Nort America
at STELCQ's Hilton Works 'D' Blas Furce in 197410.

Staves provide an excellent protection for the shell plate thoughout their service
life (which is extended while the interior brick ling remai in place). Stave
application has been implemented in al areas of the fuace from hear wall up to and
including the upper stack.

Whle some people (priary non-stave users) maitan that stave leak
detection and stave cooler replacement is complicated, in fact simple and effective
means have been developed.

5-23

¡
 One drwback for conversion of an existig plate cooled fuace to stave
cooling could be the cost of a new shell. However, if the existing shell is aleady in
distress and must be replaced in any event, the conversion cost is not a major factor.

CASTHOUSE

The casthouse (Figue 26) is the area or areas at the blast fuace where
equipment is placed to safely extct the hot metal and slag from the fuace, separate
them and direct them to the appropriate hadling equipment or facilities.
-
As mentÍoned earlier, the iron and slag are removed from the fuace though
the tap hole (iron notch). Only inuently today is slag flushed from the slag or cinder
notch.

Tap Hole Equipment

Tap hole equipment mus be reliable and require mium maitenance.

Furaces typically cast eight to eleven ties per day.

Mud Gun

The mud gu (Figue 27) is us to close the tap hole afer casing is complete.
A quatity of clay is pushed by the mud gu to fill the worn hole and to maintain an
amount of clay ("the mushroom"') \\itb the hear. The mud gu is usualy held in
place on the tap hole until the tap hole clay cures and the tap hole is securely plugged.

A "hydrulic" mud gun uses hydrulic power to swig, hold and pus the clay.
Typical clay injection pressure is in the order of 3,000 to 4,000 psi, permtting it to push
modem, viscous clays into fuaces operatig at high pressures. The hydrulic gun is
held agait the fuace with the equivalent of 15 to 35 tons of force. Ths style of mud
gu can be swug into place in one motions.

An "electromechacal" gu ha thee separate electrc drves for unt swing,

barel positionig and ramg. Hence several separate motions are required to
accurately position the gu at the tap hole. Clay injection pressure is in the range of
only 600 to 1150 psi. The electromechacal clay gu is latched to the fuace to keep
it in place durg plugging!5.

Tap Hole Drill

A tap hole drll (Figue 28) is used to bore a hole though the tap hole clay into
the hear of the fuace. A dr unt is swug into place hydraulically and held
hydraulically in the workig position. A pneumatic motor feeds the hamer drll unt ~

5-24
 (with an attched drll rod and bit) into the hole. Compressed ai is fed down the center
ofthe drll rod and the drll bit to cool the bit and blowout the removed tap hole clay.
When the tap hole ha penetrted into the hear the drll rod is retrcted and the drll
swigs clear of the hot meta steam.

Soakg Bar Technque

In recent year, the application of the soakg bar practice has improved the
casting process. Whle the tap hole clay is stil pliable afer plugging, a steel bar is
drven int~ the tap hole by the tap hole dr. Wlle the bar sits in place durg the time
between casts, it, heats up via conduction from the hear iron. Ths permts curng of
the tap hole clay along its entir lengt (as opposed to curg with the fuace and
setting at the outside near the fue cooling elements). The cured tap hole clay is
more resistat to erosion durg taping, thus improving cast flow rate control. Less
clay is requied to replug the hole. When the tap hole is to be opened, a clamping
device and back hamerig device on the tap hole drll extact the rod. The timg for
tap hole openig can be more easily controlled (predicted) than by conventional drlling.
Ths featue is importt for smooth fuce operation and for scheduling of hot meta
delivery to downstream facilities5.16,¡-.

Same Side Tap Hole Equipment

Mud gun and drlls have normly been instaled on opposite sides of the tap
hole. More recently, design development has permtted instalation of the unts on one
side of the tap hole. The drll S\\ings over the mud gu or vice versa. Ths type of
intalation faciltates improved access for tap hole and trough maitenance and the
improved application of trough and ta hole area fue collection.

With the advent of tuyere access platforms to facilitate tuyere and tuyere stock
inspection and replacement, the headoom available for the tap hole equipment has
dimshed. However, same side ta hole equipment intalations (Figure 29) can be
achieved with low headroom (for exaple 7.25 feetr

Trough and Runer System

Typical hot meta and slag taping rates are in the range of four to six and thee
to five tons per miute, respectively. The trough and ruer systems must be designed
to properly separate the iron and slag and to convey them away from the fuace for
flow rates with the normal flow rate rage and for unusua peak flow rates.

5-25
 Trough

The iron trough (Figue 30) is a refrtory lined tudish located in the casthouse
floor and designed to collect irn and slag afer discharge from the fuace. The iron
flows down the trough, under a skier and over a da into the iron ruer system.
The iron level in the trough is dictaed by the da. Proper dam design submerges the
lowest porton of the skier in the irn pooL. The slag, being lighter than the iron,
floats down the trough on top of the irn pooL. Since it canot sin into the iron and
though the skier openig, it pols on top of the iron until sufcient volume collects
to overfo~ a slag da and ru do\'\TI the slag ruer. At the end of the cast, the slag
ruer da height is lowered to dr off most of the slag. The residua iron is retaed
in the trough to prevent oxidation and thermal shock of the trough refractory ling.
When maitenance of the troug lig is requied, the iron pool can be dumped by
removing the iron da (Baker da t: -p), or by openig a trough drai gate, or by
drllng into the side of the trough (at its lowest point) with a drai drlL.

The trough bottom is usuay designed with a 2% (minum) slope for effective
drainig. Trough cross-section and lengt design are important for effective iron and
slag flow pattern, retention and separtion. A "good trough" design results in hot metal
yield improvements17.

Effective trough ling and coolig technques are importt for ling life, hot
meta temperatue, and casous Stctual steel and concrete heat protection
considerations.

Trough traditionally were contaed in steel boxes "bured in sand" in the

casouse floor system. ImprO\'ed trough design incorporate forced or natual ai
convection or water-cooling.

Runers

Modem practice requis tht the ruers be as short as possible. Ths

mies iron temperatue loss and reuces ruer maitenance and fue generation.
As well, shorter ruers can resut in reduced capita outlay for casouse building
intalation or modification.

Since the ruers mus slope away from the fuace, the casthouse floor
generally follows the same slope as the ruers. Steep floor slopes can result in diffcult
access and workig conditions. Now, where possible, operators and designers tr to
incorporate relatively flat floors to enhance casthouse operation and improve safe
workig conditions.

5-26

'I
 Slag Runers/Handling

Slag ruers are usuay designed with a 7% (minium) slope. Slag can be
diected to:

(a) Slag pots for raway or mobile equipment haulage to a remote site for
dumping.
(b) Slag pits adjacent to the fue for air cooling and water quenchig prior to
excavation by mobile equipment. Pit ru slag is used for backfll or can be
_ cruhed for use as aggrgat.
(c) Pelletzig or grulation facilities adjacent to the fuace for conversion of
the slag to material sutale for backfll, aggregate or Portland cement
replacement. Graulation unts (Figue 31), in paricular, can be provided
with systems to elimte fue emissions associated with envionmenta or
industral hygiene problems.

Iron Runers/Handling

Iron ruers are usualy designed with a 3% (minum) slope. Iron is usualy
directed to hot metal tranfer ladles (torpdo cars/ bottles/ etc.) for movement to the
steel shop (or iron foundr, pig caser, iron granulation unt, etc.). Whle normal
practice used to have one iron ruer system with diverter gates directing the iron to
different pourng positions, each \'\ ith a ladle, more recent application of the tilting iron
ruer practice has been beneficial (paricularly in the "shortened ruer" benefits
mentioned earlier).

A tilting ruer (Figue 32) diyerts molten iron to either of two torpedo cars
afer collecting it from the iron ruer. Often provided with an electrcal motor-drven
actuator (with a manua handwheel back-up), the ruer is tilted at about 5° to divert the
iron. A pool of iron is held in the titig ruer to minmize splashig and refractory
wear. When one torpedo car is fied. the ruer is tilted to the opposite side to fill the
other ladle. If required, a 10comotIye removes the full ladle and spots an empty ladle in
its place. Ths operation can be done \\1thout plugging the fuace. When the cast is
fished, the tilting ruer is tilted an additional 5° to dump its pool of iron into a
torpedo car.

Fume Collection

Fume collection requirements and applications appear to var signficantly in

Nort America. Furaces curently have ful, parial or even no casthouse fue
collection.

5-27
 Exhaust fan and bagouse capacity in the order of 325,000 to 400,000 acfI
(depending upon operation and design prctices) is tyical for ful fue captue of a two
tap hole casthouse intalation (for troug. ruers and tilting ruers).

Proper design and application of fue collection ruer covers canfaçilitate

casthouse access (i.e. flat floor confguon using steel slabs or plates) for personnel
and mobile equipment crossover. As well, ruer covers can reduce hot meta
temperatue loss and improve ruer refrctory longevity.

So-ie fuaces employ flame supression which elimiates the oxygen in the
ai diectly over-lhe trough and iron ruers. Products of combustion prevent oxidation
of the iron surace reducing visible parculate and fues.

DESIG~ EXALE
Introduction

The eight sections above are a cataog of the blast fuace equipment and
designs. Followig is a shortened and simplified example of the application of some of
that equipment in the development of a proposed blast fuace modification. Ths
example is limited to the fuace proper. For a ful study every element in the process
chai, from raw materials delivery to hot meta consumption must be checked to see
that there are no "bottle necks" in the system which will prevent the fuace from
meeting the goals set out for the modifcation. Ths example will use an existing Nort
American blast fuace. It's lines and cooling are typical of the blast fuces built
around the Great Lakes. Ths is a short yersion of one of the many thought process that
could be used to develop the improvements and some alternatives that would be
considered for an upgrde of ths fuce. The goal of ths study is to search out the
operation and equipment that will best meet the futue needs of ths operation.

Financial justification is the oyerrding consideration for all improvements.

Depending on the size of improvement the economic study can become very extensive.
Some of the larger fiancial stdies reach from the ore mine to the customer. Since
costs are confdential and site speifc they will not be presented here. There is one cost
aspect to keep in mid, that is the effect of the fuce reline on the cost of
improvements. Relines are a major capita expense. At the same time they present an
opportty to improve profitabilty thugh facility improvements. Ths opportty
comes first from the obvious fact tht the improvement's cost will not have to include a
down time penalty. Also, facility maitag costs, that money which would have
been spent to repai or replace the less cost effective materials, equipment or fuace
designs can be deducted from the cost of the improvement. The benefit resulting from
the upgrade need justify only the differnce in cost between repair and improvement.

5-28
For ths reason studies which seach out those opportties should be performed well
in advance of any planed relie.

Background

Ths example if bas on one of the tyical small Nort American blas
fues with 30,000 cubic feet of workig volume. (Figue 33) Ths fuce's
performance has been very good (Table 1).

Example Blas Furace

~. .

Present Opration

Smelting 2,491 NTHday

Production 2,428 NTHday
Percent delays 2.5 %
Operations 1 ,404 mi./day

Coke 835 #/NTH

Oil 115 #/NTH
Pellets 3,096 #/NTHM
Slag 53 #/NTH
Slag Volume 410 #/NTH
Cu. ft./NTH 41,119 SCFM
Wind 71,143 SCFM
Oxygen 21. 04 %
Temperatue OF
1,729
Grais 2.79
OF
RAT Calculation 3,706

Hot Metas Si 0.44 %

Sul 0.062 %
Mn 0.43 %

Table 1

5-29
 However the reline history leaves somethg to be desired.. Campaign have
been short and uneliable (Table 2).

1. CAMPAIGN TOTAL Tons 2.2MM

2. Tons 2.4 MM
"
~. Tons 2.4 MM
4. Tons 2.8MM
5. Tons 3.2MM
6. Tons 1.6 MM
7. Tons 2.0MM
8. Tons 1.5 MM

Averae Campaign 2.2 MM

Average Hear 4,623,791

Table 2

The causes for the fuce' s ling life problems are hardware specific. The
cooling is inadequate for today' s operation and the fuace lines are not good.

The post reline production requiement is 3,000 NTHMday which is 120% of

that achieved by today's operation (Table 3).

Production
. Production Daiy Intaeous. . . . . . . . . . . . . . . 3,000 ton/day
. Scheduled Do\\TI Time .... .................... 2 %
. Unscheduled Do\\TI Time..................... 1.5 %
. Production, Daiy A \"erae .................... 2.895 ton/day
. Anua Production . . . . . . . . . . . . . . . . . . . . . . . . . . ... 1,000,000 ton

Campaign Life

. Campaign Lengt. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 years

. Interi Stop At . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 years
. Duration of Interi Repai.................... 30 days

Hot Blast Delivered

. Hot Blast Flow Rate (Max.) .................. 80,000 scfi
. Hot Blas Tempenr ........................ 1,850 0

Table 3. Rebuild Objectives

5-30
 With some fuce upgres and chages in the operation ths production goal
is achievable.

The more diffcult problem wil be to assure that our fuace can achieve the
desired eight year campaign. Furce campaign must be looked at for both lengt of
tie and tota tonne The expted campaign tonnage is 8.8 millon tons when the
fuace is ru for eight year at the new production rate. Ths is approxiately four
times the average campaign tonnage previously achieved. In light of the fuaces
previous performance an eight year campaign with the existig fuce design fuace
would be ~ikely.
,to .

Taken together, the performance improvements needed are quite large. If

today's technology can be implemented they should be achievable.

Comparson

How do the key pareters for the present operation of the example fuace
compare with the rest of Nort America? If the proposed operating goals are imposed
on the existig facility how would it compare?

Stack Productivity

The fist measure to be looked at is stack productivity. Ths is rated in tons per
day per hundred cubic feet of \vorkig volume. Working volume is the fuace's
internal volume calculated ben:veen the tuyeres and the stocldine. Workig volume is a
measure of the volume of materials actuly in process. Therefore, tons per day per
hundred cubic feet of workig volume is a measure of the specific productivity of a
blast fuace (production rate per unt volume). Ths measure makes it possible to
compare the workig rate of diferent size fuaces.

As shown in (Figure 34) the present production is with the normal range of
operating fuace. The proposed 3,000 NlHday production imposes a very high rate
of productivity, on ths older fuace, when it is compared to other fuces. With ths
in mind we will now look at other pareters which will afect the modification of the
fuce proper.

Hear Productivity

Hear productivity is rated in tons per day per hundred cubic feet of active
hear volume. Active hear volume is the fuace's internal volume calculated
between the tuyeres and the ta hole. Active hear volume is a measure of the
fuace's holding capacity for the liquids melted in the workig volume (above the
tuyeres). Therefore, the tons per day per hundred cubic feet of active hear volume is a

5-31
 measure of the specific capacity (thoughput per unt volume) of a blast fuace's
hear.
At the present operatig rae. the existing hear is at the highest productivity of
any of the fuaces sureyed in ~ort America (Figue 35). It may be second in the
world. At 3,000 NHTMday. me existg hear could be operating at a world record
rate (Figue 36). The existg hear design with internal staves (Figue 37) tyically
wears well but, it is way too smal. At 3,000 NTHMday, the fuace would be
untable. It would be diffcult to operate because the hear liquid levels will change
rapidly wtich would cause varatons in gas flow pattern, gas utilzation and blast
pressure. Also/because of mes rapid changes in liquid level, the fuce would not
come otIwid easily and, it ''iould probably be a tuyere burer.

The best opportty tòr improvement on ths fuace is to remove the present
hear thoughput limitation by increasing it's volume. It may be considered to be a
requiement for production of 3.000 NTIday. To provide the additional volume
needed a shower cooled carbon hear. could be intalled inside the existing shell
(Figue 38). Hear volume "il increase from 2,965 ft to 3,467 ft. At 3000 tons per
day specific productivity is deceased to the point of being manageable (Figure 39).
Ths hear design has been extemely successfu in Nort America, having no elephant
foot and litte or no penetrtion beyond the cup. The size increase makes the hear
more manageable at 3,000 t/d. However, it is stil in the upper range of normal fuace
operation, so the operators mus exercise very good hearh liquid is level control. Ths
high thoughput would requi around 90% time spent casting. To make ths time
casting possible, cast floor modcations will be needed. They are not included in ths
example.

Campaign Life

Looking at the previous campaign life (Table 3) and the fuace's wear lines
(Figure 40) presents another opportty to make signficant improvements on ths
fuace. The bosh and mantle ara are badly worn and are too smalL. On a mantle
fuace, the mantle's protection and stability is crucial for long campaign. Ths
fuace's lines violate Bercz:nski's 4 x 4 rule, British Steel's 12 x 5 rule and
Carenter's 21.5 degree aw-shIt lie (Figue 41). Since the geometr is bad It is
essential that the mantle protection be able to handle intense activity, varability, and a
high heat load. A row of copper stves at the mantle is capable of dealing with these
conditions. It is also the thest option, thereby improving the geometr and openig
up ths area for better process operation (Figue 42). Copper staves are expensive but
they are very economical in comparson to the alternative which, in ths case, would be
to enlarge the mantle.

5-32
Production

The production needed usy deteres a facility's size. However, in sizg

a blast fuace, raw materials, product chemistr and even operating philosophy enter
into the determation of specifc prouctivity for a fuace and therefore the size
needed to produce a given amount of hot meta. In Nort America specific productivity
(Figue 34 J vares from six tons per hundred cubic feet per day to twelve tons per
hundred cubic feet per day. From th wide range of possible operating rates the
workig volume of the fuace mus be caculated. Productivity and therefore fuace
size will be based on the fuel rate. The expected fuel rate is determed by comparg
the proposed operation to a knO\\T ba operation and adjusting it's fuel rate for
differences between the two opetions in raw materials, hot blast pareters, iron
quaity and even the operatig phiosophy. One example comparson of the operating
pareters afected by these modifcations is shown in Table 4.

Opration
Present I Proposed Difference Coke Rate Effect
Smelting 2A9 i 3,000
I
Production 2A28 2,924
Delays - Minute 1.056
Percent 2.53% 2.53%
Operation - Minute/Day i A04

Coke 835 782 -53.4

Oil 1 15 150 35 -28.3
950 932

Pellets 3,096
Slag 53
Scrap or DR Slag Volume 410 375 -35 -7.1

Cu. ft./NTH 41.19 38,403

Wind 71.-B 80,000 8,857
Oxygen 21.0 22.06 1

Temperature 1. 729 1,850 121 -18.7

Grains 2.79 4.00 1.21 3.6
Horn. H2O 0.02 0.263 0.8

RAT Calculation 3)06 3,726

Hot Metal Si 0.+4 0.4 -0.04 -3.7

Sulf. 0.062 0.062 0.00
Mn. OA3 0.43 0.00

Table 4

5-33
 Increasing production on an existing facilty nearly always requies that
somethg be chaged. Opors would produce more if there were not constrts
imposed by some existg condition. Decisions as to how to increase production are
fudaenta. Either incras the fuce size (workig volume) at the same
productivity, or increase the fue's productivity, or depending upon individua
circumces both.

INCREASE PRODUCTnTY
,t.-

1. MORE OXYGEN

(a) Wind
(b) Enrchment

2. LO\VERFULRATE

(a) Increa.-: Hot Blast Temperatue

(b) Increa.-: Injectat
(c) Impro\'ed Burden

- Scrap or HBI
- Decree Burden Moistue
- Impro\'e Coke Quaity

(d) Hot :\leta

- Lower Silcon
- Higher Sul

(e) Impro\'e Opration

- Dry Hean
- Burden Distbution

Table 5

Increase size is a capita cost. Increasing productivity increases operatig cost.

Increasing both increases both capita and operatig costs.

5-34
 Lookig to a comparson \\ith others (Figue 34), it seems that there are 3 tiers
of blast fuace operation. In gener the first tier are rug above 9 tons per day per
100 cubic feet of workig volume. These are high cost operations using a high
percentae of scrap in the burden and signficant amounts of oxygen in the hot blast.
Furace's operating at the very highest productivity also need more robust equipment.
Hence, high capita and high COSL The second tier of fuaces operate between 7 and
8.5 tons per day per 100 cubic feet of workig volume per day. These fuaces will
have either high cost or high capita In general, the cost of operation goes down as
more capita is spent. The best option for most operators will be in ths area. Blast
fuaces operating in the thd tier, at less than 7 tons per day per 100 cubic feet of
workig volumè' are being ru at ths low rate for reasons specific to their operation.
Perhaps a niche market is smaler th the fuace capacity, or there is a supply of very
low cost raw materials that makes low productivity cost effective.

With these examples of productivity in mid, look at the stack refractory lines
for the example fuace (Figue 33). It is, at best, marginal in refractory thckness and
the shell is likely to be distessed durg its present campaign. Solutions designed to
keep the existing stack shell and make the next campaign longer by as adding to the
existing coolers or inertg st\"es between the existing cooling plates are likely to
result in a costly intallation. Alost all options which keep the same shell will result
in the same workig volume so the fuace must be ru at high productivity in order to
reach the desired 3,000 NTf/day. Ths would be a high cost, high capital operation.
Another solution might be bener. With the normal wear lines (Figure 39), shell
damaged is expected. Since the shell should be replaced, enlarging the workig volume
becomes practical. Ths option also fits in with the mantle stave scenaro (above) which
is necessar to achieve the spifed ling life. With a shell change there are a number
of options all which are aied at reachig a specific productivity which allows the
:face to produce 3,000 ~-llday goal without exceeding 8.5 tons per day per 100
cubic feet of workig volume. In Figures 43, 44, and 45, five options are presented.
The options will be subjected to operatig and economic analysis to determne which is
the best fit for ths paricular location.

Operation

Only one option to increase hear volume was developed. Ths is because
either an increase in diameter or height will require changing the whole :face. Even
with the increased hear volume, ths fuace's hear has the greatest constraint on
increasing production. To operate reliably at 3,000 NTHMday, the operators must be
given equipment to allow a very high percent time casting.

On the other hancL chagig the stack with the limts of the mantle rig and
the fuace top lip rig afects only the stack. Since the example's stack has been
distressed in every previous campaign, it will require extensive repais. So, five options~

5-35
(Figues 43, 44, and 45), al of \vhich fit with the above consaits have been
developed. Each of the stck options have a different capita cost and a different
workig volume. Each option wi operate at a different rate when producing the
requied 3,000 NTHday (Figu.l. The different operatig rates will requie
different operating practices (Table 5) and their operating cost will var. From the
operating and capita cost, the renr on investent (ROI) for each case will be
calculated. From ths data the fi decisions regarding capita investments are made.

CO~CLUSION
-
the blast fuace followed with a
Ths paVèr began with a gener description of

more detaled look at the blast fuce proper, it's components and it's ancilares. Ths
cataog of components is followed \\ith a short design example which demonstrates one
method for selectig from the many blas fuace options available. Ths paper is
intended to ilustrate the requirements and the "usua" practices for each area or
component and explore some of the alternative technologies which are available today.
More detaled discussion of specific components and their effect on process will be
given in another presentation.

The modem blast fuace ba dewloped over the past four centues. It exists
with a system of changing consts and technologies. The blast fuace has shown
itself to be a process that can be readily modified and improved to suit changing
requiements. It will continue to chage and improve in the futue. It provides a
challenging field of endeavor to all those people involved in the ironmakg field.

ACI00\VLEDGMENTS

I wish to acknowledge tht ths presentation is a continuation of the previous

Design II paper by Mr. Robert G. Goff.

REFERENCES

(1) Author unown "Univers Rota Distrbutor", PAUL WUTH-CRM

promotional document, 1984, pp. 2-3.

(2) Author unown, "Components for B.F. Top", MA-GHH promotional

document, undated, pp. 2-3.

(3) Author unown, ''NPPON STEEL Blast Furace Charging System", NSC
promotional document, 1987.

(4) Author unown, "NKK Typ ::Im"able Arour", NKK promotional document,
1979.

5-36
 (5) Bernard, G et al, "Modern Blast Furace Design by PAUL WUTH S.A.",
PAUL WUTH promotiona docinent, 1991, varous pages.

(6) Author unown "The Bell-Less Top Charging System", PAUL WUTH
promotiona docinent 1985.

(7) AISE Sub-Commttee No. 27, New Steel Pressure - Contag Components for
Blast Furace Inaton. AISE Techncal Report No. 27, Association of Iron
and Steel Engineers, 1984.

(8) Goff, R.G.,

,~. ,
"Six Yea of Maitenance Experience on STELCO's Lake Erie
Blast Furce", Iron and Steel Engineer, July 1987, pp. 17-22.

(9) Dzennejko, AJ., G. Hoelpes et aI, "Design Considerations for Utilzig

Cylindrcal Cooling Elements in the Blast Furace", PAUL WUTH, 1988.

(10) Blackbur H.W., "Evaporave Stave Cooling in a Modern Blast Furace",

presented at the American Irn and Steel Institute, 1976, pp. 1 -8.

(11) Wak, S. et al, "Report on );SC Stave-Cooled Blast Furaces", 4th International
Stave Conference, i 986, Hamton, Ontao, pp. 1-41.

(12) Bachofen, 1.1. et al. "Copper B.F. Staves developed for Multi-Campaign Use",
presented at the AISE i 991 Exposition.

(13) Author unoWI "Coppe Stave for B.F. Cooling", MA-Gll promotional
docinent, 1991.

(14) Dercycke, 1. and M. Sohi. "Characteristics and Pedonnance of SIDMA's

Stave-Cooled Blas Furce 'A"', 4th International Stave Conference, 1986,
Hamlton, Ontao, p. I-I.,l

(15) Author unoWI "High Pressure Clay Gun", BAILEY ENGINERS

promotional docinent (Puct Cataog No. CG-2- 1 3A), undated.

(16) Author unown "Compact Tap Hole Guns and Drills", PAUL WUTH
promotional docinent 1980. pp. 2-5.

(17) Ruther H.P. and H.B. Lungen et al, "Refractory Technology and Operational
Experience with Tap Holes and Troughs of Blast Furaces in the Federal
Republic of Germany", Metaurgical Plant and Technology, V oline 3, 1980, pp.
12-29.

(18) Carenter, J.A. and D.E. Swanon, "Burs Harbor 'D-4' Reline Improvements
and Results", Ironmakg Conference Proceedings, Voline 53, 1994 pp. 351-
361.

5-37
 FIGURE 1
BLAST FURNACE PLANT

PUliPHOUSE
~Al.P

SlAG PIT

!Q
BOURS
AND
TURBO
BLOWERS
D
r-
D
HOT i.ET AL TRACKS
CAS"" OU sc S

RAW MATE~lAI.
REClilo

OMAID
ljALJ
FIGURE 2
00
PLAN OF BLAST FURNACE PLANT

5-38
 ~ OFGAS SYSTE
\
FUNAce TOP
ClRGI SYSTD

SKIP 8R1GI

HOPER CARS IIGH Ut

o 0 0 STOCKHOSK

SCALa CAR

FIGURE 3
SECTION THROUGH
BLAST FURNACE PLANT

5-39
-Sca_oua
COIC.

STORA_
BI WrT
FEERS

_RATIG
SCRDN
wmGH
HOPS
WI
FES BLAST
COLLCTIG AN CHGIG CONVEYOR FURNACe

FIGURE 4
FLOWSHEET FOR
AUTOMATED STOCKHOUSE

BLAST FUNACE DRI HOE l."fOR ft~-

8TOCHOUSE ~R~
--G eOl~

FURNACE

FIGURE 5
A CONVEYOR FED BLAST FURNACE

5-40
 ~ .P CAR

RICIIYING
SULL HOPPR
nu
ROO

SULL
BEL
SMALL BILl
HOPPR
-i
LARGE

ROD GAS SEAL

LARGe LARGI BILl

Bm.L HOPPER

TOP CONE

FIGURE 6
CONVENTIONAL
TWO BELL TOP

5-41
 SKI CAR

FIGURE 7
McKEE DISTRIBUTOR TOP

5-42
 I
,~ \
8K. CAR

FIGURE 8
CRM UNIVERSAL ROTARY
DISTRIBUTOR TOP

5-43
 ~
SKIP CAR Oi

RECEIVING
HOPPDt
DRift
AS..'" Y SEAL VALVE
RIVOL VitO
DISTRIOR SMALl BILL
HOPIR ROD
S"AU BELL LOWIR SIAL

LARGI Bil GAS SIAL

ROD

LARGE BEll LAROE

BEL
HOPIR

FIGURE 9
GHH LOCK HOPPER TOP

5-44
 DRI
TOP COl
AR
SU
MOV AB
ARMOUR
PLA ,.
ROD

SH
/ AN
STOC
ARMOU
PUTU

SUP
lUG
J

,.. .:/',jl'P£
/'
tiò

FIXED ARMOUR li V ABLI ARMOUR

FIGURE 10
STOCKLINE ARMOUR

5-45
 FIGURE 11
PAUL WURTH
TWO HOPPER BELL LESS TOP

5-46
 8K. CAR

I HOPPIR
-/FIXID RIC!lV~\
I RECEIVINO
CHUTI 8K. CAR
_~ :JHOPPER.
MOVABLI FlX!D
RECEIV"Ø
V AL VI! CHUTE
LOWER
fr 1-LOCK
CA81NO
VI BFURCA TlD LOCK
i
.¡ CHUTI HOPPER
-.
TUBULAR LOWIR
CHUTI V AL Vi
8UPPORT REPLACEMEN
FRAIIE DRIVI CA81NG

Y'
::__GOGGU__
VALVE --~
IIAIt ROTATItG
, GIARBOX BURDEN
DlaTR.unON
--ROTATItG BURDIN CHUTI
DI8TRllTIO CHUI'
DOUBI HOPPER CENTRAL FI!D COMPACT
WITH IIOV ABU! FIGURE 12
RlcilVINO CHUTI
INLIT PAUL WURTH BELL-LESS TOP STYLES
 RECEG HOP
UP II TI GA 11
UP SEAL VALVE
CO"'CTJ FOR
EQUAUZI AND
REL

LOC HOPR

LOWE II TEIA GATE

Lawn SIA VAL Q
.
r __ VAL Q CASita
.j,l~ ~ FE SPOUT
, ~ - MAlt GER BOX

TOP COt

FIGURE 13
SECTION THROUGH PAUL WURTH
CENTRAL FEED BELL LESS TOP

5-48
 ./,'
..

FIGURE 14
BURDEN DISCHARGE FROM PAUL WURTH
BELL LESS TOP DISTRIBUTION CHUTE

MAH£

S~RT
COlUMN

--
IIANT SUPPORTED FURNACE FR£! aT ANDINO FURNAC!
FIGURE 15
MAIN FURNACE STYLES

5-49
 TOP RIG

OF AKa!

THOA T

DlII
THOAT ARMOUR

STOCKLJNE
· i ii
l . 3
LEVE
! õ2
c -' Ii
~ =- Gl
SHE
I Š = ST ACX
- t' e
;.
e IL
e l-
~
!¡ i~ ~
Gl :&
- Ii
:z Gl TUYERE
I i BE y
BU8T1 PIPE
TUYERE STOCK

CINDER NOTCH
IRON NOTCH
IRON NOTCH
- - ELEVAnON
HEARTH WALL

UNR HEARTH

FUC! PAD
OR FOA nON

FIGURE 16
FURNACE NOMENCLATURE

5-50
 STACK

8PRA yJì'
DEFLECTOR-l
B08H "
BAND8

Vi
i CpOLOLlNO~t:J
Vi ATE8
.. WATERJL
RETURN
TUYERE TROUGH

BANDED 8PRA Y OR 8HOWER JACKET PLATE STAYE

PLATE COOLED COOLED COOLI!D COOED
COOLlD 8EALED B08H 8EALED 8EALI!D aEALeD
B08H B08H B08H BOaH

FIGURE 17
BOSH CONFIGURATIONS
 aT AVE
TUVERE COOER
RaT AIMIG BAR

~
TUveRE COOER
HOLDeR
TUVERe COOLER
aTOC TUveRe

BLOWPIPE
RAM
RoaETTE
aHEU FURNACE UN"G

FIGURE 18
TUYERE BREAST

5-52
 CARBON BLOW If
BRICK
tEARTH UNING
WALL ~ If=ll ;;~'PRAY
CARBON BLOCK ~. COOLlfG
TAPHI ~
mYÄnON - -~_ - CERA"C
- t-__~ . TAPHO!
__ -lliION
IHn 0' ~_ _L .L BLOW IN LIlING
i,:i:-:i:l. =: ~~I~ARBON -----(
Vi
i
)mM~ll==;; ~RAM
Vi aT A VI
w -- II I \I I I I ~,.!-, ~ Ii I I I I i ~i/ IHILL
RAil __ IIII \ / I I I I II~ l:0l
II II ~ CARBON BEAM \ \ I ..",...".,. I I Ii ILL _ PANEL
INDCI!
DRAFT AIR
UNDRtEARTH
COOL"Ø PlHS in CERAMIC CERAMIC '- WATER
CERAMIC ~ LIVILUNG \ CERAMIC LEVELING COLLEcnON
TROUGH
CONCRI! "- BOTTOM / BOTTOM
PAD PLA TE CONRETe
\ PAD PLA TE

FIGURE 19
HEARTH CONFIGURATIONS
 FIGURE 20
DENSE COPPER
COOLING PLATE
INSTALLATION

FIGURE 21
COPPER COOLING
BOXES AND
CYLINDRICAL
(CIGAR) COOLERS

5-54
 RI
~

_ IICD
aFACTORY

COa~G pp "/ ~ BODY CAS,..

PRoncnoN
Pt
, ¡

COLD FAC! 8l HOT FACB

IR RECE RlCTORY
0-1 1 cu FOR CLIT

FIGURE 22
ST AVE COOLER

FIGURE 23
LOOKING UP
INSIDE A
STAVE COOLED
FURNACE

5-55
,

'/ i
 r:~ ~:;' rwli
Ii I0/1
I /I
/II II0 III
&);'1 &);'1 r 1 ¡J
i I I
i
I I
I I
I
II
I I /I I I
/I /I I I
I I II
/I /I III III
/I
:ii
i! /: 0
'=~
,; I
: ii
I I
I 0 101
I
I

II i I i I

/I I I II II
/I /I I I I I I
! ! !
I
I

I11011
II I /I
I' 1011
II :Ii I

o 0:
i I I I
I, '
II II
/I /11/1
I II =~
I

'~
:; ¡ i
11..
l~ D
"~-t '~\I .. ,~~ d~
COL FACI SI)I COL FAC! SiDe
FIST SECOND
LEGE P!E
COR 'ft
(~---------7'
~ -r"T'" "" r. (~---------7'
-r -r "T'" "" r.
I.) 1 r(-~). ì r (. I.) 1 r(-~). ì r (.
i I I I T Ll.H'~ i I I I T Ll.H-i
I ~ ....1"11 ii-~ ....1"11
!~t-f II II II
iIIiir......ïi-i j
I~ .. ~.J. ~ .1 ~ .L-u II II II
!/i~t-f i I I I I I
1~""rJ..i..u'''T-i.1
I I II I I I I i COl FAC!/ i1111111111
iir......ïi-Il
111111111111
ii/ii-_U-
I I I I I I i.i- I
..-H- i-¡- ~I SER..1
COONG
111111111111
I i I I I I I I i.i-l
i/i-_U- i-¡- ~I
..-H-
"1111"" 'PI 11111111111
I~-i"JJ-ri-L._1 i
II~...,
i-, JJ
I i-n-~-fl
1-i"' II II I I II
-ri-LJ_I I
I i-, - ..-~J
1-1"' I I I I I I I I

:: :: I~_~_~,. 1111"1,11
I I I ~.,-ri-
II II
II 1-..1"1
I )~-i7J1-i-i i
v) 'Y I I : : ii )"1"" 7J1-r--i i
I 1-..1"(\3 'Y ¡ I ::
s: \~-r(~-:I ,.,
(~:_____::c ~ \~""("":I "
(~:_____::c
HOT FACZ
BODY COOUG ~ FlS ANCHOR
'FE CAST-lN WALL BRICK
TNRD FOURTH
ICOLD FAC! DETAI OMll~ (COLD FACE DET ALS OWTTED
FOR CLARrrJ FOR CLARITJ

FIGURE 24
GENERATIONS OF
STAVE DEVELOPMENT

5-56
 STEAM
VENT STEAII
VENT
S!PARA TOR
MAKE" DRUII
180llR
QUALITY)
WATER
I!MERGENCY
UPPER ~ COOL-lO
COLECTION WATER
MANIOLDS SUPPL Y

FORcm
RI!CIRCUlA TION
\J ~~PS
\Ji INnRCON_CTIfO
.. STICH
PIPES \øoWNCO."S
I
LOWER ~
SUPPL Y
MANIFOLDS

NATURAL EV APORAnVE FORCED EV APORA TIVE

STAVE COOLING FORCID COLD WAnR COOL-lO
INEVC) STAVE COOLING
IFEVC) .FCWC)

FIGURE 25
STAVE COOLING SYSTEMS
 3

\- \-
5

SINGLE TAP-HOLE FCRNACE-CASTHOUSE LAYOUT USING SLAG POTS

1) Troug.ii 4) Slag Runner
2; Iran .~:..r... 5) Slag POl Track
3) He: ,l,~:a:: Tracks

::::~:::~::::~::::;::::::::':.,:.:
!~r~~*~~~j~IJ~l¡
:~;~:tWt¡tftf~¡l~ :.:-:-:-:.:...:.;.:-:.:.,:-.:-:.;.:.:
ttrf:~~t~t~~~;ff~ jf~~~~~j~~~iIDt
....._..-'....,...............~....
..u...._...........
~:~:~:~~~~:~~:jf:~~::f:~:~:~ .......................................
.;.:.:.:.;.:.:.;.:.:.:.:.:.;.:.:.:.;.;.
....-.....-............................
::::::::::::;:::;::::::::::::::::::::
................--.
:::~:::::~::::;::::?,:~::;~::::::
~~?tf:~~~ff:tt?; :.:.:.:.:.;.;.:.:-:.:.:.:.:.:-;:.:.:.:
.;.:.;.:-:.;..:.:.:.:.;.:.:.:.:.:.;.;.
....................
.:.:.:.:.:.:-:.:.:.:.:.:-:-;.:.:.:.:.:-
1ti~~~¡j~j¡j~~~¡~~~~1~~~j¡¡~!I~ ..............-....
:.:.:.:.:.:.:.:.;.:.:.:"ù:.:.:.:-:.:
.................,.....................
::;::::::::::;:::::::;:::;:;;::::::::::
:::::::;:~::;::::::::::::;:::::::::;::
.................... ;..:.;..;.:.:.;.:.:.:.:.:.;.:-:.:-:.:
...................
....................
:~:~:~:~:~~:::~t:~:f~~:~;- ~: ~:~:~:
...................
~t;~;~:r=~r~:!:~~:~t:;;;:~
....................
rt~~~~~;~r?~~;¡~?~~¡~::;:::::::;:::;;::::';1-:::=::::::::;;:
::::::::::;:;;:;:::::;;:;:;::::;::;;:
r!iWï~ttiltjt tit:~:t;:::~;1~:m~:

MULTIPLE TAP-HOLE FCR.VACE CASHOUSE LAYOUT USING SLAG PITS.

1) Trough 4) Slag RUlner
2) Iron R:u.. 5) SLag Piis
3) Tiiing R;vU
FIGURE 26
CASTHOUSE LAYOUTS

5-58
-I

FIGURE 27
MUD GUN

FIGURE 28
TAP HOLE DRILL

5-59
 FIGURE 29
SINGLE SIDE T APHOLE EQUIPMENT
INSTALLATION

J 5 4
::. ;::::::''::::''

~t~~m~~t~::::..
t;~:~r~:~:~;~~:t\:::::;::;::;~~:::;:::;m:;~:::;~:;:::: ~~::;;:::'~:::::::;:.::~.: ..:.". .' ~ . .

1) Taphol~ 4) Skimmr
2) Iron Poal 5) Slag Runnr Takoff
3) Siag Lan 6) Iron RUJr Takeoff

FIGURE 30
TYPICAL IRON TROUGH

5-60
 ~Hot Metal Cars

l
Slag Granulatior

FIGURE 31
PAUL WURTH -INBA SLAG GRANULATION
FACILITY

5-61
 CASTHOUSE FLOOR

nL nNG RUNNR
ru ~
SUPPORT CRADLE

~OUGH TORPIDO
DRAIN
RUNNER
I LADLZ
GRADE

FIGURE 32
TIL TING
IRON RUNNER
INSTALLATION

5-62
 SHE" :!"J _"E

SHE" ei..;: _"E

3 l -g. 10. SHELL

MATLE

::.
~
..:;~

14
23

FIGURE 33
EXAMPLE FURNACE LINES

5-63
 Tons per 100 C.D ft Working Volume

11.0

10.5 .10
10.0

9.5 f- ~3,OOO
· 9 · T/d
9.0

8.5 .8
8.0
. 5 . 6 · Exam pie
7.5 .4
.3
7.0 .2
6.5
1
6.0

Data Labels are Furnace I.D.

FIGURE 34

5-64
 Tons per Day per 100 cu ft Hearth Volume

98.0 + .13

88.0 -
. Exam pie

78.0
.5 .4
68.0 -
. 12. 7

58.0 - .1
.6
48.0

38.0 · 2

Data Labels are Furnace I.D.

FIGURE 35

5-65
 Tons per Day per 100 cu ft Hearth Volume

98.0 -- .13.3,000T/d

88.0 -
. Exam pie

78.0
.4
.5
68.0 -
.12.7
58.0 - .1
.6
48.0

38.0 ~ 2

Data Labels are Furnace I.D.

FIGURE 36

5-66
 \J - f -

21.. 6N DIA.

-_.. ..-
~ .._.., '_e_

i
--
I

FIGURE 37
EXAMPLE FURNACE HEARTH

5-67
 ~'\ '
~"'\ii ,.; .X~""~_ !! ~
Il!~ ',;
"Eo~-.

,\,
__I,'!--"
i 1 /,.-

~~.;_.._--
r :'./.. ÃI ¡p ~ .:a õl

ó.. t-=.~': - ~.- -~"'-i"

. '..: : -_~-~
- --
~~rn -¿.."~~..' - - ,.--.. i

~
~£ .Ii"
I
S :1'lt 3Lci i
I

ii-;: ~ "" ' .: .. .i

__ A. :: _
~ i.l ...~JI l l
.'"
2 -~~ '-i --
: r .I '. 1/1

~-: .
.
l
r
,
r
~
ii

9l ':,""I l ¡.

_...- _____~!O ~

,~i: ,
UC :i 'G :
~TD ..
CH lU ~ ,)C:
L~"'~
I d

Cc: y
SI
t" ~ ii :/~

~
~ :....~ JI
==
A. :; ; ::
'! ... .. -
.c,::..~
~ ,.. .. -

:: :c-!"

s- ~NC~~: wiL~

FIGURE 38
EXAMPLE FURNACE
PROPOSED HEARTH

5-68
 Tons per Day perIOO cu ft Hearth Volume

98.0
.13.3,000111

88.0
. Proposed
. Exam pre

78.0
.4
.5
68.0
. 1,.7
-

58.0 .1
.6
48.0

38.0 ; 2
Data Labels are Furnace I.D.

FIGURE 39

5-69
 FIGURE 40
EXAMPLE FURNACE WEAR LINES

5-70
 \
\

\
\
BOSH STAVES \ MATLE

R'\ 12'xS' POINT

~~
'~'l
i \ I

~ ~'
~.~~,,'
,
I
4'x4' POINT

T-lYERES

IRON NQICH_

HEATH STAVES

CONE 14

CONE 23

FIGURE 41
CRITICAL WEAR POINT
APPLIED TO EXAMPLE FURNACE

5-71
,/
i
 FIGURE 42
COPPER STAVE

5-72 '
 ,r

JI'.. a. --
il....--

OI
'r

r ..__
-~
~

.
~

BASE CAS:: ALT. "A"

EXISTING FL.~\.AC=- Stack Integrated Lining\Cooling
Bosh Stove Cooled
-WORKING VCL~ME -WORKING VOLUME
J1.948 CU.FT. .:5.64.: CU.FT.
-USEABLE HEARTH VOL:.i.E -U5EABLE HEARTH VOLUME
2,965 CU.F;. .:.199 CU. FT.

FIGURE 43

5-73
 .. ..
.; ..

, '..
~
I¡
..
.; .. IT-,°i-
I ~.

-,. .. . il-r Go St
~ .. "-J~OI

ll-"G.~
~

..
..

. ¡¡. ~ ~ I
L,
~
1"\

;. :
~~ ~
II_It ALT. "c"
ALT. :=
Generation 3 Stoves
Generatior -.'I S:cves with Movable Armor
-WORKING VCL;Ji.E (Lower Stockline 3' -0")
34.575 C~.r. -WORKING VOLUME
3.3,.318 cU.FT.
-USEABLE HEAR7~ VCL~ME -USEABLE HEARTH VOLUME
3,199 C:".r7,
FIGURE 44 .3,199 CU.FT.

5-74
 ~
r-, \
---. //;/~'

~-
--
(///';'
: 1I
l!.ç Q. tI ri.. G. HU

..
..
.....~
is.r ra

I~
:'1
..
.1-
. s ~
~I .
:'1
..
'ur
~
..
~ "l
jl:
:'1
..
~ : !.
:1
..i
.. '0'
JI... a. 'C
j
_..i
"-l 'I'":M

.. ,
..
I
~

I'" ;
i -ii
.
'~

I~
.
i'" ~

ALT. "On ALT. nEtt

Generation 3 Steves Generation 2 & Coooer Staves
with Paul Wurth ic:: with Paul Wurth Too
-WORKING VOLUME -WORKING VOLUi.E
.:5,424 cU.FT. .:6,723 cU.FT. (1 157.)
-USEABLE HEARTH VOLUi.E -USEABLE HEARTH VOLUME
.:,199 cU.FT. .:,199 CU. FT. (10a7.)
FIGURE 45

5-75
 Tons per Day per 100 cu ft Working Volume

11.0

.10

10.0 -

.9. Ex ~3,000 T/d

9.0 . Alt C

. 8. Alt B
. A~Att D

. AU E
8.0
.5.6 . Example
.4
.3
7.0 .2

.1
6.0
Data Labels are Furnace 1.0.
FIGURE 46

5-76
 BLAST FURNACE PRODUCTION

7500

6500
~
=
e
:E
==
~ 5500 l--
i

Z i
C1
~
I-
i

o.
= n
;.
~
~
~
4500

3500 :
70 80 90

YEAR

FIGURE 47
,i

5-77
 BLAST FUR~ACE FUEL RATE

1190

1090
~
==
~ ~C
~i:
-i
.. -D
990

890
70 80 90

YEAR

FIGURE 48

5-78
 LECTURE #6

BLAST FURACE DESIGN II

Neil J. Goodman

K vaerner Metals
Pittsburgh, Pennsylvania

Abstract: - Blast Furnace Design II covers air (blast) and gas system designs for modem blast
furnace operations. Increases in hot blast pressure and temperature during the past thirt years,
together with the need to improve operating and maintenance effciencies, and corresponding cost
reductions, have resulted in design improvements in the air and gas system designs. The subject
wil be covered in the following areas:

Functional Layout and Design of Hot and Cold Blast Systems

Hot Blast Stove and Ancilary Design
Optimization of Stove Operation and Control
Functional Layout and Design of Gas Cleaning Systems
Optimization of Gas Cleaning System Water Usage
Top Pressure Control and Energy Recovery Turbines

INTRODUCTION

Among the major inventions and progress achieved in blast furnace technology in the 19th century
were the production and use of coke, and heating of the blast air. In 1828 James Beaumont Neilson
introduced blast air heating in recuperative form at Clyde Iron Works, Scotland. In the course of
only a few years the equipment used for heating blast air developed from makeshift installations to
well thought-out heating apparatus. It was possible to achieve the blast temperature up to 930°F
using recuperative iron cylindrical hot blast stoves.

This was the state of blast heating when Eduard Alfred Cowper made his patent application for a
brick-type hot blast stove in 1857. From this point in time onwards there has been a steady further
development of the "Cowper" design, lasting through to the present period.

The development of the "Cowper" hot blast stove has been a function of advances in combustion
technology, refractories qualities, etc. and not least the evolution of the blast furnace process. In the
last 30 years, the progress of blast furnace technology and support ancilary plant has been
paricularly rapid.

6-1
 Hot Blast Stove Operation

Overview

The operation of hot blast stoves involves the sequential heating and cooling of a regenerative
mass of refractory. The regenerative heating is performed by the combustion of gasses and the
passage of the waste gas products of combustion through the refractory (gassing). Figure 1
shows the gassing operation.

The cooling of the regenerative refractory is performed when the blast air is introduced into the
stove (i.e. the stove is on blast, see Figure 2).

The intermediate position with the gas system and blast system isolated, the stove is "bottled"
or "boxed" (see Figure 3).

Gassing Control

Typically, a stove can provide hot blast at 100°F less than the flame temperature, and the
typical flame temperature available from 100% blast furnace gas is 2000°F. Therefore any
blast temperatue greater than 1900°F wil usually require an enriching gas to be mixed with
the blast furnace gas (natural gas or coke oven gas are usually used).

During the gassing cycle, the refractory in the stove is heated and the temperature at the top
dome of the stoves rises. Eventually the dome temperature wil achieve a target temperature
(typically 50°F below the flame temperature) and wil be kept at this value by the addition of
excess combustion air. With the dome temperature controlled to a constant level, excess heat
from the combustion gasses wil heat up the refractories lower down the stove and increase the
temperature of the waste gas. Eventually, the waste gas temperature could exceed a pre-set
maximum (typically 650°F) and the gassing wil be automatically stopped to protect the waste
gas system from thermal damage to the mechanical valves and duct work..

Ideally, the gassing should be stopped just before the stove is switched over to "blast".
Typically, stove switchovers are performed either at pre-set times, upon operator initiation or
high waste gas trips. In modern stove installations, computer models are used to match the
heat input from the gas with the heat output from the blast. These models therefore reduce the
energy losses associated with bottling and excess waste gas temperatures.

Blast Control

The cold blast (typically at 300°F) is heated by the refractory and exits the hot blast stove at
almost the same temperature of the dome refractory. As the refractory cools down, the
temperature of the hot blast also cools down until it approaches the temperature required by the
blast furnace. At this point the next hot stove wil be taken "off-gas" and then put "on-blast".
When the second stove is "on-blast", the cold stove wil be put "on-gas" to be reheated.

6-2
 Figure i - Stove On-Gas

,
i
. I

CHECKERS

HOT
BLAST

COLD
GAS BLAST

AIR
WASTE
GAS

6-3
 Figure 2 Stove On-Blast

CH ECKERS

COLD
GAS BLAST

AIR WASTE
GAS

6-4
 Figure 3 Stove Bottled or Boxed

CH ECKERS

HOT
BLAST

COLD
GAS BLAST
I
, !

AIR
WASTE
GAS

6-5
 ,. "" ..

ON BLAST ON GAS ON GAS

Figure 4 Prior to Changeover

,. '" ,.

ON BLAST BOTTIED ON GAS

Figure 5 Stove 2 "Gas" to "Bottled"

6-6
 ,. AIR
,.

ON BLAST ON BLAST ON GAS

Figure 6 Stove 2 "On Blast"

AIR AIR ..

BOTTLED ON BLAST ON GAS

Figure 7 Stove i "Bottled" from "Blast" prior to "On Gas"

6-7
 COLD BLAST SYSTEM

For a modern blast furnace practice the air requirement at the tuyeres broadly ranges from
30-40,000 scf/NTHM. This value must be supplemented by a margin suffcient to allow for
losses in the blast system, particularly stoves pressurization. The actual cold blast (blower)
demand depends on the levels of oxygen and tuyere injectant being practiced, together with the
amount of scrap or DR! in the burden.

The blower specification for a blast furnace producing an average of 5000 NTPD, to meet the
range of operating practices currently adopted in North America would incorporate an
operating window which accommodates blowing rates (including losses) of 100,000-
160,000 scfm, at the design furnace top pressure.

Blower pressure requirements are generally set by the following guidelines:

Loss thru the Cold & Hot Blast System 2 psig max.
Loss thru the Furnace 28 psig max.
Furnace Top Pressure 6-35 psig

Turbo Blowers

Figure No.8 is a schematic diagram of the cold and hot blast system from the blower station to
the bustle main.
STOVE PRESSURIZING LINE
VENTURI UHEA REGULATING VALVE

COLD BLAST WAIN

DRAfT
CONTROL

SHUT -OFF
VAlVE
ÐACKORAfr
STACK

BUSTLE
PIPE

Figure 8 - Typical Cold and Hot Blast System

The cold blast system starts at the air inlet to the blowers and ends at the entrance to the hot
blast stoves and blast mixing chambers. A two blower configuration, as it applies to a single
blast furnace plant is shown. Inlet fiters are shown for protection of the blower intemals,
venturi meters for machine control and relief (anti-surge) valves in the discharge lines for back
pressure control. Check valves and isolation valves are not shown.

The centrifugal blower has been the predominant means of delivering wind to the blast furnace
in North America. Figure 9 is a cross section thru a centrifugal blower. The machine shown
has five stages, each stage compressing air to a higher pressure.

6-8
 The air enters the eye of each impeller and leaves at the periphery; travels thru all five stages of
compression in a series manner. The most blowers of this type stil in service have a design
pressure rating of 40 psig (actual 37-38 psig). A number of machines are being upgraded to
45-46 psig conditions.

ioii PRESSURE HI~H PRESSURE

AIR IN AIR OUT

Figure 9 - Centrifugal Blower Cross Section

Traditionally, a centrifugal blower is coupled to a steam turbine, hence the term turbo blower.
The volume of air delivered is controlled by the rotor rpm which in turn is regulated by the
steam flow to the turbine.

Since the middle 60's, axial blowers have been more widely used due for their greater
effciency and lower horse power requirements. They are more suited for compressing large
volumes of air at the pressures (60 psig) required for modem high production blasts. Axial
blowers are also smaller and lighter than centrifugal machines for the same capacity and
pressure ratings. As shown in Figure 10, the air in an axial machine travels longitudinally,
parallel with the rotor shaft.

HIGH PRESSURE LOLL PRESSURE

AIR OUT AIR IN

Figure 10 - Axial Blower Cross Section

Axial blowers are coupled to either steam turbines or electric motors. Wind volume is
controlled by rotor speed, or stator blade setting.

6-9
 COLD BLAST MAIN

Cold blast temperatures and pressures vary greatly, 300°F and 38 psig for a "typical" low top
pressure operation, to 550°F and 65 psig for a high production furnace. Facilities operating at
higher temperatures generally insulate the cold blast main to conserve energy in the system.
The cost of cold blast main insulation is harder to justify at lower temperatures particularly
when the blowers are remote from the blast furnace.

In Figure 8, the first component in the cold

blast system after the blowers is the "snort"
valve, which is used to regulate blast
pressure and flow rate to the furnace during
checking or "on -blast/off-blast" procedures.
The snort valve is actually a combination of
two valves (Figure i i).

The discharge valve which opens to the

atmosphere is linked to a butterfly valve in
the cold blast main. As the discharge valve
opens to atmosphere, the valve in the main
closes, thus diverting cold blast from the
furnace.

The snort valve is fitted with a silencer to

control the noise level of the discharge.

From a process standpoint, the snort valve

should be located upstream of the blast
metering device and any subsequent
conditioning of the blast.

Figure i i - Snort Valve

Located downstream of the snort valve are two lines in paralleL. The larger of the two is a
continuation of the cold blast main, while the smaller is used to pressurize the stoves. Located
in each is a venturi meter. Ideally each time a stove is pressurized, the total flow of cold blast
is increased to meet stove pressurization requirements without upsetting blast air flow to the
furnace. However, this is difficult to achieve even with a modern installation. Stove
pressurization (or fillng) requires 2 to 4 minutes and up to 8,000 scfm depending upon stove
design.

There are also relief or blow-off valves included in the system. These valves are intended to
depressurize the stove from blast pressure to atmospheric, prior to putting the stove on gas.
Depressurizing time can be up to 4 minutes depending on the level of blast pressure. The relief
valve exhausts to atmosphere and is fitted with a silencer.

Another branch line is taken from the cold blast main downstream of the stove pressurizing
line to supply mixer air. Mixer air is added to the hot blast air that comes from the stove to
control the hot blast temperature.

6-10
 There are three concepts for mixing of hot and cold blast:

· Individual mixing via the lower combustion chamber.

· Individual mixing between the stove and hot blast valve.
· Central mixing in the hot blast main.

Individual mixing where the cold blast entry port is sited in the lower combustion chamber is
gradually being replaced. The thermal cycling in that area of the combustion chamber results
in high refractory maintenance.

Where "mushroom" type hot blast valves are used (Figure 12), the mixer connection is at the
base of the valve. This system eliminates the temperature variations in the stove and ensures
little temperature variation in the hot blast main. Unfortnately the "mushroom" type valves
are not generally applicable to hot blast systems supplying temperatures greater than 2000°F.

~ VALVE STE/1
/' (WATER COOLEOI

VALVE DISC HOT BLAST

(WATER COOLED) CONNECTON

VALVE SEAT
(WATER COOLED)

STOVE

REFRACTORY
LINING

o /1IXER
CONNECTION

Figure 12. Stove Mixer Connection via a Mushroom Valve

When "gate" type hot blast valves are used, the mixer connection is usually located between
the stove and the hot blast valve in the trunk connection (Figure 13A).

The alternate to individual mixing is to install a central mixing station located in the hot blast
main just before the bustle pipe. There are a number of design variations for this mixer.
However, good mixing can be achieved with a ring mixer as shown in Figure 13B.

6-11
 This arrangement subjects the major portion of the hot blast main to temperatures up to the
level of the dome temperature which must be accounted for in the refractory design.

1 Stove

-1-,
"

,
r:::""..+ ~
I

I
,

I Ò HOT BLAST HAIM

1---- -
,

I
,

I
.",J¡
HIXER

Figure 13A - Mixer- Connection Figure 13B - Mixer Connection to Hot

to Hot Blast Outlet Blast Main via Ring Mixer

Cold Blast Conditioning

The cold blast system includes the facilties to condition the blast air with moisture and
oxygen. Current blast furnace practices are based on high injection rates of pulverized coal, or
natural gas at the tuyeres. The loss in raceway adiabatic flame temperature is corrected by
oxygen enrichment of the blast. Twelve percent (12%) oxygen enrichment of the blast has
been safely practiced for a number of years. The safe handling of oxygen calls for the use of
stainless steel fittings and seal tight isolation valves when the furnace is "off-blast".

Blast moisture additions are only used as a secondary control of flame temperature. Steam is
added to cold blast on the basis of dew-cell measurements of the ambient moisture in the air.

HOT BLAST SYSTEM

The hot blast stoves are a regenerative heat exchange system used to preheat blast to the blast
furnace. The hot blast stoves utilize the top gas from the blast furnace as their source of
energy. The blast furnace gas used to fire the stoves is often enriched with natural gas or coke
oven gas to attain the flame temperature required to meet the specified blast temperature. The
flame temperature is normally 125°F higher than the dome temperature.

The hot blast system shown in Figure I starts at the entrance of the hot blast stoves and ends at
the blast furnace tuyeres. The main components of the system include the hot blast stoves, hot
blast main, bustle pipe, tuyere stocks, tuyeres, back-draft stack and auxiliar fuel injection
system. Most hot blast systems include three hot blast stoves with some plants having the
availability of a fourth stove.

6-12
 Internal and External Combustion Chamber Hot Blast Stoves

Two types of stoves are in use in North America; i.e. the internal and external combustion
chamber designs. As hot blast temperatures increased above 1700°F so did the incidence of
major problems with the internal combustion chamber stove design of the 1950's.

The major area of failure was the dividing wall between the combustion chamber and the
checker chamber, which is the most critical part of the refractory construction. Due to the
uniform temperature in the combustion chamber and the decreasing temperature in the checker
work, the dividing wall, particularly at the lower level of the stove, is subject to thermal
stressing and differences of movement in individual layers.

As the flame temperature was raised, the thermal stresses and the differential expansion in the
dividing wall increased, resulting in bending and destruction of the wall, short circuiting of the
combustion gases, and damage to the checker work.

Other problems included:

· Dome refractory failures.

· Failures at nozzle connections.
· Checker system failures (subsidence, flue misalignment & bottom checker crushing).
· Checker support failures caused by above problems.

In the early 1960's, the solution to the problems experienced with older internal combustion
chamber stove design lay in the development of the external combustion chamber stove. When
the combustion chamber and the checker chamber are completely separated, the foregoing
damages can be avoided. The popular approach in the late 60's and early 70's for designing for
a blast temperature of 2500°F (1350°C) and dome temperature of 2825°F (1550°C) was to
adopt the external combustion chamber stove.

There are currently three designs of external combustion chamber stove:

The Davy Krpp Koppers (DKK) design shown in Figure 14 is basically two separate
chambers each with its own dome. The two domes are connected with a pipe incorporating
two expansion joints. Differential movement is taken up in the expansion joints. The two
domes are tied together with I-beam links to contain the internal pressure force.

The M&P design (Figure 15) is similar to the DKK design except that the dome connecting
pipe does not contain an expansion joint. Differential expansion is catered for by pre-stressing
the vessel before installation of' refractory. During initial warm up the stresses are relieved by
the differential expansion between the two chambers. As the shell temperatures increase above
the average, the vessel stresses increase, but stay within permissible levels at maximum
temperature and pressure.

The Didier design (Figure 16), incorporates a heavy dome steelwork arrangement which is
carried by the checker chamber. The combustion chamber including the refractories and
burner are parially suspended in a cantilever arrangement from the dome steelwork. The base
of the combustion chamber is mounted on a "hydraulic foot", which also partially supports the
combustion chamber and absorbs differential expansion.

6-13
,r

¡
 "l'kOfC;fl~

"MOlu,,, SwC4i

Figure 14. Figure 15. Figure 16.

Davy Krupp Koppers Martin & Pagenstecher Didier
Hot Blast Stove with Hot Blast Stove with Hot Blast Stove with
External Combustion Stove External Combustion Stove External Combustion Stove
The external combustion chamber stove has not generated great momentum in North America.
Inland Steel BF 7 is the only installation of this design of stove.

The internal combustion stove design was not abandoned. In the late 1960's the super high
duty (2825°F dome temperature) design became an alternative to the external combustion
chamber stove.

The survival of the internal combustion stove design was based on improvements to the
partition wall design, which included:

· The introduction of an insulating layer in the partition wall with dense refractory on either
side. This concept minimized the temperature gradient across the dividing walL.

· The use of "sliding joints" which allowed individual layers of refractory to expand
vertically, independently of adjacent layers, thus avoiding wall bending.

· The adoption of gas sealing concepts by stainless steel sheets or concrete panels in the
dividing walL.

International competition between external and internal combustion chamber stoves has
become a commercial rather than a technical issue.

6-14
 The North American hot blast requirements lie in the range 2000°F to 2200°F with
corresponding dome temperature of2250°F and 2450°F.

Figures 17 and 18 ilustrate the styles of internal combustion stoves seen in North America.

Jr Conispherlcal Dome Design

Silica Brick Hemispherical Dome Design

Combustion Chamber

Combuslion Chamber

Iniertocklno Hlugonil Ch_i"

Hol BiasI +-

Ceramic Burner ~
Air +
.- Allo Iron Gri Support
Gas+ -- Wasie Gas
.-old Biai
Alloy Iron Support Columns

Figure 17 Hot Blast Stove with Figure 18 - Hot Blast Stove with
Internal Combustion Chamber Internal Combustion Chamber
New Dome Rebuilt Inside Existing Shell

Figure 17 represents the design adopted when a new dome shell, or a new dome and vessel
shell, is included. In both cases, the dome refractory is independently supported from the stove
shell.

Figure 18 represents a stove re-built within an existing shell. In this case, dome refractory is
supported by the refractory ring wall.

The important features to consider in the construction of a hot blast stove are:

· The stove structure must be designed to withstand stresses due to thermal expansion and
contraction.

· There must be sufficient mass of bricks to deliver required stove duty and the brick
materials must be of correct quality.

· The grid at the bottom of the stoves must be able to withstand the weight of the checker
work, and misuse due to overheating.

· Stove materials must be able to withstand chemical attack from the gases used for heating.

· The burner gives good effcient burning characteristics in order to save energy. Also, the
flame must not impinge on the dome or the checker work in order to avoid damage to
refractory brick or lining.

6-15
 The following aspects have to be taken into consideration when choosing materials for
refractory bricks for a hot blast stove:

· The maximum temperature that the materials can withstand.

· The mechanical strength to withstand required loads.
· The need for resistance to chemical attack.
· Creep characteristics of bricks.
· The cost of bricks.

A typical stove construction wil have 4 or 5 differing grades of refractory (Figures 17 and 18).
The use of thin wall checkers gives a high heating surface/mass relationship, which provides
for high effciency of heat transfer, and in many situations, permits an upgraded stove to be
built within an existing shelL. However, physical characteristics have to be considered due to
temperature and load conditions within a stove. For example, creep resistance is required -
creep being deformation of a material with load and temperature over a period of time.

Silica checkers are chosen for the higher temperature areas due to their excellent creep
resistance, and near zero expansion at temperatures above 1380°F. Care must be taken when
heating and cooling silica below this temperature to absorb volume changes associated with
crystalline phase transformations in the temperature range 480°F to 1 070°F.

Checkers used in both stove designs are a hexagonal shape with circular or hexagonal flues.
Each checker is interlocked with the course below by a series of male and female connections
and are laid up in an overlapping pattern as shown in Figures 17 and 18. The interlock design
provides structural integrity to the checker mass, making it unnecessary to rely on the ringwall
and combustion chamber to maintain checker positioning. By providing clearance allowances
between the individual checkers, expansion due to heat-up takes place between them.

The checker work is supported by a

number of iron columns, girders and
grids. Figure 19 shows a typical
checker support system used in both
internal/external combustion chamber
stove designs. All components are of
a low alloy cast iron suitable for
temperatures of 850°F. The system
shown is designed to maintain a flat
support beneath the checkers thereby
preventing checker deterioration due
to an uneven supporting platform.
The column, girder and grid support
are interlocked with each other to
prevent movement of girders, and grid
supports in relation to one another
during operation.
Figure 19 Modem Hot Blast Stove
Checker Support System

6-16
 The system can also include a series of tie rods between columns to improve rigidity. When the
internal and external combustion chamber stove designs were put into service for dome
temperatures in excess of 2460°F (1500°C) and operated for a number of years, a new
unexpected problem with the shells of the stoves occurred; i.e. intercrystalline stress corrosion
cracking or, for short, stress corrosion cracking.

Stress corrosion cracking is the failure of a metal resulting from chemical attach in the
presence of tensile stresses. The essential ingredients necessary to promote stress corrosion
cracking are:

Tensile stress.
Acidic environment.
Susceptible metal.

The following methods have been employed to control stress corrosion cracking:
· Application of a protective coating of the inner surfaces of the steel shell to prevent the
acidic environment from coming in contact with the susceptible metal under stress.

· Dew point control by applying insulation to the external surfaces of the shell thereby
eliminating condensed water vapor from combining with NOx to form the acidic
environment.

· Limiting stove dome temperatures to below 2460°F thereby, eliminating NOx formation
resulting in a lack of the acidic environment.

Hot Blast Stove Ancilaries

While all hot blast stove ancilaries are important to the effective performance of the stove, the
two most important are the stove burner and the hot blast valve.

A typical stove combustion air and gas flow diagram is shown in Figure 20. Many hot blast
i
stoves in North America are equipped with mechanical burners, external to the stove itself.
I
ì Figure 2lA ilustrates a design which consists of two concentric tubes separating air and gas
which mixes in the stove combustion chamber. This type of burner pedormed adequately on
low effciency stove operations, i.e. dome temperatures up to 2 100°F. Improvements in the
design of mechanical burners Figure 21B have extended the dome temperature range which
can be attained by this type of burner to 2350°F.

COMBUSTION CHAMBER

BLAST ISOLATION VALVE

VENTURI METER

Figure 20. Combustion Gas and Air Flow Diagram

6-17
 Water Cooled Gas Burner Valve

Burner No:i:ile

Conventional low-Energy
With Shutoff Valve High Energy

Figure 21a Typical Mechanical Stove Figure 21 b Typical Mechanical Stove

Burner Designs Burner Designs
To overcome the problem associated with mechanical burers, vertical firing ceramic burners
have been developed and are in use successfully throughout the world. Ceramic burners have
the following advantages:

The limit in dome temperature of2350°F is eliminated (increased to 2825°F).

Combustion chamber target wall failures due to thermal shock are eliminated.
The need for a high temperature burner isolation valve is eliminated.

The basic design features which should be incorporated in a ceramic burner system are:

· Air and gas chambers function as plenums to provide uniform gas and air entry at the point
of mixing. The gas chamber should also act as a low velocity separator to drop out any
substantial portion of entrained moisture, which should be drained on a periodic basis.

· Gas and air- should enter their respective chambers at the lowest elevation of the burner.
This will reduce temperatures in the gas and air inlet ports to the lowest possible leveL.

· At the point of burner exit, the air and gas should be mixed while flowing at velocities in
the turbulent flow range to insure a uniform mixture.

In the burner shown in Figure 22, the uniformly distributed alternating parallel streams of
turbulent fluids provide for effective gas and air mixing as they are blended into each other
when rising through a three level ceramic grid configuration. This ceramic grid is placed
above the slots and functions like many individual nozzles. Each nozzle is served by a
minimum of one pair of parallel slots. Therefore, gas and air are thoroughly and uniformly
mixed prior to entering the stove combustion chamber.

By having a completely combustible mixture prior to entrance to the combustion chamber, the
flame wil be stable and short. This wil prevent the combustion chamber from being subjected
to severe differential temperatures or the effects of incomplete combustion.

6-18
 A pilot burner should be provided to assure ignition
immediately after the gas and air combustion
mixture exits from the burner.

A low pressure drop style of ceramic burner is

shown in Figure 23. The flame produced by this
burner is several times longer than that produced by
the burner represented in Figure 22.

The design of ceramic and mechanical burners must

lead to complete combustion over a range of gas
calorific values. Incomplete combustion gives rise
to pulsations which result in refractory damage. The
principal cause of pulsations is related to the
harmonics of the combustionchamber and external
gas and air main systems. When combustion does
not take place uniformly, low frequency pulsations
are initiated which can be amplified by an
interrelationship between combustion chamber and
gas and air main harmonics.

Figure 22 Internal Combustion Chamber

Ceramic Burner Design

A physical device which creates a pressure drop across the burner system is often designed into
the system to act as a decoupler of combustion chamber and gas and air main system
harmonics.

The hot blast valve is the most critical valve in the

entire hot blast system since it is exposed to the
highest temperature. In North America the
mushroom type valve has been the standard for
many years for hot blast temperature applications up
to 2000°F (1 ioO°C). See Figure 12. However, as hr
hot blast temperatures have been increased beyond
this level, more and more interest has been directed
toward gate type hot blast valves.

Gate type hot blast valves have undergone radical

changes since their inception. Early designs were
made of cast iron and later of cast steel with water
cooled seat insert rings of electrolytic cast copper.
This design had the disadvantage that under certain
operating conditions, particularly in the case of
increased temperatures, leaks caused by distortion of
uncooled valve components develop. This problem
led to failures requiring repair or replacement of hot
blast after only a few weeks of operation.
Figure 23 Internal Combustion Chamber
Lower Pressure Drop
Ceramic Burner Design'

6-19
 To alleviate these problems, a fabricated hot blast valve was developed utilizing a water cooled
body with integral steel seat rings instead of insert rings, thus avoiding the disadvantages of
independent copper seats. Also, the HEV cooling passages were redesigned to increase
cooling water velocity and eliminate "dead spots".

Figure 24 ilustrates a current hot blast valve

for use up to 2800'F. A summary of the hot
blast valve specification is:

· Hot blast valves are water cooled,

refractory lined gate valves. The valves
are suitable for working temperature and
pressures of 1500°C and 4.5 bar g
(70 psig).

· The paddle is faced with refractory on --

'"
both sides and is water cooled. Cooling +-
water flow is arranged in a spiral
arrangement to minimize differential
temperature and consequent distortion
across the paddle.

· The seat, body and bonnet are also water

cooled and refractory lined.

Figure 24 Hot Blast Valve

Figure 25 shows a cutaway section of body and paddle showing water passages and refractory
lining.

Figure 25 Hot Blast Valve BodyfPaddle

(showing Cooling Water Passages and Refractory Lining)

6-20
 Hot Blast Main

Most operating, problems with hot blast mains are the result of inappropriate design for the
expansion of the refractory lining and the steel shell. The design basis of mains includes
provisions to:

· Allow movement due to thermal expansion and pressure forces in steel and brickwork.

· Keep loading on supports to a minimum by designing mains to prevent high pressure

forces being transmitted into structures, and incorporating slide bearings into the supports to
reduce friction loads.

· Keep loading at branch connections and through valves to a minimum.

When designing, a main, the route and location of supports are established. The system is then
analyzed to determine the optimum location of fixed points and expansion joints.

Figure 26 shows a typical hot blast system layout. It indicates the location of fixed points in
the system, location of valves, expansion joints, tie rods and supports.

PRESSURE BALANCE

Figure 26. Hot Blast Main AnchorÆxpansion System for Gate Valves

In this case, the centerline of the BF is effectively a fixed point as the bustle pipe is held
concentric with the furnace. The centerline of the stove/hot blast branch is fixed in plan and
therefore the main is restrained axially at the intersection of main and each branch centerline.

Expansion joints are required between each fixed point to accommodate thermal expansion in
each section. Expansion joints must be restrained to prevent them "blowing out" or
straightening under the integral pressure force.

This can be done by either making the anchors suffciently robust to resist the pressure end
force, or by installing a tie rod system. In case of large diameter mains at relatively high
pressure, the forces become too great to economically restrain them with support brackets and
support structure. Tie rods are therefore used to contain this force.

A pressure balance expansion joint is required in the main to cater for the movement of the tie
rod bracket adjacent to the bustle pipe and also extension of the tie rods due to temperature
fluctuation and tensile loading due to pressure forces in the main.

6-21
 This expansion joint ensures that tie rods always remain in tension and that pressure forces are
not transferred into the fixed or anchor points.

The hot blast branches are equipped with twin expansion joints. Twin joints are necessary in
this location as lateral movement is required in the joint. The stove branch moves upwards due
to expansion of the stove shell when it is heated up. The hot blast main elevation is relatively
stable as it is supported on structural steelwork which is subject only to fluctuations in ambient
temperatures. The branch expansion joints also give flexibility for the valve flanges to be
separated to facilitate changing of hot blast valves.

Figure 27 shows a hot blast restraint! expansion system for a set of stoves with McKee blast
valves. The older hot blast mains designs use fabricated, diaphragm type expansion joints
located between stoves and between the stoves and the bustle pipe. These hot blast mains are
often anchored to the cast house frame and to the stove furthest away from the furnace. Visible
twisting of the mushroom hot blast valves at their seals ilustrates the inadequacy of this
design.
2-BELLOWS EXP. JOINTS
2-HANGERS PER STOVE W/2-CAST HOUSE COLS.
SUPPORTED HANGERS
I.USHROOI. TYPE H.B.
VAL VE UPPER TRUNK Anchor to C.H. Structure
1.0TION CONTROLLED i
I

Figure 27. Hot Blast Main AnchorÆxpansion System for Mushroom Valves

Figure 28 shows the refractory configuration of the hot blast main, together with a tyical hot
blast main expansion joint. The lining should consist of "hard" refractory and insulating brick.
The use of compressible insulating layers has been discontinued. The weight of the working
lining caused this material in the lower section of the main to compress, leaving a gap in the
upper section. This resulted in hot spots on the shell, and distortion of the shell if close to an
expansion joint.

2300' F Insula lion

2600' F Insula lion

60% AI ,03

Figure 28. Hot Blast Main Expansion Bellows

6-22
 Common practice is to control the hot spot by either grouting the area and/or use water sprays
both these methods provide only temporary benefits. Grouting wil eliminate any future
expansion capabilty and can result in further refractory problems. While water sprays induce
very high stress levels in the steel often resulting in cracking of the steel.

The relative position of hot blast valve to the branch expansion joint is determined for each
stove installation when the layout of stove vessel, hot blast valve and hot blast valve changing
hoist is established. The ideal location of the hot blast expansion joint is on the hot blast main
side of the hot blast valve. In this position there is minimal thermal load changes.

Tuyere stocks (Figure 29) make the connection between the bustle main and the blast furnace.
Centering of the bustle main is important for reaching optimum working conditions for the
tuyere stocks. The double Cardan units compensate for all relative movements between the
bustle main and the blast furnace due to thermal expansion. Movements are controlled by
restraining straps that form a gimbel-type joint. The assembly is gas tight from the bustle pipe
nozzle to the tuyere. Changing the blow pipe is accomplished by unbolting the flange joint at
the top of the elbow.
Bulllt
P'øe

furnac.
W..II

Figure 29. Tuyere Stock Arrangement

Tuveres

The ability to maintain long tuyere life results in significant reductions of furnace downtime.
The primary reason for losing a tuyere is bum out of the nose of the tuyere. There has been a
great deal of improvements in both copper casting quality and cooling systems to help reduce
the problem of nose burnout. The addition of hard surfacing to the tuyere nose has further
improved tuyere life.

The use of high quality castings with a "high velocity" tuyere design wil provide optimum
tuyere life. There are many "high velocity" designs available some with separate nose circuit
cooling, they all maintain minimum velocity of 50 fts in the nose circuit. Tuyere burnout
problems can be identified before adding excessive mounts of water to the furnace by
monitoring the tuyere nose with thermocouples.

Attention should be paid to the water system supplying the tuyeres so each tuyere is supplied
with the proper amount of water. Many piping systems do not provide any control of water
quantity. Variations in supply pipe length can result in wide variations at individual tuyeres.
Some operators are utilizing refractory tuyere liners to reduce heat loss in the hot blast,
improving overall system effciency.

6-23
 Backdraft Stack

The primary purpose of backdrafting is to ensure safe working conditions around furnace
openings made as a result of changing tuyeres, coolers, blow pipes, etc.

The backdraft stack is connected to the hot blast main between the first stove and the bustle
main. The connection varies between 30" and 48". Backdrafting is started by opening a gate
valve (identical in design to a hot blast valve ). Various devices are used to control the draft
from the stack. Too much draft can draw coke as far as the bustle main. A butterfy
arrangement at the base of the stack (not exposed to heat) is the principal means of controlling
the draft. After the butterfy valve, the stack extends upwards to a point nearly level with the
bleeders. In some areas steam injection is used rather than an air draft.

Backdrafting places the hearth and bosh region of the furnace under a small negative pressure
ensuring that any carbon monoxide and hydrogen formed after the blast has been taken off the
furnace is drawn out of the furnace.

At the start of the backdrafting temperatures can reach 3,000°F in the bustle pipelbackdraft
connection as any carbon monoxide coming from the furnace wil combust as air is also drawn
into the system. Sufficient excess air should be admitted to dilute the combustion gases and
reduce the flame temperature.

Energy Efficiency

During the past decade increasing effort has been made to improve the effciency of the hot
blast system. Typically the improvements are based on the following actions:

· Accurate metering of blast furnace gas and combustion air.

· Analysis of the blast furnace gas plus any enrichment gas. Subsequent determination of
the correct proportion of the gasses for the required flame temperature.

· Measurement of the excess oxygen in the stove waste gas. Note: Some operators are now
using both CO and O2 to trim the fuel/air ratio.

Gassing management models are available which allow combustion conditions to be controlled
to blast heat requirements and can develop either maximum stove effciency or minimum
enrichment gas usage for a given hot blast temperature. The use of such automation avoids
"bottling" of stoves and minimizes stove changeover time contributing to improvements in
overall stove effciency.

The sensible heat remaining in this stove waste gas can be used via a heat exchanger to preheat
either combustion air or blast furnace gas. The energy saved by this type of technology can be
used to reduce the dependence on enrichment gas or to increase stove dome temperatures
(Figure 30). There are four primary methods in use:

Fixed Plate Type Element Rotation Type (Ljungstrom)

Hood Rotation Type (Rothemule) Heat Medium Recirculation Type

6-24
 3.5

3.0

2.5

2,0

1.5

1.0

100 200 300

PREHEAl TEnPERATURE 'F

Figure 30. Effect of Gas and Air Preheat on % Enriching Gas Required
to Maintain 2350°F Dome Temperature

The first three can only be used to preheat the combustion air as they do not completely
separate the two gases, whereas the heat medium recirculation type can be safely used to
preheat both combustion air and blast furnace gas. A schematic arrangement of the
recirculation system is shown in Figure 31.

('m......_¡¡¡l¡¡¡..~:::n. .,

1........._._...-1 j ~:~..-

Figure 31. Hot Blast Stove Heat Recovery System Using a Heat Transfer Medium

An installation of this type can increase stove effciency by up to 3%. However, to date, the
low energy costs in North America rarely allow a waste heat recovery system to be
economically viable.

6-25
 BLAST FURACE GAS CLEANING SYSTEM

The gas produced in the blast furnace is an important energy source in the effcient operation of
an integrated steel milL. The gas is used to preheat the blast air in the stoves and as the
principal fuel in the boiler house.

Efficient operation of the furnace will produce gas within the following range of analysis:

Chemical Analysis (dry) CO 20-25% by Volume

CO2 20-25% by Volume
H2 4-12% by Volume
N2 Remainder

Calorific Value (net) 80-105 BTU/scf

Temperature 250°F to 350°F
Dust Content 5-10 gr/scf

The size range of the dust is a function of the screening effciency of the stockhouse and the
moisture content is dependent on the water content of the material charged to the furnace.

The primary purpose of the gas cleaning system is to produce a clean gas that can be burned in
the stoves without causing the stove effciency to deteriorate over time. This requires a dust
content in the clean gas of not more than 0.005 gr/scf. Since most furnaces utilze wet
scrubbing as the gas cleaning method, then the gas must also be cooled to as Iowa temperature
as the water supply wil allow, to minimize the level of saturated water of the gas, thus
improving the net CV of the gas.

Dustcatcher

The first element of the gas cleaning system is the dustcatcher. The dustcatcher is a large
chamber which reverses the direction of the gas flow while simultaneously reducing its
velocity. This results in dust particles greater than 50 microns being removed from the gas
stream. As much as 60% of the total dust content wil be removed in an effcient dustcatcher,
which should be emptied daily to assure continued effcient operations. Dust is removed
through a variety of systems which typically wet the dust to reduce the generation of a dust
cloud during the dumping operation.

Gas Cleaning Svstem ISO

130
""
'-
With the introduction of high top pressure, a
~.oo
110
..
variety of gas cleaning techniques, based on C) II
utilzation of the pressure energy of the gas,
~ Ì' '"
have been engineered.
~ 7D
..
oQ) "'"
The level of top pressure required to clean
blast furnace gas from a gas cleaning plant
=i so
In
In
Gl
õ: ..
~ ..
inlet level of 5 grlscfd is shown in Figure 32. :i
0.00 0.00 O.oo 0.00 0.00 0.008 0.01 0,02

Outlet Dust Loading ( gr/SCF)

Figure 32

6-26
The most common type of scrubber in North America is the variable throat type as shown in
Figure 33. The variable throat venturi is normally sited prior to the gas cooler, Figure 34. The
units give excellent performance up to a top pressure of 8-10 psig. Above that level of top
pressure, maintenance costs have risen and gradually this type of gas cleaning system is being
overtaken by the annular gap scrubber.

The current generation of annular gap scrubbers, Figure 35, are based on two stage scrubbing.
This figure ilustrates the gas passing through a conditioning unit where it is contacted with
water from centrally located sprays, causing cooling, saturation and partial cleaning of the gas.

CLEAN
-GAS

Brick Lining
OIRTV_
GAS
M'ñillE! WATER
INlEl
X 6PI1
lOWER
PACKING

Figure 33 Adjustable Gas Venturi Figure 34 Variable Venturi

Scrubber Cross Section Followed By Cooler

COtuIITIO_INO VEIU£l

DlIMlaT1!I1l ,.iCIlIJlO

MYDIlAULIt. AC1'UA TOfl

Figure 35 Annular Gap Scrubber Cross Section

6-27
 The annular gap scrubbing section consists basically of a fixed cone within which a movable
cone operates. Raising and lowering of the movable cone decreases or increases the annular
gap between the two cones through which the gas passes for final cleaning. A hydraulic
system controlled by signals from the furnace top pressure controller adjusts the gap to create a
gas pressure drop required for turbulent gas scrubbing and for furnace top pressure control.
Water for the gas cleaning is applied through radial and tangential sprays positioned
just above
the annular gap. A good "rule of thumb" to estimate the water requirements for gas cleaning is
10 gall 1,000 scf of blast. This does not allow for gas cooling requirements.

Effciency of dust removal is dependent principally on the degree of turbulence created by the
scrubbing section. The scrubber is able to achieve the required level of gas cleanliness over a
wide range of gas flow rates and furnace top pressure by controllng the cross-sectional area of
the annular gap between the inner and outer cones of the scrubbing section.

Moisture carrover is inherent in gas cleaning/cooling systems and a mist eliminator is

normally installed at the outlet of the system.

Gas Conditioning Svstem Water Treatment

Most blast furnace gas conditioning plant water systems are ofthe closed circuit type as shown
in Figure 36. By their nature, closed circuit systems reduce the amount of blowdown and the
quantity of contaminants discharged. However, this must be balanced with increased
contaminant concentration. In some cases, contaminant concentration may be self-limiting,
which is the case with suspended solids. Additionally, closed circuit systems reduce the
amount of blowdown requiring treatment and makeup, the cost of waste treatment facilities and
operation is minimized, and in some cases, the actual need for any waste treatment facilties
may be eliminated. Some plants are limited in the amount of makeup water that is available
either because the water is scarce or the cost of purchasing water from a municipal authority is
prohibitive. Water reuse is helpful in reducing actual water need.

Cool Water
8300 USGPM

Recirculation
1900 USGPM

Recirculation
Cooling Pump
Tower

Overflow

Slowdown
700 USGPM

Figure 36. Closed Circuit Variable Annular Gap Scrubber Gas Cleaning System

Efficient operation of the solids removal equipment is probably the most important part of a
successful closed circuit system. Suspended solids in the recycled water should be reduced to
at least 25 ppm to prevent problems of deposition in low flow areas such as cooling tower_
sumps and pipe manifolds. -
6-28
 The recovery turbine is usually placed after the final scrubbing units as shown in Figure 37 and
it may be arranged to recover a fixed quantity of energy, the balance of pressure being lost
either over the scrubber or a suitably positioned septum valve, or it may be aranged to recover
the maximum amount of energy available in which case there must be certain provision on the
output side of the turbine to cope with variations in output.

SEPTUM VALVE
GAS SYSTEM FOR ALTERNATIVE
PRESSURE CONTROL

\----..---
TO DISTRIBUTIO
SYSTEM

GENERATOR

'V

FURNACE SCBBR RECVERY

DUSTCATCHER TURBINE

ElCTRICAL POER

Figure 37. Energy Recovery Electrical Power Generation

The provision of a bypass to the turbine is always made to maintain the independence of
furnace operation. However, availability of these turbines usually exceeds that of other furnace
equipment.

The most common method of utilizing recovered energy is by generation of electrical power
using a generator directly coupled to the shaft of the recovery turbine.

With the advent of top pressure recovery systems, the type of gas cleaning system to be used
needs to be re-evaluated. As stated earlier, a good scrubbing system wil utilize 80-100 ins.
H20 of pressure. Japanese operators are utilizing electrostatic precipitators and bag fiters as
the primary cleaning device in order to save this energy for the recovery turbines.

Top Gas Recoverv

High top pressure furnaces are equipped with a gas lock chamber through which raw materials
are charged into the furnace. With each charge of raw material, the pressure in the gas lock
chamber is equalized with furnace top pressure using blast furnace gas and then reduced down
to atmospheric pressure. During this process the gas in the gas lock chamber is discharged to
atmosphere. The volume of gas discharge depends on the chamber volume, charging cycle and
magnitude of top pressure. In the Pacific Basin, where the majority of furnaces are operated at
top pressures in excess of 30 psig, the total quantity of gas released is roughly 2% of the total
top gas volume generated.

With fuel costs escalating, means to recover the heat value in top gas released is attracting
more and more attention. Systems have been developed to recover the gas from furnaces
operated at elevated top pressure.

6-29
The gas released from the gas lock chamber is effused naturally by action of the pressure
difference between the gas lock chamber inner pressure and the gas main pressure when the
relief valve is opened, permitting recovery of the gas in the main. When the gas lock chamber
and the gas main pressures are close to being equal, the primary relief valve closes and the
secondary relief valve is opened to discharge residual gas in the gas lock chamber to
atmosphere.

CONCLUSION

In summary, even though dramatic improvements have already been made in North American
Ironmaking facilities, opportunities stil exist for improvement in overall energy effciency.
Unquestionably, furnaces that are to remain in operation wil be those employing advanced
technological features to reduce iron production costs. Methods such as hot blast stove
systems designed to generate higher blast temperatures, charging systems which wil permit
higher pressure and more effective burden distribution, and effective means of handling the
high volumes of high pressure gas generated can all contribute to more effcient operation.

The implementation of these new technologies wil be dictated by escalating costs of energy.
Blast furnace operators and designers wil have to bring about effective ways of saving energy
by not only being aware of the fuel management program within the blast furnace plant, but
also knowledgeable in the disposition of fuel and energy on a plant-wide scale. Since this is an
ever changing scene, only effective, forward planning will reduce energy consumption to a
practical minimum.

6-30
 LECTURE #7

BLAST FURACE DESIGN III

Steve Sostar, General Foreman

Blast Furnace

Lake Erie Steel Company

Nanticoke, Ontario, Canada

Abstract: This paper will cover the basic features of an

ideal blast furnace and ancillary equipment. Consideration
will be given to equipment which will operate consistently
and reliably, thus providing the operator with a minimum
operating cost per ton of hot metal.

Recognising current trends in the Iron and Steel

industry, it is highly unlikely that a furnace as described
would be builtin the near future in North America11.
Consequently, also covered in this paper will be some of the
actual improvements incorporated into the last reline at
Hilton Works "E" Blast Furnace and some recent improvements
at Lake Erie Steel #1 Blast Furnace.

INTRODUCTION

Many hours can be spent debating the merits of various

designs of the "ideal" blast furnace based on:

(a) Operator preference as a result of past

experiences i both good and bad, with existing
equipment and processes.
(b) Types of raw materials used in an individual's
blast furnace plant.

(c) Specific preferences for types of equipment based

on detailed skills of those personnel who will
maintain the equipment.

(d) Geographic location of the plant which may affect

the use of various raw materials, especially
fuels, based on transportation costs.

7-1
 The "ideal" blast furnace as described by this operator
will reflect the most current equipment available to provide
the plant with an efficient, consistent operating furnace.
One of the aims of this furnace design will be to provide an
installation which can be maintained by a series of short
maintenance stops i. e. (less than 24 hrs.) i in combination
with periodic interim repairs (3-5 week duration) to replace
or overhaul specific areas of the furnace every 8 to 10
years.
When designing detailed layouts of blast furnaces, many
engineering hours are spent putting the most economical
equipment into the smallest possible space for the least
cost.2,3 In many cases this saving in initial capital
investment is lost in maintenance costs, furnace production
delays and costly changes required at the next reline or
interim repair. To avoid some of these costs it is wise to
install spare, critical equipment as an "on line" running or
standby spare rather than having this equipment sitting and
deteriorating on a shelf in a warehouse. Equipment does wear
out and fail, so one of the tasks of a good design engineer
is to provide methods to isolate and remove equipment in a
timely, safe fashion for maintenance while having little or
no effect on the balance of the blast furnace plant.

Legislation, either currently in place or being

considered, also has a part to play in our "ideal" furnace
design. Safety of workers, short term and long term health
effects on workers and environmental restrictions for air,
water and noise must all be considered when selecting
equipment.
In an attempt to satisfy all of the above concerns we
will now discuss both our "ideal" blast furnace and recent
repairs, improvements and changes at Stelco and Lake Erie
Steel under the following topics:

Stockhouse
Top Charging Equipment
Furnace Design
Furnace Cooling Systems
Furnace Refrpctory
Stoves and Hot Blast System
Gas Cleaning Plant
Casthouse
Instrumentation and Control Equipment

7-2
 I STOCKHOUSE
Handling of raw materials for the blast furnace process
has changed dramatically over the past 20 years. Traditional
scale cars and skip cars are being replaced with vibro-
feeders, belts and weigh hoppers.

Our ideal stockhouse will be configured in two

identical halves on either side of the main furnace feed
conveyor. (Figure 1) Each half of the stockhouse will
contain:
( a) Two sets of two ferrous material bins.

(b) One set of 5 miscellaneous bins.

(c) One large coke bin.

Each set of ferrous material bins will have two vibro-
feeders per bin discharging onto a collector belt, over a
screen and then into a weigh hopper located over the main
furnace feed belt.

Material from each miscellaneous bin will be discharged

into its own weigh hopper located directly below the bin.
These bins require their own weigh hoppers because when small
quantities per charge are required, they cannot be accurately
weighed in a central holding hopper. Material from each
weigh hopper is then transferred to one holding hopper over
the main feed belt, where total weight for all miscellaneous
materials per charge are checked.

The coke bin will have 5 vibro-feeders feeding onto a

bel t, over a screen and into its own weigh hopper.

All belts in our stockhouse system will be covered and

a dust collection system will be installed to capture dust
generated at transfer points. It is very important to ensure
that well designed covers and side skirts are installed over
the full length of all inclined belts. This will stop the
rollback of individual material particles, particularly
pellets, thus keeping walkways clean and safe to use. In 1997
a new belt fed stockhouse was installed at Rouge Steel 12
which incorporated a circular conveyor gallery for the
furnace feed conveyors. This unique design incorporates a
large area for spillage accumulation, off the walkways and a
series of drop legs for cleanup. This concept will be
incorporated into our furnace.

7-3
 T Y PI C AL AUTO MATE D STaCKHOUSE
iAG LE~ND PROPERTES CF ~w MATE~IALS
illl i.i I -l
I
IJ I ÐOCl..nCl i
C IC_ I
i
hlTPlIAL
~~ ~~
¡~- I~~SI W'lAl.
I ~;'%Ui
!w¡rr~-~':f~
L U"I
JI "'-L
I COl r- J i "'''0 i
LQ-l 25 115 '" ;~z . L...UTllHI
!~ ~__!:o~L
15 ¡ 'O~
~.t. i "'aGEY CIo fI 1"l.£r 'Loa , ~Q I'" '" c: i "'CL."..;r
~
: 100. '0"
J5 51_
i~:40~
I1:\0: f~ ~ "... . ., i . :;Z
H.'. , HfeK£ FIHI
NoH. i HO i. li_"- ~ i-ClTT FIll" /001 l.-l. I
i 1.'0 i is I io ~ $1 i i"fLLET FI..ES

L- ~
KHJS ~
M. MIS( _TEA
1'. ~ tNI ~ SLI :;: 3S i /0.-_7. . COlt '1.."
Y.. ylalt 1M I'foe
lI .. .
IC Y_T1 sawn HI'" SlCA i 1,~, .::
F. a.
.. 11 Q

CD
'T11
C:~Y&'l '--"~--,'...
c:
--- ....'1

!U.A:!~
, ........ce

Figure i

7-4
 Moisture gauges will be installed in the four pellet
weigh hoppers and the two coke weigh hoppers. Belts and
screens will run continuously while the starting and stopping
of the vibro-feeders will be done by two PLC i S (programmable
logic controllers), one operating and one standby. These
PLC i s will sequence material, and compensate for belt run
off, material discharge speeds and moisture content. Timing
of our stockhouse will be such that we will have 140% feed
rate to the furnace to ensure that any minor problems, which
may cause delays in the stockhouse system, do not affect
furnace filling.

Two optional improvements were investigated for "E"

Furnace, a conventional skip-fed furnace. One was to
eliminate the scale car and install vibro-feeders and belt
systems to existing ore holding hoppers. This was ruled out
due to economic considerations. Another option was a
combination of replacing pneumatic cylinders with a hydraulic
system and also making the scale car a remotely operated
unit. This was considered in an attempt to improve the
working environment for the car operator. This was also
ruled out because of the concern that a hydraulic system
would not be fully reliable with the existing bin gate
arrangement. Consequently, there were no maj or changes or
improvements to the existing stockhouse on this reline.

The Lake Erie Steel #1 Furnace was started in 1980 with

half of the ideal stockhouse because of low start-up
proj ected production rates and cost considerations at time of
construction. Since then this existing stockhouse has been
able to sustain production rates of over 6600 NT/day. This
was accomplished by:
(a) Manipulating vibro-feeder rates to optimise
filling of weigh hoppers.

(b) Altering equipment-sequencing logic to move

material quicker.

(c) Setting the stockhouse to be proactive and look at

burden level and burden decent rate rather than
just the operation of top filling equipment.

7-5
 II TOP CHAGING EOUIPMENT
Our choice of furnace filling equipment will be a dual
lockhopper bell-less top.4 (Figure 2) There are a number of
reasons for this choice:

(a) Flexibility in burden distribution is far greater

with this system as opposed to standard two bell systems
wi th or without movable armour.

(b) Individual components are much smaller and lighter

than conventional two bell systems, thus reducing furnace
top loading.

(c) Maintenance of individual components can be

planned and carried out in short stops.

(d) Lockhoppers will be fitted with load cells to act

as check weights on all of the various weigh hoppers in the
stockhouse as well as to control accurate burden
distribution.
(e) Twin lockhoppers will provide increased filling
capacity, as well as continued filling availability in the
case of a failure in one lockhopper.

At the last reline "E" furnace was fitted with a single

lockhopper bell-less top capable of holding 3 skips of
material. A receiving hopper was constructed above the
lockhopper, capable of holding 2 skips of material. This
design was chosen because it had less structural impact on
the furnace top than a two-lockhopper system. (Figure 3)

7-6
 DUAL LOCK H O.PPE R
BE LL-LESS TOP

TRAVELLING
CHUTE

SEAL VALVE

FLOW ,
CONTROL __ SEAL
GATE VALVE
~ HOUSING

SEAL
VALVE

MAIN DRIVE
GEARBOX

REVOLVING
CHUTE

...::''';..: .,~;::: . ."' -.- .

. ~ '.' . .
.. . 'f'

Figure 2

7-7
 SINGLE LOCKHOPPER
BELL - LESS TOP
SKIP CAR

RECEIVING
HOPPER

UPPER SEAL
VALVE

HOPPER

LOWER SEAL
VALVE

r- GEAR BOX

ROTATING
CHUTE

Figure 3

7-8
 I I I FURNACE DESIGN

Our II ideal II furnace would be a free standing furnace

with a 33 ft diameter hearth. This hearth diameter has been
chosen because of:
(a) Personal experience with two furnaces of this
size.
(b) Ability to control low production levels ( as low
as 1700 NT/day) and high levels (up to over 6000
NT/day) depending on plant requirements and market
conditions.
(c) Much of the equipment used will have been
produced, tested and used on existing
installations.
A free standing furnace design has many advantages over
the traditional mantle supported furnaces.

(a) Access to the tuyere breast and bosh area of the

furnace is far less restricted.

(b) Top equipment has its own support and is totally

independent of the furnace shell.

(c) Piping configuration in the furnace area can be

simplified.
(d) Maintenance walkways around the stack of the
furnace can be enlarged for ease of access and
maintenance.
(e) The complicated designs necessary to cool the
mantle area are no longer required.

7-9
 iv FURACE COOLING
In this section we will look at three areas of the
furnace: shell cooling, tuyeres and tuyere coolers, and
underhearth cool ing .
Shell Cooling

Our furnace will be completely stave cooled from hearth

to the underside of the stockline armour. The cooling medium
wiii be forced recirculated, boiler quality, treated
feedwater. Boiler quality water with 02 scavenger additions
and corrosion inhibitors is necessary to prevent oxidation or
build-ups inside the cooling pipes, which ultimately cause
reduction in cooling efficiency. (Figure 4)

Stave design will incorporate some of the following

features: (Figure 5)

(a) Ductile iron will be used in low heat load areas

of the hearth, tuyere breast and upper stack of
the furnace. Ductile iron has better high
temperature crack resistance than grey iron. In
the high heat load areas of the bosh, lower and
mid stack of the furnace we will use copper
staves13,14.
(b) Intense corner cooling pipes will be incorporated
in the bosh, lower and mid stack staves.

(c) A serpentine pipe will be cast behind the four

body pipes, as a backup to protect the cast iron
in case of a body pipe failure.

(d) Periodic spacing of water cooled ledge staves will

be installed to hold the initial brick lining in
place for as long as possible.

(e) Alumina/Silicon Carbide brick will be embedded in

the ribs of the bosh and stack staves to protect
the staves after the loss of the refractory lining
in front of the staves.

Supply of water to the staves will be done by four

separate pumping systems. Each system will feed one of the
main body pipes in each stave. This has been done to provide
for cooling in all staves, even if one system fails for any
reason. Ledge and corner pipes will be staggered and balanced
among the four systems to equalize cooling requirements.

7-10
 STAVE COOLED FURNACE

CERAMIC

TUYERE
JACKET
HEARTH CARBON
WALL

Figure 4

7-11
 ,,IiIiil
I I I _~_----.:==.===~l iii
.._..~ _~..__" .......... ~i
I ..

....~_.. _.... ...---.. ~----

l ~_..~--~-_~----..-=.._..~_..~..-
I l . -.4l..-- ............... .
I.., .l...._~...,-i

V \
.I
l I
tI
..

~UcI
i-
i:
i:
- i

cI
G, C
a.
.- U
0. en
cI Q)
C1

.-0.
C1
c:
-i:
.....- C-
o. Q.
i-
(I C)
0.
i. ;) i:
-C i-
G, :
(/' o o
, CO U

Figure 5

7-12
 Flow meters will be used to monitor feed and discharge
volumes to each of the four systems as a primary means of
leak detection. Thermo-couples in the same location as the
flow meters will also be used to measure total heat flux on
this portion of the furnace.

With a cooling system of the design as described, our

furnace should run gas tight and trouble free for at least 10
years and probably up to 20 years if required.

On the last rebuild of "E" furnace, 9 rows of staves

with some corner cooling were installed. Three rows of bosh
staves replaced the existing shower cooled bosh. Six rows of
stack staves replaced the existing plate cooled stack. All
staves were made of ductile iron. The existing tuyere breast
and shower cooled hearth walls were not altered as they have
been proven to supply adequate cooling in these portions of
the furnace.

Tuyeres and Tuyere Coolers

Tuyeres on our "ideal" furnace will be dual chamber,

with high intensity nose cooling. This is required to
ensure that the most vulnerable area of the tuyere gets the
best cooling possible. If a tuyere nose does fail during
operation and the body circuit is still intact, the water to
the nose circuit can be turned off and the tuyere changed at
the next regular maintenance stop. Tuyeres will be externally
coated to protect them from liquid iron splashes and
internally insulated for energy efficiency.

As with the stave system, the water supply to the

tuyeres and coolers will be recirculated boiler quality
treated feed water. There will be two independent systems i
one very high pressure system for the nose circuits and one
of slightly lower pressure for tuyere body, tuyere coolers
and stove valves. Accurate feed and discharge flow
measurement will be installed on each circuit to monitor
leaks.
This system was contemplated for "E" blast furnace but
was ultimately put on the back burner because of overall
economic restrictions.

7-13
Under Hearth Cool ing
Our furnace will have an induced draft under hearth air
cooling system installed in the hearth refractory. This
cooling has proven successful in reducing erosion and thus
extending the campaign life of the hearth bottom. This type
of cooling is currently in use on all of Stelco i s furnace
hearths.

V FURACE REFRACTORIES
Having chosen our basic furnace design and cooling
system, we will now look at the refractories for inside the
furnace.5 Above the under hearth cooling tubes, which are
already embedded in high conductivity carbon, we will install
four 28" thick layers of carbon beams. (Figure 6) Our
hearth wall will be a 27" thick layer of hot pressed carbon
brick, using alternate layers of 3 - 9" and 2 - 13.5" bricks.
Of prime importance is to ensure good contact between the
staves and the hearth carbon, by using a tar based, high
conducti vi ty rammed refractory in the gap between the hearth
beams and the staves. Carbon in the hearth wall will be laid
tight to the staves. The hearth wall will be corbelled to
54" in the area of the iron notches to account for the excess
wear expected in this area of the hearth. Refractory around
the tuyere coolers will be hot pressed carbon bricks, pre-cut
and pre-glued for ease of installation.

To extend the life of the staves we will install a

carbon lining in the bosh and tuyere breast of the furnace
and a high alumina lining in the stack.

Stockline armour will be installed in the top 6 feet of

the furnace to protect the shell from the abrasion of the
burden. This armour will be installed in a concrete fill and
will be supported by a steel ring and tie rods back to the
shell behind the armour.

VI STOVES AN HOT BLAST SYSTEM

Our "ideal" blast furnace will be built to provide the
operator with the capability of running with a straight line,
2200 degrees F hot blast temperature at a maximum wind rate
of 200,000 SCFM.

7-14
 /TUYERE
'CARBON BRICK'

SHELL
SLOW IN
UNING

STAVE

CARBO
CARBON
.-UNDERRnRTH-- , , CARBON
COLING
CARBON .'
TUBE CARB
; 'I.; RE BRICK
, FIRE BRICK-

SECTION - FURNACE HEARTH

Figure 6

7-15
 To achieve this temperature requirement we will install 3
stoves with the following features: (Figure 7)

(a) Internal ceramic burners because of their high

efficiency and trouble free operation as opposed
to conventional external burners.

(b) A waste heat recovery system on the stove stack to

preheat the combustion air.

(c) High silica checkers capable of providing a dome

temperature of 2400 degrees F.

(d) Provision for enrichment of blast furnace gas with

either natural gas or coke oven gas, depending on
current plant needs for fuels.
(e) Gate type hot blast valves because of their
trouble free operation and their reliability at
high temperature and high pressure operation.
(f) A back draft stack system capable of burning the
gas at the bottom, prior to emission to the
atmosphere. This stack is absolutely necessary to
protect the stoves from being exposed to very
dirty and sometimes very hot products of
combustion during back drafting.
The design of the hot blast main and bustle pipe is
very critical to the operation of this system. Because of
the volume and temperature of hot blast anticipated there
must be sufficient allowance for expansion and movement of
the stoves, hot blast main, bustle pipe and pent stock .
Double cardan tuyere stock will be provided to allow for
movement between the bustle pipe and tuyeres.

Provision will be made for oxygen inj ection, of up to

15% of the wind, into the blast system so that if economics
dictate, oxygen enriched hot blast can be provided to the
operator. Steam inj ection for blast humidity control will
also be provided.

7-16
,'
i
INTERNAL COM-
BUSTION CHAMBER

::'STOVE '
GAS MAIN

HIGH TEMPERATURE STOVE

Figure 7

7-17
 Our blowpipes will be designed to provide the operator
with the ability to inject natural gas, oil or coal or a
combination of these fuels. Because of the environmental
restrictions on coke ovens, inj ected fuels are a necessity.
Very critical to the efficiency of high volume inj ection of
oil, gas or coal is the need to know the oxygen availability
at each tuyere so that fuel rates can be adjusted accordingly
to provide for proper combustion of these fuels. Venturi
tuyere stock will be provided to give this information.

A hot blast main isolation valve will be provided

to allow stove maintenance without having to drop blowpipes
and seal all the tuyeres.

In an effort to get at least another 2 a years out of

the stoves at "E" furnace, all three stoves were completely
rebuil t during the last reline. In addition, a back draft
valve and stack was installed to protect the stoves during
furnace stops. No other major changes to the existing
equipment were done. E furnace currently has the capability
to inj ect natural gas and pulverized coal.

VI I GAS CLEANING PLAN

Our "ideal" furnace will have a gas cleaning plant
capable of handling 280, 000 SCFM of blast furnace gas at
three atmospheres top pressure.

Primary dust removal will be done in a dustcatcher.

Dust removal will be done on a continuos basis using a pair
of seal valves and a surge hopper. A spherical shutoff valve
will be installed at the top of the dustcatcher. This will be
used to isolate the furnace from the gas cleaning system
during furnace stops. Because of the valve design, it can be
utilized to provide maintenance access to all of the gas
cleaning equipment.

Following the dustcatcher will be a variable annular

gap scrubber. This unit will provide for final gas cleaning
and cooling and for top pressure control. To give the
operator the most flexibility based on the required wind rate
and top pressure the vessel will be designed with 3 variable
gap cones in parallel. (Figure 8)
To ensure that as much moisture as possible is removed
from the gas a mist eliminator will be installed downstream
of the scrubber.

7-18
 INLET l
i

PRE-WASH
SPRAYS

HIGH PRESSURE
SIDE
GAS
OUTLET
"
~R.S. SPRAYS

R.S. ELEMENT

R.S.
ELEMEN
HOUSING
LOW PRESSURE
SIDE

ANNULAR GAP
SCRUBBER

Figure 8

7-19
 Supporting the gas cleaning plant will be a
recirculating water system comprised of thickeners, cooling
tower and lagoons to remove the dirt particles from the
process water.
On the last "E" reline, the gas cooler, septum valve
and precipitators were replaced with a single unit annular
gap scrubber and a packed bed cooler. The cost of this
installation was approximately equal to the cost of
overhauling the existing equipment. This new design operates
with a much lower maintenance cost and still provides good
top pressure control, and cool, clean gas for the stoves and
central power station.

VIII CASTHOUSE

The final maj or area to deal wi th is the cas thouse . In

our "ideal" blast furnace, this is the area that is the most
dynamic and can create one of the most disastrous situations
if not handled properly. Our aim in designing a cast house
is to create a truly continuous process by giving the
operator the ability to remove the slag and iron from the
furnace as it is made. This philosophy has many advantages:

(a) The most stable operating blast furnaces run with

a dry hearth practice.

(b) With a dry hearth practice the furnace can be shut

down safely at any time for maintenance.

(c) The metal generated is stored in torpedo cars, not

in the hearth and is thus available to the
steelmaker when he needs it.

The best way to accomplish the above requirements is

wi th a furnace that has 4 tapholes, 90 degrees apart. Each
taphole will have its own convection air-cooled trough,
feeding a tilting runner. Slag from each pair of tapholes
will be run to a slag granulation plant, with slag pits as a
backup to the granulator.

The taphole drill and mudgun will be located on the

same side of the trough so that mobile equipment can be
utilized for cleanup and tearout when trough maintenance is
required. Our mudgun will be a hydraulic powered unit,
capable of handling the best taphole clays. The taphole
drill will have a reverse jackhammer capability so that a hot
bar practice can be utilized if the operator so wishes.

7-20
 Each taphole with have remote cameras to moni tor
casting from a central control room. Continuos iron
temperature measurement will be done using either immersion
thermocouples of infrared monitors. Each ladle position will
have automatic ladle fill monitors as well as automatic
probes for iron sample and ladle temperature measurement.

There are a number of labour-saving devices which will

also be included on the cast house floor. One is a rod
changer to install a new drill rod or soaking bar on the
taphole drill boom and also to remove the old, hot i spent rod
after opening of the taphole. 6 Another device is an oxygen
opener to allow for remote oxygen lancing of the taphole
wi thout having to place a person adj acent to the hot trough. 7

Access to the casthouse will be via ramps from ground

level - one positioned for each pair of tapholes. A second
mezzanine level will be located above the casthouse floor
below the tuyeres. This will be accessible by ramp from the
casthouse floor. With this arrangement, mobile equipment can
be utilized for tuyere, cooler and penstock changes.

The overall casthouse floor will be large enough to

store spare equipment and spare casthouse refractories yet
still allow for full access by mobile equipment such as
backhoes, small front end loaders and fork lift trucks.
Because of this mobile equipment the casthouse floor will be
made as level as possible.

To eliminate casthouse emissions and improve the

environment for the casthouse crew, all runners will be
covered. A baghouse will be constructed to draw the fumes
from the runners, both during casting and during rebuilds of
the runner-work.

A remote control crane, large enough for tilting runner

changes will be available on each casthouse.

Currently "E" Blast Furnace has two tapholes, each with

its own casthouse. During the last reline this area was
completely revamped. Included in the revamp was:

(a) New air-cooled, carbon lined troughs.

(b) New hydraulic mudguns.

(c) New taphole drills.
(d) Drills and guns mounted on the same side of
the trough.

~21
 (e) Installation of a tuyere mezzanine level,
accessible by forklift from the casthouse via
a ramp from the one casthouse.

(f) A ramp from grade level to the casthouse

floor.
(g) A complete rebuild of the existing casthouse
floor to make it level for equipment access.

(h) Runner covers designed for fume capture and

also to carry mobile equipment in critical
locations.

The above improvements resulted in a reduction in

refractory costs, improved casthouse availability and a
reduction in physical "bullwork" required by casthouse and
maintenance crews.

IX INSTRUMENTATION AN CONTROL EOUIPMENT

The final requirement of our "ideal" blast furnace is to
be able to monitor, analyse and control both the iron making
process itself and all the various mechanical systems
previously described. When choosing control equipment, keep
in mind that we have three requirements:

(a) To know instantaneously the status of the

process;
(b) To collect and display data in a format which
will be useful to the operator for decision
making and also to indicate to him subtle
changes or trends in the process;

(c) To collect sufficient data to analyse past

performance in an attempt to improve future
performance.
To monitor and control the various mechanical systems we
will select modular units. Each unit will control its own
specific area such as stove system, stave cooling system, gas
cleaning plant, stockhouse and bell -less top. 8

7-22
 Because this is an "ideal" blast furnace we would put
in all of the available equipment to allow us to monitor the
process variables including: (Figure 9)

(a) Top gas analyzer complete with a good sample

preparation unit to measure CO, CO2, H2, N2, 02
and thermal value of the top gas. (Note: 02
would only be used during maintenance stops
or a complete furnace blowdown.)

(b) Four stockrods to monitor filling. (Three

micro wave and one mechanical rod as backup.)
(c) A profile meter to allow for a periodic
measurement of the burden profile.

(d) Four above-burden probe for temperature and

gas analysis across the furnace radius, 900
apart.
(e) In-burden gas and temperature probes at the
upper stack and mid stack.
(f) Vertical probe.
(g) Refractory thermocouples in the bosh and
stack to monitor rate of wear and ultimate
loss of this refractory.

(h) Thermocouples in the staves to monitor their

performance.
(i) Hearth wall and under hearth thermocouples.
(j ) Pressure taps on the stack to measure stack
pressure drop.
(k) Top pressure and temperature measurement.
(l) Hot blast pressure, temperature and humidity
measurement.
(m) Cold blast volume measurement. (i. e. Turbo
blower flow minus snort valve bleed)

(n) Flow measurement of wind to each tuyere.

(0) Flow measurement of inj ected fuel to each
tuyere.

7-23
 STOCKRODS
VERTICAL PROBE
PROFILE
METER
GAS ANAL YSER

INFRA RED
CAMERA

THERMOCOUPLES

PRESSURE
TAPPINGS

BLAST
INSTRUMENTS

FURNACE PROCESS SENSORS

Figure 9

7-24
 (p) Flow, inlet temperature and outlet
temperature measurement to all furnace
cooling elements to determine the wall heat
flux.
As one can see, with all of the above equipment to
monitor and analyze we will have to develop a supervisory
computer system to monitor and display the data in a format
that is useful to the operator.

Once we have collected sufficient information for a good

data base, and have gained some confidence in the reliability
of our equipment, we will look to incorporate some on- line
computer control of our process. These systems have been
developed and proven in operation and are currently being
used on some furnaces in Japan and Finland. (Figure 10) 9,10.
During the last reline on "E" Blast Furnace the
following equipment was installed:

(a) PLC with on line backup for stockhouse and

bell - less top control.
(b) PLC with basic relay logic for stave system
control.
(c) Data logger with computer display of the
various thermocouples in the stave cooling
system.
(d) Provision for above-burden probes.

x SUMY
I have approached this topic with the same perspective
as "the kid in the candy store". As one can see it would be
impractical to build an "ideal II blast furnace with all of the
recommended "goodies
II . This becomes very evident when we
consider the amount of equipment and changes we initially put
on our "wish list" for "E" Blast Furnace and then compare it
to the final list of equipment selected based on our economic
rationalisation.
As good blast furnace operators, we must each look at our
own operation, available capital for new equipment and our
operating budgets before we can justify each individual
expenditure for new equipment. i would hope that this
lecture will help you in the future when you look at
improving your blast furnace operation.

7-25
 ti
z
a:
II
E
:&
II
I;
Oa:
II II
-0
Cl ¡:
~
ic fi
II
~U II z
io . iu~o
3c
t; %

.. z a:
C

iI- a: Q
a: :I ~
z
~
:) 8 ou
.. II
a: .~i-
" ti

:I :E a:
uo Ii :)

Z ! =
~
fA
! :E

ou II
U
o
f
..
cu
-
a.
~
~
æi¥
C...
OR Z II
oi 0 Cl
U

Figure 10

7-26
 REFERENCES

1. HILL, R. N., "Blow-in and Low Level Operation of

Stelco Lake Erie Works No.1 Blast Furnace" - A.M.I.E.
Ironmaking Proceedings Vol.40 Toronto, Ontario. 1981
(pages 300-307)

2. HOLDITCH, J. E. "Blast Furnace Design III II - An

Intensi ve Course - Blast Furnace Ironmaking Vol I
McMaster University, Hamilton, Ontario. May 1989

3. BERCZYNSKI, IIBlast Furnace Design 11111 - An Intensive

Course - Blast Furnace Ironmaking Vol I - McMaster
Uni versi ty, Hamil ton, Ontario. May 1985
4 . BERNARD, G. i & CALMES, M., II Modern Blast Furnace
Design" - Paul Wurth CY Luxembourg publication.

5. VAN LA, J., II Ironmaking Refractories" - An Intensive

Course - Blast Furnace Ironmaking Vol I McMaster
University, Hamilton, Ontario. May 1989
6. Nippon Steel "Rod Changer" - Nippon Steel Corporation
publication
7. Nippon Steel "Oxygen Opener" - Nippon Steel Corporation
publication
8. BEST, C. L., & CARTER, G. C., "Application of modern
technology to the design of a large blast furnace"
Davy McKee publication.

9. Nippon Steel "Outline of Ironmaking Division" - Nippon

Steel Corporation, Oita Works publication
10 . Rautaruukki OY "The Rautaruukki Blast Furnace
Supervision and Control System" - Rautaruukki OY
Engineering publication.
11. Jo Isenberg-O'Loughlin , "Banking on Blast Furnaces" -
33 Metal Producing 11/97
12. Thomas A. Obrecht, David M. Armstrong, Anthony Bridges,
David V. Walnoha, John A. Carpenter, "Automated Raw
Material Handling System and Blast Furnace Charging
i
System at Rouge Steel - AISE Annual Convention and
¡
I
Mini-Expo, Pittsburgh, September 1998

7-27
13. Luc Bonte, Heli Delanghe, Maarten Depamelaere, Bert
Speleers, "Installing Copper Staves and Operational
Practice at Sidmar" - AISE Annual Convention and Mini-
Expo, Pittsburgh, September 1998

14. Robert G. Helenbrook, Paul F. Roy, Hartmut Hille

"Correlation of Experimental Data with Analytical
Predictions for Blast Furnace Copper Staves" - AISE
Annual Convention and Mini-Expo, Pittsburgh, September
1998

7~8
 LECTURE #8

IRONMAKIG REFRACTORIES: CONSIDERATIONS

FOR CREATING SUCCESSFUL REFRACTORY "SYSTEMS"

Albert 1. Dzermejko
Hoogovens Technical Services Inc.
Pittsburgh, Pennsylvania, USA

Abstract: Successful lifetimes of refractories utilzed in the blast furnace are

dependent upon a variety of factors. The factors that directly influence performance
and lifetime can be categorized as external or internal to the refractories. "External"
factors are those influences that have nothing to do with the refractories themselves
such as furnace productivity, operating practices, burden material type and quality,
furnace availability, furnace geometry and cooling capability, yet tremendously affect
pedormance potentiaL. "Internal" factors are those influences that have a direct
bearing on refractory pedormance such as configuration, wear mechanisms, stresses
and thermal movements, heat transfer capabilities, material type, characteristics and
/1 properties. The success or failure of refractories wil be determined by how these
external and internal factors are addressed or ignored. The paper reviews these
i
i ¡ significant factors, with the intention of providing guidelines for creating successful
refractory "systems" in the blast furnace.
i

ìI
INTRODUCTION

Optimizing blast furnace productivity and effciency demands high rates of tuyere
I
injected fuels, oxygen injection and higher hot blast temperatures. Profitability
optimization often requires rationalization of facilities and concentration of production
i
in fewer, highly productive furnaces. These factors result in increased thermal
loading, more frequent and intense temperature "peaks" and higher potential for
destructive effects on blast furnace refractories.
f

8-1

I
 The concentration ofproduction in fewer blast furnaces often in single-furnace plants,
increases the need for reliable, uninterrpted operation. Unscheduled stops to repair
or replace damaged linings or reduction of production intensity to "nurse" sick linings
to permit continued, albeit reduced production levels, are simply ttnacceptable in
today's competitive environment. Furthermore, capital intensive relines often result in
crippling production interrption and adversely affect profitability. Consequently,
long campaigns with minimum reline periods are essentiaL.

Today, it is possible to design and configure blast furnace lining/cooling "systems"

that provide the potential for continuous, uninterrpted service and allow for the so-
called "endless" campaign. The cost of relining and providing the required
components can represent a large proportion of the capital available for the entire
plant. The capital available is often limited because of the cyclic nature of the
business and by higher priorities such as the finishing end. These capital investment
limitations often dictate compromises in design, configuration and materials with
consequential pedormance penalties. However, to achieve the endless campaign
requires cooling capability and protection for it, utilizing an appropriate refractory
system.

Pedormance and lifetimes of refractories are dependent upon a variety of factors, both
external and internal to the lining/cooling "system". The success or failure of
refractories wil be determined by how these external and internal factors are
addressed or ignored. The actual refractory product comprises only one par of a
complex, interrelated system of components and features affected by these external
and internal factors. External factors are those influences that have nothing to do with
the refractories themselves such as furnace productivity, operating practices, burden
material tye and quality, burden distribution capability, furnace availabilty, furnace
geometry and cooling capability, yet can adversely affect performance potentiaL.
Internal factors are those influences that have a direct affect on refractory performance
such as configuration, wear mechanisms, stresses, thermal movements, heat transfer
capabilities and refractory material type, characteristics and properties.

There is no ideal or "pedect" refractory which possesses magical powers to guarantee

long life. The very best refractory material for a paricular application wil fail
miserably if consideration is not given to these external and internal factors.
Refractory selection based solely on properties wil not assure successtul pedormance
or long life. It is imperative that expected operating conditions be identified, wear
mechanisms evaluated, and a comprehensive analysis conducted of all of the external
and internal factors which wil impact refractory performance. Only then can the
"system" be properly configured and refractory materials selected, appropriate for the
configuration and application.

8-2
 BLAST FURNACE HEARTH

One of the largest users of refractory materials is the blast furnace hear. Worldwide,
the 'configuration and design of this large volume refractory system varies
considerably, with major differences in performance. This zone of the blast furnace
probably exhibits more varied designs, conflicting practices and vastly different
performance histories than any other. Technical aricles from certain countries
continually describe the hear as the zone of the blast furnace most responsible for
the termination or interrption of the campaign. Contrasting this experience are the
apparent success stories from other countries of trouble-free, long campaign lives of
blast furnace hearhs (1). Figure 1 depicts these historical wear pattern differences.
There are many reasons for this different pedormance history, especially when the
designer analyzes the internal and external factors which affect the hearh refractory
"system". Especially important are behavioral differences of the various refractory
materials utilized, resulting from these factors.

Refractories

Traditional hear refractories have been carbonaceous in nature. Various grades of

carbons, graphite containing carbons, semigraphites and graphites are utilized. Often,
various grades of ceramic refractories are combined with these carbonaceous materials
to form composite linings. It is also common to utilize several types and grades of
carbonaceous refractories in these composite lining configurations to utilize specific
properties or characteristics of each type to their best advantage.

The words "carbon" and "graphite" are often used interchangeably in the literature,
but the two are not synonymous. Additionally, the words "semigraphite" and
"semigraphitic" are also misused. Compounding the problem is the fact that there are
no industry-wide standards or specifications to define carbonaceous products. Each
worldwide producer manufactures unique products exhibiting unique properties and
characteristics. This is the result of raw material differences, proprietar product
ingredients, additives, manufacturing methods and techniques. This is important to
recognize because the behavior of these unique products can be very different in the
same application. These behavioral differences can result in major system
performance differences, despite experiencing identical external and internal wear
factors. This is especially true when refractory configuration is not compatible with
material characteristics, properties and behavior resulting from these factors.

The following describes the major differences and characteristics of the carbonaceous
material types used as refractories in the blast furnace. Please remember that the
general nomenclature of these material types represent an entire family of materials,
from a variety of manufacturers exhibiting unique compositions, characteristics and
properties.
I

8-3
 Carbon

The terms carbon, formed carbon, manufactured carbon, amorphous carbon and baked
carbon, refer to products that result from the process of mixing carbonaceous filler
materials such as calcined anthracite coal, petroleum coke or carbon black with binder
materials such as petroleum pitch or coal tar. These mixtures are formed by molding
or extrusion, and the formed pieces conventionally baked in furnaces to carbonize the
binder at temperatures from 800° to 1400°C (1500° to 25500P). The resulting product
is comprised of carbon paricles with a carbon binder.

Typically, conventionally baked carbon is manufactured in relatively large blocks. As

the binders carbonize and the liquids volatilize they escape through the block,
resulting in porosity. This porosity results in a permeable material that can absorb
elements from the blast furnace environment such as alkalies. These contaminants use
the same passages that the volatilizing binders used to escape the block to enter the
carbon and chemically attack the structure.

Conventionally baked carbon can be densified and thus permeability improved and
pore sizes reduced. This can be accomplished by the introduction of additional
binders impregnated into the baked carbon under a vacuum and the resultant product
rebaked to carbonize the impregnation. Multiple impregnations are also possible to
double or triple densifY the end product. Each densification however, adds additional
cost and results in a higher priced product.

Some manufacturers also add special raw materials to the carbonaceous mix prior to
baking to improve the end products' properties. Silicon carbide, alumina powders, or
silicon metal can be added to improve permeability, reduce pore sizes and improve
abrasion resistance. Arificial or natural graphite can also be added to improve
thermal conductivity. Some manufacturers also impregnate the baked carbon to
improve thermal conductivity. However, each of these steps also results in a higher
priced product.

Conventional carbon is manufactured in large blocks and can be machined to precise

tolerances. Grain structure however, can be different depending on the manufacturer,
which can result in moderate paricle pull-out at sharp comers when paricle sizes are
very large.

Hot-Dressed Carbon

A North American manufacturer developed a unique proprietar method of

manufacturing carbon which is called the BP process or hot pressing. In this method
of manufacturing carbon which, as previously described, is a product comprised of
carbon paricles with a carbon binder, a special pressing/carbonizing operation is
utilized.

8-4
 In this process, carbon paricles and binders are mixed as with conventional carbon,
but are then introduced into a special mold. A hydraulic ram then pressurizes the
mixture while simultaneously an electric current passes through the mold, carbonizing
the binders. Unlike conventionally baked carbons that take several weeks to properly
bake the binders, this proprietar process carbonizes the binders in minutes. More
importantly, as the liquids volatilize, the hydraulic ram squeezes the mixture together,
closing off the pores formed by escaping gases. This forms an impermeable carbon
compared to conventionally baked carbon, usually at least 100 times less permeable.
This low permeability makes it diffcult for blast furnace contaminants such as alkalis
to enter hot-pressed brick.

Special silica and quar additions are also added to improve alkali attack resistance.
These additions are made because sodium or potassium in the blast furnace react
preferentially with silica, forming compounds that do not swell in the carbon.
Normally, the reaction of these alkalis with carbon would form lamellar compounds
which do swell, causing volume expansion spalling of carbon. However, the
combination of hot pressing and raw material composition results in an improved
alkali-resistant carbon.

Hot pressing also results in a higher thermal conductivity than conventional carbon
which helps promote the formation of a protective skull of frozen materials on the
lining hot face. High conductivity linings have the ability to maintain a hot face
temperature that is below the solidification temperature of iron and slag. The resulting
skull protects the wall from chemical attack and erosion from molten material flow.

Because of the special manufacturing process required for hot pressing, the product is
limited to sizes not exceeding approximately 500 x 250 x 120mm (20 x 10 x 5 in).

Graphite

The term graphite, also called synthetic, artificial or electrographite, refers to a carbon
product that has been additionally heat treated at a temperature between 2400° and
3000°C (4350° and 54000P). This process of graphitization changes the
crystallographic structure of carbon and also changes the physical and chemical
properties.

Graphite is also found in nature in flake form as a mined mineraL. It can be added to
various carbonaceous or ceramic refractory products to enhance thermal conductivity.
It is also utilized as the major component of graphitic ramming materials. The soft,
flake form of natural graphite is unsuitable as a refractory lining material however.

Arificial or synthetic graphite refractories begin as a baked carbon material, similar in

manufacture to the carbon refractory material described previously. However, after
carbonizing of the binder is completed, this baked carbon is then loaded into another
furnace to be graphitized at a high temperature. Graphitization changes the structurç;

8-5
 not only of the carbon paricles, but also the binder. The resulting product is
comprised of graphitized paricles as well as graphitized binder.

There is no industry-wide system for designating the various grades of graphite that
are commercially available. Each manufacturer has a method and nomenclature to
describe the available grades and varieties which are made for specific purposes or
properties. These grades differ with regard to raw materials, grain sizes, purity,
density, etc. For denser versions, the porosity of the material can be filled with
additional binder materials such as tar or pitch by impregnation under a vacuum. Then
the impregnated material is regraphitized, forming a less porous product. Multiple
reimpregnations and graphitizations can be pedormed to provide additional
densification and higher thermal conductivity.

Purification can be utilized to reduce the ash levels of graphite for special
requirements. In addition, proprietar manufacturing methods and techniques can also
be used to minimize ash or iron contamination of graphites. Since iron is a catalyst for
certain chemical attack of graphite in a blast furnace, graphites intended for use as a
refractory should contain relatively low iron.

Graphite products are manufactured in large blocks or rounds and must be cut and
machined into shapes for use as a refractory. Precise tolerances can be maintained
with machined graphite components due to its easy machinability.

Semie:raDhite

The term semigraphite is used to describe a product that is composed of arificial

graphite paricles mixed with carbonaceous binders such as pitch or tar and baked at
carbonization temperatures of 800° to 1400°C (1500° to 2550°F). The resulting
product is comprised of carbon bonded graphite particles in which the graphite
paricles had previously been manufactured at temperatures close to 3000°C (5400°F),
but with binders that have only been baked in the 800° to 1400°C (1500° to 2550°F)
range. The resulting product, a true carbon bonded graphite exhibits higher thermal
conductivity than the carbons but, because of the carbon binder, not as high as 100%
graphite. Thermal conductivities wil var with baking temperature and can be
increased by baking at higher temperatures.

These products are also conventionally baked (as described for carbon), which results
in a relatively porous materiaL. However, these conventionally baked semigraphites
can also be densified and rebaked to carbonize the impregnated binder. Thus porosity
and consequently, permeability can be reduced. Some conventionally baked
semigraphites are also impregnated with or combined with silicon metal and silicon
carbide for greater abrasion resistance and lower permeability. These products
however, are usually intended for use in the bosh and stack.

8-6
 Semigraphite products are manufactured in large blocks or rounds and must be cut and
machined into shapes for use as a refractory. Precise tolerances can be maintained
with machined semi graphite components, but the carbon binder makes it a harder
material than graphite, which can affect machining pricing.

Hot-Dressed Semi2raDhite

One North American manufacturer also utilizes its proprietar hot-pressing method to
make a true semigraphite refractory product. The resultant product is considerably
less penneable and has a higher thennal conductivity than conventionally baked
semigraphites.

Two distinct products are available for a variety of applications. One grade is
composed of crushed graphite paricles, which were previously processed at
graphitization temperatures, with a carbonaceous binder and the addition of silica and
quarz materials for alkali resistance (as previously described for hot-pressed carbon).

The other grade is a silicon carbide containing hot-pressed semigraphite refractory. It

is composed of the same graphite component as the first product and the same
carbonaceous binder. However, silicon carbide is substituted for the silica and quar.
The resultant product is more abrasion resistant and even less penneable than the first
product. It has proven especially resistant to thennal shock and cyclic operation.

Because of the special manufacturing process required for hot pressing, the resultant
products are limited to sizes not exceeding approximately 500 x 250 x 125 mm (20 x
10 x 5 in.).

Semi2raDhitized

The tenn semigraphitized material refers to a carbon product that has been baked at a
temperature between 1600° and 2400°C (2900° and 4350°F). This high baking
temperature begins to change the crystallographic structure of the carbon and alters its
physical and chemical properties. However, because this heat treating occurs at
temperatures below graphitization temperatures, the product is considered to be only
semigraphitized. It is comprised of carbon paricles with a carbon binder, which are
both semigraphitized during baking. (This is different than a semigraphite product
which is composed of true graphite paricles with a carbon binder). Semi graphitized
carbon has a higher thennal conductivity and resistance to chemical attack (alkali or
oxidation) than carbon or semigraphite. This is because the binder is usually
preferentially attacked and the semi
graphitized binder is more resistant to attack than
the carbon binder of a semigraphite.

These semigraphitized products are manufactured in large blocks or rounds and must
be cut and machined into shapes for use as a refractory. However, because of their
semigraphitized bonding, they are more difficult to machine than a true graphite.

8-7
. ¡
 Discussion

These groups of carbonaceous materials form the basis for a full range of specialized
products that are intended to enhance performance in the blast furnace. As discussed,
various additives such as graphite paricles, alumina, silicon carbide or other ceramics
are included by some manufacturers to improve properties, or multiple impregnations
are used to improve permeability or reduce pore sizes. However, the general
description of each material classification does not change. For convenience, these
classifications are summarized in Table i.

TABLE I

Classifvinl! Carbonaceous Materials

Product Baking
Classification Temnerature. °C Particles Binder

Carbon 800 - 1400 Carbon Carbon

Hot-pressed carbon , 1000 Carbon Carbon

Graphite 2400 - 3000 Graphite Graphite

Semi graphite 800 - 1400 Graphite Carbon

Hot-pressed , 1000 Graphite Carbon

semigraphite

Semigraphitized 1600 - 2200 Semigraphitized Semigraphitized

carbon carbon

Currently, there is a large variety of carbonaceous refractory material on the market,

produced by different manufacturing techniques, exhibiting unique properties. It is
diffcult to provide material properties for these products without referring to specific
manufacturer's grade designations because as was noted before, each manufacturer
produces products that are unique to that manufacturer and thus exhibit unique
properties. However, a representative listing of some of these materials' properties are
summarized in Table II.

In general, carbon or semi graphite materials are used for the hot face lining materials
that wil be in contact with molten iron. Usually, graphite materials are reserved for a
backup lining to take advantage of their high thermal conductivities and because they
are more easily dissolved by the iron. In addition, many ceramic materials such as

8-8
 TABLE II

Representative Carbonaceous Hearth Materials

Conventionally Conventionally
Proprietary Conventionally Conventionally baked, big block baked, big Low ash,
hot-pressed baked, big baked, big micropore ' Semi graphitied block low iron
Propert carbon brick beam carbon block carbon carbon carbon bi~ block semie:raphite e:raphite
Bul density, wcc 1.62 1.6 1.7 1.6 1.65 1.62 1.67
Crushig strengt, 30.5 17.9 35.5 44 27 25 28
Mpa
Ash, % 10 * 8 4.8 13 * 0.4 0.4 0.2
00 Permeability, 9 800 -200 -21 -150 -150 N.A
i
\0 m' Darcys
Theral conductivity,
W/moK

at 600° C 18.4 10.4 4.3 II 45 42 120

at 800° C 18.8 10.4 5 15.4 N.A 38 N.A
at 1000° C 19.3 10.5 5.5 16.5 32 32 70
at 1200° C 19.7 10.9 N.A N.A N.A N.A N.A

* Contains deliberate additions of non-carbonaceous ingredients.

 high alumina, mullite and chrome corundum are used in the hear pad as a wearing
sudace to minimize exposure of the carbonaceous materials of the hearh to molten
materials. Some designers also configure a lining of ceramic materials on the face of
the hearh walls for wear protection and to minimize heat losses, mainly because of
poor historical performance with some large, conventionally baked, carbon block
designs.

Ceramic materials used for the hear pad are normally inexpensive super-duty
fireclays of 40 to 50% alumina or a variety of high alumina products in the 60% range.
The objective is to provide a lining that wil melt and vitrify (or fuse together) on its
hot face in the presence of liquid iron, effectively sealing the surface to penetration
and preventing potential brick dislodging and flotation.

In another philosophy, refractory materials such as arificial mullites or chrome

corundum are chosen which are resistant to melting. These materials however, require
joining techniques such as interlocking, tongue and groove or roll-lock interfacing to
prevent joint penetration by molten materials and resultant flotation of bricks.

Whichever ceramic materials are utilized in the pad, the effect is that the iron remains
in contact with the ceramic, which is more resistant to abrasion from moving liquids.
The carbonaceous material in the pad thus forms a cooling member instead of a
crucible, until late in the campaign when the ceramic may totally wear away by
abrasion. The high conductivity of carbonaceous materials, especially if underhearh
cooling or a graphite cooling course is utilized, enables penetration of the iron into the
pad to be arested in the ceramic layer. This provides a long-wearing hearth design,
combining the properties of two or more different refractory materials to optimize the
performance of each, in the zone to which they are most suited.

There is also a growing belief that the incorporation of a ceramic hearh pad in high
productivity blast furnaces can often result in accelerated hear wall wear. This can
be especially true if the hear well volume is less than desirable and if poor coke
quality is utilized. These conditions tend to result in higher peripheral flow of hot
metal. This problem is intensified by a high melting point ceramic pad layer, which
prevents formation of a bowl shaped "salamander" penetration and its consequential
well volume increase which reduces hot metal velocities. Alternatively, a carbon
hearh pad would quickly form a bowl-shaped "salamander" depression from
dissolution by the iron, increasing active well volume and consequently increasing the
iron buoyancy effect on the coke deadman and decreasing hot metal velocities. Figure
2 depicts the effects on hearh well volume of an all-carbon pad versus a high melting
point ceramic pad. An explanation of the resulting damaging effects is described later.

8-10
 Representative properties of ceramic materials used in the hearh are shown in Table
III. They can be combined in various layers such that the more economical materials
are located on the hot face, where they wil be consumed more easily until thermal
. equilibrium is reached. The more expensive, hot metal resistant materials are then
located next to the carbonaceous materials where they can be more easily cooled for
longevity. The tendency is to utilize specific grades of refractories in each hearh zone
that can best withstand the attack mechanisms prevalent in that zone. The result is a
hearh lining composed of not just one grade of refractory, but sometimes even six or
eight different tyes of materials, both carbonaceous and ceramic.

TABLE III
!
¡
_I ReDresentative Ceramic Hearth Materials

Material
Hard-burned
superduty 60% Artificial Chrome
ProDertv fireclav Alumina mullte Corundum
Density, glcc 2.24 2.40 2.45 3.43
Crushing 31 35 85 78
strength, MPa
Porosity, % 13 22 19 8
Thermal
conductivity ,
W/moK
at 5000 C 1.9 2.0 N.A. N.A.
at 10000 C 0.9 1.7 1.8 2.3

r Wear Mechanisms

In the hear, refractory survivability depends upon proper uninterrpted cooling. The
hearh bottom pad and walls are cooled on their cold face and almost exclusively
utilize various conductive refractory materials such as carbon, semigraphite,
semigraphitized carbon and arificial graphite alone or in combination with each other,
or combined with ceramic materials. The pedormance of the hearh lining system is
totally dependent upon effective and uninterrpted heat transfer through the refractory
configuration because it is cooled on its cold face. All chemical attack mechanisms
that affect hearh refractories are temperature dependent chemical reactions. This
means that if refractory temperatures can be maintained below the temperature at
which a particular chemical reaction begins, attack by that mechanism cannot occur.
This threshold temperature at which the chemical reaction begins is called the "critical
reaction temperature". The only way that the refractory hot face temperature can bç
maintained below the critical reaction temperature for the various wear mechanisms

8-11
 encountered is to provide an effcient and unchanging heat transfer path from the hot
face to the cold face (1,i).

If conductive refractory hot face temperatures are allowed to exceed approximately

1150°C (21000P), these carbonaceous materials wil be chemically attacked by
dissolution by the iron and wil be subjected to erosion and wear by the movement of
molten materials. This is because the refractory hot face temperature would be above
the solidification temperature of the iron and thus would be in constant contact with
the molten materials. Consequently, these molten materials may also be forced into
the pores of the refractories due to ferrostatic head and high furnace operating
pressures.

If conductive refractory temperatures are allowed to exceed approximately 870°C

(16000P), these carbonaceous materials wil be chemically attacked by alkalis and zinc
which preferentially destroy the refractory binder system. As the binder system is
attacked, material strengt and properties are destroyed, most notably thermal
conductivity. Thus, as chemical attack progresses, the ability of the refractory to
transmit heat is lost, which then results in even higher hot face refractory temperatures
and intensified attck.

If conductive refractory temperatures are allowed to exceed approximately 450°C

(8400P) for carbon and approximately 500°C (950°F) for graphite, steam oxidation
from cooling water leaks wil occur from the chemical reaction:

C + HiO-+ CO + Hz

which results in carbon loss and disappearance as it dissociates to form the two gasses
carbon monoxide. and hydrogen. The resulting carbon loss can form irregular
"ratholes", tunnels, chambers or similar cavities in the lining from the flow of these
gasses as they escape into the furnace.

If conductive refractory temperatures are allowed to exceed approximately 450°C

(840°F) for carbon and approximately 650°C (l200°F) for graphite, carbon monoxide
degradation wil occur. This reaction is catalyzed by iron contamination in the
carbonaceous materials and intensifies as iron content increases. The presence of
steam and hydrogen from cooling water leaks wil also dramatically intensify carbon
monoxide degradation as shown in Figure 3(3. This degradation results in deposition
of carbon within the molecular structure of the refractory formed during the chemical
reaction:

2 CO -+ C + COz

As this carbon deposit increases with time, it causes a volumetric expansion which
results in swelling, cracking of the refractory and destruction of strengt. The
cracking also interrpts the heat transfer path from hot face to cold face, resulting in

8-12
 increased hot face temperatures and consequential intensified chemical attack from
alkali and zinc and carbon dissolution by the iron.

Blast furnace designers generally utilize computer modeling techniques to locate the
critical reaction temperature isotherms in the hearh lining. Table iv summarizes
these critical reaction temperatures for various hear refractory attack mechanisms.
The location of these isotherms permits an evaluation of the potential chemical attack
zones and "salamander" penetration into the hearth due to thermal considerations. As
I noted earlier, the 1150°C (21000P) isotherm wil define the star of dissolution of the
carbon by iron, the 450° or 650°C (840° or 12000P) isotherms will define the start of
carbon monoxide degradation depending on the material, and the 870°C (16000P)
isotherm wil define the star of alkali and zinc attack. Additionally, the 1250°C
(22500P) isotherm wil define the softening point of some ceramics that would be
expected to be eroded away by molten material movement. This computer modeling
tool can therefore be utilzed to provide an estimate of chemical attack zones and iron
penetration, once the hear refractory mass reaches thermal equilibrium.

TABLE IV

Critical Reaction TemDeratures for Hearth Refractorv Chemical Attack

Chemical Attack Refractorv TVDe Critical Reaction

Mechanism TemDerature. °C

Dissolution in Iron Carbonaceous 1150

Alkali / Zinc Carbonaceous Stars (c 870, stops (c 1100
Alkali / Zinc All Ceramics / Ceramic 560, intensifies as temp.
Additives increases
Alkali / Zinc Silicon Carbide Additives 870, intensifies as temp.
increases
CO Degradation Carbons, High Iron Graphites / stars ê 450, stops ê 750
Semigraohites
CO Degradation Low Iron Graphites / stars ê 650, stops ê 750
Semigraphites
CO Degradation Ceramics 400
Steam Oxidation Carbons / Semigraphites 450
Steam Oxidation Graphites 500

It should be noted that the major causes of chemical attack in carbonaceous materials,
notably alkali / zinc attack and CO degradation, only occur within a specific
temperature range. The "critical reaction temperature" defines the temperature at
which the chemical attack mechanism begins. Attack severity is dependent upon
temperature, increasing, then gradually decreasing until the chemical reaction ceases
as it reaches its upper temperature limit. This attack behavior often results in the

8-13
 critical reaction temperature and the corresponding reaction temperature upper limit
both being located within a specific refractory zone. The material located between
these two isotherms wil be chemically attacked. This zone of chemically attacked
carbon could be located between two unaffected zones of carbon, at the hot and cold
faces of the affected zone. This chemically afected zone, located between two
unafected zones, is referred to as the "brittle" zone or "mushy" zone. This is shown
in Figure 4.

Actual hear pad penetration and wall deterioration is a function of more than just
thermal effects from temperature dependent chemical reactions. Mechanical stress,
lack ofthermal expansion provisions and erosion from molten material movement also
contribute to hear wear. For this reason, many computer models utilize historical
wear line data to establish boundary conditions.

These boundary conditions allow the computer model to simulate hearh wear in the
profie that historically results from that particular lining concept. However, a
paricular hearh model may have no significance for a different hearh concept or
configuration, or for a furnace exhibiting a different historical wear pattern. For
example, hearh computer models that are developed to estimate the inverted
mushroom shaped wear pattern or "elephant's foot", with severe wall material loss at
the wall/pad interface, wil not accurately predict the expected wear pattern of a
furnace that exhibits a historical bowl-shaped wear pattern, with little or no wall
material loss. Therefore, the designer must consider the concept, historical wear
performance and other "internal" and "external" factors which affect refractory
performance when designing a heart(1.

External Factors: ODerations Effects

The blast furnace hearh lining can be adversely affected by furnace productivity,
operating practices, burden material quality especially the coke and furnace
availability. The effects from these external factors can often be intensified by other
factors external to the refractory system such as furnace geometry and cooling
capability. Another external factor which can dramatically affect hearth refractory
performance is leaking cooling water from cooling plates, tuyeres or staves. No
carbonaceous refractory can survive steam oxidation caused by cooler water leaks.
These leaks can also result in sudden loss of protective skull accretions as they
explosively separate from the wall hot face due to pressure forces from steam
formation. Therefore, cooling system maintenance practices can also play an
important role in refractory lining longevity and survivaL.

The following summarizes critical operations effects which are major external factors
in blast furnace hear performance:

8-14
 Productivitv

High productivity increases the amount of molten materials flowing through the hearh
and out of the tapholes. Hearh productivity is normally expressed as an index in
tonnes of hot metal per unit heart "well" volume, per day. The increased throughput
of molten materials may accelerate hearh wear because of increased hear activity(4).
There are many interacting factors involved including the effect of geometry. Hearth
activity intensity can be reduced by increasing the holding capacity (well volume) and
increasing the taphole-to-pad distance. Figure 5 ilustrates the effect on molten
material velocity by increasing hearh well volume. Multiple tapholes can also be
utilized to decrease throughput and distribute potential erosive wear.

Coke Quality

Coke must be stable and strong to support the burden weight without mechanical
failure or degradation. Coke eventually forms a "deadman", an inactive zone in the
furnace center from the hearh upwards, to above the tuyeres in the bosh. Strong,
properly sized coke tends to result in a permeable deadman with suffcient voidage
between the individual coke pieces to permit molten metal flow. If the deadman is
permeable, it allows the liquids to flow completely across the hearh diameter. If coke
quality and sizing is poor, deadman permeability is decreased as the voidage between
pieces is reduced or disappears. This loss of permeability forces molten metal flow
around the perimeter of the deadman, in an anulus created between the refractories
and the impermeable coke mass. This peripheral flow can intensify erosive loss and
heat flux. Additionally, if the hear well volume is too small and cannot provide a
buoyancy effect by the molten iron, the coke deadman wil "sit" and rest on the hear
pad instead of "floating" and providing additional flow area for the metal under the
deadman. This is depicted in Figure 6.

Iniected Fuels

High levels of tuyere injected fuels especially coal, reduce the proportion of coke
charged into the furnace. Consequently, coke quality becomes extremely important to
hearh wear as high rates of injected fuels are utilized. It is also theorized that high
rates of injected coal have a deleterious effect on deadman permeability because coke
voidage is blocked by the by-products of combusting coaL. Operating practices must
then be adjusted to increase the proportion of coke in the furnace center. Center coke
charging practice requires furnace top burden distribution capabilities. This center-
charged coke gradually works its way downward and replenishes the coke in the
deadman increasing permeability and thus decreasing molten material flow velocities.

Availabiltv

Whenever the furnace is off-wind, furnace stability is interrpted. Refractory damage

¡
can occur from the results of coming on and off blast, refractory temperature cycling
and consequential fatigue and erratic operation that might occur during recovery from
I

8-15

'I
 the stoppage. Long campaigns are most likely with virtally continuous operation
without lengthy planed or many, short duration, unplaned shutdowns. However,
some operators theorize that off-wind periods during planned maintenance outages
encourage the formation of protective accretions (skulls) on the refractories, which are
beneficial to achieving long life. This is especially true if titania bearing materials are
charged or injected to assist with protective skull formation.

Hot Metal Chemistry

Higher hot metal silcon levels decrease the probability of hearth carbon dissolution(4).
This is because the carbon saturation level decreases with increasing silicon content.
Additionally, an increase in hot metal silicon increases the hot metal liquidus
temperature and reduces its fluidity. This results in increased metal viscosity, reduces
flow velocity and encourages accretion (skull) formation. At lower hot metal
temperatures, the carbon saturation level of the iron is lower and is more easily
achieved. Conversely, loss of hearh carbon from dissolution wil be more likely as
silicon levels are reduced.

TaDhole Lell~th and Practice

Long tapholes allow withdrawal of metal from deeper in the hearh and reduce the
probability of wall flow as the molten materials flow towards the open hole(4).
Taphole clay quality and clay gun capability play important roles in determining
taphole length. The taphole clay forms a "button" or "mushroom" where it exits the
tap hole at the wall hot face and can be progressively increased in size as the number of
taps increases. Increased taphole lengths generally result in lower refractory wall
temperatures. Short taphole lengths, especially in single taphole furnaces, generally
result in more intense sidewall activity, higher wall temperatures and increased
probability of skull loss and erosion damage. Poor taphole clay or inadequate clay gun
capability can prevent the achievement of long tapholes and their benefits. Multiple
tap holes often can be utilized to spread tap hole wear more evenly and allow for longer
clay curing times between taphole uses. Alternate casts from tapholes located on
opposite sides of a furnace also result in more effective hearth drainage. Lower
casting rates also decrease hot metal flow velocities in the hearh but increase the
casting time. Decreasing the number of casts per day also increases casting duration,
which can have an adverse effect on wall wear if taphole clay quality is lacking.

External Factors: Geometrv

The greater the volume of the hear holding capacity (well volume), the more likely
the hot metal buoyancy effect can "lift" the coke deadman. Consequently, the
available flow area for the molten material increases, which decreases their velocity
and destructive effects. Deep well hearhs provide a taphole-to-pad distance of 2m
(6.6 ft) or greater which also allows longer tapholes and decreased flow activity at the
walls. However, the positive benefits of a deep well hearth can be negated by poor

8-16
coke quality and consequential deadman impermeability. The effects on velocity of
well volume and coke quality is depicted in Figure 6.

External Factors: Coolinf! Capabilty

All hearh refractories must be cooled on their cold face which requires an
uninterrpted heat flow path. Any disruption of this heat flow path or loss of cooling
efficiency wil result in elevated refractory temperatures. The most common
disruption of cooling capability results from loss of cooling effectiveness from
sediment or mineral deposits or corrosion of the cooling elements. These effectively
insulate the water from the heat source and can result in refractory degradation and
loss. Another potential cooling capability loss can occur because of separation of the
cooling element from contact with the refractories. This most often results from high
temperature differentials across the cooling element and consequential differential
thermal expansion between the cooled element and the refractories. The resulting "air
gap" wil reduce heat transfer, significantly increasing refractory temperatures and
increasing the probability of chemical attack, skull loss and material loss. Pressure
grouting a conductive material into the separation anulus formed between the cooling
element and the refractories is a proven corrective action that reestablishes the heat
transfer path.

Desif!n Considerations

Hearh walls comprised of large carbon blocks exhibit problems that can be traced to a
combination of factors: lack of thermal expansion relief, high thermal gradients across
the wall block and the inability to accommodate differential thermal expansion. All of
these factors promote cracks with subsequent hot metal and chemical attack (5). Attack
of the wall by hot metal and chemicals most often is a result of the cracking problem.

Proper wall design requires a high thermal conductivity refractory that minimizes
thermal gradients through the wall and consequently promotes the formation of a
protective accretion of solidified materials on its hot face. Proper wall design also
incorporates provisions for radial thermal expansion of the wall but more importantly,
incorporates provisions to accommodate differential thermal expansion of the wall
thickness(6) .

Differential expansion occurs because the wall hot face temperature is higher than the
wall cold face temperature. This differential is at least 1450°C (2650°F), especially
when an accretion of solidified materials is absent. As a result, the hot face of the wall
grows at a faster rate than the cold face. The differential growth induces high stresses
in the blocks which are restrained from bending or bowing. Cracks result, parallel
with the hot face.

8-17
 Thermal spalling and cracking of the hot face can also be induced by the rigors of a
blow-in, especially when the wall design canot accommodate radial expansion and
the refractory thermal conductivity is low. This type of cracking also occurs parallel
to the refractory hot face.

Cracks interrpt the ability of individual blocks to convey heat and facilitate cooling
because each crack acts as an air gap which is a barier to effective heat transfer.
Once the ability to convey heat is lost, the protective accretion may no longer form
and therefore, carbon could be attacked by the hot metal and chemicals. This is
because the carbon temperature wil be above the critical reaction temperature for
attack by these mechanisms.

Additionally, the ramed layer required between the shell ( or stave) and the cold face
of a large block carbon wall also insulates the lining from the cooling system. This is
because ramming materials shrink when cured and possess thermal conductivities that
are significantly lower than baked carbon. The lower conductivity and shrinkage
combine to provide additional bariers to heat transfer and result in high hot face
temperatures, often higher than the solidification temperature, so that skulls canot
form on block walls.

Proper wall design not only accommodates thermal growth, expected differential
movements and utilizes a carbon refractory with high thermal conductivity, but also
uses a carbon refractory possessing an extremely low permeability. The low
permeabilty minimizes chemical and hot metal attack by preventing penetration into
the refractory.

It has also been demonstrated that a hearh refractory that possesses a low elastic
modulus, combined with a low coeffcient of thermal expansion, results in low
mechanical stress at the important pad/wall intedace. American big beam blocks and
hot-pressed carbon as well as graphite and semigraphite, fulfill these requirements.
Because of the elastic properties of these materials, expansion stresses are easily
accommodated which prevents cracks from occurring in the wall. The opposite is true
for the strong, large blocks that typically are used in Europe and Asia in an attempt to
increase the life of the hearh wall.

Because of differential expansion and bending and the tight fit due to precision
machining and the lack of thermal expansion provisions, these stronger blocks are
prone to stress cracking, pinch spalling and thermal shock. Thermal shock is
paricularly size dependent so that the larger the exposed hot face cross-section, the
more likely thermal shock wil occur. Walls composed of smaller cross-section pieces
are usually unaffected by thermal shock.

Expansion relief is also a requirement for preventing pinch spalling and stress
cracking. This relief can be provided by specially designed expansion joints between
blocks or by the use of special heat setting, carbonaceous cements. Ideally, these
cements should be installed in a suffciently thick layer to provide expansion relief

8-18
before curing. Afer curing, they should provide a strong carbonaceous bond to seal
the joint. Multiple layers and rings provided by small brick also permit differential
expansion without cracking.

High thermal conductivity hot-pressed carbon and semigraphite refractories promote

the formation of a protective skull of frozen material on the hot face of hearh walls.
This protective skull prevents wear of refractories due to erosion from gases or molten
materials. Additionally, rammed layers should not be utilized to maximize heat
transfer to the stave or shell.

A single, full-thickness block canot accommodate the differential growth

experienced and consequently it cracks, thus interrpting heat transfer. The crack
prevents the hot face of the block from being cooled below solidification temperature
so a protective skull canot form. Thus, the large block carbon is continually exposed
to molten materials at high ferrostatic pressure and high gas pressure. These high
pressures tend to force the molten materials into the pores of the big block materials.
Hot metal impregnation results in damage to the carbon and additional cracking and
spalling.

In an attempt to prevent hot metal impregnation of large carbon blocks, many

manufacturers have introduced densified or reimpregnated carbon blocks with low
porosity and minimal pore size. These "micropore" carbon refractories are designed to
limit the amount of molten materials that can enter the structure of the material
through its porosity. This solution is contrar to that employed with hot-pressed
carbon or semigraphite concepts which utilize high thermal conductivity and the
prevention of cracking to promote a hot face temperature that is maintained below
solidification temperature. Thus, in the case of the latter, cooler wall concept, a skull
quickly forms on the wall hot face and impregnation by molten materials is prevented.
The resulting skull thickens over time to form an insulating layer once thermal
equilibrium is achieved. Wall hot face temperatures at the back of the skull in these
systems are typically in the range of 200° to 300°C (400° to 570°F). Another
advantage that this cooler wall provides is that other temperature dependent reactions
such as carbon monoxide degradation, alkali and zinc attack canot occur as long as
the wall temperatures remain below their critical reaction temperatures. Typically,
these critical reaction temperatures are between 450° and 1100°C (840° and 2000°F)
as previously discussed. As long as wall temperatures can be maintained below these
critical reaction temperatures, attack by these mechanisms canot occur. However, if
stress-induced cracking, deterioration of ram layers or any other disruption of heat
transfer occurs, wall temperatures wil increase, usually above these critical reaction
temperatures. This results in chemical attack of the wall material in the zone of the
wall that exceeds these critical temperatures. As was also previously discussed, some
chemical reactions do not occur above 11 OO°C (2000°F). Consequently, a deteriorated
band of material can be formed within the wall thickness. This brittle zone is usually
sandwiched between sound carbon on both the hot and cold faces which is defined by
the critical reaction temperature isotherm locations.

8-19
 As was also previously mentioned, some designers are utilizing a ceramic hot face
layer on the carbon walls to prevent wall erosion. In addition, because of the low
thermal conductivity of these materials, it is believed that wa11 heat losses wil be
reduced. Several furnaces in Europe and Asia have been lined using this concept. The
longevity of the ceramic is dependent upon good thermal contact with the carbon and
maintaining uninterrpted heat transfer capability through the large block carbon for
the life of the ceramic. For reasons previously discussed, large block carbon walls are
prone to cracking and loss of heat transfer capability. Thus, if cracking does occur,
high temperatures result in the ceramic, hastening their demise.

In Europe and especially Japan, it is a1so common practice to create hearh protection
by adding significant amounts oftitania bearing ores in the burden or directly injecting
through the tuyeres. This addition allows the formation of a protective layer of
titaium carbide to temporarily form on the hear walls, as long as the titania is
charged into the furnace. Once injection is stopped, the protective layer quickly wears
away. This is an expensive method of heart preservation since the titania bearing ore
is expensive and the furnace coke rate increases since energy is required to release the
titania from the ore(7. However, this operating cost penalty is often accepted because
of the high financial pena1ty that would be incurred if the protection layer were not
induced and the hearh refractory failed. Thus, some operators are forced to add these
titania ores to artificially induce accretions on the linings, as a result of the failure of
the lining "system" to naturally provide the ability to freeze process materials on the
wall hot face.

All carbonaceous products are resistant to chemical attck as long as they are properly
cooled. However, because all hearh wall cooling is dependent upon heat transfer
through the entire wall thickness and then to a stave or furnace shell on the wall cold
face, it is imperative that contact be maintained with the cooling system at all times.
Properly designed and configured hearh staves are unaffected by differential thermal
movements between the furnace shell and refractories. Therefore, staves provide a
more certain cooling contact with the wall, with virtually no separation. Externally
cooled steel shells, especially the sprayed tye, often experience high differential
temperatures and are prone to separation from the refractories. Often, high
conductivity grouting materials must be injected between the shell and wa11 to re-
establish contact with the refractories, thus assuring heat transfer. Otherwise, the
small air gap that forms between the shell and wall wil result in high wall
temperatures and consequently, chemical and hot metal attack wil occur.

When rammed layers are used between the cooling elements and refractories, care
must be taken to insure that ram materials are utilized that exhibit little or no shrinkage
and are installed utilizing the highest densities possible. Preferably, no rams should be
used because a rammed layer is never as good as the refractory material adjacent to it.
This is because the density, porosity and thermal conductivity of the ramming material
wil always be inferior when compared to the carbonaceous refractory product.
Shrinkage of the rammed layer over time wil also result in a loss of heat transfer
capability of the wall, shortening its life. When combined with other problems such as

8-20
block cracking or low thermal conductivity, this combination of problems results in
severe hearh wall deterioration, cutback in an inverted mushroom shape and ultimate
failure.

Summary - Blast Furnace Hearth

Blast furnace hearh design concepts, materials, configuration and wear patterns vary
greatly throughout the world. Hearhs are generally composed of varing grades of
carbonaceous and ceramic materials, zoned to take advantage of the properties of each
grade, to minimize wear.

Historically, severe hearh wall erosion problems are minimal in North America, but
are a major source of downtime and termination of campaigns in Europe and Asia.
Solutions to hearh wall problems must consider all of the external and internal factors
responsible for refractory wear such as operations effects, burden materials, thermal
shock and stress, mechanical stress, differential thermal expansion as well as
traditional mechanisms such as chemical attack and erosion.

Creating the successful hear refractory "system" requires comprehensive analysis.

A variety of design concepts, configurations and materials are available that enable the
blast furnace operator the capability to extend hearh campaign life. Survival depends
upon recognition of and reaction to all of the external factors that can destroy any
refractory system, even one that properly addresses all of the internal factors in its
execution. With the proper combination of operating practices and expertise,
performance monitoring and control and the appropriate lining/cooling system, the
"endless" hearh campaign can truly become a realistic goal.

8-21
 BOSH. BELLY AND STACK

The refractory systems comprising the bosh, belly and stack of the blast furnace
proper probably are the most critical in terms of their impact on the operating
capability of the ironmakng complex. No other refractory systems in the ironmaking
plant have a greater effect on the day-to-day survivability and integrity of their process
containment vesseL. Indeed, if the furnace must shut down because of bosh, belly or
stack lining problems or failures, iron production ceases and the need for any other
refractory lined system in the complex ends.

In North America, this is especially true because of the scarcity of major heart wall
problems. Many other worldwide blast furnaces suffer a myriad of hear wear
problems. These problems are often more critical to campaign termination than those
experienced in the bosh, belly and stack, because there are no easy ways to correct
them once they appear. However, once hear refractory problems are minimized or
eliminated by adopting proper operating practices, raw materials, lining concepts,
configurations and refractories, the bosh, belly and stack becomes the critical system
for attention.

The cooling aspect of the lining/cooling "system" is a most critical factor which can
determine the success or failure of a refractory product. If refractory temperatures rise
above their "critical reaction temperature" for chemical attack, refractory failure and
loss are inevitable(8). Additionally, severe blast furnace gas flow pattern changes can
thermally shock certain types of refractories, even if they are properly cooled. Thus, it
is imperative to consider not only the cooling method, but how to configure a
"system" which incorporates refractory materials that are appropriate for the expected
wear factors. The properties and characteristics of these refractories must work in
combination with the cooling, to achieve the intended performance.

Refractory Materials

Bosh linings are comprised of various conductive refractories such as carbon,

semi graphite and graphites, varing tyes of ceramics or sometimes combinations of
both. Historically, belly and stack linings were comprised of various types of
ceramics. Lately however, conductive refractories such as graphite and semigraphite
have proven superior, used alone for their excellent chemical attack and thermal shock
resistance, or for their cooling capability in combination with various ceramics.

Carbonaceous (Conductive )Refractories

Table I in the preceeding BLAST FURNACE HEARTH Section lists the

classifications of the various types of conductive carbonaceous refractories. It is
diffcult to present material properties of these products, without referring to specific
manufacturer's grade designations. That is because each manufacturer produces
unique products that exhibit unique properties. Tables V and VI however, present a
representative listing of conductive carbonaceous materials that are used' as

8-22
refractories in the bosh, belly and stack and typical properties. The iron content of
carbonaceous refractories is critical because iron catalyzes carbon monoxide
degradation. Therefore, the lower the iron content of the material, the greater the
potential to resist CO degradation.

Ceramic Refractories

The properties and characteristics of all ceramic refractories depend upon the raw
materials utilized and their size consist. The fine paricles in the mix form the ceramic
bonding of the larger paricles as the material is fired at high temperature. The fired
refractory contains larger crystallne paricles bonded together with glass or other
smaller crystalline paricles that have fused together during firing.

Crystals composed of silica or alumina form strong bonds in materials such as fireclay
or high alumina. Glass bonded refractories tend to have good strengt but soften and
deform under load. Additionally, impurities such as iron oxide or lime promote the
formation of glass. Therefore, manufacturers try to limit the amount of impurities in
these types of products.

Table V
GraDhite Material ProDerties

Gra hite Material Descri tion

Standard Standard Medium High
density, density, density, density,
Pro er Std ash Low ash Low ash Low ash

Bulk densi cc 1.63 1.67 1.72 1.80

Porosi , % 21 16 14 12

Cold crushing
Stren h, MPa 20 28 40 51

Thermal conductivity,
W/moK 0) 20°C 150 140 15 160
1000°C 70 70 75 80

Ash, % 0.5 0.2 0.2 0.2

8-23
 TABLE VI

Semigraphite and Semigraphitized Material Properties

Semigraphites Semigraphitied Hot Pressed Semigraphites

High Fired, High Fired, High Fired, High Fired, Graphite Graphite
Conventional, Conventional, Conventional, Conventional, With Silca With Silcon
ProDert Med. Density Hi!!h Densitv Med. Densitv Hi!!h Densitv Addition Carbide Add.
Bulk density, wcc 1.62 1.73 1.65 1.75 1.8 1.87
Porosity, % 19 15 18 15 18 -
Pereabilitv, m Darcvs - - - - 8 0.6
Cold cruhi~ strenir, MPa 25 36 27 38 31 48
Thermal conductivity, W/moK
00 At 2000 C 45 45 50 50 60 60
i
tv At 6000 C 32 32 25 25 40 48
.¡
Ash, % 0.4 0.2 0.4 0.2 9.5* 20**

* Ash content includes quar and silica addition to control alkali attack.
** Ash content includes silicon carbide.
 When designing ceramic lining "systems", significant refractory properties are the
the refractory and its porosity
bulk density, which reflects the heat caring capacity of

or permeability which provides a means to determine the ability to resist penetration

by molten materials and gases. Bulk density also influences thermal conductivity and
the chemical resistance of the brick to such wear mechanisms as alkali, carbon
monoxide degradation and slag or hot metal attack.

Refractory service temperature is an important issue in any ceramic refractory system.

For each chemical attack wear mechanism, there exists a "critical reaction
temperature" for that refractory grade, which defines the point at which chemical
attack commences. If refractories can be cooled below this critical reaction
temperature, chemical attack could be prevented. This is because all chemical attack
mechanisms are thermochemical reactions and as such, the rate of reactions is
temperature dependent. Therefore, the refractory designer must provide a "system"
that provides refractory temperatures consistently below the critical reaction
temperature for each grade of refractory in the system. Cooling enhancement or
highly conductive refractory components can assist with this effort.

Thermal shock resistance is also a critical issue when investigating refractory system
design. Thermal shock or "spallng" is caused by thermal stresses which develop from
uneven rates of expansion and contraction within the refractory, caused by rapid
temperature changes. There are no standard tests for evaluating thermal shock
resistance because shock is also a fuction of size and shape. A qualitative prediction
of the resistance of materials to fracture by thermal shock can be expressed by the
factor:
ks/aE, where:

k = Thermal conductivity
s = Tensile strength
a= Coeffcient of thermal expansion
E = Modulus of elasticity

The higher the value of this factor, the higher the predicted thermal shock resistance of
the materiaL.

Erosion and abrasion are important issues, especially in the top of the furnace and
areas of high gas flow. Erosion of refractory particles results when the bond of the
refractory is destroyed by impact or impingement of process materials or dust-laden
gasses. The high-density materials exhibit higher resistance to abrasion or erosion.

Fireclav

Fireclay refractories consist of hydrated aluminum silicates and minor proportions of

other materials. Examples of fireclay refractories are super duty, high duty or medium
duty. These materials contain between 18 to 50% alumina and 50 to 80% silica,
depending on the grade.

8-25
 Super duty fireclay refractories exhibit good strengt and volume stability and have an
alumina content of approximately 40 to 50%. Often, super duty fireclays can be high
temperature fired to enhance the high temperature strength of the brick, stabilize
volume and prevent damage by carbon deposition. These materials are often used as
economical bosh, belly and stack linings in low production facilities or as a sacrificial
lining material on the hot face of the refractory wall. Tar impregnation can be used to
reduce porosity and permeability to improve resistance to chemical attack.

High duty or medium duty fireclays are normally utilized in areas subjected to
moderate attack mechanisms. In the bosh, belly and stack, they are often used as a
blow-in protection lining and as a low cost sacrificial hot face lining materiaL.

Hi2h Alumina

High alumina refractories are available with alumina contents of 45 to 99+ %. They
are limited to a maximum service temperature of approximately 18000e (3300°F).
They exhibit high refractoriness and chemical resistance at high temperatures. Mullite
and corundum materials are also considered as high alumina refractories. These
materials are often used as a bosh, belly and stack refractory in low to moderate
production facilities or where budgets are limited. They can also be utilzed as the hot
face or cold face lining layers in "sandwich" lining configurations. These materials
also can be tar impregnated to improve permeability and thus improve chemical attack
resistance.

Silcon Carbide

Refractories comprised of silicon carbide are used in the bosh, belly and stack due to
their higher resistance to chemical attack, abrasion and thermal shock as compared to
fireclay or high alumina refractories. Silicon carbide refractories can utilize several
different bond types which change the physical properties of the refractory.

In general, silicon nitride (ShN4) bonded silicon carbide has proven to be preferred
over various direct bonded, self-bonded or carbon silicon bonded materials. Recently,
sialon (SkxAIOxNs-x) bonded silicon carbide (as well as sialon bonded high aluminas)
have been used in the bosh, belly and stack for their improved alkali resistance.

The bonding system used in silicon carbide refractories can be affected by the various
wear mechanisms encountered in the blast furnace. For example, oxidation can be a
problem to the self-bonded or direct bonded silicon carbides, which causes a
"swelling" of the materiaL. For this reason, most ironmakers utilize either silicon
nitride bonded or sialon bonded silicon carbide for the bosh, belly and stack.

8-26
 Silicon carbide refractories, because of their abrasion resistance, can also be utilized in
the stockline area where impact from falling charge materials is severe. Phosphate
bonded high alumina materials also are successfully utilized in this erosion prone
zone.

Ceramic Properties

Historically, many different grades of ceramic refractories have been used in the bosh,
belly and stack with varing degrees of success. Often, capital restraints or the
existing cooling system and its capabilities limit the choice of refractories.
Sometimes, very economical grades of refractories are chosen for their "sacrificial"
use as a blow-in protection lining or for a stave cooling system that is intended to
operate "naked", that is, without a refractory lining. Sometimes, budgetar limitations
preclude the use of exotic ceramics or silicon carbide refractories that could improve
performance. However, there are a large and varied group of ceramic refractories
available, at varing price levels, to achieve the intended lifetime goals. A
representative listing of some of these materials' properties are summarized in Table
VII.

TABLE VII

Representative Bosh. Bellv and Stack Ceramic Materials

Superduty High Alumina High Alumina Silcon Carbide Silcon Carbide

Property ("'48%Ali03) (",60%Ali03) ( ",90%Ali03) (SbN4 Bonded) (Sialon Bonded)

Density, wcc 2.4 2.55 2.95 2.58 2.72

Crushig strengt,
Mpa , 60 80 120 140 180
Porosity, % 11 15 15 15 14
Thermal conductivity,
W/moK (t 1000°C 1.7 1.9 2.9 13 12

The many permutations possible from the range of different materials available today
offer the designer challenging opportunities for optimizing the lining design. The key
of course, is to recognize that the success of the "system" wil depend upon how the
i I
!
many internal and external factors which affect the refractory system are addressed.
Additionally, lifetime improvements can also occur if various grades and types of
refractories are combined in one system to take advantage the best properties or
characteristics of each of the products used.

8-27
 Refractory Performance Factors - External and Internal

Bosh, belly and stack refractory performance is a function of many factors, most of
which are determined by the presence of chemicals and high temperatures in the blast
furnace process. A major contributing factor is the location, direction and velocity of
the counter-current gas flow as it permeates through the burden materials. Also
contributing are the abrasion affects of the descending burden materials and
ascending, dust laden gasses. Review of historical furnace refractory "wear lines" wil
show evidence that the areas that experience the most intense gas flow patterns and
high heat load are the bosh, belly and lower stack.

The location of this severe "wear" zone can var slightly from furnace to furnace and
is dependent upon many operating variables. These variables can include type and
quality of the burden materials, burden distribution capability, quality and quantity of
tuyere injected fuels, quantity of wind blown, furnace availability and many other
operations related factors.

The purpose of this discussion is to review all of the external and internal factors
which can affect performance with the intention of optimizing refractory life. This
wil permit the configuration of an initial lining with the best chance of survival and
thus wil postpone inevitable repair until very late into the furnace campaign.

External Wear Factors

Blast furnace operations can destroy any refractory system, even one that is comprised
of the most appropriate refractory material for the application. The intense chemical
reactions that occur in the blast furnace, coupled with high velocity, high temperature,
dust laden gasses often impinging directly against refractories, results in relentless
attack.

If the geometry of the lining, the so called "furnace lines" or the refractory
configuration are incorrect for the application, refractory loss wil be hastened. In
particular, the cooling system type and effciency is a most critical external factor
which can determine the success or failure of any refractory product. Some experts
have even claimed that "cooling water is the best refractory". However, it should be
recognized that cooling water removes heat energy from the blast furnace process and
that properly engineered lining/cooling systems provide a way to optimize lining
performance and minimize wall heat losses over the full campaign.

The wear mechanisms encountered in the furnace vary by type and intensity, by zone.
Recognizing which mechanisms of wear wil be encountered and gauging their
intensity allows the designer to configure a lining system best able to resist the attack
mechanisms for the longest time. Each of these external factors must be considered
individually to properly create a successful refractory system.

8-28
 External Factors: ODerations Effects

The goal of any blast furnace operator is to make a good profit for the owner. Each
blast furnace wil have productivity, fuel rate and performance goals unique to the
owner's paricular situation. This means that in some cases, furnaces must be operated
at intense levels for maximum production, even at the expense of fuel rate or
refractory life. Others must try to achieve a balance of high productivity with a
minimum fuel rate to keep costs low. Stil others must operate their furnaces
conservatively at moderate production rates, especially if total ironmaking capacity is
out of balance with steelmaking capacity. Others are limited by raw materials tye or
quality or physical limitations such as blowing capacity. Stil others are able to
achieve very high production rates at minimum fuel consumption and high effciency
because of the type and high quality of the burden materials and modem physical
plant.

It is because of this broad variation in furnace performance and capabilities, which are
functions of external influences unique to each furnace owner, that it is impossible to
"standardize" refractory configurations. What might work for one furnace might well
be a total failure when applied to a similar furnace operating with different variables.

Productivity

High productivity intensifies the destructive factors which affect lining life. Wind
rates are high which results in high process gas volume and velocity. High wind rates
require high tonnages of charged raw materials, increasing abrasion. High
productivity also intensifies heat loads on refractory walls and can often intensify
variations in wall heat loads resulting in severe temperature "peaks".

Conversely, low or moderate productivity can result in minimal wall gas flow, smaller
temperature "peaks" and low wall heat load. Thus, these conditions would be more
friendly to refractory life and may permit a much different lining configuration and
quality than that required for a high productivity furnace.

, \
Injected Fuels
!

The type and quality of tuyere injected fuels can also play an important role in
refractory life. This is especially true if high rates of oxygen are injected. The
chemical reactions from tuyere injected fuels can be endothermic (where heat is
absorbed) or exothermic (where heat is released). Operators try to control raceway
flame temperatures, depending upon whether endothermic reactions or exothermic
reactions wil occur. This can result in different raceway conditions and consequential
gas flow pattern differences.

8-29
 Burden Distribution

Burden materials are deposited in predetermined, premeasured layers. The coke layers
are configured to allow gas flow through the mass, because they are reasonably
permeable compared to the ore component of the charge. This is graphically depicted
in Figure 7. There is a "science" of burden distribution techniques utilizing a variety
of furnace charging hardware which permits the operator to control gas flow patterns.
This is done by changing the ore-to-coke ratio at various locations across the burden
surface and by physical placement of ore or coke at specific locations to either retard
or enhance gas flow. Thus, it is possible to control productivity, fuel rate and wall gas
flow to achieve the goals established for maximum profitability. Conversely, if
burden distribution capabilities and techniques are unavailable or minimal, wall gas
flow, heat load and temperatures would be erratic. This would adversely affect long
term refractory performance.

Burden Materials

High quality burden materials especially coke, can improve furnace performance and
effciency and thus have a positive influence on refractory life. This is especially
important in the lower stack, belly and bosh where the combusting coke must allow
unimpeded gas flow, and yet have suffcient strengt to support the great weight of the
burden column above. It has been proven time and again that poor coke quality wil
adversely affect furnace permeability and thus performance, with consequential lining
life penalties.

A vailabiltv

Blast furnaces pedorm best when they are operated continuously with a minimum of
stoppages or disruptions to driving rate. Every reduction of wind volume, material
charge delay, casting delay or maintenance stop disrupts "smooth" operation and gas
flow. Thus, furnaces which are plagued with maintenance problems which force
shutdowns or productivity reductions, wil be paricularly hard on refractories.
Conversely, blast furnaces which operate smoothly and at a reasonably constant
production rate with a minimum of shutdowns and delays, wil be easier on
refractories.

These operations effects can result in changes to the shape and location of the
"cohesive zone" where the liquid metal droplets form and changes in the intensity and
direction of the hot process gasses, as they pass through the permeable coke "slits"
that are layered in the burden mass. The shape of the burden profie can vary from a
"Y" to a "W", depending upon burden distribution capability and practice. The "Y"
shaped profile results in more central gas flow and reduced wall working, and the "W"
shaped profie results in less central flow and more wall working.

8-30
There are many operating philosophies regarding the "optimum" burden distribution,
gas flow patterns and cohesive zone shape and location that maximize productivity
and minimize fuel rate. However, many of these practices are paricularly severe on
refractory survivaL. Many operators are also familiar with burdening practices that can
minimize wall working and protect refractories. The problem in this age of
competitiveness is that most producers must optimize production and minimize fuel
rate at the expense of the refractories. This compromise results in shorter lining life
and requires periodic furnace stoppages to repair linings by the injection of grouting
materials or by the application of sprayed-on, "gunned" refractory, both of which self-
set in the furnace to form a "consumable" lining. However, this practice requires
periodic re-application to be effective. The goal should be to provide an initial
refractory system which can optimize life and postpone repair.

External Factors: Geometrv Effects

Ifthe physical geometry or so called "furnace lines" are improperly configured, severe
refractory wear can result. This is especially true in the bosh where raceway action
can result in impingement of high velocity gasses and entrained paricles on the
refractory wall. An example of this is shown in Figure 8. Proper geometry wil
consider expected gas flow patterns and directions as well as historical performance to
eliminate potential problems. Even the most appropriate refractory for the application
cannot survive the relentless and never ending actions of the raceway if the geometry
results in impingement. Furnace designers the world over can provide the proper
relationship required when determining furnace lines (geometry) for a new or rebuilt
vesseL. However, the problem most users must address is how to incorporate proper
geometry into existing facilities if there is a shortage of capital. However, no
refractories can correct the problem of bad geometry.

External Factors: Confiiwration Effects

The actual refractory configuration can affect refractory life. For example, a
refractory wall cooled on its cold face such as with sprays or staves, can have a very
high hot face temperature ifthe wall is configured excessively thick. Accretion (skull)
formation could prove diffcult or the resulting accretion might be very thin. This is
because the heat must travel completely from hot face to cold face and if the thermal
resistance of the wall is very high (thick wall, low thermal conductivity), it would be
diffcult to remove the heat fast enough to prevent a high hot face temperature. Figure
9 depicts a comparison of hot face temperature and skull thickness for a thin versus
thick wall configuration, utilizing the same refractory materiaL. You could improve
the situation by increasing the thermal conductivity of the wall or reducing the wall
thickness (or both) to lower the thermal resistance and thus lower wall hot face
temperature. Another way to decrease the thermal resistance of a thicker wall is to
utilize a composite construction of a very high conductivity cold face material to
decrease the thermal resistance of the entire wall. This is depicted in Figure 10. The
object is to obtain wall hot face temperatures that are low enough to condense vapors
and form thick protective accretions of solidified process materials. -

8-31
 External Factors: Coolin!! Considerations

Cooling system capability is affected by the type of cooling employed. Typically,

bosh, belly and stack refractories are either cooled externally on their cold face or
internally cooled by a multiplicity of copper elements installed in rows, which are
inserted radially within the wall. Figure 1 1 ilustrates arangements of the various
cooling system types used in the bosh, belly and stack.

Coolin!! Tv De - External Water Film (SDrav Coolin!!)

External cooling is accomplished by several methods, each of which has good and bad
features. The earliest application of external cooling is merely the introduction of a
series of water nozzles, aranged circumferentially around the furnace jacket, which
provide cascading water film on the jacket surface. The water is then collected in a
trough at the bottom of the jacket. This arangement is often called "spray" cooling,
but in fact is really "fim cooling" since a thin fim of water actually performs the heat
removal. The advantages are simplicity, low cost, efficient heat removal and the
pressure containing jacket remains visible for easy "hot spot" or crack detection.
Disadvantages are that the open water collection system is easily contaminated by dust
and debris, water flow can be disrupted or inadvertently stopped by obstructions in the
water flow path and instrumentation or other vessel connections disrupt water flow
and are hard to install during furnace operation. External cooling also results in
adding thermal stresses to the pressure containing jacket and can result in differential
thermal expansion between shell and refractories, causing a loss of cooling contact
with the refractories.

Coolin!! TVDe - External Panels

Another form of external shell cooling is the use of water containment jackets,
chanels, angles or other steel weldments to form water flow passages on the shell
cold face. The advantage is the ability to totally close the water system to prevent
contamination by dust or debris. The disadvantages of this type of jacket cooling or
panel cooling as it is called are many. First, it is very difficult to arange the flow
passages to achieve high water velocity throughout the flow path. There can be areas
of low velocity or eddy currents which impede heat removaL. Additionally, if
untreated river or lake water is utilized, organic, mineral and sediment build-ups can
insulate the water from the jacket, interrpting heat transfer. Another problem is that
the panels completely hide the sudace of the pressure containing vessel, which can
prevent the discovery of shell hot spots or cracks until damage occurs. This type of
jacket.
cooling also adds thermal stresses to the pressure containing

8-32
Coolin!! TVDe - Staves

A third type of external cooling utilizes cast iron or copper cooling elements called
staves. Water flow passages are integrated into the cast iron elements using steel
pipes, or in the copper elements by using integrated pipes or machined or cast flow
passages. The advantage of stave cooling is primarily the ability to cool the
refractories from within the pressure vessel, thus eliminating thermal stress of the
jacket. Additionally, the jacket is totally exposed for inspection purposes and since
the cooling system intercepts heat before it reaches the pressure containing jacket,
thermal stresses in the jacket are low. The disadvantages of staves are their high cost,
inability to be easily changed in case of wear or damage and they generally require
chemically treated water to prevent mineral build-ups and maintain effectiveness.

Cast iron staves also require various types of "insert" materials, installed within
horizontal grooves located on the stave hot face. The iron stave "ribs" which form
these grooves, contain and support the insert materiaL. These insert materials can be
conductive refractories for best heat removal capability, or be insulating materials in
case the stave is intended to operate "naked", that is without a refractory lining in front
of the stave.

All external cooling systems are sensitive to any interrption of the heat flow path
from the hot face of the refractory to the water. Any degradation, disruption or loss of
contact in this heat flow path wil adversely affect refractory temperature and hasten
degradation. These factors can include poor water quality and low velocity, which
result in corrosion, mineral build-ups, sediment deposits and organic build-ups on the
cooling element. Once these deposits form, they provide an insulating layer between
the cooled surface and the water, preventing heat removaL. Consequently, refractory
cooling is adversely affected and chemical attack results.

Another serious potential problem with any externally cooled refractory system is that
separation or loss of cooling contact can occur from differential movements. Once
refractory cold face contact is lost, the resulting "air gap" provides a very effective
barier to heat transfer, causing a disruption of cooling and high refractory
temperatures. All externally cooled refractory systems should be equipped with
provisions for injecting conductive grouts between the refractory cold face and the
cooling surface. Periodic injections of this grout can fill these air gaps and re-establish
heat transfer, thus improving refractory pedormance.

Coolin!! TVDe - Inserted Coolers

Another type of cooling system utilized in the bosh, belly and stack is the use of
inserted cooling elements. Most often, especially in high heat load zones of the
furnace, these inserted elements are comprised of cast or forged copper. Water flow
paths are formed by cast-in-place or machined passages, tubes or combinations of
these methods in the same element.

8-33
 The elements are generally arranged in rows, radially inserted into the lining and
configured such that the number of elements per row and the row spacing are
sufficient to provide the required cooling effect on the refractories. The "density" of
these cooling elements, that is, the number of elements per row and the spacing of the
rows can be such that intense refractory cooling can be provided. Generally, the hot
face of the cooling element (the nose) should be located at or very close to the
refractory hot face.

Heat transfer to the cooling elements is effected through the use of a conductive
rammed anulus between the cooler and the refractories or by intimate contact with
the refractories. Care should be taken to carefully configure this critical detail to
minimize interrption of the heat flow path.

The advantages of inserted cooler elements are primarily the ability to cool the
refractory wall internally from hot face to cold face and the ability to change
individual cooling elements if they are damaged in service. They also provide
physical support of the refractory mass, which is very important for refractory wall
integrity.

For optimum pedormance, the connection where the cooling elements penetrate the
pressure vessel must be gas-tight. This can be a problem on older furnaces where
capital constraints prevent gas-tight connections from being incorporated. This can
result in "gas tracking", consequential loss of heat transfer capability, high shell
temperatures and of course, a safety hazard to personneL.

The disadvantages of the inserted cooling elements are mainly related to economic
considerations. In order to incorporate a modem, densely spaced cooler pattern on an
existing furnace, a new steel jacket may be required. Additionally, water quality and
cooling element design are important considerations that should be incorporated for
optimum pedormance. However, inserted cooling elements offer the highest cooling
effciency, longest life potential and easiest maintenance capabilty of all of the
available cooling systems.

Often, bosh, belly and stack linings are cooled with both internal and external cooling
types, especially the combination of staves with inserted coolers. This can be done by
zone, such that the bosh might be stave cooled and the belly and stack cooled by
inserted plates. Another concept utilizes stave coolers between rows of inserted
coolers, when the existing cooler rows are too far apart for proper refractory cooling.
The object of any cooling system configuration is to provide the desired cooling
effect, while recognizing the strengts and weakesses of each, and making
allowances for them in the refractory configuration.

8-34
External Factors: Wear Mechanisms

The severity of attack by the mechanisms of wear in the bosh, belly and stack can be
different from furnace to furnace even in the same plant, due to variations in furnace
geometry, burden materials and distribution and furnace operation. The lower zones
of the blast furnace, from the bosh to the mid stack are most affected by thermal
shock, high head load and chemical attack. These zones are the real "trouble areas" on
virtually all blast furnaces and are most responsible for termination of furnace
campaigns or lengthy repair downtime.

From the upper middle stack to the stockline, mechanical wear and impact from
charging become the main contributors of wear, along with chemical attack.

Thermal attack includes exposure to high temperatures over time, severe temperature
fluctuations and fatigue. Chemical attack includes attack by alkali vapor and
condensate, carbon monoxide degradation (carbon deposition), oxidation and attack by
slag or molten metal. Mechanical wear includes erosion from ascending dust laden
gases, abrasive wear of descending burden materials and impact loads from falling
burden materials.

A summar of these wear mechanisms, by severity and by furnace zone, is listed in

Table VIII and is graphically portrayed in Figure 12.

Wear Mechanisms - Thermal Shock

It is universally agreed that the predominantly pellet charged, typical North American
blast furnace wil be subject to more intense high temperature fluctuations at the wall
than experienced by predominantly sinter charged, European and Japanese blast
furnaces.

It has been demonstrated by Hoogovens, an ironmaker in the Netherlands, that these

temperature fluctuations increase dramatically as the pellet charge exceeds
approximately 15 to 20% of the total metallics charged. Actual temperature peaks
experienced by a 50% pellet - 50% sinter charged furnace, have been shown to be
typically up to 1000°C (1850°F) over a 6 to 7 minute period, or approximately 150°C
(300°F) per minute temperature change. However, the predominately sinter charged
furnaces consistently experience wall temperature fluctuations of only approximately
40°C (100°F) over the same 6 to 7 minute period, or approximately 7°C (20°F) per
minute temperature change (9).

8-35
 TABLE VIII

Comparison of Furnace Wear Mechanism Severity by Zone

Severity of Attack

Wear Mechanism Bosh Stockline

Thennal Shock Moderate - Hi Low
Heat Load Moderate - Hi Low - Moderate
Slag Attck Extreme None
Alali/Zinc Attack Hi Low - Moderate
Oxidation Low - Moderate Low
Abrasion Low - Moderate Extreme
00
I
I.
0\
This means that whatever refractory is chosen, it can experience exposure to
temperature changes of up to 150°C (300°F) per minute if pellets are charged and only
approximately 7°C (20°F) per minute if predominantly sinter is charged. Typical North
American operation utilizing predominantly pellet burdens, thus exposes wall
refractories to the more severe temperature fluctuations due to gas flow changes in the
pellet burdens.

It has been demonstrated that all ceramic refractories, including silicon carbide
materials, wil spall and thermally crack if they experience temperature fluctuations of
this severe magnitude. Critical spalling rates have been discussed in several technical
papers by Hoogovens of the Netherlands (9). A list of typical critical spallng rates for a
variety of materials is shown in Table ix.

TABLE IX

Critical Spallng Rates for Various Materials(9)

Material °C/Min. OF IMin.

High Duty 4 7
High Alumina 5 9
Chrome Corundum 5 9
Cast Iron 50 90
Silicon Carbide 50 90
Carbon 200 400
Semigraphite 250 450
Graphite 500 900

These critical spalling rates define the maximum temperature variations (heating or
cooling) that the hot face of the refractory materials can survive without cracking.
Beyond these rates, cracking and spalling wil occur. As can be seen from the table, the
only materials which can withstand the normally occurring 150°C (300°F) per minute
temperature excursions of a typical pellet charged furnace are carbonaceous materials,
the so called "conductive refractories".

The thermal shock failure effect is most severe when refractories are cooled from one
side, like with staves or externally cooled jackets. This is because thermal shock cracks
occur parallel to the refractory hot face, which result in three problems. First, these
cracks permit the alkali vapors and condensate to be exposed to a greater refractory
surface area including the interior of the refractories, hastening chemical attack.
Second, because these cracks occur parallel to the hot face, air gaps form which
interrpt heat transfer to the cooling system, thus increasing refractory hot face

8-37
 temperature. Consequently, since chemical attack is temperature dependent, this
increase in refractory temperature at the hot face wil assure that the hot face is
chemically attacked as temperatures exceed approximately 600°C (1100°F) for high
aluminas and fireclays and 800°C (1500°F) for silicon carbides(8).

Third, as' accretions (skulls) form and fall off as wall temperatures increase, the fallng
skulls pull away this cracked layer of refractory which adheres to the skull, thus again
exposing a new hot face to be thermally shocked, repeating the cycle. This sequence is
graphically depicted in Figure 13.

As the thermal shock/chemical attack/scab pull-out of material cycle repeats itself over
time, refractory lining thickness is reduced continuously until the stave or furnace jacket
is completely exposed to the furnace environment. In the case of the cast iron stave
cooled wall, exposing the staves to the same temperature fluctuations as the refractory
before it wil cause cracking and spalling of the cast iron surface, shortening stave life
dramatically. Wall temperatures can sometimes by controlled by burden distribution
and charging techniques. However, these measures usually result in production and fuel
rate penalties, which may prove unacceptable to plant goals.

Virtally all refractory/cooling system design improvements have historically

concentrated on finding refractories which were resistant to chemical attack. The
effects of thermal shock were either unkown or ignored, until failures or minimal
lifetime improvements were experienced, even with effciently cooled silicon carbon
linings. The Japanese and Europeans in paricular began to study the thermal shock
phenomenon in detail and many technical papers have been published on the subject.
What was leared is that it was not enough to have a chemically resistant lining when
severe thermal shock wil be experienced.

In stave cooled or other externally cooled boshes, this is especially important because
the continuous vast expanse of hot face refractory sudace is exposed to many
temperature differentials over this surface. This can result in severe localized spalling
and subsequent loss of refractory support. Once support is lost, entire "panels" of
the wall unit and is the
refractory can fall out in "sheets". This destroys the integrity of

reason many stave designers include refractory support "shelves" integral to the stave,
to support the refractories at various levels. The inserted copper cooler system also
provides this valuable support function.

Wear Mechanisms - Chemical Attack

The chemical attack mechanisms in the bosh and stack are identified as oxidation,
carbon deposition, alkali, slag and hot metal attack. Oxidation can occur by steam
formed from burden moisture, hot blast moisture or leaking coolers. Oxidation can also
occur from carbon dioxide formation, leaking outside air during backdrafting or from a
"lazy" raceway which is too close to the refractory hot face.

8-38
Carbon deposition can occur especially when iron is present in the refractories, which
breaks down the CO to CO2 and C. The carbon builds up within the refractory, causing
cracking.

Alkalies, most notably potassium and sodium, attack the refractory by destroying the
bonding mechanisms which hold the refractories together. This attack causes refractory
swelling and cracking.

As was previously discussed, the temperature at which a paricular refractory is attacked

by a paricular mechanism is called its "critical reaction temperature". Critical reaction
temperatures can be different for each refractory type and are different for each attack
mechanism. Many factors can affect the actual value of a critical reaction temperature
for a paricular refractory, such as the presence of a tramp element or contaminant
which catalyzes the chemical reaction. In general, Table X lists the typically
recognized critical reaction temperatures for various attck mechanisms and refractory
types.

TABLE X

Critical Reaction Temperatures

Attack Alumina/ Silcon Hot Pressed Low Iron

Mechanism Fireclay Carbide Semigraphite Graphite

Alkali 590°C 870°C ~900°C ~900°C

Oxidation None 800°C 400°C 500°C

CO 480°C 600°C 450°C 650°C

As was previously discussed, all of these chemical reactions are temperature dependent
reactions. This means that if the refractory can be maintained at a temperature which is
below the "critical reaction temperature" for chemical attack of that refractory, the
chemical reactions canot occur. One of the diffculties of trying to maintain a low hot
face temperature of a stave or other outside cooled refractory wall is that all heat must
travel through the wall to the cooling medium. Any interrption of the heat transfer
such as an air gap between the stave or brick due to differential growth or a stress crack
parallel to the refractory hot face, assures that the refractory wil be chemically attacked
because it cannot be cooled below its critical reaction temperature.

8-39
 Additionally, in the event the refractory is effectively cooled, the formation of
accretions is accelerated. However, as furnace temperatures fluctuate, these scabs fall
off exposing cool refractory to the hot gases, thermally shocking them and cracking the
refra:cory as previously described.

The important point to consider is that if the refractory configuration is able to be

cooled below its critical reaction temperature, the material chosen must be compatible
with the cooling capabilities. In the case of externally cooled refractories, it must be
recognized that periodically, protective accretions wil fall off the refractory hot face.
Consequently, if insulating ceramic refractories are thus exposed, the cooling effect wil
be slow and thermal shock effects wil occur to the refractory hot face. Therefore, the
refractory lifetime wil be shortened, unless shock resistant refractories are utilized.

It is also important to provide a refractory configuration that is also compatible with the
refractory materials that are considered. Refractory walls cooled from one side such as
with staves wil be diffcult to maintain below their critical reaction temperature, if they
are configured excessively thick or if material conductivity is too low. The key to
success is to configure the refractories to withstand the expected wear mechanisms and
to select the best available materials to do the job. This sometimes requires composite
linings of two or more materials.

Finite element computer modeling can be utilized to locate critical reaction temperature
isotherms and identify zones of potential chemical attack in the refractory. Figure 14
shows examples of an externally (stave) cooled lining and an inserted plate cooled
lining. Two critical reaction temperature isotherms are located in each example. The
590°C (11000P) isotherms define the star of alkali attack of alumina and the 870°C
(1600°F) isotherms define the star of alkali attack of silicon carbide.

If the linings were alumina, all of the material from the 590°C (1100°F) isotherm to the
hot face wil be chemically attacked. The situation could be improved by either
intensifying the cooling (which is impossible when the lining is cooled from its cold
face) or by choosing a material that exhibits a higher critical reaction temperature, in
this case, 870°C (1600°F) for silicon carbide. Thus, potential refractory loss from
chemical !attack would be reduced as shown in the figure.

Wear Mechanisms - Abrasion

Mechanical abrasion and erosion also contribute to bosh, belly and stack wear but at a
much smaller magnitude than thermal shock or chemical attack in the lower zones.
Most abrasion in these zones is the result of dust laden ascending gasses and descending
burden materials. As was previously discussed, furnace operations and geometry can
greatly affect erosion of refractories. If burden distribution results in excessive wall gas
flow or if furnace wind is often reduced so that tuyere velocity is low or if the furnace is
"fanned" for extended periods or the bosh is allowed to "flood" due to improper casting,
severe wall working can contribute to excessive wall wear. Also, if furnace geometry_is
not appropriate for the intended productivity and operations intensity, long..teim

8-40
impingement effects wil destroy any refractory, even the most appropriate type for the
application. The important point is that erosion or abrasion effects canot be stopped
only by refractory properties.

Internal Wear Factors

Afer review and consideration of the external factors which affect refractory
performance, a review of those wear factors internal to the refractory system must be
conducted. These internal factors can be critical in determining the success or failure of
a bosh, belly and stack refractory system. Even the most appropriate refractory for the
application wil fail if these factors are ignored or improperly taken into account. It
should be remembered that merely selecting appropriate refractory properties is not
enough to assure survival or long life.

Internal Factors - Accommodatin2 Expansion

One of the most critical internal factors in any refractory system is to assure that the
arangement and configuration of the individual components allows for thermal
expansion without damage as temperatures increase. This means not only must
refractory thermal expansion provisions be included, but an examination of the effects
on the pressure containing vessel must also be conducted. Failure to properly allow for
thermal expansion compensation can result in destructive cracking of refractories and
the vessel, deformation of the vessel or the lining and premature failure. The use of
heat setting cement, installed in joints with suffcient thickness to compensate for the
expected movements, is one way to provide for thermal expansion. Another is to utilize
compressible layers of refractory fibers or layers of organic materials that wil bum
away to compensate for expected movements. This is especially important when the
refractory configuration wil encounter abrupt changes in diameter or shape and at
nozzle projections.

Internal Factors - Accommodatin2 Differential Movements

It is also important to recognize situations which wil result in differential thermal

movements in a refractory system. These can be caused by utilizing refractory
materials with different coeffcients of thermal expansion in the same lining thickness.
These differential movements can also result from configurations with excessively thick
walls with high hot face and low cold face temperatures. This high temperature
differential across the wall can cause cracking if the wall if comprised of a one-piece
material thickness. Accommodating differential movements is thus mandatory to
prevent cracking and displacement of the refractory components, which interrpts
cooling.

8-41
 Internal Factors - Accommodatim! Stresses

The successful refractory system wil properly compensate for the expected thermal
expansion of components. It wil also consider all expected differential thermal
expansion from all sources and provide required movement compensation. If properly
done, accommodating these movements wil result in a refractory system free of
damaging mechanical and thermal stresses. These stresses can result in "pinch"
spalling ofthe refractory hot face, caused by two adjacent components squeezing tightly
together until the contact surfaces literally explode, displacing large hot face pieces of
the refractories. Insuffcient compensation can also result in highly stressed refractories
that are restrained by rigid structural weldments such as a furnace mantle (lintel) ring
girder. The successful refractory system wil minimize stresses on the refractory
components, which prevents cracking and consequential loss of heat transfer capability.

Internal Factors - Effective Heat Transfer

Proper refractory system configuration requires that the heat transfer path from the hot
face to the water be as effcient as possible. This means that the more direct is the
contact between components, the more effective the heat transfer path wil be.
However, practical considerations require compromise to this direct path philosophy.
The object is to minimize bariers to effective heat transfer. Thermal resistance should
be optimized by utilizing highly conductive materials, eliminating or minimizing
rammed anuli, preventing cracks and the resulting "air gaps" and by taking steps to
periodically grout the gaps which occur at externally cooled refractory contact surfaces.
Another important factor is to correct sources of cooling system ineffectiveness from
mineral or organic deposits and sediment build-ups, as well as separation from
refractory contact. As cooling effectiveness deteriorates, refractory temperatures
increase, intensifying chemical attack and preventing the formation of protective
accretions.

Internal Factors - Refractorv Properties

Afer considering all of the external and internal factors which affect refractory system
pedormance, the last internal factor that must be considered is the refractory itself. In
some ways, refractory properties are the least important of all the factors considered so
far. As was mentioned several times previously, even the most appropriate refractory
for the application wil fail if the other external and internal factors are not properly
accommodated. You can't overcome with refractory properties, the effects of poor
operations, poor quality burden materials, improper geometry or configuration, poor
cooling, lack of thermal expansion provisions, high thermal and mechanical stresses and
bariers to effective heat transfer. However, you can stil have failure even if you do
properly consider all of these other external and internal factors, if you choose an
inappropriate refractory material for the application.

8-42
The important point is that the refractory material chosen be appropriate for the
expected wear mechanisms. They must also be compatible with the type and effciency
of the cooling system employed. And most importantly, the properties of the materials
selected must be compatible with the refractory configuration being considered. No
refractory properties can overcome the effects of excessive wall thickness when cooled
from one side or resist continual exposure to impingement by high temperature gasses
with entrained solids. However, proper selection of refractory type, possessing the
characteristics and properties desired, wil provide the best opportnity for achieving
the intended service life.

Desie:n Considerations

Bosh and lower stack wear is a combination of factors, primarily of thermal shock
induced cracking, which accelerates chemical attack by exposing more refractory
surface area to alkali and by increasing refractory temperature by interrpting heat
transfer.

Upper stack wear is a combination of different factors, primarly chemical attack and
abrasion, especially at the stockline working zone. Thus, when analyzing the zones to
determine suitable refractory materials, often the best potential for success will be to
combine several different grades or types of refractories in each "system". Thus, the
best properties or characteristics from each type used wil contribute to the overall
success of the system.

For example, in a stave cooled system, the refractory "inserts" in the stave face can
utilize highly conductive semigraphite or graphite to optimize cooling effciency. This
permits one hundred percent of the stave face to effciently cool the refractories, thus
lowering their temperature and consequently lowering the rate of chemical attack.
Insulating tyes of refractory stave inserts reduce the stave's ability to remove heat by
limiting the heat pick-up area to the exposed cast iron rib sudaces only. This results in
higher refractory temperatures and increased chemical attack.

Another case would be the use of a refractory "sandwich" consisting of three different
grades of refractory in the same wall thickness. For example, a silicon carbide layer of
refractories could be "sandwiched" between a cold face lining of lower cost high
alumina or highly conductive semigraphite and an economical fireclay on the hot face
to absorb blow-in and initial thermal shock damage.
'the rigors of

Another example would be to utilize lintel blocks of highly conductive graphite or

semigraphite to form the bridged opening for copper cooling plates or to form "passive"
cooling bands or rings to enhance the cooling effect of widely spaced cooling plates.

The possibilities are endless but the important point is to remember to consider all of
the important internal and external factors that wil affect the "system" pedormance
such as expansion provisions, differential movement, mechanical stresses, integration
with the cooling elements and analysis of the wear mechanisms to be encountered.' -

8-43
 Bosh ADDlIcations of Conductive Refractories

The application of carbonaceous materials as "conductive" blast furnace refractories for

the bosh and lower stack are many and varied. The most common usage of carbon and
semigraphite has been as a bosh lining, primarily cooled on its cold face by staves or by
external shell cooling such as sprays or enclosed panel type cooling. External cooling
was originally preferred to eliminate any possibility of water leaks that would result in
oxidation of the carbonaceous refractories.

These traditional arangements share a common requirement for satisfactory

performance. That is, cooling effectiveness is totally dependent upon good surface
contact between the refractory cold face and the stave or steel shelL. Some users prefer
to utilize a high conductivity ram between the refractory and the shelL. However, a ram
material wil always possess lower thermal conductivity, lower density and higher
porosity than the refractory materials and thus results in a weak point in the heat
transfer capability. Therefore, it is usually best to design these lining systems so that
minimal or better yet, no rammed joints are utilized. Instead, heat curing cement or
expansion joints can be utilized to accommodate the thermal expansion of the lining,
while maintaining a tight fit between brick and shell or stave, which is coated with a
layer of high conductivity, heat setting cement. However, even with this arangement,
it is often desirable to inject a high conductivity carbonaceous grout between the
refractory and the shell as the furnace campaign progresses, to fill-in any gaps which
may develop due to differential thermal expansion between the shell and the lining or
from localized shell heating.

Conventionally baked carbon blocks or hot pressed carbon bricks are usually used for
low cost linings of this type. The benefits include good thermal shock resistance and in
the case of hot pressed carbon, excellent resistance to alkali attack. These materials can
readily promote an accretion of solidified slag and iron because of their good thermal
conductivity and thus achieve reasonable life at minimum capital cost.

Conventionally baked semigraphite, semigraphitized blocks or hot pressed semigraphite

brick are also utilized as improved materials in linings of this type. They offer even
higher thermal conductivity, resistance to thermal shock and in the case of hot pressed
semi graphite, excellent resistance to alkali attack. Various additions such as silicon
carbide can also be incorporated to increase abrasion resistance and lower permeability
and thus improve resistance to chemical attack.

The main drawbacks of these lining/cooling configurations are the total dependency on
good contact with the shell or the stave to maintain cooling and the lack of periodic
refractory support along the bosh height to prevent the loss of refractories above, if a
localized failure occurs. These drawbacks can and often do result in premature loss of
refractory due to insuffcient cooling or sudden loss of entire sections of wall due to loss

goals. -
of wall integrity in a small, localized wear area. However, many furnaces worldwide
have had excellent success with bosh linings of this type for moderate campaign life

8-44
Cast iron stave users are especially concerned with the refractory/stave interface.
Water cooled shelves or even inserted cooling plates are utilized to provide physical
support of the refractories at various levels in the lining. Thus, a localized loss of
material below such supported linings would not cause the collapse of the lining above,
as would be the case without the supports.

Another concept rapidly finding favor is to attach lining materials directly to the stave,
either by casting them in place or by the use of special cements. This arangement
allows the simultaneous installation of stave and lining in prebricked assemblies.

These configurations also permit improved cooling effciency when conductive

refractory stave inserts are utilized with either a cold face conductive refractory lining
combined with a hot face layer of silicon carbide or an all conductive refractory lining.
These conductive materials permit low hot face temperatures for good skull formation
and excellent thermal shock and chemical attack resistance.

Another improved bosh design utilizes semi graphite and/or graphite linings, sometimes
combined with silicon carbide, in combination with densely spaced, copper cooling
plates, which offer solutions to the weakesses of the traditional configurations
described previously. This improved design concept provides intensified cooling of the
very high conductivity linings throughout their wall thickness, while providing the
critical physical support of the wall. Thus, the cooling of the wall is no longer
dependent upon tenuous contact of the vast expanse of shell or staves with the equally
vast surface area of the refractory cold face. Instead, individual "fingers" of copper
coolers penetrate into the wall in a densely spaced pattern, thus maintaining lower
overall refractory temperatures.

The heat removal capabilities of such systems are dependent upon the contact of the
lining material with the copper coolers. Two methods are used to maintain contact with
the linings. The most commonly used method is to provide an anulus between the
refractory and the cooler, which is filled with a high thermal conductivity ramming
material. However, another proprietar design has also been adopted, utilizing
machined refractories in contact with machined copper coolers, which provides intimate
contact between cooler and refractory for good heat transfer. This method eliminates
the possibility that a poor ramming job wil adversely affect heat transfer to the coolers,
assuming of course, that a good machining job and consequently no air gaps are
allowed to exist at installation or during operation.

The lining materials used in these plate cooled designs can be all graphite materials or a
combination of semigraphite and graphite or all semi graphite, or sometimes combined
with silicon carbide depending upon the intended campaign life and anticipated wear
mechanisms. Some blast furnaces exhibit a history of minimal bosh wear and can
utilize more economical ceramic materials. Others, especially if high rates of injected
fuels such as pulverized coal are used, wil be affected by a lowered cohesive zone and
thus intensified bosh wear mechanisms, requiring intensified cooling and higher quality

8-45
 materials. As with any lining/cooling system design, the individual furnace operation
characteristics and campaign life goals would dictate which combinations are required.

Bellv and Stack Applications of Conductive Refractories

The application of graphitic and semigraphitic materials to the belly, lower and mid-
stack has been gaining in acceptance. It was previously feared that the use of
carbonactous materials above the bosh was a risky proposition because of the
propensity of water leaks from low quality coolers and the possibility of contact with air
from non-gas-tight cooler holders during backdrafting. Advancements in cooler plate
design and manufacturing as well as modem cooling system leak detection and gas tight
furnace jackets, have eliminated potential risks and offer new opportunities for
extending campaign life in these critical areas.

Graphite and semigraphite, because of their high shock resistance, resistance to

chemical attack and high thermal conductivity, can be combined with intensified
cooling from densely spaced, copper cooling plates, to provide a solution to severe wear
areas of the belly and lower stack. As was described for the bosh, contact between the
cooling plates and the refractories can be achieved with high conductivity rams or with
the proprietar machined contact system. Because chemical attack of all materials is
temperature dependent, the high conductivity of the refractory and the intensified
cooling combine to provide refractory walls that are too cool to be chemically attacked
and thus readily form protective accretions on their hot face.

The materials utilized can be combinations of various grades of graphites, varing in

density and properties depending on the zones where they wil be utilized, or
combinations with semi graphite or even ceramic materials and silicon carbide, as was
previously discussed.

Whenever the cooler plate spacing canot be optimized, conductive carbonaceous

refractories can be utilized to help conventional refractories work better in the blast
furnace. These conductive refractories can cool the hot face ceramic refractories much
more effectively than if the entire lining was composed of the lower conductivity
ceramic material, by directing heat to the back of the cooling plates, which are normally
under-utilized at the back side. Another benefit of this arrangement is that when
insulated from the steel shell, this conductive layer directs heat effectively to the
coolers, preventing overheating of the shell as the hot face ceramic lining becomes
thinner over time. Three dimensional, finite element analysis can be utilized to
determine the improvement of the location of the critical reaction temperature isotherms
in the ceramic material by the inclusion of the conductive zone. Additionally, if the
vertical cooler plate spacing is exceptionally great, "coolers" of graphite can be located
between the existing copper cooler rows to act as "passive" coolers, by directing heat to
the copper coolers.

8-46
Cooler plate spacing wil var depending upon furnace zone and expected heat load.
Typically, vertical spacing of between 250 to 380mm (10 to 15 in) is utilized for most
effective cooling and horizontal spacing is aranged so that "overlapping" of coolers
from row to row is achieved. Thermal modeling is especially effective for predicting
isotherm locations for optimizing cooler spacing. However, often other external factors
such as required capital costs or available reline time dictate design parameters that
result in less than optimum spacing. It is in these situations that arangements using
conductive refractories to enhance cooling effectiveness can provide compromise
solutions.

It should also be mentioned that the use of conductive refractories, especially when they
are used alone in a wall, do not necessarily result in excessive process heat loss. This is
because their high thermal conductivity provides a cold refractory hot face, which
promotes a build-up of an insulting accretion. This skull protects both the cooler and
the refractory from abrasion and reduces the total heat loss through the walL.

This is especially true when compared to the situation when cooler plates or staves are
completely exposed in the furnace due to the loss of a ceramic lining. The resulting
heat losses are much higher during this situation than when the coolers are covered by
even a high conductivity graphite materiaL. The concept is to utilize materials which
wil remain in place for long periods of time, to maximize life. Carbonaceous materials
can provide the means to achieve this end, alone or in combination with ceramic
materials, when combined with an effective cooling system.

Summary - Bosh. Bellv and Stack

Blast furnace bosh, belly and stack lining/cooling concepts are many and varied. Wear
mechanisms differ from furnace to furnace and zone to zone and must be thoroughly
analyzed before any refractory selection can be made.

The lining is only one par of a complex, interrelated system of components and
features, ~nc1uding influences by internal and external factors. Wear mechanisms such
as thermil shock, high heat loads, chemical attack, abrasion, erosion and impact, are
some of these factors. Another important factor is the cooling type and effciency,
which when combined with the proper refractory products, can significantly improve
campaign life.

The materials available are many and varied, with a full range of desirable properties
and characteristics. Proprietar design concept systems are available to the user, as well
as more conventional designs, utilizing various cooling methods.

Carbonaceous refractories can be combined with a variety of other refractories to

achieve optimum performance of each product used in the system or to minimize the
effect of a cooling deficiency.

8-47
 HOT BLAST SYSTEM

A very large user of refractory materials in the ironmaking complex are the hot blast
stoves and the related hot blast delivery system. The technology and design features of
these complex, refractory lined entities are often comprised of proprietar, sometimes
patented know-how. An entire volume can be written on proper stove design and
configuration including combustion chamber concepts, internal ceramic burner
technologies, metallc gridwork and checker supporting systems, dome configuration,
concepts of differential expansion accommodation, checkerwork and flue design as well
as nozzle lining concepts and configuration. However, some general comments can be
made in regard to proper refractory "system" design.

Stove refractories must be designed to accommodate the expected differential thermal

growth that cycles continuously as long as the stove operates. This heating and cooling
cycling, especially in the combustion chamber and checker mass, results in the constant
"moving" of refractories. During operation, this cycling can destroy refractories and
insulation and open up joints to allow hot gas short-circuiting and hot spots in the steel
shell. Expanding refractories can "grab" the steel shell and by the force of friction,
actually lift the shell from its foundation.

Insulation in critical areas can be crushed, abraded away or destroyed by short circuiting
hot gasses. Lack of thermal expansion provisions can dislodge nozzle brick, which
allows gasses to penetrate into the insulating layers.

Improper stove firing, incomplete or non-working instrumentation, gas explosions,

entrained moisture in combustion air or gas, backdrafing of the blast furnace through
the stove proper, dirt blast furnace gas and a myriad of similar occurrences, can
dramatically affect refractory life and pedormance.

Additionally, stresses from thermally expanding steel mains and shells can result in
deformation and/or cracking of the steel containment vessels, which can adversely
afect the refractory contained within. Therefore, it is imperative that stove shell and
main configurations, support and anchoring systems be analyzed and engineered by
competent stove and hot blast system designers.

Hot blast stove and the related hot blast delivery system refractory designs and
concepts, as well as material selection, is a specialized field, best conducted in
collaboration with professional hot blast system engineers and suppliers.

8-48
 TROUGH AND RUNER SYSTEMS

The casthouse trough and runner systems represent a large consumable refractory
demand that requires periodic maintenance and replacement. Additionally, a portion of
the refractory lining materials may be semi-permanent, such as insulating, back-up or
safety linings.

Trough and runner lining life is usually determined by operating practices such as
number of casts per day and in the case of the trough, how often it is drained and if it is
cooled. Severe thermal shock of refractories is experienced whenever a lining system is
allowed to cool down between casts.

Lining life also can be extended by remedial repairs between casts using gunite
application, ramming or hot patching techniques. Often, maintenance contracts are let
by the furnace operator, whereby all responsibility for material selection, installation
and maintenance of the linings are assigned to a subcontractor. However, operating and
maintenance practices of the furnace operator can have a major effect on lining
performance. For example, water sprayed onto hot refractory surfaces wil result in
thermal spalling and cracking. Taphole driling angles also can adversely affect impact
wear in the trough and taphole practices and poor clay quality can result in high casting
rates as the taphole erodes.

Refractory life in the trough can also be prolonged by the cooling of the exterior of the
trough enclosure. This can be forced or induced draft air cooling, water cooling or
natural convection.

Refractory life is also affected by the physical layout of the trough and runner system.
Flow velocities, impingement areas, turbulence, "eddy" currents and the like can
quickly cause erosion of even the best refractories.

Additionally, thermal expansion of long runs of refractory linings can result in

displaced runner offakes and similar connections, causing cracking and breakouts. The
system designer must take care that proper anchors are used and provisions made for
thermal expansion and differential movement of the branch connections whenever the
refractory containment "boxes" or forms are configured.

Trough and runner refractories can consist of a variety of different materials including
low moisture castables, dry vibratables, rams, precast shapes, carbon and graphite
i
blocks and many combinations of each other. It is beyond the scope of this paper to be
able to consider all of the possible combinations and configurations for discussion. It is
best to consult with experts in the field regarding proper trough and runner design
configuration before embarking on any refractory design or selection. Proper system
configuration can eliminate many of the wear points due to impact, impingement, high
velocities and turbulence, which can destroy even the best refractory available for the
intended application.

8-49
 SUMMARY

Successful refractory systems are dependent upon consideration of a variety of external

and internal factors, which can affect wear. Furnace operation, geometry, lining
configuration, cooling type and capability and the wear mechanisms encountered, all
can adversely affect refractory life. Improper refractory configurations which canot
accommodate expected thermal movements, differential expansion and stresses,
ineffective heat transfer or inappropriate properties, wil not survive long.

The refractory systems designer wil also recognize that properties alone cannot assure
long life and that refractory survival depends upon utilization of the most appropriate
concept and configuration for the application. This often requires the utilization of two
or more different materials in the same configuration, to take advantage of the best
properties and characteristics of each. There is no "perfect" refractory that can
overcome the effects of poor cooling, rough operation or abuse. Nor is there a
"standard" refractory system that is appropriate for every worldwide blast furnace. The
successful refractory system is one that considers each furnace as a unique problem,
demanding a unique solution, by examining its particular external and internal factors
which affect refractory performance.

It should also be remembered that refractory survival is totally dependent upon

recognition of factors external to the lining/cooling system. How these factors are
addressed or ignored wil determine whether or not the refractory system that was
created can be considered truly "successful".

8-50
 REFERENCES

1. Dzermejko, Albert 1., "Blast Furnace Hear Design Theory, Materials and
Practice", Iron and Steel Engineer, December, 1991, pp. 23-31.

2. Dzermejko, Albert J., "Design Considerations for Utilizing Graphitic Materials

as Blast Furnace Refractories", Ironmaking Conference Proceedings, ISS/AIME,
VoL. 49, 1990, pp. 361-377.

3. van Stein Callenfels, Egenolf, et.al., "Intermediate Repairs for Hearh, Bosh and
Lower Stack", Iron and Steel Society Svmoosium on Blast Furnace Campaign
Life Extension, Myrtle Beach, North Carolina, November, 1997.

4. Jameson, D., et.al, "Prolonging Blast Furnace Campaign Life", Technical Study
into the Means of Prolonging Blast Furnace Camoaign Life, European
Commission on Technical Steel Research Final Report, pp. 5-16, 1997.

5. Robinson, G.c., et.al., "Alkali Attack of Carbon Refractories", Ceramic

Bulletin, Vol. 58, No.7, 1979, pp. 668-675.

6. Bongers, Uwe, "Improving the Lifetime for Furnace and Runner Linings with
Carbon and Graphite Products", Sorechsaal, Vol. 117, No.4, 1984, pp. 332-340.

7. Higuchi, Masaaki, "Life of Large Blast Furnaces", Ironmaking Conference

Proceedings, ISS/AIME, Vol. 37, 1978, pp. 492-505.

8., R. M. Bucha, A. 1. Dzermejko and 1. G. Stuar, "Combining Equilibrium Theory

with Three Dimensional Heat Transfer Analysis, To Predict Blast Furnace Stack
Cooling and Refractory Performance", Ironmaking Conference Proceedings,
ISS/AIE, VoL. 42, 1983, pp. 673-679.

9. DeBoer, J., et.al., "History and Actual State of Lining and Cooling Systems at
the Estel Hoogovens IJmuiden Blast Furnaces", International Ceramic Review,
VoL. 32, 1983, pp. 16-18.

8-51

I
 CONVERSION FACTORS

TO CONVERT: INTO: MUL TIPL Y BY:

glee 1b/ft3 62.43

MPa psi 144.928

W/moK BTU - ft 0.578

Ft2-hr-op

W/moK K -ea1 0.861

m-hr-oC

°C 0p 1.8 + 32°P

8-52
-- - -- - --_.,- ---- --- - --------

LARGE' BLOCK CARBON WALL

TYICALLY 1.. 2m THICK

rHOT-PRESSf CARBON BRICK WAL.

i- TYiCAY 600BOm lHlCK

CHEMICALLY ATTACKED
-BRITT- ZONE
PROTECllVE
-SKULL-
CONICAL
TYICAL
CYLNDRICAL SHELL
00 SALAMANDER
I SHELL
Vi PROALE
l. I

-~ ¡ORIGINAL
PROALE

ORIGINAL HEARTH
---L ------------ _==~~i- __~~lH PROALE

HEARTH PAD HEARTH PAD

:EPHANrs FOOi PATlRN

FIGURE 1 - HEARTH WEAR PATTRN COMPARISON

 TAPHOLE TAPHOLE
ORIGINAL -WELL- DEPTH FIXED -WELL- DEPTH
ORIGINAL PROFILE

EQUIUBRIUM PROFILE AFTR 1500"C MOLTEN IRON I

00 I
I CARBON DISSOLUTION LOSS
Vl
l:
ADDITONAL 1900"C MEL llNG POINT CERAMIC PAD
WELL VOLUME
POTENTIAL FOR INTENSIFIED WEAR

CARBON PAD 1150.C SOUDIFICA TlON

CARBON PAD
TEMPERATURE ISOTHERM

ALL CARBON PAD HIGH MELTING POINT CERAMIC TOP LAYER

FIGURE 2 - WELL VOLUMLEFFECTS COMPARISON

,,
0
:ic-
+
:ic- 0u
+
u0 00
CO

oo
,.

o
o
co

0
0 SJ
10
LJ
ct
~
~
ct
00 I.
. a.
~
~
;)
ct
~
U
0 ~
0 I.
,.e:
ct

o
o
N

o
o
.-

· J4l6 'NOLLISOd30 N08èlVO

8-55
8-56
 ~
a.
I

f2

~
z
a.
~~(.
~~
~~

oc(
a.
I

f2
i

~ F!
z
a.
a:
oc
~~(. I.
:i
~~
il~ I

d..
, ¡
e
I

8-57
 zc(
- ::
C
~bc(
mZLa
c(Fc
La c( La
~g~
La u. 0
0.. U
51 ~
(3
o..
~
~
-
a:
.

)-
ß:
L&
::
L&

~ ~
g-i ~
::
~ c(
G ~
:E Z
J~ ~
c(
~ ~
~~ o
:i
La
-i, c( c(
Lab
::
a. cC " . 0: Z~
.. (.
- i-
0: La
i= La
La i- ~.
a. ëñ 0
lt :: i
~.
. ~ qu z
C)
U)
i= L&
ëñ a:
~
~

8-58
 i

LUMPY ZONE

SOFTENING
FRONT

MELTING
FRONT

..
I

~
COKE
SUT

COHESIVE
ZONE

HOT HOT
"BST · · BLAST

t HEATH

FIGURE 7 - BLAST FURNACE GAS FLOW PATTRNS

8-59
 ~
Z
~ ::
LL
-i ~
~~~
~~~ \,
~
8-60
 LOW HOT FACE 1150°C HIGH HOT FACE l 1150°C
TEMPERA lURE l" TEMPERA lURE

600°C 850°C
THICK SKULL THIN SKULL
AT EQUIUBRIUM AT EQUIUBRIUM

00
I
0\
..

J
THIN WALL IlICK WALL

~E 9 - REFRACTORY CONAGUBA~QN-ECTS
 SAME LOW l 1150°C SAME LOW l 1150°C
HOT FACE HOT FACE
TEMPERA TURE 6000C TEMPERATURE 6000C

00
I
0\
N

THIN WALL HK WALL W /HI-CONDUCTlVlTY MAJf

FI
 H20
OUT

H20
FLOODED
CHAMBER

H20
WATER IN
H20
COlLCTION IN

TROUGH

JACKET TYi STAVE TYE

SPRAY TY-E
00
I
0\
W

H20 OUT .. III c: c:

.. !II c:
c: c:

ii
~
~
IN R P A
EIGURE 11 - REEACTORY COOUNG SYSm4 TYES
 f0 SKULL f0 SKULL

UPPER
~ ~
Q NO SKULL Q NO SKUll
STACK

:~~:: +
.:::: +

..:.:.+

(:::: +
..:.:.:. +
LOWER
::::::;.: +
STACK
00 ::::::::: +
I
0\ :::::::::. +
.j :::::::::: +
:::::::::: +
BELLY ::::::::::: +
::::::::::: +
::::::::::: +
:::::::::::. +
;:::::::::: + +
:::::::::::: + i
BOSH ::::::::;.:: + +
::::::::::::: +
::::::::::::: + +
::::::::::::: +
TUYERE
------- ::::::;.::::: + + TUYERESJ
i:::::::;'::: +
ZONE ;::::::::::: + +
-------1
;:::(:::::: + +
TAPHOLES ;::::::;.::: + + J ~~H9L__s ~
::::::::::::
':::::;.;.:' ++ +
++
1// '" I
HEARTH

I THERMAL I HEAT I CO I CHEMICAL I AION

FlUCTUATIONS LOAD DEGRADATION ATTACK

FI
 CHEMICALS
ATTACK NEWL Y

CRACK THREE EXPOSED

SURFACES SURFACE

,/
t
00
I
0'
VI

.FIGURE 13 - "lERMAL SHOCK F AIUÆ SEQlNCE

 $.
0
"ex (/
::
l.
a=
F!
0en
a=
~
-i
l.
a=
a.
::
~
z
F
u
-i
l.
a=

" --- ------ --- -- ---

$.
o
ex

$.
o
0)
&(
I

..V
+ +
+ + + +
+++++++++++++
+++++++++++++
a=
;:
C)
i:

8-66
 LECTURE #9

IRON-BEARING BURDEN MATERIALS

Madhu G. Ranade
Inland Steel Company
East Chicago, Indiana, 46312 USA

INTRODUCTION

Iron-bearing materials are natural and synthetic materials with a

making blast
sufficient iron (Fe) content to be economically usable in an iron

furnace. Iron-bearing materials may come from a mine, from steel-mill

wastes (fines and dust), or from discarded metallc products of iron (Le.,
scrap). This chapter wil focus on iron-bearing materials originating from a.
mine, Le., "virgin" or "natural" Fe units in iron ore, which are processed into
various forms, such as lumps, pellets, and sinter. The use of steel-mil waste
oxides in the blast furnace will be discussed briefly. Since it is most
economical to recycle metallic scrap directly to pneumatic or electric
steelmaking processes, it will not be discussed in this chapter.

Pellets are roughly spherical, thermally- and/or chemically-bonded

agglomerates 5 to 15 mm in diameter. Sinter consists of irregularly shaped,
partially fused agglomerates in the size range of 5 to 30 mm. Lump ore
consists of irregularly shaped, large ore particles in the size range of 5 to 30
mm. Briquettes are mechanically- and/or chemically-bonded pilow-shaped
or cylindrical agglomerates, in the size range of 20 to 75 mm. Other

9-1
 miscellaneous materials include processed steelmaking slag and siliceous
ore trim. These materials are irregularly shaped and sized 5 to 30 mm.

Iron-bearing materials should enable the reliable production of hot

metal (also called pig iron) from the blast furnace in the desired quality and
quantity and at a minimum cost. While these requirements seem simple
enough, differences in geographical, geological, and commercial factors
have required the customization for each iron-bearing material supplier and
blast furnace user.
Iron ore mines and blast furnace operations are often located several
thousand kilometers from each other. The transportation of iron ore
represents about 15% of the dry cargo trade in the world. In 1990, this
amounted to 350 milion tonnes. The major iron producers and consumers
are listed in Table 1.(1) Typically, 95 millon tonnes of iron ore products are
produced in North America each year, with more than 80% coming from the
Minnesota, Michigan, Quebec, and Labrador regions.

Iron-bearing materials should be able to survive transportation and

handling from the mine into the blast furnace. Since iron-bearing materials
are charged downward from the top of the blast furnace through gases
moving in an upward direction, they should be free from fines that can be
carried out of the furnace by the gases. Also, the materials should not be so
large as to cause difficulties with conveyor transfers, bins, and charging
equipment. In general, materials originating from iron ore should contain a
minimum of 50% Fe, with particles mostly in the size range of 5 to 30 mm to
meet these basic requirements.

Iron ore mined from the ground does not meet the basic requirements
stated above. Invariably, further processing is required. A few high-grade
(;: 50% Fe) ores can be easily converted into blast furnace feed through
simple crushing, washing, and screening. Most iron ores require finer
crushing, grinding, and mineral dressing to separate the impurities (gangue)
from Fe minerals. The extent of crushing and grinding depends on the
"liberation size," Le., the size to which an ore particle must be crushed in
order to break apart iron minerals from the gangue minerals, such as
quart, silicates, carbonates, and aluminates. This concept of liberation size
is schematically shown in Figure 1. When the gangue minerals are fine and
intimately mixed with iron-bearing minerals, the ore particle must be crushed
to a very fine size in order to "un-lock" the iron minerals from the gangue.
Further processing may stil be required to physically separate iron minerals
from the finely crushed mixture. These beneficiation (up-grading) operations
often produce iron mineral particles that are too fine to meet the size
requirements mentioned earlier. Therefore, agglomeration techniques, such
as pelletizing, briquetting, and sintering, are used to increase the apparent _

9-2
size. Depending on the as-mined ore grade and liberation size, appropriate
processing schemes (Figure 2) are practiced to produce a suitable blast
furnace feed.

The above discussion makes it clear that liberation size plays a very
important role in determining whether an iron ore can simply be sized as a
lump ore, or would require sophisticated processing to produce pellets and
sinter. In addition to the simple transportation and size requirements
mentioned above, there are a number of metallurgical aspects that govern
the suitabilty of an iron-bearing material as a blast furnace feed. After briefly
reviewing these aspects, the production of various iron-bearing materials
wil be examined.

IRON BURDEN PROPERTIES

In the blast furnace, iron-bearing materials are subjected to reduction

and smelting reactions. Reduction reactions involve the removal of oxygen
contained in the iron oxide minerals in order to produce metallc iron. A
typical reduction sequence entails:
Fe203 (Hematite) --). Fe304 (Magnetite) --). FeO (Wustite) --). Fe

The smelting reactions consist of the melting of metallic iron and the
reactions with other non-iron-bearing minerals to produce liquid slag and
hot metal. All of these reactions are dependent on the temperature and gas
compositions prevalent in the blast furnace. The dissection of quenched
blast furnaces in Japan led to a new understanding of the internal state of
the blast furnace above the hearth region in terms of the five zones shown in
Figure 3. (2-4) Also indicated in Figure 4 are the reactions occurring in these
zones.(5) Various laboratory tests are used to estimate the likely behavior of
iron-bearing materials given the time, temperature, gas composition, and
stress prevalent in these zones, as shown in Figure 4.

Perhaps the single most important finding from the dissection studies
was the existence of the cohesive zone. It was deduced that most gas flow
in the cohesive zone occurs through the coke layers or "slits." The cohesive
zone represents the zone in which iron-bearing materials undergo solid to
liquid transformation. The location, shape, and size of this conical zone,
consisting of alternate layers of "cohesive" or fused iron-bearing materials
and coke, had a profound effect on hot metal productivity, composition,
operating stability, and lining wear.(2-5) The properties of iron-bearing
materials and coke, as well as blast furnace operating practices, such as
tuyere variables and burden distribution, determine the configuration of the
cohesive zone. (2-5)

9-3
 In reviewing the burden properties, it is essential to remember that the
blast furnace is a moving bed reactor in which gases move in a direction
counter-current to the movement of solids and liquids. Thus permeabilty of
the bed to ascending gases is extremely important to ensure good gas-solid
and gas-liquid contact so that the necessary reactions can take place. It is
also important to consider that reaction occurring in one zone of the furnace
can cause changes in burden characteristics that can affect behavior in
succeeding zones.

The iron-burden properties shown in Figure 4 can be divided into four

groups according to the testing temperature:
· Ambient Temperature: Chemical composition, size
consistency, compression
strength, and the tumble index.

· Low Temperature: Low Temperature

Disintegration (L TD) or
Reduction Degradation Index
(RDI).

. Intermediate Temperature: Swelling, reducibility,

compression strength after
reduction (CSAR).

. High Temperature: Contraction, softening, and

melting characteristics.

Previously, an extensive analysis was made of the data reported in the

literature concerning the effect of various burden properties on blast furnace
performance.(6) Table 2 summarizes the results of this analysis. It is
apparent that, although iron burden properties can have a significant effect
on blast furnace productivity and fuel consumption, the extent can differ
considerably from one furnace to another, as evidenced by the 95%
confidence interval.

Among the ambient properties, chemical composition directly affects

hot metal and slag compositions, and indirectly affects many other
properties. Some of the direct effects are summarized in Table 3. As
indicated, the chemical composition can affect hot metal composition as
well as blast furnace operation. The Si02 content of iron-bearing materials is
particularly important as it determines the slag "volume" (actually, mass of
the slag) produced in the blast furnace. Some slag is necessary in the blast
furnace for removing impurities, such as S, K20, and Na20. Excessive slag
volumes, however, represent unnecessary thermal load on the furnace and_

9-4
lead to a greater fuel consumption. To ensure a fluid slag with sufficient
desulfurization and alkali removal capabilty at ironmaking temperatures, its
Si02, CaO, MgO, and AI20S contents are controlled through proper
burdening of the furnace. The Si02 content also affects physical and
metallurgical properties of iron-bearing materials. The CaO/Si02 ratio of
pellets and sinter is often designed to yield the desired combination of
physical strength, LTD/RDI, and the intermediate and high temperature
properties.
The compression strength and tumble index primarily indicate the
generation of fines during the stockpilng, transportation, and handling of
materials. When the materials are screened in the stockhouse, the fines
represent a loss. If screening is not employed, the fines affect permeabilty
in the granular zone. The effect of the tumble strength of sinter on blast
furnace permeabilty is shown in Figure 5. (7)

Size distribution can affect the void fraction contained in a packed bed
of particles; fine particles tend to occupy voids between large particles,
thereby reducing the overall void fraction. Thus, size distribution can affect
blast furnace permeabilty. Size distribution also reflects the surface area to
volume ratio, and can affect the rate of reduction in the furnace.

The LTD/RDI and the CSAR indicate the tendency of the materials to
breakdown and generate fines in the granular zone. Consequently, they can
affect permeabilty, gas distribution, and the flue dust generation rate. The
L TD value represents the tendency of pellets to disintegrate due to stresses
generated in the iron oxide lattice during the reduction of hematite to
magnetite. The effect of sinter RDI on blast furnace permeabilty is shown in
Figure 6. (7)

The intermediate temperature properties, reducibilty and swelling, are

important in the lower part of the granular zone. Reducibility can affect the
utilzation of the reducing potential of CO and H2 gases in the furnace. A
high reducibility also leads to less FeO reduction in the high temperature
zone, and, therefore, improved softening and melting properties. Swellng
can affect burden movement, permeabilty, and gas distribution. The effect
of pellet reducibility on the blast furnace coke rate is shown in Figure 7. (8)

The high temperature properties, such as the softening and melting

temperatures, influence the location and geometry of the cohesive zone. It
is important to remember that burden materials in the lower zones are
subjected to the weight of the materials in the zones above. In the upper
part of the cohesive zone, where liquids begin to form inside the particles,
the incident load leads to plastic deformation blocking voids and restricting
gas flow within the particle. This aspect is characterized by the contraction _

9-5
 test. The effect of the contraction of pellets on the blast furnace operation is
shown in Figure 8.(9)

With the exception of contraction and softening-melting, all burden

propert evaluation procedures have been standardized by the International
Organization for Standardization (ISO) and adopted by the American
Society for Testing of Materials (ASTM). The test details can be obtained by
referring to the appropriate documents and will not be discussed here. A
brief summary is provided in Appendix i. The typical equipment used in
these tests, as well as in the contraction and softening-melting tests, is
shown in Figures 9 and 10. The testing conditions used in Kobe Steel's
contraction test and in a softening-melting test in use in North America are
shown in Table 4.

The burden distribution characteristics of iron-bearing materials are

also important. Pellets, being spherical, tend to roll, whereas sinter and lump
ore, which are angular and irregularly shaped, tend to remain where
deposited. This is evidenced by the angle of repose of pellets (28-32°),
which is shallower than sinter (32-36°). Therefore, sinter and lump ore have
more predictable distribution characteristics when charged using
conventional equipment, such as multiple bell tops with movable armor. At
one time, this was considered to be a serious disadvantage for pellets, and
it was doubtful that a large blast furnace with a significant proportion of
pellets in the burden could be operated successfully. However, the
development of a bell-less top which employs a series of lockhoppers and a
rotating chute for distributing materials has enabled effective burden
distribution control with pellets. As a result, the distribution characteristics of
pellets are no longer considered to be a major technical disadvantage.

IRON BURDEN COMPOSITION

Typical burden compositions for North American, European, and

Japanese blast furnaces are shown in Table 5; pellets predominate in North
America and sinter predominates in Europe and Japan. A brief historical
perspective is helpful in understanding the reasons for this difference.

In the early 1900's, iron ore producing mines, wood/charcoal sources,

and ironmaking operations were located relatively close-by. The United
States, United Kingdom, France, and Germany were the major producers.
Generally, ores with ;: 50% Fe content ("high-grade" or "direct shipping")
were mined and used with minimal processing. As the demand increased,
spurred by the industrial revolution and wars, high grade ores were
depleted and new ore sources had to be found. Also, with the advent of
coke-based iron making, operations often had to be located away from ore _

9-6
 mines, closer to steel customers and sources of other raw materials, such
as metallurgical coaL. During 1930-1945, many steel plants were constructed
in the Lower Great Lakes area relying on the supply of high-grade,
hematitic, "direct shipping" lump ore mined in Minnesota, Michigan, Ontario,
and Quebec. With the advent of the Bessemer steelmaking process,
phosphorus content in the ore had to be maintained below 0.045%. High
grade ores in Australia, Asia, Africa, and South America were known to exist
at the time, but with a few exceptions, their use in North American and
European iron making operations was not economical at the time. The
Japanese steel industry did import ores from Asia and South America.

In the United States, high-grade ores in Minnesota and Michigan were

depleted towards the end of the second world war. A totally new processing
technique had to be developed during the 1950's to make use of low-grade
(25-30% Fe) Taconite ore, which was considered as a "waste rock" earlier.
The liberation size for this ore was very fine (80% -325M or -45lL). New
mining and grinding techniques were developed to process this hard rock,
and magnetic separation and flotation techniques were applied to produce a
magnetite concentrate. Since the concentrate was too fine to be transported
or used in the blast furnace, pelletizing and thermal induration equipment
and processes were invented to produce pellets containing more than 60%
Fe. Figure 11 shows the transition from direct-shipping lump ore to pellets
for the Minnesota region. (10) In this time period, sintering was employed in
North American steel plants to recycle accumulated stocks of blast furnace
flue dust, which was becoming a major storage nuisance. With the
construction of major pelletizing facilties in the 1960's, North American blast
furnace operations transitioned from high-grade lump ore to pellets as the
major iron-bearing burden materiaL. Some low-grade hematite ores in the
Quebec-Labrador area could also be upgraded through fine crushing and
hydraulic separation techniques. The liberation size for these ores is
somewhat coarser than Taconite. Thus, these mines can produce pellet
feed or coarse concentrates which can be used in limited quantities for
sintering. The majority of production is pelletized and indurated.

In Europe, high-grade ores were exhausted between 1955-1965 and,

with the exception of Sweden, no other major iron ore reserves were
available. Therefore, iron ore imports were inevitable. As ocean shipping
became feasible and economical, ore reserves in South America, Australia,
and Africa were developed. As these ores could be liberated at a relatively
coarse size (-10 mm), they were suitable for sintering. Therefore, steel
plants built or re-built after the second world war utilized sintering facilties.
Thus, European and Japanese blast furnaces transitioned in the 1960's
from high-grade lump ore to sinter as the major iron-bearing burden
materiaL. During the upgrading of South American, Australian, and African

9-7
 ores, a small portion could be liberated at the size of lump ores. Therefore,
European and Japanese furnaces do use 5-20% lump ore in the burden.

As explained above, pellets have become the major iron burden

material for North American blast furnaces while sinter predominates in
European and Japanese blast furnace burdens, however there are some
interesting exceptions. During the 1970's, some steel plants in the United
States were built or expanded in anticipation of a great surge in steel
demand. Some of these plants were specifically designed to use ore
imported from South America in the form of lumps, pellets, or sinter feed.
Bethlehem Steel's Sparrows Point plant is one such example. The presence
of an ore similar to Taconite has also been the basis for pellet plants and a
pellet-based blast furnace operation in Sweden. During the 1970's, there
was a great debate on the optimum feed for the large, high temperature,
high top pressure blast furnaces built at the time. After evaluating economic
and technical factors, Hoogovens in Netherlands and Kobe Steel in Japan
decided to employ pelletizing and sintering operations on-site, using
imported ores.
In light of this background, the data in Table 5 can be interpreted as
follows: (1) the majority of hot metal produced in North America is from
domestically produced iron-bearing materials, (2) the majority of hot metal
produced in Europe and Japan is from imported iron-bearing materials, and
(3) South America, Asia, and Australia export most of their iron-bearing
materials to Europe and Japan.

At present, there are 20 iron ore mines, 13 pelletizing plants, and 12

sintering plants (Table 6) in North America.(11,12) The pelletizing capacity of
87.1 milion tonnes far exceeds the sintering capacity of 17 milion tonnes.
The production of lump ore in North America is practically non-existent.

This historical perspective serves as a background for the following

description of the production of pellets, lump ore, sinter, and miscellaneous
materials. In producing iron-bearing feed, it is necessary to upgrade it
through removal of gangue and to impart the physical and metallurgical
properties desirable to the blast furnace in an economical fashion.

PRODUCTION OF LUMP ORE

Lump ore is produced mainly in South America, Africa, Australia, and

India. These ores are typically high-grade hematite and often are readily
accessible and relatively uncontaminated outcrops. The production
process is, therefore, straightforward. Stripping, drilling and blasting
operations are similar to those described later in the production of pellets, _

9-8
 albeit easier. Coarse crushing and fine crushing operations are also similar.
At this stage, the ore is screened to produce lumps sized in the range of 10
to 50 mm. Water is sprayed on the screens to eliminate "piggy-back" fines
adhering to the ore lumps. The "clean" or "washed" lump ore containing
;: 60% Fe is railed to the shipping docks, stockpiled, and loaded on large
ocean-going vessels for transport to Europe and Japan.
Lump ore recovered as described above usually consists of a small
portion of the ore mined (5 to 20%). A large amount of -10 mm ore is
I produced in the process. This fraction can be easily upgraded using simple
, ¡ hydraulic separation techniques to ;: 60% Fe content and a size range of 1
to 10 mm. After dewatering (and sometimes thermal drying in rotating
drums), this fraction becomes suitable for sintering. This fraction of ore is
often referred to as "fine ore," "sintering fines," or "sinter feed." If a large
quantity of -1 mm ore particles are produced in the process, they are further
ground to -100 M or -150 ¡i, upgraded if necessary, and used in production
of pellets. For most South American and Australian mines, the production of
lump ore and sinter feed is sufficient for profitable mine operation as the
proportion of -1 mm particles produced in the process is relatively smalL.

Product characterization for lump ore and fine ore involves systematic
sampling and testing according to standard (ISO/JIS/Proprietary)
procedures specified in the customer supplier agreement. These tests
typically involve only size, moisture, and chemical analysis and bulk density
characterization. Typical properties of lump ore are listed in Table 7.

PRODUCTION OF PELLETS

Pellets are commonly produced and used in North America. The

production of iron ore pellets consists of a sequence of operations involving
the removal of ore from the ground, size reduction, upgrading,
agglomeration to produce spherical pellets, and thermal induration to impart

steps is described below. .

the necessary physical and metallurgical properties. A typical sequence of

Stripping

The process of converting ore in the ground with 25-30% Fe to

narrowly sized pellets with 60-65% Fe begins with ground preparation for
mining. "Overburden," the earth covering the ore body, is removed with
shovels and trucks, creating the ore pit. For each tonne of crude ore, 3 to 4
tonnes of overburden may have to be removed.

9-9
 Driling and Blasting

A large rotary dril is used to blast holes in a precisely engineered

pattern. Each hole can be 0.4 m diameter X 15 m deep and spaced 6 m.
Explosives are pumped into the holes and detonated by a blasting cord. A
slight delay between the detonation of successive rows of holes causes the
progressive fracturing of ore into crude ore chunks that can be easily
scooped up by an electric powered shoveL. For each tonne of pellets, as
much as 3 to 4 tonnes of ore have to be processed. The broken crude ore
is loaded into trucks or rail cars and delivered to a crushing plant.

Crushing

A large gyratory crusher reduces ore to 150-200 mm chunks and,

subsequently, additional gyratory crushers reduce the ore to gravel-size
pieces.
Grinding

In this operation, ore is reduced to its liberation size to faciltate the

subsequent separation of gangue from iron minerals. The grinding process
takes place in large rotating mils. Grinding can be autogenous--the ore is
broken up as it tumbles against itself, or exogenous--grinding media, such
as steel rods or balls are used. At this stage, the ore consists of a mixture of
liberated iron mineral particles, gangue particles, and still locked gangue-
iron mineral particles in a water slurry.

Concentrating
For Taconite ores, magnetic properties are exploited to achieve the
physical separation of liberated iron ore minerals from the rest of the ore.
For hematite ores, the difference in density between iron ore minerals and
gangue is often exploited to achieve the separation. In both cases, the
differences in their settling velocity in a fluid is exploited by using spiral
classifiers, hydrocyclones, and hydroseparators. Other techniques include
fine screening and flotation. Often a large proportion of silica-bearing
particles is present in a relatively coarse size fraction of the feed from the
final stage of grinding. Wet screening can effectively remove this fraction
without overgrinding the ore. In flotation, differences in surface
characteristics of iron and silica-bearing minerals are exploited to "float" the
latter by attaching an air bubble while the iron ore concentrate slurry is
drawn-off from the bottom of the flotation celL. These operations are
uniquely coupled for each plant, depending on ore characteristics and
equipment. Two different flowsheets are presented in Figures 12 and 13

9-10
which highlight the wide variations in processing schemes and equipment
used.(13,14)

Flux Preparation

Often a mixture of fine limestone and dolomite is used to produce

"fluxed pellets." In most cases, coarse flux in the size range of 5 to 50 mm is
received from limestone and dolomite quarries. It is blended in the desired
proportion and then crushed to about the same size as the concentrate by
dry crushing in a roll or gyratory crusher, followed by wet grinding in a ball
mill circuit incorporating fine screens (Figure 12). The flux slurry is stored in
a slurry storage tank.

Dewatering and Filtration

Iron ore concentrate slurry from the concentrator is partially dewatered

in thickeners and pumped to a slurry tank. A limestone-dolomite slurry is
added at this stage if "fluxed" pellets are to be produced. The final
dewatering stage is in the disc filters where water is removed from the
concentrate-flux mixture by a vacuum through a series of cloth covered
discs. Remaining on the disc is the filter cake containing about 9% water
and 60-65% Fe with a particle size of roughly 80% -325 M or -45lL.

Balling

Balling is the process in which the filter cake with a proper moisture
content is mixed with a binder and rolled into spherical pellets. Uniform
mixing of the binder with the concentrate is important for a stable ballng
operation. Bentonite clay or an organic binder is used to assist in enhancing
pellet growth and strength. Bentonite does contaminate the concentrate
with silca and alumina. To compensate, the ore has to be upgraded a little
more than is necessary, strictly based on the final pellet composition
specification. This represents an additional cost and also represents a
source of variability in pellet chemistry. Organic binders avoid this
contamination, but may be expensive or unavailable. The ballng operation
is performed using rotating drums or discs. Green pellets discharging from
the disc or drum are screened prior to their being loaded on the indurating
machine.

Induration

An indurating system may consist of a travellng grate alone (the

"straight grate" system) or a linked travellng grate-rotary kiln-circular cooler
system (the "grate-kiln" system). In rare cases, shaft furnace modules are
used as the indurating system. The green (unfired) pellets are loaded onto a _

9-11
 travelling grate which is a moving metal conveyor. In the straight grate
system, it contains a series of pallets. The pellets are dried, preheated, and
heat-hardened, as they pass through the furnace heated by a series of
burners fired by natural gas, pulverized coal, or oiL. Sometimes external
combustion chambers replace some or all of the individual burners
arranged along the side of the indurating system. In a system employing a
rotating kiln, a burner is provided at the discharge end. Typically, pellets
experience a peak temperature of 1280° to 1320°C for a pre-determined
time. Following induration, pellets are air-cooled to an ambient temperature
on the travellng grate itself or in a circular cooler. The air, thus pre-heated,
is used for combustion in other parts of the indurating system. Cooled
pellets are screened; fines are discarded or recycled to a re-grind mill and
added back to the concentrate slurry tank. The coarse size fraction
represents product pellets for shipping to steel mils.

Some product pellets are re-sized on a coarser bottom screen for use
as a hearth layer. This layer is deposited on the strand prior to the
deposition of green pellets. Thus, the grate bars in the pallets used on the
straight grate system are protected from experiencing excessive
temperatures.
The induration process has a significant and irreversible effect on the
physical and metallurgical properties of pellets. Until this stage, all changes
undergone by an ore particle are physical in nature. During induration,
chemical changes involving a series of phase transformations take place.
These include the exothermic oxidation of magnetite to hematite, the
calcination of limestone and dolomite fluxes, a reaction between iron oxides,
gangue, binder, and fluxes which produces silicates and aluminates of iron,
calcium, and/or magnesium, and the sintering and recrystallzation of iron-
bearing phases. The proper selection and control of the temperature-time
cycle experienced by pellets in the indurating machine is, therefore, critical
for producing pellets with the desired physical and metallurgical properties.

Product Characterization and Shipping

Fired pellets are systematically sampled and tested using standard

(ASTM/ISO/Proprietary) procedures specified by the customer-supplier
agreement. These tests typically include size, moisture, and chemical
analysis, bulk density, physical strength, and a variety of metallurgical
properties. These tests are used as a cross-check on the pellet production
process and as an aid to the blast furnace operator. Typical properties of
pellets are shown in Table 8.

Pellets are loaded onto railroad cars for short- or long-term storage at
the shipping docks and subsequent transport on ships to the steel mills _

9-12
located in the Great Lakes region. At the steel mills, pellets are directly
unloaded and stored in the ore field for subsequent use in the blast furnace.
In some instances, additional barge, rail, and/or truck transfer may be
involved. A few steel mils without convenient waterway access receive pellet
shipments by rail directly from the mine.

Recent Developments

The development of fluxed pellets has been the most significant

change in pelletizing practice in North America in recent years. It was found
that although "acid" pellets, produced without the addition of
limestone/dolomite fluxes, have good ambient and low temperature
properties, their intermediate and high temperature properties are relatively
poor. This is because at elevated temperatures, wustite combines with
silicious gangue in the pellet, forming a low-melting fayalite. This leads to a
high contraction, low melting temperature, and a large softening-melting
temperature range. The addition of appropriate fluxes to achieve a
CaO /Si02 ratio of 0.9 to 1.2 with an MgO content of 1.5 to 2.0% yields a
vast improvement in pellet properties as shown in Table 8. As a result, it
becomes possible to achieve a more favorable cohesive zone configuration
with fluxed pellets than with acid pellets as shown schematically in Figure
14. A number of blast furnace trials in North America have conclusively
proven that significant productivity, hot metal composition, and fuel rate
improvements can be achieved by using fluxed pellets. As a result, the
production of fluxed pellets has been rapidly rising in North America as
shown in Figure 15,(15)

Another related development has been the use of synthetic, organic

binders to replace bentonite used in forming green pellets. The main
justification appears to be a lower and less variable silica content in the
pellets and improved reducibilty in the case of acid pellets. However, the
physical strength of these pellets appears to be weaker. Therefore,
significant changes in indurating conditions and/or the addition of limestone
are used as countermeasures. The growth of these organic binder-based,
partially fluxed pellets can also be seen in Figure 15. The relatively high cost
of synthetic binders appears to be a limiting factor in their wider application.

PRODUCTION OF SINTER

Sinter is the most commonly used burden material in Europe and

Japan. In North America, it usually serves as a supplemental iron-bearing
material to pellets.

9-13
 As mentioned before, sintering operations are usually located at the
steel milL. Raw materials used in sintering consist of virgin ore ("fine ore"), as
well as a number of waste products from iron and steelmaking operations.
Sintering operations thus serve as an important and profitable method for
recovering valuable iron, manganese, magnesium, calcium, and carbon
units from these wastes, while minimizing the environmental liabilties of
waste disposaL. In many North American steel plants, sintering operations
are based exclusively on recycled steel mil wastes, such as mil-scale, flue
dust, coke breeze, flux fines, fine steel slag, and pellet fines (blast furnace
stockhouse screenings). As these waste materials are heterogeneous with a
bulk composition that tends to vary over time, considerable efforts have to
be invested in order to produce a relatively homogenous sinter feed of a
known composition, through extensive bedding, blending, and mixing
steps. Sinter plants using fine ores tend to purchase several brands of ores
and also require a similar treatment.

Typically, limestone and dolomite are used as fluxes in sintering to yield

the desired product sinter composition. Sinter is often classified on the basis
of its basicity, B/A = (CaO + MgO)/(Si02 + AI203):
Acid B/A -= 1.0
Fluxed B/A = 1 to 2.5
Super-fluxed B/A :: 2.5
Acid sinter is now rarely used. Blast furnaces using sinter as the major
burden component use fluxed sinter. Blast furnaces using sinter as a
supplemental feed to pellets use either fluxed sinter (for fluxed pellets) or
super-fluxed sinter (for acid pellets) in order to achieve a chemically-
balanced blast furnace burden.

The sintering operation is schematically shown in Figure 16.(16) In a

typical sinter plant, raw materials are received by ship, rail, and truck. A
bedding and blending yard is used to prepare and reclaim a relatively
homogenous feed. Other materials are added to this feed in the blending
yard and in the sinter plant, and composition adjustments are made as
necessary. Coke breeze is added as a fueL. Hot and cold in-plant sinter
return fines are also added at this stage. A drum or a pug mixer is used to
increase feed uniformity. This is often followed by a drum or a disc to
achieve the micro-pelletization of the mix. Water is added at this stage to
promote the adhering of fine particles in the mix to coarse particles called
"nuclei." This ensures that the fine particles wil neither "plug-up" the bed on
the sinter strand, thereby interfering with the sintering process, nor be lost
to the off-gases exiting from the strand. There is an optimum mix moisture
for maximum pre-ignition permeability (Figure 17).(17) The optimum value

9-14
has to be experimentally determined for each mix and micro-pelletization
conditions.

The sinter mix is then carefully deposited on the sinter strand, which is
essentially an open travelling grate machine with a series of pallets. An
ignition furnace is provided at the feed end. The ignition furnace employs a
matrix of burners fired with coke oven gas, natural gas, oil, or pulverized
coaL. When the mix enters the ignition furnace, coke breeze in the top layers
is ignited. The quantity and size distribution of coke breeze play an
important role in achieving proper ignition. Subsequently, suction is applied
under the strand and air entering from the top of the bed sustains the
combustion of coke breeze within a narrow layer.

While the flame front formed in the top layer of the bed moves
downwards in the bed, the air pre-heated above the flame front ignites the
coke breeze in the lower portions of the bed. At the discharge end of the
sinter strand, the flame front should be near the bottom of the bed,
indicating the completion of sintering. The exhaust gas temperature in the
last few wind boxes is measured to predict this "burn-thru" point, peak gas
temperature. Strand speed and bed height are controlled to maintain the
box.
"burn-thru" point just before the last wind

After being discharged from the strand, the sinter passes through a
sinter breaker, which breaks large chunks of sinter, and moves on to
screens capable of handling hot sinter (called "hot screens"). The coarse
hot sinter is discharged into a circular cooler. Hot fines are recycled back to
the sinter feed. Sinter is cooled by an updraft flow of air. The air, thus pre-
heated, can be used in the ignition furnace for combustion. After cooling,
sinter is screened and sized for transport to blast furnace. Fines generated
during screening are recycled to the feed.

When the sinter bed comes out of the ignition furnace, it contacts with
ambient air and undergoes rapid cooling. This results in a glassy, friable
sinter in the top region of the bed. In some plants, a hood is placed on the
strand after it comes out of the ignition furnace, to reduce the cooling rate,
thereby improving sinter yield.

Coke breeze additions to the mix are controlled to maintain a relatively

stable circulating load of hot and cold sinter fines in the plant. The quantity
of the coke breeze used determines this circulating load. Too little coke
breeze can cause a large fines circulating load and low productivity. Too
much coke breeze can produce a rock-hard sinter with very low reducibilty;
also, bed slagging problems could curtail productivity. Any exothermic
reactions that may take place in the mix during sintering must be considered
in determining the coke breeze addition rate. This is particularly true of _

9-15
 mixes containing a significant quantity of mil-scale, which oxidizes liberating
heat during the sintering process.

A portion of the product sinter is re-sized on a coarser bottom screen

for use as a hearth layer. This layer is deposited on the strand prior to the
deposition of the sinter mix. As the flame front does not enter the hearth
layer, grate bars in the pallets used on the strand are thus protected from
excessive temperatures. The hearth layer also helps in minimizing the loss
of sinter mix into the windbox through the openings between grate bars.

Product sinter is systematically sampled and tested according to

standard (ASTM/ISO/JIS/Proprietary) procedures specified in the
agreements between sinter plant and blast furnace operations. In North
America, this characterization is generally limited to size, moisture, and
chemical analysis and to physical strength. In Europe and Japan, where
sinter is the major iron burden component, several metallurgical properties
of sinter are also measured. Typical sinter properties are shown in Table 9.

Recent Developments
Major developments have taken place recently in the selection of
sintering ore blends, micro-pelletization technology, energy recovery, and
the use of sensors and control techniques at the sinter plant. The most
remarkable development, however, is the development of Hybrid Pellet
Sinter (HPS) at NKK.(18) This is an attempt to combine the desirable burden
distribution characteristics of sinter with the desirable metallurgical
characteristics of pellets. The HPS process uses a mixture of sintering ore
that provides "nuclei" and ore concentrates that provide "adhering" fines.
This mixture is pelletized on a disc pelletizer and coated with fine coke
breeze in a rotating drum to yield mini-pellets, about 5 mm in diameter,
which are then sintered on a travelling grate. A commercial plant has been
recently put into operation by NKK.

Another significant change has been the use of olivine instead of

dolomite at some sinter plants to obtain MgO. This has the advantage of
lower coke breeze consumption, and increased strength and productivity.
Some plants are also employing burnt lime additions to act as a binder at
the micro-pelletization stage. The resultant improvements in mix
permeabilty lead to increased productivity.

In North America, a number of sinter plants are being operated solely

for the recycling of steel-mill wastes, such as mill-scale, flue dust, and fine
steel slag. This provides a low cost feed material to the blast furnace while
minimizing waste disposal costs.

9-16
 MISCELLANEOUS IRON-BEARING MATERIALS

For reasons of economy and environmental protection, a growing

emphasis is being placed on recycling the waste products of iron and
steelmaking operations. In general, when little processing is involved and
the material can be recycled to downstream operations (e.g., primary or
secondary steelmaking), economic benefits are maximized. This is
exemplified by processing and recycling of steelmaking slag as shown in
Figure 18. Through this processing, materials for direct use in the BOF,
blast furnace, and sinter plant are produced. However, not all materials can
be recycled in this fashion. Chemical bonding (or "cold bonding") is then
employed through briquetting (Figure 19) or cold pelletizing (Figure 20).(19)
The binders used in these processes are often proprietary, however, most
involve materials containing lime, silica, and cement.

Briquetting can handle relatively coarse materials. The process

involves mixing feed materials and binder in proper proportions and passing
them through rolls containing briquette molds. Pressure applied on the rolls
results in dense, compact pilow-shaped briquettes. Green briquettes
usually have to be cured for 12 to 48 hours to develop the strength sufficient
to withstand further handling. While relatively strong briquettes can be
produced, there is insufficient experience in their handling behavior using
conventional ore field and stockhouse equipment. Recent trials at U. S.
Steel have shown that the briquettes may be used in up to 10% of the
burden. (20)

When fine, wet dusts that cannot be used directly in sintering are
involved, cold pelletizing can be used to produce pellets with physical and
metallurgical properties comparable to conventional pellets. The process
used by NSC is shown in Figure 20.(19) In this process, binder consisting of
finely crushed blast furnace slag and cement is employed. The process has
been in operation for several years at the Nagoya Works. Michigan
Technological University (MTU) developed a hydrothermal agglomeration
process in the late 1970's. In this process, a calcium-hydrosilcate bond is
formed in a steam autoclave.(21) The process is now in use in a pilot plant in
Michigan.

While cold pelletizing and briquetting appear to be technically feasible,

their high cost relative to sintering has prevented wider applications.
However, the situation may change if stricter environmental regulations lead
to a shutdown of sintering operations or if fine, wet blast furnace and BOF
dust disposal costs escalate.
Some blast furnaces using high basicity pellets and/or fluxed sinter
require silica additions in order to achieve a chemically balanced burden _

9-17
commensurate with the aim hearth slag basicity. In the past, silca gravel
was used as a burden trim for this purpose. Recently, sized Taconite ore is
being used instead. While the quantities used are very small (1-3% of the
burden), this silicious ore does bring in relatively inexpensive iron units
without alkali or phosphorus contamination. Taconite, after the fine crushing
stage (Figure 12), is re-sized to 5 X 25 mm for this purpose.

DISCUSSION

Thus far, we have covered raw materials, processing techniques, and

the important physical and metallurgical characteristics of iron-bearing
materials used in the blast furnace. While applying this information in real-
life, either to select an iron-bearing material, or to identify whether a
particular burden material propert is affecting blast furnace operation, three
important aspects must be kept in mind.

Firstly, a definite hierarchy exists among the various physical and

metallurgical properties. This is pictorialized in Figure 21. When faced with a
burden material that does not meet several physical and metallurgical
property requirements, it is important to focus first on improving those
properties towards the base of the pyramid shown in Figure 21. There is
hardly any merit to being concerned about low reducibility if a material has
high levels of undesirable impurities, e.g., alkali, zinc, or phosphorous.
Furthermore, when more than one iron-bearing material is used in the
burden, the compatibility of high temperature properties becomes
important. If the softening and melting characteristics of the materials are
significantly different and each material is charged as a nearly separate
burden layer, the cohesive zone configuration may be worse than if any
material was used alone. It is good practice to ensure that the majority of the
iron burden constituents have similar high temperature properties. lron-
bearing materials with different high temperature properties should be mixed
prior to (or during) charging in the blast furnace to minimize adverse effects.

Secondly, it should be noted that properties are measured under

constant and "idealized" conditions in the laboratory. Re-circulating species
inside the furnace, such as sulfur and alkali, could greatly affect the actual
behavior of burden materials inside the furnace.
Thirdly, blast furnace performance is a composite of interactions
between iron burden, coke, and operating variables, as shown in Figure 22.
To utilze the full capability of high-temperature hot blast stoves, high
temperature properties of iron-bearing materials must be adequate to
sustain a high flame temperature operation. Coke provides permeability
below the cohesive zone and provides CO for reduction through the _

9-18
 solution loss reaction (Figure 4). If coke properties are inadequate or if
burden distribution is poor, iron burden reducibilty may not play any role in
affecting blast furnace performance. Similarly, when the blowing rate (Le.,
wind rate) is increased for production, the furnace is likely to be more
sensitive to the burden material that generates fines inside the furnace. On
the other hand, the same material may perform adequately at lower wind
rates.

ACKNOWLEDGEMENTS

In preparing this chapter, I have liberally used material from my

predecessors, Messrs. Gladysz, Limons, and Cheplick. My colleagues at
Inland Steel and in the industry have also provided valuable information and
insights on the subject. I would like to thank Inland Steel for giving
permission to present this lecture.

9-19
 REFERENCES

1. "World Steel in Figures," 1991, International Iron and Steel Institute.

2. Kanabara, K., Hagiwara, T., Shigemi, A., Kondo, S., Kanayama, Y.,
Wakabayashi, K., and Hiramoto, N., Trans. Iron & Steel Institute of
Japan, Vol. 17, 1977,371-380.

3. Shimomura, Y., Nishikawa, K., Arino, S., Katayama, T., Hida, Y., and
Isoyama, T., ibid., 381-390.

4. Sasaki, M., Ono, K., Suzuki, A., Okuno, Y., and Yoshizawa, K., ibid.,
391-400.

5. Ishikawa, Y. and Yoshimoto, H., Proceedings of the Metal Bulletin's

First International Iron Ore Symposium, Amsterdam, March 1979,
142-155.

6. Ranade, M. G., Proceedings of the 57th Annual Meeting of the

Minnesota Section AIME. and 45th Annual Mining Symposium, 5-1 to
5-32.

7. Nishio, H., Yamaoka, Y., Nakano, K., Yanaka, H., and Shiohara, K.,
lronmaking Proceedings, ISS-AIME, Vol. 41,1982,90-97.

8. Blattner, J. L., Ranade, M. G., and Ricketts, J. A., Ironmaking

Proceedings, ISS-AIME, Vol. 43, 1984,267-271.

9. Saeki, 0., Taguchi, K., Nishida, I., Fujita, I., Onoda, M., and Tuchiya,
0., Agglomeration 77, AIME, Vol. 2, 803-815.

10. Minnesota Mining Tax Guide, October 1991,4.

11. 33 Metal Producing, May 1991, 23-24.

12. Personal Communications with Mining and Steel Plant personneL.

13. Minorca Mine, Brochure, Inland Steel.

14. MinnTac Mine, Brochure, U.S. Steel.

15. Minnesota Mining Tax Guide, October 1991, 22.

9-20
16. Ball, D. F., Dartnell, J., Davison, A., Grieve, A., and Wild, R.,
Agglomeration of Iron Ores, Heinemann Educational Books Limited,
1973.

17. Balajee, S. R., and Wilson, G. S., lronmaking Proceedings, ISS-AIME,

Vol. 43, 1984,59-71.

18. Niwa, Y., Komatsu, 0., Noda, H., Sakamoto, N., and Ogawa, S.,
lronmaking Proceedings, ISS-AIME, Vol. 29, 683-690.

19. "Recycling of Dust and Sludge," Nippon Steel Brochure, 1984.

20. Wargo, R. T., Bogdan, E. A., and Myklebust, K. L., lronmaking

Proceedings, ISS-AIME, Vol. 50,1991,69-87.

21. Goksel, A., Coburn, J., and Kohut, J., ibid., 97-112.

9-21
 LIST OF FIGURES

Figure 1: The Concept of Liberation Size

Figure 2: Processing Routes for Iron are

Figure 3: Internal State of the Blast Furnace as Deduced From

Dissection Studies(2,5)

Figure 4: Blast Furnace Reactions and the Relevant Raw Material

Properties(5)

Figure 5: The Effect of Tumble Index of Sinter on Blast Furnace

Permeabilty (NKK)(7)

Figure 6: The Effect of the Sinter RDI on Blast Furnace Permeability

(N
KK) (7)

Figure 7: The Effect of the Pellet Reducibility on Blast Furnace Fuel

Rate (Inland Steel)(8)

Figure 8: The Effect of the Pellet Contraction on Blast Furnace Fuel

Rate (Kobe Steel) (9)

Figure 9: Schematic of the Low and Intermediate Testing Equipment

for Iron-Bearing Materials

Figure 10: Schematic of the High Temperature Testing Equipment for

Iron-Bearing Materials

Figure 11: Transition from Lump are to Pellets Shipped from

Minnesota(10)

Figure 12: Flowsheet for Processing the Minorca Pit Ore(13)

Figure 13: Simplified Flowsheet of the MinnTac Plant(14)

Figure 14: Conceptual Sketch of the Effect of Acid and Fluxed Pellets
on the Blast Furnace Cohesive Zone and Performance

Figure 15: Trends in Pellet Production in Minnesota(15)

Figure 16: Schematic of the Sinter Plant Operation(16)

9-22
Figure 17: The Concept of Optimum Sinter Mix Moisture (Inland
Steel) (17)

Figure 18: An Example of Processing Scheme for Basic Oxygen

Furnace (BOF) Steelmaking Slag (Inland Steel)

Figure 19: A Schematic Flow Diagram of A Briquetting Operating

Figure 20: Cold-Bonded Pellet Plant of Nippon Steel(19)

Figure 21: A Hierarchy in Iron-Bearing Material Properties

Figure 22: Inter-relationships Between Iron-Bearing Material

Properties, Coke Properties, Operating Conditions, and
Blast Furnace Performance

9-23
APPENIX I: standard Testinq Methods for Iron-Bearinq Materials

DETERMINATION OF CRUSHING STRENGTH OF IRON ORE PELLETS

Reference ISO/DIS 4700

Document ASTM E382-72/1978 (Revised)

Sample
Particle Size* 1 0 x 1 2.5/9 .5 x 12. 5 mm
Number of Pellets )60
Configuration Load appl ied to a single pellet

Testing Conditions
Loading Method Constan t speed
Platen Speed 15+5 mm/min

Test Measurement Max imum compressive load at

which each pellet breaks
compl etel y

Test Repor t** ' 1) Crushing Stength = Arithmatic

mean of the test measurements

2) Standard deviation of the

test measurements

· Al ternately, particle siz~ as agreed upon between the

interested parties may be used.
*-Relative frequency of pellets which break at less than a
specific compressive load (e.g., 80 or 100 kgf or daN) is
also reported in some cases.

N: G,., Ranade
2/15/83
Inland Steel Company

9-24
 APPENIX I: standard Testing Methods for Iron-Bearing Materials
(cont. )

TULER TESl FOR IRON ORE, PELET, AN SINlR

Reference iso AS
Document 3271 E279-69 (1979)

Sale
Particle Size (nu x nu)
i,
i Pellets 6.3 x 40 6.3 x 38.1
! Ore/Sinter 10 x 40 9 . 51 x 50.8

Weight (kg) 15 :! 0.15 11.3 :! 0.23

Testing Conditions

OIU
10 x Width (nu x nm) 1000 x 500 914 x 457
Shell Thickness (mm) 5 6.3
Lifters 2 ~ SO x 50 x 5 DI 2 8 SO.8x50.8x6.3S nm

~er of Reolutions 200 E! 25 :! 1 rp 200 8 24 :! 1 rp

Treatmnt after React ion Screen Anlysis Screen Anys is

Test Report

1. Tumler Index (Tl) i +6.3 nu \ +6.3 nu

2. Abrasion Index (AI) \ -0 . 5 nu \ -0.6 mm (30 Mesh)

M. G. Rae
Inlan Steel Compy
2/23/83

9-25
 STANDA~n PROCEOURES OEVELOPEO BY INTERNATIONAL ORGANIZATION FOR STANDAROIZATION (I.S.O)
FOR TESTING, IRON ORE. PELLETS, ANO SINTER
(' ~
o '"
;; ~
H
X
H

tf
rt
II
::
p,
II
1'
p,
i-
CD
tI
rt
...
::
\Q
:i
CD
rt
=r
o
p,
tI
Hi
o
1'

H
1'
o
::
i
ti
CD
II
1'
...
::
\Q
:i
II
Test RepQrt (-) Averige ROI Averige of
rt
Averige of Averige Swe 111 ny CD
Rounded off Two tests (dlscird Two or Four ind Reduction 1'
to one dec1mil mInImum ind maxIm.. tests rounded rounded off to ...
place Vi i ues from 4 tests) off to In one declmil place II
rounded off to 1 nteger I-
two declmii places tI

*Generaiery 10 mIn.
*:Averige of the pilred results 1s found to, be sufflc1ently preciSe.
++Usuilly N2.
+++ Tumble Drum: 130 in. x ZOO in, Z lifters ZOO x ZO x Zin; 10 mln ~ 30 rpm.
Optlonii for measurIng CSAR. H. G. Ranade
Ipermitted virlltlon In compositIon,. +o.~i ibsolute.
ZHaxlmum o.ozi HZ' ii ternitely zoi co-"žoicoz-zi HZ-5BiNZ gas mixture miy be used. Inland Steel Company
2/23/82
TABLE 1: MAJOR IRON ORE PRODUCERS AND CONSUMERS(1)
(1989 Milion Metric Tons)

Apparent
Production - Exports + Imports = Consumption
North America 105.7 34.1 27.6 99.2
South America 185.7 138.4 4.8 52.1
Western Europe 50.7 25.0 150.6 176.3
Africa 60.6 38.5 1.5 23.6
Asia (inc!. Japan) 52.5 38.3 172.3 186.4
Australia & New Zealand 111.6 109.6 1.0 3.0

9-27
TABLE 2: ANALYSIS OF THE REPORTED EFFECTS OF IRON BURDEN
PROPERTIES ON BLAST FURNACE PRODUCTION AND COKE RA TES(6)

Increase in Decrease in
Relative Production Rate Coke Rate,
Change 95% C.I. * 95% C.I.*
Property (%) (%) (%)

Fines -1 ** 0.8 to 2.3 0.5 to 1.3

Low Temperature
Disintegration +10 1.3 to 5.9 -1.4 to 2.0+
Reducibility +10 NR 2.2 to 4.3
Softening and Melting *** 3.8 to 33.8 + + 3.2 to 14.1 + +

* Calculated 95% confidence interval

** Absolute change in the fines content of burden materials
*** Not quantified due to differences in measured test parameters
+ Indicates that no change in coke rate is possible
+ + Reported absolute range of effects
NR Not Reported

9-28
TABLE 3: EFFECT OF CHEMICAL COMPOSITION OF IRON-BEARING
MATERIAL

Composition
Descriptor Effect
Fe Reports to hot metal (95-97%)

p Reports to hot metal (90-95%)

Mn Reports to hot metal slag; and contribution to hot

metal (60-80%) affects steelmaking

Si02 Reports primarily to slag, but contribution to hot metal

affects steelmaking
AI203 Reports to slag (90-95%)

CaO Reports to slag (90-95%)

MgO Reports to slag (90-95%)

S Reports primarily to slag but contribution to hot metal

affects steelmaking
Na20, K20 Re-circulate causing scaffolds, report primarily to slag

Zn Reports to flue dust, but can penetrate the furnace

lining

Ti02 Reports primarily to slag affecting viscosity; controlled

accumulation in the hearth decreases wear

As, Cu, Sn, Ni Report to hot metal (90-95%); not desirable for
carbon steel production

Cr Reports primarily to hot metal; not desirable for

carbon steel production
H20 Reports to off-gas; represents additional thermal load
on the furnace

9-29
TABLE 4: CONTRACTION AND SOFTENING-MELT DOWN TEST

Softening
Contraction Melt-Down
Sample 500 g, 9.5 x 12.7 mm 500 g, 9.5 x 12.7 mm

Reactor 75 or 78 mm 10, Stainless Steel 75 mm 10, Graphite

Packing None Coke Breeze (Top & Bottom)

Load 2 kgj cm2 0.5 kgjcm2

Gas Flow 15 LPM 20 LPM

Gas Composition (%) Gas Composition (%)

Reduction Temperature Temperature
Program (CO) C02 CO N2 (CO) C02 CO N2
Up to 800 a a 100 Up to 400 a a 100
800-1100 a 30 70 400-600 22 18 60
600-800 16 24 60
800-1000 12 28 60
Above 1000 a 40 60
Heating Rate -8-9°Cjmin to 800°C 5°Cjmin
1.67"Cjmin from 800-1100°C
Measurements Bed Height Differential Pressure
Sample Weight after Cool-down Exhaust Gas Composition
Weight of Drippings

9-30
TABLE 5: BLAST FURNACE BURDEN COMPOSITIONS EXAMPLES

IRON BURDEN
% Pellets % Sinter % Other

North America 65 25 10

Europe* 15 65 20

Japan 10 70 20

*Excluding Sweden and Netherlands

9-31
TABLE 6: PELLET AND SINTER PLANTS IN NORTH AMERICA(11,12)

PELLET PLANTS

Capacity*
Plant Location Millon tonnes/year
Cyprus Northshore Silver Bay, Minnesota 4.0
Empire Palmer, Michigan 8.0
Eveleth Forbes, Minnesota 5.4
HibTac Hibbing, Minnesota 8.5
IOC-Carol Lake Labrador City, Newfoundland 10.3
L TV-Erie Hoyt Lake, Minnesota 8.0
MinnTac Mountain Iron, Minnesota 15.0
Minorca Virginia, Minnesota 2.5
National Keewatin, Minnesota 4.7
Pea Ridge Missouri 1.2
Tilden Tilden, Michigan 6.7
QCM Quebec 8.3
Wabush Point Noire, Quebec 4.5
Total 87.1

SINTER PLANTS

Capacity*
Company Location Millon tonnes/year
Armco Ashland, Kentucky 0.8
Middletown, Ohio 0.8
Bethlehem Burns Harbor, Indiana 2.6
Sparrows Point, Maryland 3.6
Geneva Orem, Utah 0.6
Inland East Chicago, Indiana 1.0
LTV East Chicago, Indiana 1.2
Stelco Hamilton, Ontario 0.5
USSteel Gary, Indiana 4.0
Warren Consolidated Youngstown, Ohio 0.5
Wierton Wierton, West Virginia 1.0
Wheeling-Pittsburgh Steubenvile, West Virginia 0.4
Total 17.0

* The reported capacity numbers could vary from year to year.

9-32
 TABLE 7: EXAMPLES OF LUMP ORE PROPERTIES

Chemical Analysis (Wt%) Australia Brazil

Fe 64.6 67.5
p 0.054 0.05
Mn 0.07 0.30
Si02 3.6 0.50
AI203 1.5 0.90
CaO 0.15 0.10
MgO 0.15 0.10
S 0.004 0.005

Size Analysis (%)

+ 10 mm 67 75
-5mm 8 8

)
, ¡

i
¡

9-33
 TABLE 8: EXAMPLES OF PELLET PROPERTIES

MINE EMPIRE MINNTAC CAROL LAKE WABUSH

Pellet Type Acid Flux I Flux II Acid Flux Acid Acid
Chemical Analysis (W~k)
Fe 65.4 59.5 61.4 65.5 63.8 66.4 65.6
Si02 5.6 5.5 5.5 5.4 4.0 5.0 3.7
AI203 0.4 0.4 0.4 0.2 0.2 0.2 0.4
CaO 0.3 7.1 4.9 0.4 3.5 0.4 0.1
MgO 0.3 2.1 1.6 0.3 1.1 0.2 0.2
Mn 0.07 0.1 0.1 0.1 0.1 0.1 1.9
P 0.016 0.013 0.015 0.014 0.014 0.01 0.01
S 0.002 0.002 0.002 0.001 0.002 0.003 0.004
CaD /SiD2 0.05 1.30 0.9 0.07 0.9 0.09 0.02
\D Size Analysis (%)
i
v. + 12.7 mm 4.5 1.8 5.0 7.5 10.0
.t 1.5 14
-6.3 mm 0.7 1.0 0.8 1.0 .15 1.5 0.5
Properties (ISO / ASTM)
Tumble Index (%) 96.5 97.2 97 97 96 95.5 96
Compression (kg) 232 228 205 245 210 201 222
LTD (+6.3 mm) 86 90 85 93 89 94 92
Swellng (%) 22 9 12 - - 16 17
Reducibilty l%/min) 0.7 1.3 1.1 0.8 1.2 0.7 1.0
Contraction (%) 25 7.6 12
Softening (OC) 1220* 1410* 1400* - - 1190** 1245**
Softening-Melting 270 150 110
Range CO C)

1 Kobe Steel Method

* Inland test
** Dofasco Test
TABLE 9: EXAMPLES OF SINTER PROPERTIES

Chemical Analysis (Wt%) Japan Europe USA

Total Fe 57.0 56.5 51.5

FeO 6.0 7.0 9.0
Si02 5.4 6.0 6.25
AI203 1.9 1.5 1.2
CaO 9.6 9.7 13.5
MgO 1.4 1.5 2.5

Size Analysis (%)

-5mm 5 8 12

Properties

Tumble Index (%) 75* 74* 80***

LTD/RDI (-3 mm %) 35* 25** 25**
Reducibilty 65* 1.4** 1.0**

* JIS
** ISO
*** ASTM

9-35
 LIST OF FIGURES

Figure 1: The Concept of Liberation Size

Figure 2: Processing Routes for Iron Ore

Figure 3: Internal State of the Blast Furnace as Deduced From

Dissection Studies(2.5)

Figure 4: Blast Furnace Reactions and the Relevant Raw Material

Properties(5)

Figure 5: The Effect of Tumble Index of Sinter on Blast Furnace

Permeabilty (N KK) (7)

Figure 6: The Effect of the Sinter RDI on Blast Furnace Permeability

(N KK) (7)

Figure 7: The Effect of the Pellet Reducibilty on Blast Furnace Fuel

Rate (Inland Steel)(8)

Figure 8: The Effect of the Pellet Contraction on Blast Furnace Fuel

Rate (Kobe Steel)(9)

Figure 9: Schematic of the Low and Intermediate Testing Equipment

for Iron-Bearing Materials

Figure 10: Schematic of the High Temperature Testing Equipment for

Iron-Bearing Materials

Figure 11: Transition from Lump Ore to Pellets Shipped from

Minnesota(10)

Figure 12: Flowsheet for Processing the Minorca Pit Ore(13)

Figure 13: Simplified Flowsheet of the MinnTac Plant(14)

Figure 14: Conceptual Sketch of the Effect of Acid and Fluxed Pellets
on the Blast Furnace Cohesive Zone and Performance

Figure 15: Trends in Pellet Production in Minnesota(15)

Figure 16: Schematic of the Sinter Plant Operation(16)

9-36
Figure 17: The Concept of Optimum Sinter Mix Moisture (Inland
Steel) (17)

Figure 18: An Example of Processing Scheme for Basic Oxygen

Furnace (BOF) Steelmaking Slag (Inland Steel)

Figure 19: A Schematic Flow Diagram of A Briquetting Operating

Figure 20: Cold-Bonded Pellet Plant of Nippon Steel(19)

Figure 21: A Hierarchy in Iron-Bearing Material Properties

Figure 22: Inter-relationships Between Iron-Bearing Material

Properties, Coke Properties, Operating Conditions, and
Blast Furnace Performance

9-37
MR53204.VR Sun Jun 15 01:14:50 1980

o re/G a n g u e After Grinding

Assemblage

Coarse ~-l ,. "

Il~_'
Liberation 'I ll
ll' ~~""
._.... .., ..l tl:1h :!l, i- '-i .
Size ~.tltllltl-__lH
\0
i
W
00

Ore Gangue Mixed Particle

/-"l-r
Finer mriIr.IL
. '. ..
dI lI.h i
~ i. .. "ii
. ..
Liberation i- ' 1l '- .-
"ii
.11'n-l
:~ìir.''''_.,
II.., .. ' " f il' _ .
Size II
.. II, II..i !l
. ii
,"II

MR53204.PS Wed Mar 04 10:56:30 1992

--"-- --"----_._-".~ _...._~.-' -'~'--'-- ~-----'- -- ..--_.,.-- ---~

MR532021.VR Mon Mar 16 09:30:34 1992

Run - of - Mine Ore

I

Liberation and
Beneficiation

I I
Lump Ore ("Coarse Ore") Sinter Feed ("Fine Ore") Pellet Feed ("Concentrate")
- 100 mm - 10 mm - 0.1 m m
\0
i
W
\0

Crush & Screen Sinter Plant Pellet Plant

5 x 30 mm

5 x 30 mm 5 x 15 mm
Blast Furnace

MR532021.PS Mon Mar 16 10: 13: 38 1992

 CD
u -0
c:
-0 ..
(1
as c:
~
C
at
- ..
cø

.. C)
.-c:c: ù: (1
~ ~
~ (1 )( -0
(1
en C)
-
c: en
cø
==
-
..(1 ~0 ~0 (1 (1 cø
LL .c () C)
..en 0 ()
(1
a.
c: CJ ()
(1
~ 0
()
cø
c:
cø
CJ
(1
~
-m
as

C)
c
--
~as
E
c0
-c..
as -0 (1

--c
c:
0 (1 0
N
N c: c: N
0\
(1 (1 0 0\
0\
0\
.-
0
en c:
0 0
c:~(10N .-
\0
N N ~(1
0 ()
N..
LO CD
..
\0
C .. (1 ..
~ () c:
(X
LO
""
..
(Y 0 cø
:: en (1
rn
-
.c
..
(X
0
.-
(Y
.- N s:
..
cø
(1
.c ~
0 () - s:
C)
..
rn
..
cø
(1
\0
.-
I-
ct
CJ () c: CJ :c I-
ct
~
~
c:
.r! 0
I-
¡. ~
p:
:; (/
i:
LO LO

0.-
N 9-40
.-
0
N
(Y (Y
LO LO
p: p:
~ ~
MR532014.VR Wed Mar 1810:16:201992

Relationship of Burden Properties to the

Internal State of a Blast Furnace
Processes Properties
Pellets & Sinter Coke
* Size consistency I * Size consistency
Charging, Drying,
Preheating Tumble index
I ** Compression
strength
* Low temperature I * Stabiliy
Granular I 3Fe203 + CO ~ 2Fe304 + CO2
disintegration
Zone I Fe304 + CO ~ 3FeO + CO2 * Reducibility
FeD + CO ~ Fe + CO2 * Swelling
""", "......,
\0
i l:j¡¡ljHjjjj
* Compression
.l FeO + CO ~ Fe + CO2
.. \1 strength after
C02 + C ~ 2CO
reduction I * Reactivity
FeD + CO ~ Fe + CO 2
Cohesive I
* Contraction * Coke strength
C02 + C ~ 2CO
Zone after reaction
I

Gas-Metal Reactions (CSR)

* Softening
Active Coke * High
Gas-Metal Reactions
Zone * Melting temperature
Stagnant Gas-Metal, strength
Coke Zone Slag-Metal Reactions
Raceway I 2C + Oi ~ 2CO
1-0 + C ~ H2 + CO I * Reduction I * Combustibilty

MR532014.PS Wed Mar 18 14:53:56 1992

MR5 3 2012 . VR Sat May 31 08: 2 6 : 56 1980

Relation Between Sintering Strength

and Permeability of Blast Furnace

3.8
0. ,,/." -- -- ~ ---
~
'~- /.// .. . .
..~ li
.-
.- 3.4 l //
\0
i ;/ .. il · II .
.t
-.c
N
Cd . . .""' --..--
Q)
E
~ I. //
Æ 3.0
11 /
. //
. /'
/
52 56 60 64 68
TI + 10mm (%)

MR532012.PS Fri Feb 28 12: 05: 11 1992

-. --"'- - --- ..-- -"'-'-"- ~"--~ ---~

MR532010. VR Sun Jun 01 23: 35: 34 1980

Relation Between RDI and

Permeability of Blast Furnace
\
\
- I
5.8 .. \\ \
(\
o I Co 5.6
--. It \
+ -i
:: -
\'...~.Ie \
\\\ \
-~ 5.4
,e e e 'e,
'-i +"
.- , , , ,e _6 , ,
.¡
W .- 5.2
-.0 ,
~ ~, rll ..
Q)
..\ e;r
\.
E 5.0 ,
Q)
,
a. ,,
4.8 ,
\
,

38 40 42 44 46 48 50
RDI (0/0 - 3 mm)

MR532010 .PS Fri Feb 28 12: 31: 59 1992

MR53207 . VR Fri Mar 13 13: 27: 12 1992

Relationship of Relative Coke Rate

to Pellet Reducibility
600 - Relative Coke Rate =
638.5 - 57.1 x Reducibility (R2 = 0.94)
\,
,\
Minimum reported slope
Q)
+o
\\
co 590 Est. confidence interval at 90%
a: _ \ \\ I confidence level about the mean
~o ~I
Ü C) ~
\0
I 580
l:
l: Q) ~
;: ..
.- 0)
ëa ..
-- \,
Q) \
a: 570 "'
\\
\
\
560
0.8 0.9 1.0 1.1 1. 2 1.3 1.4
Reducibility (%/min)

MR53207.PS Mon Mar 16 08:28:55 1992

MR532028. VR Thu Mar 19 18: 02: 46 1992

510
fI
- 500
~
:i 490
C)
.../.
,)~
~ 480
-~ ~~ LFuel Rate
Q)
470 Effect of Contraction
+-
æ 460 of Pellets Containing
Q) g.
Lt 450 )~. MgO on Coke Rate
"C and Fuel Rate
c 440
~
, ~.~- . /iß
\0
i
.t
Vi
~ 430
0: -sY:/'
4l. ~. /
1: ij Coke Rate
~0 420 ..'~
. 'Idll
Ü 410
J-l · No.1 SF
41IUIi
400 l ~'iUl
I1 No.2 SF
390 ·
0 10 20 30 40
Contraction % (1100° C )

MR532028.PS Fri Mar 20 08: 51: 16 1992

 0
ò -
Ol
tn ..
-
CI
CI
c:
..tn u. ('000l' 0
.. 0
CD
en
Q) .l ,~
I' l
';
n,. ~

CI
3:
LO
0) Ò
CO

C"
Q)
a. l I ~¡
en ..
co
..
~ ..--~-
LO

..(I .c
0 0 00 0 0 C)
0
..:: :: .q 0
CO LO
LO
"0 0) LO
CI
a:
..
CO

(I . c:
c. 3:
C. E 0
E ~~~ 000 0 0
C\ 0
0 C)
0
03: CI
..
C\ C\ CO LO CO 0
~ .. co
LO

..(I c:
-0 -:: -U
-- --. -
CO c:
en
.. 0a. N a. 0 0
-c
(I
- EOO
CI

o uuz
:: :: ::N 0 ~
a. c:
o '"
.-
Q)
N
en
+-
as
~
CI
E
~ o 0 0 0 - - '"0-
E
Q)
:¡ E
:J Q)
Q)
a.
:J
-0
C)

..E ..
~
a.
en
as
~
en
as
en
as
~ (! a: l-
E
Q) .-
E
as
CJ
c:
U
..c:
(I
N
0\
0\
rl
--c 0Q)
as
E
N
0\
0\
rl
r-
CX
l/..
CX
c: ir l/..
0l/
0.. CO
0 l/
rl
rl
CX
rl 3: - CX
rl
i-
l\
~
0
-. 0
Q)
==
::
i-
l\
~
'Ö
'Ö U)
Q)
:s
as ~
c;
ii U)
:; n.
0\ 0\
0
N 9-46 ¡
0
(Y N
(Y
l/ l/
ii ii
~ ~
MR53208.VR Wed Mar 18 08:22:58 1992

High Temperature Tests

Load Device
Displacement
II Gauge (åV)

Gas
Analyzer ~
Test Variables

1. Gas composition

\D Pressyre
2. Gas flow rate
i
.t
.. Gaug Furnace 3. Gas temperature
(åP)

4. Reduction time
5. Sample size
Gas
Pre-heater 6. Load

MR53208.PS Wed Mar 18 15:41:02 1992

 o
0)
0)
T-

LO
CO

o
CO

co
+- LO
r-
o
U)
"-
ro
R~
Q)
cc LO

-- CO

o(0
:E
LO
LO
0)
N
0\
0\
o
o o
CO
o
CO
o
~ o
C\
O~ N
0"
0"
M T- M

1. i-
qo
N
Lf qo
N
N paddl4S aJO lo % SB CD
(Y
M o
(Y
M slaiiad alluooBL 1.
M
¡.
¡.
nl II
~ ~
.,. c:
¡. o
¡. ~
0: U)
t; 0.
(Y
o
N 9-48
(Y
o
(Y N
Lf (Y
0: Lf
~ 0:
~
MR1HEAD.VR Fri Mar 13 08:22:42 1992

Minorca Mine Flow Diagram

.ECOIDA TERTl
CRSHER CRHER

COAE TAILS TO DUMP

FlUXSToNE UNLOAING
\0
i
L
\0 TO TAILINGS Wli

FLUXSTolE
SLURR
STORAE
TAN
FI.TER CA PW
STOCJ ST
,.
ORGAIC
BIIIDER
BIll
,,,,, \\\\\
flUX
CoNC.
SLURRY
,\
.._-----
STORAGE
FILTE
TANK

.IlTl ~

MR1HEAD .PS Fri Mar 13 08:29:14 1992

 CRUDe OR

Minntac Mine Flow Diagram

CO
CRSHER

SEDARY
alUSHER

FWX ADDITON
(GROUND LIMESTONE)
FLOTATIO
CELL
BENTONITE

CONCENTTEt:~INS
SLURRYf i .
BALLING
FILTERS 1 DRUM

~XJ
\0
i
V'
o

GRATE KILN
. ~-~
MAGNETIC
II SEPARTOR
(FINISHER)
~ ~
~ l!
~ ~
¡: Q.
W 0
II i-
/
 - ~- --- ---- --._--- -- ~
MR532022.VR Wed Mar 18 14:23:30 1992

Cohesive Zone Configuration with Acid & Fluxed Pellets

Acid Pellets Fluxed Pellets

· Thick, fewer coke slits for gases to · Thin, more coke slits for gases to
"pass-thru" IIpass-thru"
(Permeability., Rough Operationl) (Permeabilityt, Rough Operation.)
· Outer surface closer to walls, more · Outer surface away from walls,
lIdead-end" coke slits causing
fewer "dead-end" coke slits, less
gas impingement, (Heat Fluxl, gas impingement (Heat Flux.,
\0
Lining Wearl, Rough Operationl) Lining Wear" Rough Operation,)
i
VI
.. · Longer dripping distances (Si+) · Shorter dripping distances (Si.)
· Smaller lumpy zone · Larger lumpy zone
(CO Utilization., Fuel Ratel) (CO Utilizationl, Fuel rate, )
used ore layer
1 - lIDead-end" coke slit
2 - IIPass-thru" coke slit Tuyere
Slag
Hot IMetal Tap hole

MR532022.PS Wed Mar 18 15:18:07 1992

 MR532023. VR Wed Mar 18 14: 30: 34 1992

Growth of Fluxed Pellets

(Minnesota)
15
_ Fluxed 13.5
Partially Fluxed
(/
c 10
\0
i
~
VI
N c:
--0
-
-- 5
~
o
1987 1988 1989 1990

MR532023. PS Wed Mar 18 16: 21: 52 1992

 +' .-Q) 1 - ~~
UJ Q)
c: ~ ~ ê ,ß ~
CI .. ~
~co
õ .s.2 í~
E c: ~
Q)
(! :: ~ (I
CI co .c
C) a: ~ Q) .~
c: Ii c:
-co
o 0
0: 0
-
0
~
UJ
Q)
'"
co CI :;(,o c:
~ .. -
as
~
~
ic .: )(
- U)
co
-::
Q)

UJ c:
Ii co C (I
CI
c:
~o (I
~
o
-
UJ
-
CI Q) o UJ
:r
~ o
co
c:
~ -
Q)
c:
c:
æ -~~::
c:
.2
:t -~
Q)

co UJ
-Q)
c:
~ .-
.- c:
UJ
2.c
co (I Q)
.~
:! C) ~
UJ

3: ~
co 0
Q)
'"
c:
0: ei c: ~
N
-0
-:i -0~
co
N
0'
0' 0'
0' UJ rl
rl )(
"i Cl Q) 00
rl..
L!
.. ~
rl
rl ~(, "i
..
L!
rl -co
en
c:
Lf
~
rl
(Y
(Y
rl rl
~ ~
II
II
~ ~
.r! .r!
~ ~
ri ri
i
ir CI
I :: 0.
00 00
I 0rl rl
o
N
(Y
9-53 N
(Y
L! L!
ir ir
~ ~
MR532027.VR Mon Mar 1616:41:201992

Sinter Mix Moisture - Permeability Curve

êñ
1.0 1 Â Ore fines
E
-~ ~£
0.9 l . No ore fines
-..0.-..- /
as
Q)
\0
i E
VI "-
.t
æ 0.8 l l mix
Nearmoisture
optimum(95%)
.x
~
./-'\
"- 0.7
Q)
..
.-c
en
0.6
4 5 6 7 8
Sinter Mix Moisture

MR532027.PS Tue Mar 17 11:32:06 1992

- -,.,.,- ~ - - - - - - - - -- - _._---
MR532029. VR Tue Mar 17 09: 06: 36 1992

Steelmaking Slag Magnetics Preparation

Cooled BOF Slag In Pit

sc:iping Screen ~ Oversize + 250.4 mm

Matnetic Drum ~ Non-Magnetic Slag (To Land Fill)

l
Magnetic Slag
l
Screen ~ 76.2 X 250.4 mm (To BOF)
\0
i
Vi
Vi l
-76.2 mm
l
sCíeen ~ 76.2 X 12.7 mm (Coarse To BI. Fce.)

-12.7 mm
l
Screen ~ 12.7 mm X 4.7 mm (Fine To 81. Fce.)
l
- 4.7 mm (To Sinter Plant)

MR532029.PS Tue Mar 17 11:37:19 1992

MR532024. VR Fri Mar 13 13: 53: 10 1992

Ranade's Burden Property Hierarchy

Compatibility
Melt-Down
Contraction
\D
i
VI Reducibility
00
Swelling
LTD / RDI

Weathering
Size
Strength
Chemistry

MR532024.PS Mon Mar 16 08:53:57 1992

MR53206. VR Mon Jun 02 04: 19: 18 1980

Inter-relationship Between Charge Materials,

Operating Conditions and Blast Furnace Performance
Carbonization Ore Pelletizing/Sintering
Variables Characteristics " ~ Variables

Raw Material
Coal Properties Misc. Material
1 Ths~d ~ Characteristics 1 /Characteristics
Tested Coke '¥
Properties ~ __ Burden Behavior
\D
i
Vl
\D
'\ Coke ~ Blast Furnace
Behavior Performance
Blast Furnace
Conditions"
EQUiPment/ . Practices
t t
Strategic Tactical
SF Targets SF Targets

MR53206.PS Fri Feb 2813:29:14 1992

 LECTURE #10

BLAST FURNACE CONTROL - MEASUREMENT DATA

and STRATEGY

R. J. Donaldson
B. J. Parker

DOFASCO INC.
Hamilton Ontario
Canada

Abstract: - The most desired operation of a blast furnace is through the use of quality
raw materials and dependable control strategies. This offers a dilemma for operators as
costs increase when trying to satisfy both criteria. Consequently, Ironmakers must walk
a tightrope between obtaining adequate raw materials while ensuring their control
strategy is predictive enough to eliminate process upsets that may affect corporate.
profitability.

This paper wil deal with the measurement systems and control strategies that are
curently in use in today's blast furnace operations. It wil make special note of the
financial impact of model based control strategies and the measurement systems
required to successfully operate these systems. Emphasis wil also be placed on "what
makes sense" for a facility in designing and implementing a furnace control strategy.
In providing this summar, the fundamentals of model design and operation wil also
be discussed.

MEASURMENT SYSTEMS
Blast furnace measurement systems can be broken into two areas; Process Control
and Monitoring and Process Optimization. The former is the measurement and control
systems that wil:
. Ensure plant safety.
. Ensure plant operation is within specifications
· Ensure that plant remains in control.

10-1
 Process optimization measurements provide data for higher level models that
deliver feedback that will;

. Validate sensor measurements.

· Ensure plant economics are on target.
· Assist in developing new control strategies and strategic direction.

To discuss either of these areas in detail would require a book and is outside the
scope of this paper. What will be presented, are some details on the most crucial
measurements for supporting a dependable blast furnace control strategy.

Process Control and Monitoring

Process control and monitoring sensors consist of "Discrete" and "Continuous"
Measurements. Discrete measurements are "on/off' readings such as pump star and
stops, proximity switches, and zero speed switches. Continuous sensors are those that
provide an ongoing measurement of a process variable. Items such as temperature,
pressure, and flow are the most common in this category. A modem blast furnace relies
on both types of measurement systems to safely operate the plant and equipment.

Discrete Sensors:

The most obvious discrete sensors are limit and proximity switches as shown in
Figures i and 2. The proximity switch produces an electrical flux that wil detect the
presence of a magnetic object as it moves into its field. The limit switch on the other
hand has a mechanical striker that wil move when struck by a moving piece of
equipment. These sensors are usually used for logic applications were safety of the
plant is of importance. For instance, valve position, piston locations or sequential
control.

Some discrete sensors take a continuous measurement of a process variable and

convert them to a digital output. These include pressure, temperature and level
switches, vibration sensors and resolvers. The switches and vibration sensors are
usually tied to primar measurement devices that continuously monitor the related
parameter of the body or fluid. When a preset limit is obtained, a digital output is
generated that is then used to initiate some action (i.e. alar or shutdown). Resolvers
wil take the circular or angular movement of an object and convert it to a digital output
at a prefixed setting;

10-2
 In all the above, discrete sensors are generally used to control a sequential
operation or to protect the plant. They are tied into a PLC (Programable Logic
Controller) or DCS (Distributed Control System) to ensure rudimentary control is
always available. These systems form the "level zero" control structure of a facility
and are typically found in stove control and furnace filling applications. In this
capacity, these sensors must be simple, repeatable and reliable. High maintenance is
not acceptable and if it is required, the sensor application should be re-examined.

Continuous Measurement Systems:

These sensors are not only important for safe operation but are critical for the
success of many level 2 control systems. Most process calculations depend on the
basic measurements of Flows, Temperature, Pressure, and Composition,

Flows:

The measurement of gases, water and auxiliar fuels is of extreme importance to

process
the Blast Furnace process. Gas flows (air too) are required for the calculation of

models and the proper control of injected fuel rates, water flows for plant safety or
environmental concerns and injected fuel flows for thermal control.

There are many types of flow meters available for gas measurement as shown in
table I, and to cover each one would require a book and a good cigar. However, the
meter of choice is the venturi (See BF Control - Two Stage Heat and Mass balance).
Based on the Bournelli principle of operation, the venturi correlates the pressure drop
through the device with the flow measurement. The venturi is preferred due to its
averaging of the entire flow stream to produce a very accurate measurement. Its design
also provides less of an unrecoverable pressure drop. The downside of the venturi is its
cost. Whereas an orifice plate could cost $5000 installed, a comparable Ventui can
easily run i 0 times this figure.

The preferred flow measurement for water is the magnetic flowmeter (See BF
Control - Two Stage Heat and Mass Balance). The meter, based on Faradays principle,
is simply a wire coil that surrounds a pipe. The coil has a known field power and as
shown in Figure 3 and in equation onel. The flow of the fluid through the field wil
develop a proportional electrical output called the electro-motive force (emf). This type
of flow meter is very desirable for water or slur applications as there is no pressure
drop, (i.e. no increase in pumping costs) and the accuracy of the device is extremely
good. The only drawback is that the pipe must be full and the fluid electrically
conductive.

emf = - BDV * 10-8 (1)

10-3
 One of the most important flow measurements in the Blast Furace is that of
auxiliar fuel injection. Whether it is natural gas, bwier C oil or coal, it is extremely
important that this measurement is as accurate and as repeatable as possible. From a
process control standpoint, it is preferred that this measurement is mass flow and not
volumetrically based. To do this with gas is a matter of adding pressure and
temperature compensation and, as long as the gas composition is known, the mass flow
correction can be made. The principles get slightly more diffcult for higher specific
gravity substances such coal and oiL. In this case, load cell based injection systems or
mass flow meters are used.

One of the most repeatable, linear, and accurate mass flow devices is the Corio lis
meter. Figure 4 illustrates the operation of this meter 1. To summarize, as the material
goes through the oscillating pipe, the pipe wil twist, this measured twist will be
proportional to the mass flow rate of the materiaL.

Temperature:

There are thee common methods of temperature measurement in the blast furnace
area - Thermocouples, Resistive Thermal Devices, and Optical sensors (See BF
Control- Two Stage Heat and Mass Balance, Burden Distrbution, Auxilar Systems).

The basis of the thermocouple is the Seebeck effect and is ilustrated in figure 52.
If two dissimilar metals are heated at the same time an electron flow (emf) is produced.
This voltage versus temperature relationship can then be used to establish a calibration
curve. There are several types of thermocouples available as shown in table 2.

A Resistance Thermal Device (RTD) relies on the principle of the Wheatstone

bridge. Graphically this is shown in figure 6 2. Ifresistors are placed in this
arangement, the resulting voltage is proportional to the temperature increase. The
same principle is used in strain gauges for weight and force measurements. RTDs are
more sensitive and more accurate than thermocouples (+/- O.3°C versus +/- 2°C).
However, they are limited in temperature range (-260° to 630° C), more expensive
(platinum based) and do require an external power source. In the iroruaking facility,
R TDs are predominantly used for water measurement while thermocouples are used for
higher temperature applications such as hot blast or fuace gas streams.

The third most common blast fuace temperature measurement method is using
optical pyrometry. There are two basic methods of measurement, infrared and spectral
radiation. Infra red systems, which include optical pyrometers and fibre optic systems,
have an infrared source that is compared to the infra radiation of the body being tested.
The difference in the readings is then used to determine temperature.

10-4
 Radiation type pyrometers depend on the radiant heat transfer principles of the
tested body. This measurement is based on the simple formula 3;

E+R+T=l (2)
where:
E =Emittance - Ratio of energy released by a body relative to a black body.
R= Reflectance - Percentage of total radiant energy falling on a body that is
reflected without entering the body.
T= Transmittance - - Percentage of total radiant energy falling on a body that
passes through.

A black body is defined as an item with an Emittance of one (see figure 7) 3. A

stove dome is a perfect example. The optical sensor reads the energy level of the body
and comparing this energy level with that of a black body (using an emissivity factor)
determines the temperature.

Both infra red and radiation systems are somewhat slower to respond to
temperature changes due to their non-contact nature (i.e RTD's and thermocouples sit
right in the gas stream). On the other hand, their life expectancy is far superior when
properly maintained.

Level and Pressure:

Level and Pressure measurements are extremely important in the daily operation
of a Blast Furnace. The reason that these are mentioned together is that in several
applications the same principles of measurement are used (i.e. in many cases fluid head
pressure is converted to level).

Pressure sensors (See BF Control- Cohesive Zone Model) come in several

these devices would take forever to review. However, the most
varieties and full list of

common is the capsule based pressure transmitter (Figure 8). Typically these
transmitters are connected to the process vessel via a stainless steel pipe or impulse
line. One pipe is connected to the high-pressure side or low-pressure side or both sides
of a vessel to get an absolute pressure or a differential pressure. In either case, the
I
pressure can then be used to either indicate the system pressure or converted to flow or
level using basic mathematical relationships.

In the level measurement area, bubbler devices are very common for blast furnace
water and slurr systems. In this application, the level of a body of fluid is indicated by
the amount of air pressure required to allow a bubble to be transmitted. This is
proportional to the level of the liquid based on its specific gravity.

10-5
 In the area of stockline and torpedo car measurement, the use of microwave
systems and in some cases laser systems is gaining more and more popularity. In many
installations, the ever-reliable mechanical gauge system (Figure 9) is stil in service.

Analyzing

Moisture

There are two basic measurements in the blast furnace for moisture measurement,
the nuclear moisture analyser for coke and pellet moisture and the measurement of gas
and air stream moisture in the blast furnace (See BF Control - Heat and Mass balance,
Coke Rate).

The nuclear moisture gauge is based on the principle that neutrons wil be
thermalized or slowdown by hitting hydrogen atoms. The number of slow moving
neutrons is directly proportional to the number of Hydrogen atoms that are present.
Consequently, the moisture content of a coke or pellet bed can be determined from the
resulting rate count. In many cases, density compensation is also added to this control
system by using gamma sources. The combination of the two readings is then used to
control the dry coke unit of the blast furnace.

Blast moisture is very important in the determination of furnace hydrogen

utilization and therefore the identification of leaks. There are two basic types of
measurements the dewcell and the chiled mirror. The dewcell operates on the basis
that a lithium chloride salt solution wil create an ionic curent as it gets wet. The
resulting current heats the probe, and this temperature is an indication of dewpoint
temperature. From this dewpoint temperature, the relative humidity can then be
determined. The chilled mirror works on the principle that a mirror's temperature is
controlled until a slight fim of water vapour is seen on the lens. This temperature is
the dewpoint of the water and from this, the humidity level can be determined.

Carbon Monoxide, Carbon Dioxide and Hydrogen.

Although there are other gases present in the blast furnace, these three (and by
difference Nitrogen) are the most important for blast furnace control and monitoring
(See BF Control - Two stage Heat and Mass Balance, Gas Distribution). As explained
later they typically form the basis of all heat and mass balance developments and are
essential components for thermal control and leak detection.

There are two preferred analysis techniques in blast furnaces today for the
measurement of these gases, the mass spec and a combination of both infra red and
thermal conductivity measurement systems. The mass spec is based on the principle of
excitation of atoms and the resulting spectral emission. From this emission, the
concentration of the chemical species can be determined.

10-6
 In the case of Infrared technology, common for both CO and C02 species
determination, the absorption rate of infra red radiation is proportional to the
composition of the element. This type of measurement is also common for the CO
safety alar systems that are in place in most facilities. Infra Red detection is not as
sensitive for hydrogen and consequently, thermal conductivity measurements are used
for this element. In this case, the gas cools a heating element that is maintained at a
constant reference temperature. The voltage maintaining the reference temperature is
then used to determine the hydrogen content.

-¡
i Oxygen

Oxygen sensors are most commonly found in the blast furnace stove area (See BF
Control - Auxiliary Systems). They are typically based on the Nerst equation (see
below) due to the higher temperature application 3. The output of the reading is a
logarithmic function of the difference in oxygen content of the sample and reference
sources. These units are very repeatable and relatively easy to maintain.

E = RT lnr P(02RiJ (3)

nF L P(02sl

E = Voltage
R =Ideal Gas Constant
F = Faraday Constant.
n = Number of electrons in the electrode reaction
T = Absolute Temperature - Reference Temperature
POiR = Oxygen Partial Pressure of Reference Gas.
POiS = Oxygen Partial Pressure of Sample Gas

Weight:

One of the essential elements for furnace control is the measurement of raw
material arid hot metal weights (See BF Control - Two Stage Heat and Mass Balance).
The method of choice is the load celL. Based again on the Wheatstone bridge theory,
the R TD wil measure the strain put on the cell from changes in the structure weight.
Simple in construction (figure 10 ) the load cell is very reliable 3. The most common
problem comes with mechanical binds of the weighing structure.

10-7
 PROCESS OPTIMIZATION

The sensors discussed thus far are really those related to fundamental operation
and control of the blast furnace. However to improve the furnace, more intelligent
monitoring systems must be pressed into action. This includes what we would call the
smar sensor arrangement. For the sake of argument, we wil include in this category
the following sensors:

· Above Burden Gas Probe

. Profilometer
· In Burden Probes - Shaft, Vertical and Bosh or Belly.
· Tuyere Probe.

Above Burden Gas Probe

This probe can be of the permanent or the retractable design (figure 11). Several
points can be placed throughout the probe for temperature and gas measurement.
Depending on the arangement, water cooling may be required. Typically, these probes
are used on a periodic basis to sample the furnace top gas. The gas from the different
sample points is stored in pressurized bottles, and then analyzed to determine each
point's composition. This information is then used by level two models to developed
gas utilization profies and gas flow models for the furnace.

The drawback with the above burden gas probe is that there is a fair amount of
mixing on the top of the burden. As well plugging of the sample ports of these probes
is an ongoing problem.

Profiometers

These probes are used to periodically monitor the burden level of the fuace
across a radius or diameter (See BF Control - Gas Distribution). In many cases they
are combined with the above burden gas probes. They typically consist of several
mechanical probes that are allowed to settle on the burden surface. From these readings
and associated top gas or in-burden measurements, very good modelling strategies can
be developed.

In Burden Probes

The most common of these probes is the horizontal shaft probe, which can be
either fixed or retractable. As in the above burden gas probe, there are several
measurement points for gas and temperatures. In some cases, mechanical sampling
systems and cameras are used to determine size distribution of the burden materials.~
Magnetometers are also used to detect material movement 4.

10-8
 Typically they are located about 10 meters below the burden's surface and are
used in gas and burden distribution models. Fixed probes (figure 12) of this nature
mustbe of tough construction to ensure long service life 4. Retractable probes require
sturdy external support to ensure trouble free movement in and out of the furnace. In
either case, cost for these probes versus their above burden counter parts is much
higher. On the other hand, it is felt that the information provided under the burden is
much more reliable from a furnace modelling perspective. As always, there is a trade
off between cost and performance.

The vertical shaft probe (figure 13) is a more rare probe largely due to its cost,
headroom and auxiliary equipment requirements (figure 14). These probes have been
tried in Japan, Australia, and Germany. The Japanese design, which has gone to a depth
of 23 meters, has an on board fibre scope that feeds a high sensitivity colour TV used
for particle size distribution 5. As well, samples can be taken when the probe is
withdrawn from the furnace. Information from this probe can be used in the
development of furnace shaft gas flow and cohesive zone models.

Belly or Bosh probes have also been tried by the Japanese as shown in Figure 155.
As in the case of the vertical shaft probe, cameras are used to determine size
distribution. When it is withdrawn from the fuace, a hole wil develop in front of the
probe that is the positive indication of the cohesive zone location.

Tuyere Probes

Tuyere probes have ranged from pipes pushed through the tuyeres to high tech
camera systems (Figure 16) 4. The purose of these probes is to determine the size,
temperature and activity of the raceway (See BF Control - Raceway). Typically, high-
speed cameras are used to monitor the raceway through the probe. From the periodic
brightness changes, the raceway size and temperature can be determined. This
information is especially useful for coal injected furnace, where raceway collapse and
restructuring is crucial for successful high coal injection.

10-9
 BLAST FURNACE CONTROL STRATGEY
Today's control strategies are possible because of development in five areas. Figure
17 shows that optimal blast furnace performance can be visualised as a pyramid. Most
operations strive to have their operation as efficient as possible. However, there is an
increasing cost associated with this goal. Similarly, unstable operation is undesirable
since operating costs (fuel rate) will increase. The challenge to Ironmakers is to find
their optimal performance level given the measurement systems and modelling
capability that is available today.

Model Types

Today's models can be broadly classed as follows:

Mass and Energy Balances

Burden Distribution and Gas Flow
Predictive Control
Auxiliar Systems

Many of these models include the same data sets and often the outputs of one
model are important inputs to another. What follows is a brief discussion of the basics of
these models.

Mass and Energy Balances

Global Mass and Energy Balances

The original attempts at a furnace mass and energy balance were quite simple in
nature. The evolution of the global heat and mass balance is summarised very well by
Poos and wil not be covered 6. However, the operation of the blast furnace is most
easily understood through the examination of an overall furnace mass and energy
balance.

Figure 18 ilustrates the basic assumptions of some early mass balance models 9. A
somewhat more detailed representation is shown in figure 19 ~°. Quite simply, the
figures translate to the following equations 9.

ninPe = noulFe
(1)
ninC = nOUIC (2)
nino = noutO (3)

where:
nin = number of moles of each component entering the furnace.
nOUI = number of moles of each component exiting the furnace.

1 0-1 0
 All three of these chemicals enter the furnace in various forms but leave in limited
paths. The iron leaves as hot metal and trace amounts of iron oxide (FeO) in the slag.
The oxygen exits in the top gas as a carbon bonded gas and the carbon leaves as gas and
as about 4.5% of the hot metal. These simple reactions are the basis of blast furnace
operation. For a summary of common chemical balances often considered in these
models refer to Table 3.

An overall heat balance output is shown in Table 4. In summary, heat in the blast
furnace comes primarily from two sources; the combustion of carbon (coke and
injectants), and the sensible heat from the hot blast. The use of
the heat can generally be
classified as; the sensible heat of the liquids, the reduction requirements, the solution
loss reaction (i.e. the combustion of coke), heatlosses, and the top gas heat.

When reviewing the outputs some generalizations on these models are possible:

i) There is no indication of the effciency of the operation.

ii) The use of heat is concentrated on the melting of materials.
iii) An error term is required.

These models are convenient tools for assessing various operating scenarios from a
global perspective but are not robust enough to determine actual impact of burden or
process changes on the efficiency of the furnace operation.

Two Stage Heat and Mass Balances

Akerman (1866) introduced the idea of staged mass and energy balances by
differentiating between reduction and heating carbon. In the 1920's Reichhardt took this
idea and made it useful by dividing the blast furnace into temperature rather than
geometric regions. The typically accepted isothermal lines of 950, 1200 and 1500° C are
still in use today 6.

To understand the theory behind the staged mass and energy balance, we must first
introduce the zoned blast fuace. Although two stage balances make sense from a
mathematical modelling standpoint, there are at least 3 and perhaps 4 very distinct
operations in the ironmaking blast fuace 9. 11.

1) An upper zone where the burden is heated to 950° C. This zone is located from the
burden surface to 3 meters below it. In this area the gas entering the zone is as
much as 450° C hotter than the burden. This is also called the preheating area of
the furnace.

10-11
 2) An intermediate or preparation zone (also called the thermal reserve zone) where
the gas and solid's temperature remain the same (approximately 9500 C). In this
area the hematite (Fe203 ) and magnetite (Fe304 ) components of the burden are
reduced to Wustite (FeO). As well the water/gas shift reaction (H20 + C = CO +
H2) occurs in this region.

3) The lower zone of the furnace called the elaboration or processing zone where the
FeO is reduced to iron and both the iron and slag are superheated to temperatures
from 9500 C to over 15000 C. In some cases this is considered two zones, melting
and superheating. The solution loss (C02 + CCoke = 2CO) and direct reduction (FeO
+ C = Fe +CO) reactions also occur in this area. This area's size and location are
largely driven by coke quality.

In all modem two-stage heat and mass balances, the furnace is divided at a point in
the intermediate zone. Typically this is done at a point where the gas and solids are at
approximately the same temperature 950 0 C. The oxidation potentials of CO and Hz are
0.295 and 0.39 respectively, and only Wustite is present. This is represented by point W
on the Rist diagram, shown on figure 20 1 i. Figure 21 shows this point as a relative
furnace position and figure 22 gives a quick overview ofthe various reactions that occur
in the different regions 9,12.

The function of the two-stage model is to close the heat balance between the
bottom and top regions of the furnace and the furnace globally. This is accomplished by
calculating the heat generated by the sensible heat of the hot blast and by the
combustion of coke at the tuyeres. From this, the temperature difference between the gas
and solids can be calculated. This calculation is then made for the upper section of the
furnace. As shown by Table 5, some assumptions are made as to the locations of various
reactions. These assumptions can be validated by ensuring that the second law of
thermal dynamics is followed (i.e., gas temperature must be higher than solids).

Several of these models are available with varying degrees of output complexity.
One example is the Carbon Direct Reduction Rate (CDRR) model (figure 23) which
provides an estimate of the optimal furnace operating point 13 . In this model the fuel rate
is plotted against the direct reduction rate. The heat boundary (left-hand side) is based
on the heat transfer requirements from the gas and the chemical boundary (right side) is
determined by the amount of gas required for Wustite reduction. The point of i

intersection ofthe two lines is the minimum fuel rate required to support the operationl4.
i
To place the operating point on the diagram, a two-stage heat and mass balance is
calculated 16 (See Measurement Systems - Mass Spec, Moisture, Load Cells, Venturi).
As with many two-stage balances, more than one set of assumptions is used to calculate 1

the heat and mass balance. The CDRR model takes three variables; top gas analysis,
coke rate, and wind rate and uses two of these to calculate the third 15 . It does this for
each combination and then will either plot one selected or all three points from each 1

calculation on the diagram. If no instrumentation problems exist, all three calculations

10-12

, i
 lie close to each other. Table 6 provides a balanced
wil yield the same results and will

output example for one operating day. If instruentation problems exist then the three
calculations will yield different results. This cross check is a good tool for
troubleshooting the process, and establishing the required level of data quality. Figure
24 shows the misalignment in operating points due to data quality.

Another common feature of these models, is the calculation of an excess heat term.
This term typically represents the amount of heat required to superheat the hot metal.
An example is the IRSID calculation of "Wu" (thermal condition of furnace
superheating required) i I. As shown in figure 25, "Wu" when used as a control variable
to predict the silicon content of the hot metal was very successfuL. Despite this success,
use of calculated values such as "Wu" and IRM's "Ec" must be used with other models,
such as gas flow and kinetic models, to obtain full benefit 16 .

Still others, use the output of the mass and energy balance in a more direct format.
BHP's model "HBM2" produces a predicted hot metal temperature from its mass and
energy balance. A much more tangible term for most operators 17.

Another major characteristic of staged models is the calculation of the direct

reduction rate of the furnace. In most cases, the calculation is a variation of the following
, I
equation from US Steel's carbon direct reduction equation ~°.

C DR = CCoke - CFlue Dust - CMetalIoid Reaction - CBumed - C Hot Metal (4)

Two-stage energy and mass balance models are primarily used in evaluating
furnace performance and identifying faulty measurement systems. They are standard
models now for all blast furnace operations and are the basis for many higher level on-
line control systems. In all of the above cases, the two stage model began as an off-line
tool and was later rewritten to offer on-line furnace functionality.

Burden Distribution and Gas Flow Models

Burden Distribution

As early as 1850 Ironmakers recognized that certain fillng methods and material
types placed fines in the centre and promoted wall working 7. It is well known that
successful furnace operation depends on intimate contact of the reducing gas and burden
solids. The objective of all operators is to get ideal gas flow that includes 12,18 ;

1) High central gas flow to ensure furnace movement.

2) Ideal side wall flow to both reduce wall accretions and heat losses.
3) Good intermediate gas flow to maximize furnace efficiency.

1 0-13
 Today, most facilities have either developed or purchased models that simulate the
filling methods used for their fuaces.

In all burden distribution models, whether bell or bell-less type, several

assumptions are made about the raw materials. These include properties such as bulk
density, angle of repose, size fraction, discharge velocities, and shape factor. The goal of
these models is to establish the optimal ore to coke thickness that allows maximum gas
utilization and uninterrpted burden movement with minimum pressure drop.

Figure 26 is a good representation of a burden distribution model output 19. These

models give an indication of the amount of material located in the furnace relative to the
wall and centre. As the angle of trajectory, burden type and discharge times are
changed, a prediction as to the resulting burden profile is generated. Many of today's
models predict burden profiles that are remarkably close to measured values as shown in
figure 27 20 . The power of these models is greatly enhanced if validation is possible by
actual measurement of the burden profile (See Measurement Systems - Profilometer).

The importance of these models cannot be underestimated. When used with gas
distribution models, the impact of raw material changes can be predicted before
implementation. As all blast furnace operators know, if burden distribution is impaired,
production rates wil decrease, furnace operating problems wil occur, and
environmental problems may also be generated.

Gas Distribution Models

The most important process function of the blast furnace is to efficiently get the
reducing gas to contact the solids. To optimize furnace performance, researchers have
for several years probed and simulated the flow of gases and several models have been
created. Included in this category are stack gas distribution, cohesive zone, and raceway
models.

Stack gas distribution models predict a furnace's behaviour when either raw
material and/or furnace tuyere parameters are adjusted. Good gas flow models consider
the output from the burden distribution model in their assessment. In most cases, the gas
flow model determines the axial gas flow characteristics of the fuace. Most gas
distribution models are based on the Ergun equation for packed bed reactors, which
reads as follows 21.22 :

M 1. 75 * ( ~ - & J * G2 I
L & DeØ Pg
(5 )
1

1 0-14
 where:

dP = Pressure Drop
L = Length of the column or packed bed.
a = Void Fraction
De = Equivalent paricle diameter.
N = Shape Factor
G = Gas Flow Rate
Dg = Gas Density

If the burden properties are kept constant, the equation could take the form:

G2
M = k* (6)
Pg

As ilustrated by equation 6, gas density is inversely proportional to the pressure

drop. As top pressure increases, differential pressure decreases and the bosh gas wil get
more dense. Because of the gas wil increase and productivity
this, the mass flow rate of

will increase.

Researchers have used the basis of this equation to model the entire furnace by
breaking it into small patches or meshes on several levels and then calculating the gas
flows and heat transfer between meshes 23. A graphical output of one of these models is
shown in figure 2824. These models are extremely computer intensive and were first used
in an off-line capacity. However, with today's computer technology, these models are
beginning to see more on-line application.

Other researchers have also included to a certain extent the radial aspects of both
gas and heat transfer results in their modellng. By analyzing the radial gas profile (See
Measurement Systems - Above Burden and In Burden Gas Probe, ) either in the burden
or above it, an estimate of the gas distribution can be made and the radial heat and mass
flows can be estimated. One example is the "Model Super", which divides the furnace
into six circumferential or parial furnaces and calculates the inter-furnace reactions of
gas and heat transfer 25.

Top gas temperature distributions have also been used to track the evolution of the
gas distribution in the shaft 26. Polynomial approximations of top temperatures are used
to measure changes in gas flow patterns in the centre of the burden and at the walls.
These indices, centre flow and wall flow, can be plotted over time (see figure 29) to
help detect process disturbances and / or validate burden distribution control actions.

10-15
 Cohesive zone models have also been developed by several companies. One
example is that of figure 30 which gives reasonable correlation between actual pressure
readings (See Measurement Systems - Level and Pressure J and predicted mathematical
results 25. The prediction algorithm is based on the amount of coke charged to the
centre of the furnace.

This has also been shown by BHP , who use their RABIT model (originally
developed by NIPPON steel) with their cohesive zone model to predict the impact of
changes in burden distribution on gas flow 27,28. In these simulations, all other variables
other than burden distribution are held constant. The model divides the furnace into 14
grids with 65 levels. As in the other model previously discussed, the gas mass flow
and
heat transfer are then calculated between meshes. Figure 31 a shows the theorized burden
distribution and figure 31 b its effect on cohesive zone properties such as temperature
and CO utilization as calculated by the modeL. Figure 32a shows the predicted CO
utilization and temperature profiles for both a "V" and "W" profile. The furnace was
actually operating in a region between the two profiles and the measured process data
can be found in figure 32b. Figure 33 shows the results of the model versus the vertical
probe results. In all cases good correlation was experienced.

Most cohesive zone models assume that softening begins at 1200° C, melting at
1400° C, and superheating is done up to 1500° C. These temperatures are used to
develop isotherms in the model to provide a "melting line" and an indication of furnace
thermal changes to the operator.

Raceway models are another important gas flow model used in fuace
assessment. Raceway modellng was first attempted in 1952 by Ellot et al who used
wood to simulate the raceway response and high speed photography to record the results
30. Poveromo et al combined theoretical work and actual blast furnace results to produce
a very good mathematical model of the raceway penetration 31.

D
APr WHN
(38.9 Q2 TbJ %0
(7 )

where;
D= depth of the raceway in inches.
Q = wind rate in SCFM.
Tb = blast temperature in ° R
Pr = raceway pressure = 2/3 (Pb1as' -Piop) + Piop
A = cross sectional area of the tuyere opening in square inches.
W = average bulk density of burden in pounds per cubic foot.

1 0-16
 H = vertical distance from tuyere to stockline level in feet.
N = number of tuyeres.
(See Measurement Systems - Tuyere Probes)

As shown in equation 7, raceway penetration is dependent on the kinetic energy of

the blast as defined by Q and Tb. As shown in figure 34, good correlation was obtained
between several different operating blast furnaces and predicted model results. Later
research examined the impact of tuyere parameter changes on raceway profie and
penetration. In this testing dry ice was used to represent gas flow and inert particles such
as beans and sand were used to simulate particle flow 32, 33, 34.

By combining the raceway, cohesive zone, gas flow, and burden distribution
models it is possible to develop a fully dynamic model of the modern blast furnace. This
can be a valuable analysis tool when evaluating the impact of proposed raw material or
injection practice changes on the blast furnace operation. However, the developer must
be aware of the assumptions and data used to validate the models to ensure they are not
used outside their valid limits.

Predictive Control
For many years the goal of researchers has been to develop reliable models that
accurately predict the thermal state of the fuace. Operators would prefer a model that
was real time and could predict actual furnace performance. Today there are essentially
three types of models that do this fuction, statistical, thermal dynamic, and reaction
kinetics.

Statistical

Since the 1950's, several attempts have been made at furnace control using
statistical based techniques. Probably one of the most famous, was Flint's multiple
regression work in the 1950's that forms the basis of many quick furnace calculations
that are used today 8. However, Flint's factors were empirical relationships that needed
tuning for each installation. Due to the limitation of the available computer technology,
they were not ready to be used in an on-line capacity.

Attempts at using on line mass balances to predict blast furnace thermal conditions
were made as early as 194235. In the 1960's the first successful control systems were
variable
introduced using this type of technology i 1,12. Some of these balances used multi

regression techniques on such parameters as silicon, sulphur, hot metal temperature, and
wind rate to predict the movement in furnace thermal condition 35. An example of
control predictions based on these models is shown in figure 35. However, due to the
time delay of the process, the confdence in these models was limited.

10-17
 In the 1970's, other statistical techniques were attempted that considered the time
dependency or auto correlation of the process. Time related regression analysis
techniques such as the ARIMA (auto regressive integrated moving average) and ARV
(auto regressive vector) were employed 36, 37 . These models considered process auto
correlàtion by using historical readings in the preparation of the predicted varable. In
most cases, readings 3 to 4 hours old were used to help calculate the next predicted
value. Some techniques such as the DDS (Dynamic Data System) method use transfer
functions to adjust discrete measurements, such as silicon, manganese and hot metal
temperature, into the continuous domain enabling discrete control of the continuous
blast furace process 37 .

These early technologies have been furthered developed and are stil in use today.
Figure 36 illustrates the results of an ARV model in use at Rautaruukki, which
calculates every 5 minutes and stays 25 calculations ahead of the actual process
measurements. The plot of actual (solid lines) versus predicted (boxes) silicon shows very
good correlation 38.

Presently, newer techniques such as Principal Component Analysis (PCA) and

Parial Least Squares have been developed and used in several industries. These
statistical techniques condense several data points into one or two variables that can be
controlled with standard SPC techniques. The use of this technique has been limited in
blast furnace applications, but as computers become more powerful, more uses of this
technique wil develop 39.

It should be stressed that statistical models are only as powerful as the data
supplied and outside the range of the analyzed data no longer valid. As pointed out by
Thompson and Bowman, regression models can only be applicable in good quality
burden situations where the fuace operation can stay within historical values 40. If raw
material changes are made, tuning of the model is required to maintain its integrity.

Thermodynamic Prediction Models

Thermodynamic models use inferences between chemical reactions to predict hot

metal chemistry and temperature. One example is the model by Ponghis et al that uses
the chemical activities of various hot metal and slag components to predict the hot metal
manganese, silicon, sulphur, and carbon 41. As shown in figure 37, predicted values
mirrored the actual hot metal composition quite welL.

Reaction Kinetic Prediction Models

Kinetic models try to predict fuace parameters such as the furnace melting line
(cohesive zone). They use selected process data inputs and apply control loop strategies
using Kalman fiters (an electronic technique that compensates for both measurement
and model error). In essence they try to control the furnace like an analog control loop42.:

1 0-18

I
 In many cases certain assumptions such as axial gas flow only are used to allow quicker
calculations. Over the last few years these models have seen greater use as an on-line
tool with the improvement in computer technology.

All these types of predictive models, whether statistical, thermodynamic, or

kinetically based, use inferences from other measurements or process calculations to
predict the desired control parameters.

Auxilary Systems

There are several other models used today to help operators in the evaluation of
what could be called auxiliary furnace systems.

Hearth area monitoring includes drainage or tapping models that determine the best
method and timing for maintaining optimal furnace liquid levels for both slag and hot
metal 43, 44. Most of these models are tied into operator guidance systems for feedback
control. Other Hearth models monitor refractory condition by using either one or two
dimensional heat transfer calculations to determine hearth lining thickness and skull
buildup 45, 46, 47.

Most facilities have a fuace heatloss model either based on inwall thermocouples
or water flow and temperatue measurements (See Measurement Systems - Mass flow
meters, thermocouples). From this data, changes in burden distribution patterns or
scaffolding/peeling problems can be identified.

Several models have been developed to simulate stove operation. In most cases
these models break-up the stove along its height and calculate the heat transfer between
equal size layers (See Measurement Systems - Temperature, Optical Pyrometer). The
model goes through several iterations to reach steady state and then determine the final
profile. These models are used to evaluate fuel efficiency (See Measurement Systems -
Oxygen sensors) and hot blast strategy changes on stove performance.
In several cases, energy management models have also been produced that
determine and monitor plant wide energy consumption. These models are extremely
valuable in determining the economic impact of fuel changes on the plant.

CONTROL STRATEGIES
As shown in the aforementioned, there are several models, calculations,
measurements, and predictions available for today's furnace operator. Consequently,
information overload for the operator is a major concern. To address this problem,
engineers have developed control strategies that can be placed in two categories,
"Strategic" and "Process".

1 0-19
 Strategic Control

Strategic control is a long or medium term method of establishing the most

economical scenario to operate the ironmaking plant. This could be considered the
setpoint or deterministic control of the overall process of making iron. Generally
speaking, these control strategies are centred on fuel rate and material quality. For
instance, a new pellet type is proposed by purchasing to replace an existing materiaL. A
heat and energy balance, burden distribution, and gas flow model would be executed to
determine the impact of the new material on the fuel rate and furnace operation. If
acceptable, the new material would be purchased and become part of the plant operating
strategy.

Usually, this type of analysis is validated by a plant test of the materiaL. However,
most operations want models that are robust enough to give an accurate assessment of a
material's impact without the undue cost and time involved with a detailed triaL. This
can only be done if the model can freely and accurately associate material quality
aspects with its outcomes. Furnace control at this level is normally the responsibility of
the plant managers.

Perhaps one of the most comprehensive examples of this type of system is the
NICE system (Nippon Steel Ironmaking Control and Data Exchange System) 49. As
shown in figure 38, data from all plants is fed into a control computer system. At the
corporate level, daily data is available for analysis in proposed strategic changes. This
type of structure allows senior management quick access to models and results that wil
help reduce costs.

Most plants have some sort of evaluation model for management. However, the
success of the control strategy is as good as the data provided to the model from both
the field sensors and the management team. Large overview models like this can be
subject to some very inaccurate conclusions if input data is not rigorously screened.

Process Control i

Process control is really the random error portion or stochastic control of the blast
i
furace. Once the plant "setpoint" is established, operators are then responsible to run
the plant at the most cost effective rate possible. To accomplish this, today's process
control methods can be classified as "operating practice" based and "knowledge i
systems" based.

Operating Practice Based I

Figure 39 is an example of a simple operating practice based control system. In this

method, the control variable is the measured hot metal temperature. Standard statistical I

process control rules are applied to the reading when it goes outside the control limits. :
The operator actions are predetermined and shown on an accompanying decision
1

10-20

I
 making flow char. Several items in the control chart would have their own standard
practices associated with them. Similar control strategies are applied to other key
process variables (KPV's) such as slag basicity, manganese, and sulphur.

Similar variations of this theme are common in North America. Although simple,
they do not limit blast furnace production capability. Arco Middletown works, one of
the most productive plants in the world, uses a simple steam control strategy based on
hot metal temperature 50.

The big disadvantage of this method of control is that it targets only a few of the
process parameters involved. They also are reactive in nature and do not anticipate
furnace cooling or heating trends. As a result, the overall impact is greater process
variability with that comes higher fuel rates (i.e. most operators carr more insurance
with their setpoint aims) and on average, higher processing costs for a given operating
point.

Knowledge Based Control Systems

As process models developed, different control strategies based on .Jess tangible

parameters such as calculated fuel rate were developed. By the mid sixties,.these control
systems were common in both Europe and Asia. The original Nippon Kokan control
system, which used the output of calculated carbon rate as the control variable and
adjusted steam injection to control the total heat demand, is but one example 51.
However, as previously stated with the increase in complexity, a need for enhanced
decision making capability became apparent.

As a result, Operator Guidance Systems (OGS) were developed using Knowledge

Based or Artificial Intelligence (AI) control systems. AI systems consist of Expert
systems, Fuzzy Logic Controllers and Neural networks. These systems have been in
existence since the late seventies and are radically expanding in several industries.
Knowledge based systems provide the operator with control recommendations based on
programmed response to measured process data.

The purpose of a Knowledge Based system is to 52,53:

I) Preserve the experience base of the plant.

2) Allow higher level decision making at a lower level of process expertise.
3) Optimize the process.
4) Enable decision making to be more automatic and consistent.
5) To reason heuristically (by discovery), to allow qualitative assessment of empirical
data.

10-21
 Before developing a Knowledge Based system, certain criteria should be met 54,55 :

1) The problem is complicated enough to warant a heuristic control system.

2) A real expert is available.
3) The end users are interested in the system. Is there "BUY IN" ?
4) The system is cost effective.
5) The system fits in with the current computer structure.
6) System development time is reasonable.

Once these questions are answered, then picking the correct AI tool is needed.

Expert Systems (ES) are one of the most basic AI type systems available today.
Essentially they are written to try to mimic the thought process of the best possible
operators "the expert". They consist of "IF - THEN" logic programmed in such a way to
lead an operator through a series of operating data points to a solution. Expert systems
have four basic components 56:

Knowledge Base - this is a data file of previous experiences and what the
outcomes were following reactions to situations. These experiences can be
captured on-line as process knowledge develops or by triaL. The trial method
requires that the time, magnitude and source of a known process disturbance be
correlated with the effect of the disturbance. This dynamic testing method has been
used by British Steel to reduce thermal and production variation 57 .

Rule Base - this is the logic pattern that the operator followed in developing his
course of actions when troubleshooting the problems.

Inference Engine - this is the software that connects the rule and knowledge bases
together.

Operator Interface - the medium by which the information is communicated

between the operator and the expert system.

The development of an expert system is extremely tricky and requires the

assembly of the correct people:

· Knowledge engineer - to develop the software.

· Process expert - understands both the long and short-term control aspects of the
process.
· Operations expert - knows the effect of process moves and their success rate.
· Systems engineer - knows the computer system structure and capability.

10-22
 Once the correct team is assembled the fun has just begun. In an effort to capture
all related knowledge a rigorous process must be followed. Sollac reported that a total
of 6100 hours of knowledge capitalization was required to develop the SACHEM 58
system from documentation of experience to tuning the data-base with plant data.

An example of a simple expert system is shown below. In this system, a Socratic

approach known as "forward chaining" is applied. The stove operator answers a senes of
questions that have been designed based on the normal operating charactenstics of the
stove system. The expert system examines the input data to decide if the stoves are
operating normally. In this case, the time on blast is judged to be too low, and remedial
action is given.

Input:

What is the time on blast of the three stoves? - 55 minutes.

What is the range of the time on blast? 3 minutes.
What is the average wind rate? 75, 000 scfm.
What is the aim hot blast temperature? 1750 0 F
What is the present cold blast temperature? 300 0 F
Has the mixer positon been checked? Yes.
What is the diferential temperature between check and control ? 18 0 F

Output:

The control temperature is reading lower than the check temperature. The lower
time on blast indicates that the check thermocouple is correct. Change control to
the check thermocouple and re-evaluate operation in three stove cycles.

Perhaps one of the first and most famous ES is the Kawasaki GO-STOP system.
Developed in 1977, the first prototype features the recommendation to operators to
adjust furnace operating conditions based on eight indices 59:

· Total pressure drop.

· Pressure drop in the furnace shaft.
· Change in burden descent speed.
. Top gas temperature
· Shaft gas efficiency.
. Shaft wall temperatures.
· Thermal state of the furnace.
· Slag residual heat in the hearth.

By analyzing the absolute values, variation, and the combined effects of these
variables, an evaluation of the furnace is made. Based on this analysis, a
recommendation is made to the operator that typically involves a wind cut or a change:

10- 23
 in the ore to coke ratio. In the mid-eighties, the original system was upgraded to
increase the robustness of this control strategy by allowing the operators to adjust even
more parameters.

There are several other expert systems available and all have varied on the same
theme 60 to 68. Their application range from suggesting appropriate measures for furnace
cooling trends to identifying or predicting furnace disturbances such as slips and
changes in furnace gas flow.

Fuzzy logic is used in cases where simple "IF - THEN" logic is insufficient to
adequately control the furnace. Some conditions such at looking into the tuyere and
judging the reaction of the coke by its brightness is operator based and can't be
numerically identified. As a result, when a mathematical model assigns a value to this
"judgement" a certain amount offuzziness or error will be associated. To understand the
theory of fuzzy logic a brief example is required 69.

If given a set of numbers such that:

A = 0,2,3,4,5,6, 7, 8, 9, 10, 11, 12)

and we were then asked to identify all the prime numbers in this set then;

B = ~ 2,3,5, 7, 11)

Clearly a precise solution exists to this condition. Ifit was asked that we define the
set of "small numbers" from set "A", such a clear solution would not exist. What could
be stated is that the number 1 is definitely the smallest number and therefore is a
member of the sub set "small numbers". The certainty value (CV) of one could be
assigned to this condition. Similarly, number 12 is the largest number and clearly not a
member of the subset and is assigned the CV of zero. The remaining numbers can be
arbitrarily assigned any value between zero and one based on the programer's or
expert's experience. The resulting values identified with each number produce what is
called a "membership" fuction. Graphically, this is shown figure 40.

In practice, several different parameters are evaluated and assigned their own CV's.
From these, an overall certainty value of an event or condition can be concluded. One
example is shown in figure 41 that depicts the slip index certainty based on several
other "fuzy logic" determinations 70. This type of logic has been used for things like
stove control, furnace heat control, sensor evaluation, and burden distribution
diagnoses7!. Success rates as high as 97%, have been achieved using this type of control
logic in predicting the effects of furnace changes 72.

When the number of parameters becomes too excessive, evaluation by both Fuzz
logic and ES becomes unmanageable. Consider for example the application of a 6-point:
radial gas probe, along with a vertical probe with 14 measuring points all measuring

10-24
 CO, COl, Hi, and temperature. This would give the possibility of over 2500 outcomes
alone for the vertical probing. When this is combined with the outcome of the radial gas
profile, interpretation of the data becomes very complex. To handle this volume of data,
neural networks have been developed.

Figure 42 shows a neuron and a neural network 73. Several inputs go into the
neuron, each with its own weighting from 0 to 1.0 (fuzz application), a transfer
function (equation 8 75 typical example) is performed on the inputs and an output
produced. There can be several layers in such a network, each feeding a higher level of
reasoning.

1
=
f ( I)
l+e-I (8)

To tune these models two constants are typically used, the learning rate and
momentum constant. The learning rate is dependent upon the amount of error that is
acceptable for the model output and is gauged by the number of iterations required to
come to convergence. The momentum constant indicates the amount of weighting that is
applied to the previous calculated result. When the tuning parameters are selected, a
learning set based on actual operating data is used to teach the modeL. Upon "training"
completion, the model can then be used in the operating environment.

These types of control strategies have been applied off-line in silicon control loops
and have been developed for on-line control strategies such as burden distribution
pattern recognition and control 74. Once again the output of these systems are
recommendations to the operator of the condition present and the remedial course of
action to be executed. Figure 43 shows the logic flow of one such network used to
evaluate gas flow parameters.

Economic Considerations for Control Strategies

Obviously far more options exist for blast furnace control than ever before. To
help determine the most cost effective control solution, we can again refer to figure one.
While reviewing the figure, it becomes apparent that selection of a control strategy is
, ¡ plant specific. Such factors as cost of currency, governent subsidies, raw material cost,
and training requirements all figure into the equation. If a plant has very good raw
i
materials, good process measurement, and well-trained operators it probably makes
sense for them to spend less time and money in model development and advanced
process control techniques. If, however, a plant has an ageing workforce and wil
experience a "brain drain" of talent, investment in on-line process models and expert
systems would probably make sense.

Investment in expert systems, from implementation of primar models, to:

development of the expert shell, will probably cost between two and five milion
10-25
 dollars, depending on the software and sensor requirements. Although this may sound
steep, a fuel savings of about 18 to 45 pounds of carbon per net ton of hot metal, for a
5000-ton operation, would pay this offin one year.

The challenge to all furnace operators is to establish a process control strategy that
maintains consistent operation but is also cost effective. By manipulating all the
parameters found in figure 1, it is possible to devise a control strategy based on the
strengths and weaknesses of a paricular facility.

CONCLUSIONS

The improvements in both measurement systems and modelling are driven by the
need to understand what is happening inside the blast furnace to improve process
stability and product quality.

As improvements to measurement system accuracy and robustness develop so does

the ability to model and control the blast furnace. However, regardless of the control
strategy, the biggest potential gain is through the application of what we learn through
measurement and modellng. If we canot apply what we lear in a consistent and
repeatable maner there is no true measurable gain, for the company or for the operator.

Increased computing power has made it possible to provide huge amounts of

sensor and modelled information to the operator on-line. The need to manage the
massive amount of information has resulted in advanced control techniques such as
Expert Systems. It is important to note two things; these systems are only as good as the
data provided by the measurement systems and the information provided by the
operators, and that at the end of the day the people are the experts - not the system !
Wisdom, common sense, and attention to detail will always be the operator's greatest
asset when taming the giant reducing machine called the blast furnace.

REFERENCES

I) J. P. DiCarlo "Fundamentals of Flow Measurement". Instrument Society of America, Research

triangle Park, NC, 1984, pp 158,211,,213,214,228.

2) Omega Energy Corporation, "Volume 29 - Complete Temperature Measurement Handbook and

Encyclopaedia", Stamford CT, 1995, pp Z13, 226

3) B. G. Liptak et aI, "Instrumentation Engineering Handbook", Chilton Book Company, Radnor

Penn's., 1982, pp 504.

4) N. Konno et aI, "A Study of Phenomena Inside Blast Furnaces Using Vertical Probes", AIME_
lronmaking Proceedings, Volume 45, 1986, pp 43 I to 439.

10-26
 5) K. Tamura et ai, "Studies on the Inside Phenomena of Blast Furnace by Newly Developed
Sensors", AIME Ironmaking Proceedings, Volume 43, 1984, pp 407 to 414.

6) A. Poos, "Blast Furnace Control - Strategy". Blast Furnace Ironmaking Course, Lu, W. K.,
editor, McMaster University, Hamilton, Canada, 1994, pp 12- I to 12-41.

7) J. A. Ricketts, "In search of Ancient Iron: Ironmaking in 1870's vrs. 1980's", AI

ME
Ironmaking Proceedings, Volume 50, 1991, pp 39 to 60.

8) V. Flint, "Effect of Burden Mass and Practices on Blast Furnace Coke Rate", AISI Regional
Technology Meeting, Chicago, 1961.

9) 1. G. Peacey and W. G. Davenport, "The Iron Blast Furnace", Pergamon Press, Elmsfort New
York and Oxford U.K., 1979.

10) T. W. Oshnock et ai, "Utilzation of a Daily Material and Energy Balance Program to
Evaluate and Improve Blast Furnace Performance", AIME Ironmaking Proceedings, Volume
42, 1983, pp 427 t0435.

1 I) C. Staib et ai, "Blast Furnace Dynamic Behaviour and Automatic Control", AIME Ironmaking
Proceedings, Volume 26, 1967, pp66t088.

12) A. K. Biswas, "Principles of Blast Furnace Ironmaking", Cootha Publishing House, Brisbane
Australia, 1981, page 8 and page 170.

13) S. Haimi et ai, "Advances in the Application of Blast Furnace Control Models at
Rautaruukki's Raahe Blast Furnaces", AIME Ironmaking Proceedings, Volume 51, 1992, pp
149to 158.

14) K. H. Peters et ai, "Influencing the Fuel Consumption and Productivity of Blast Furnaces by
the Use of Models", AIME Ironmaking Proceedings, Volume 47, 1988, pp 545 to 564.

15) R. 1. Donaldson et ai, "Blast Furnace Process Modellng and Sensor Applications at Dofasco",
AIME Ironmaking Proceedings, Volume 50, 1991, pp 733 to 736.

16) J. M. van Langren et ai, "Burden Preparation and Computer Control of the Blast Furnace",
AIME lronmaking Proceedings, Volume 26, 1967, pp326t0335.

17) P. Goldsworthy et ai, "Application of an On-line Mass and Energy Balance to Blast Furnace
Thermal Control", AIME Ironmaking Proceedings, Volume 5 I, 1992, pp 159to 162.

18) A. S. Harshaw and R. R. Schat, "Two Bell Charge Flexibilty at Dofasco", AIME Ironmaking
Proceedings, Volume 42, 1983, pp 511 to 521.

19) J. J Poveromo, "Blast Furnace Burden Distribution Fundamentals", Iron and Steelmaker,
November 1995.

20) S. Inaba, "The Control Techniques of Burden Distribution in Blast Furnaces at Kobe Steel,
LTD.", AIME Ironmaking Proceedings, Volume 42,1983, pp 503 to 509.

21) S. Ergun, "Fluid Flow Through Packed Columns", Chemical Engineering Program, Volume 48,
1952, p 59.

10-27
 22) J. J. Poveromo, "Blast Furnace Burden Distribution Fundamentals", Iron and Steelmaker,
December 1995.

23) Y. Sawa et aI, "Mathematical Modellng of Blast Furnace Characteristics by the Precise
Layer Structure in Stock Column", AIME Ironmaking Proceedings, Volume 50, 199 I, pp 4 I 7 to
423.

24) Y. Otsuka, "A Two-Dimensional Simulation Program to Analyze the Inner State of Blast
Furnaces", AIME lronmaking Proceedings, Volume 47, 1988, pp 305 to 310.

25) G. Danloy and C. Stolz, "Shape and Position of the Cohesive Zone in the Blast Furnace",
AIME lronmaking Proceedings, Volume 50, 199 i, pp 395 t0401.

26) H. Saxen, "Interpretation of Probe Temperatures in the Blast Furnace using Polynomial
Approximations", Steel Research, Volume 67, No.3, 1996.

27) P. D. Burke and J. M. Burgess, " A Coupled Gas and Solid Flow, Heat Transfer and Chemical
Reaction Rate Model for the Ironmaking Blast Furnace", AIME lronmaking Proceedings,
Volume 48, 1989, pp 773 to 78 I.

28) K. L. Hockings et aI, "Application of RABIT Burden Distribution Model to BHP Blast
Furnaces", AIME lronmaking Proceedings, Volume 47,1988, pp289t0296.

29) S. Wakuri et aI, "Development of Cohesive Zone Control System for #2 Blast Furnace at Oita
Works", AIME lronmaking Proceedings, Volume 40, 1981, pp i 12 to 120.

30) Ellot et aI, "Physical Conditions in the Combustion and Smelting Zones of a Blast Furnace",
Transactions AIME, Journal of Metals pp 709 to 7 i 7, July 1952.

3 i) 1. 1. Poveromo et aI, "An Experimental Measurement of Raceway Dimensions in Bethlehem

Steel Corporations Bethlehem, PA. Plant", AIME Ironmaking Proceedings, Volume 35, 1975,
pp 383 to 65.

32) K. A. Fenech, " Attempts to Model Physically the Formation and Structure of the Raceway of
the Iron Blast Furnace", AIME lronmaking Proceedings, Volume 47, 1988, pp 26 i to 267.

33) M. Hatano et aI, "Aerodynamic Study on Raceway in Blast Furnace", AIME Ironmaking
Proceedings, Volume 42, 1983, pp 577 to 586.

34) C. Yamagata et aI, "Fundamental Study on the Internal State of the Blast Furnace at High
Pulverized Coal Injection", Blast Furnace Ironmaking Symposium #19, Raceway Control for
Optimum Blast Furnace Performance, W. K. Lu, editor, McMaster University, Hamilton, Canada,
1991, pp 47 to 59.

35) A. N. Pokhevisnev et aI, "Blast Furnace Process Automation Control", AIME Ironmaking
Proceedings, Volume 3 I, 1972, pp 403 to 4 I I.

36) S. M. Pandit et aI, "Modellng, Prediction and Control of Blast Furnace Operation from
, I
Observed Data by Multivariable Time Series", AIME Ironmaking Proceedings, Volume 34,
1975, pp403t041 i.

37) A. A. Nayeb Hashemi et aI, " Blast Furnace Modelling and Control by the DDS Method", I

AIME lronmaking Proceedings, Volume 36, 1977, pp 2 to 297.

10-28
 38) H. Saxen and L. Karilainen, "Model for Short-Term Prediction of Silcon Content in the Blast
Furnace Process", AIME Ironmaking Proceedings, Volume 51, 1992, pp 185 to 191.

39) 1. V. Kresta et ai, "Multivariable Statistical Monitoring of Process Operating Performance",

The Canadian Journal of Chemical Engineering, Volume 69, February, 1991,pp 35to 47.

40) C. D. Thompson and R. W. Bowman, "Multiple Regression Analysis of Blast Furnace

Operating Data", AIME Ironmaking Proceedings, Volume 41, 1982, pp 372 to 509.

41) N. Ponghis et ai, "Limits and Constraints for the Production of Hot Metal Contents in Silcon,
Sulphur and Nitrogen", AIME Ironmaking Proceedings, Volume 50, 1991, pp 457t0464.

42) A. 1. Kilpinen, "On-line Estimation of the Melting Line", AIME lronmaking Proceedings,
Volume 48, 1989, pp 793 to 796.

43) K. Shibata et aI, "Control of Hot Metal Flow in Blast Furnace Hearth by Blasting and
Drainage", Blast Furnace Ironmaking Symposium #19, Raceway Control for Optimum Blast
Furnace Performance, W. K. Lu, editor, McMaster University, Hamilton, Canada, 1991, pp 118 to
140.

44) T. Fukutake, "The Flow of Slag and Metal in the Blast Furnace Hearth During Tapping",
AIME Ironmaking Proceedings, Volume 50, 1983, pp 567 to 574.

45) G. Kiniel and G. Kolb, "Hearth Lining - the Crucial Factor for the Campaign of Blast Furnace
"A", AIME lronmaking Proceedings, Volume 50, 1991, pp 287 to 292.

46) G. Leprince et ai, "P28 Hearth Brick Erosion Measurement and Modelling", Second European
lronmaking Congress.

47) G. Leprince et ai, "Blast Furnace Hearth Life: Models for Assessing the Hearth and
Understanding the Transient Thermal State", AIME Ironmaking Proceedings, Volume 52,
1993,pp 123to132.

48) 1. G. Mathieson et ai, "The Use of Sensing Techniques and Mathematical Models to Improve
Blast Furnace Performance", AIME Ironmaking Proceedings, Volume 48, 1989, pp 587 to 595.

49) H. Kanoshima et ai, "Nippon Steel Iron making Control and Data Exchange System", Nippon
Steel Technical Report, Number 29, April 1986, pp i to iI.

50) D. A. Wise and J. F. Blattner, "The Process Control Philosophy and Strategy of ARMCO Steel
Company", AIME Ironmaking Proceedings, Volume 50, 1991, pp 435 to 443.

51) Y. Fujii et ai, "Application of Automatic Blast Furnace Control at Nippon Kokan Mizue",
AIME lronmaking Proceedings, Volume 26,1967, pp 58t065.

52) K. Noderer and H. Henein, "A Survey on the Use of Expert Systems in the Iron and Steel
Industry", AIME lronmaking Proceedings, Volume 49, 1990, pp 497 to 501 T.
53) C. Forsberg et ai, "Real-Time Expert System for Blast Furnace Control", AIME Ironmaking
Proceedings, Volume 50,1991, pp 739to 745.

54) J. Walters, "Developing Knowledge-Based Applications: Factors for Success and Failure",_
AIME Ironmaking Proceedings, Volume 49, 1990, pp 491 to 495.

10-29
 55) H. J. Bachhofen et aI, "The Application of Modern Process Control Technology in Hot Metal
Production", AIME Ironmaking Proceedings, Volume 50, i 99 I, pp 703 to 708

56) R. J. Thierauf, "Effective Management Information Systems", Merrill Publishing Company,

Columbus Ohio, 1987, pp 58 to 65.

57) M.J. Hague, "Dynamic Testing and Modellng for Improved Thermal and Production Control
of a Blast Furnace", ICSTI / Ironmaking Conference Proceedings, 1998, pp 237 to 234.

58) J.M. Libralesso et aI, "The Blast Furnace under High Supervision", ICSTI / Ironmaking
Conference Proceedings, 1998, pp 231 to 236.

59) Nagai et ai, "Go-Stop System Applied to Blast Furnace Computer of CHIBA Works,
Kawasaki Steel Corporation", AIME lronmaking Proceedings, Volume 36, 1977, pp 326 to 335.

60) P. W. Warren, "Recent Developments in Blast Furnace Control Within British Steel", AIME
Ironmaking Proceedings, Volume 54, 1995, pp 28 I to 284.

6 I) W. Kowalski et aI, "Process Control Techniques at the Blast Furnaces of Thyssen Stahl
AG",AIME Ironmaking Proceedings, Volume 54, 1995, pp 23 I to239.

62) L. G. Lock Lee and A. R. McNamara, "A Review of Expert System Developments for Primary
Processing within BHP Australia",AIME Ironmaking Proceedings, Volume 49, 1990, pp 523 to
528.

63) L. Karilainen et aI, "Interactive Control of Blast Furnaces" , Steel Technology International,
1995/96, pp 54 to 60.

64) P. Inkala et aI, "Computer System for Controllng Blast Furnace Operations at Rautaruukki",
Iron and Steel Engineer, Vol 72, #8, August 1995, pp 44-48.

65) L. Karilainen et aI, " An Interactive System for Real Time Operator Aided Control of Blast
Furnace",AIME Ironmaking Proceedings, Volume 53,1994, pp 365to 370.

66) N. Ponghis et aI, "ACCES - Model for Blast Furnace Control", AIME Ironmaking Proceedings,
Volume 48,1989, pp 523t0527.

67) H. Rusila et aI, "Control Strategy of Rhetoric's Blast Furnaces", AIME Ironmaking
Proceedings, Volume 48, 1989, pp 529 to 532.

68) D. Vanderberghe et aI, "Process Control Techniques for the Sidmar Blast Furnaces", AIME
Ironmaking Proceedings, Volume 54,1995, pp 281 to 284

69) M. Stachowicz, "The Application of Fuzzy Modellng in Real Time Expert Systems for
Control",AIME Ironmaking Proceedings, Volume 49, 1990, pp 503 to 51 i.

70) S. Hirose et aI, "Application of AI Systems to Mizushima No.3 Blast Furnace", AIME
Iron making Proceedings, Volume 5 I, 1992, pp 163 to 170.

7 I) R. Nakajama et aI, "Operation Control System of Blast Furnace by Artificial Intellgence",

AIME Ironmaking Proceedings, Volume 46, 1987, pp 209 to 216.

10-30
72) Y. Tsunozaki, "An Expert System for Blast Furnace Control at Fukugama Works", Nippon
Kokan Technical Report, Overseas, No.5 I, 1987, pp I to 10.

73) M. Bramming et aI, " Development and Application of New Techniques for Blast Furnace
Process Control at SSAB Tunnplat, Lulea Works", AIME lronmaking Proceedings, Volume 54,
1995, pp 271 to 279.

74) O. Iida, "Application of AI Techniques to Blast Furnace Operations", Iron and Steel Engineer,
Oct 1995, pp 24 to 28.

75) Z. Guangquing, et ai, "A Neural Network Model for Predicting the Silcon Content of Hot
Metal at No.2 Blast Furnace of SSAB Lulea", AIME Ironmaking Proceedings, Volume 55,
1996, pp 21 I to 221.

10-31
 Flowmeter $election

Energy Additive Spelal Techniques

P I Tàgglng A P

::1fR ,:tV. Flow markers , :!1R :: Vi

'!.ViR :t V. . ÜserDoppler :tVzR :tV.
C9rrelålion ::2R .:tVi
..Sqiiare~edge,prlf~e ' ',~.O/(S"i'.,',
'Ecêeòltlc prifce " :iJ9lS "
!¡:'!Se""
. riiáortiee
,_,_ ,000'" ,'.' .." ' ,t2S,
'.~. 'Æ:iQUäifia,'. iit,ëdo.e, orifce , ,i2S A P A
: ~\íY.ntn 'I ,".~ :0/..5. -
! t:Nô~(é' .,., ¡ ,
:0/.$
i
Time-of-flight =1R :tV. Fluidic ::25 :t Vi
Dóppler :tlS :t V. Ion :t2R :t V.

A = system accuracy/uncertainty
O~n, Chinnel Fluid Dynamic
S .. percent of full scale accuracy

'Weirs :t3R :t:i. Vortex :t¥.R :tV.

.,Aumes :5R :t:i. Precessing vortex :t2R ':tV. R .. percent of actual flow rate accuracy

fluidic oscilator :t1S :t 'I.

l p, '" precision/repeatability

Table 1: Types of Flow Meters.

 TABLE 2

Thermocouple Types and Their Range

ANSI ALLOY COMBINATION MAXIMUM TEMPERATURE LIMITS W ERROR

CODE RANGE (Whichever is Greater
+ Lead - Lead Standard Special
J Fe Cu-Ni -21010 1200OL. -346 10 2193° 2.2OL orO.75% 1. OL or 0.4%
(Magnetic)
K Ni-Cr Ni-AI -27010 1372"' . -454 to 2501 ° 2,2OL or 0.75% 1. OL- or 0.4%
(Magnetic)
V Cu Cu-Ni o to 80 32 10 176°'

T Cu Cu-Ni -270 to 400~ . -454 to 752° 1.0 or 0.75% 0.5°C or 0.4%

E Ni-Cr Cu-Ni -270 to I OOOOL . -454 10 1832° i. 70L or 0.5% 1,0= or 0,4%

N Ni-Cr-Si Ni-Si-Mg -270 to 1300OL. -450 to 2372 2,2"' or 0.75% I. ¡= or 0.4%

R Pt-13% Rh Pi -50101768"'. -58103214° I.5OL or 0.25% 0,6 orO.I%

S Pt-IO% Rh Pt -50 to 1768~. -58 10 3214° !.5OL or 0,25% 0.6"' orO.I%
U. Cu Cu-Ni o to 50~32 10 122°

B Pt-30% Rh Pt-6% Rh o to I 8200L . 32 to 3308° Not established

W W W-26% Re o to 2320~ , 32 to 4208°' 4,5 to 4250L Not established.
1.0% to 23200C
WS W-5% Re W-26% Re o to 2320~. 32 to 420lF- 4.5 to 425~ Not established
i .0% to 23200C
W3 W-3% Re W-25% Re o to 23200L . 32 to 4208'" 4.5 to 425"' Not established
i .0% to 23200C

Reference - Volume 29 Ome~a Complete Temperature Measurement Handbook 1995 Stamford CT.

10-33
 TABLE 3: TYPICAL MATERIAL BALANCE DATA

Fe C O2 Si02 CaO MgO AhOJ Ti02 S K20 Na20 Mn "2

In 1896 837,8 1546 172.9 178.0 53. 38,0 7.8 7,8 1.9 1.4 14.5 1.0
Ibs/nt

Out 1790 841.6 1532 1720 177.9 53.0 38,0 8,) 8.2 2.2 2,1 14,5 1.0
Ibs/nt

Yield 94.4 100,5 99.1 99.5 99.9 99.3 100 104.2 105.2 116,6 151.4 100 100
%

Table 4: GLOBAL HEAT BALANCE

INPUT: ( MEGACALIMTHM) VALUE

Combustion of Carbon 587.
Sensible Heat of Blast 351.8

Slag Formation Heat 15. I

Reduction by CO 51.6

TOTAL 1005.8
OUTPUT: (MEGACAL/MTHM) VALUE
Sensible Heat of Top Gas 62.1

Sensible Heat of Hot Metal 300.9

Sensible Heat of Slag 99.5

Reduction of Si 21.
Reduction Mn 5.7

Reduction of P .9

Solution Loss 333.5

Fuel Decomposition .3

H20 Decomposition 4,0

Reduction by H2 15.3

Calcination Reactions 6.3

Carbon Solution in Hot Metal 22.3

Unaccounted Heat Loss (UHL) 134.5

TOTAL 1006.5

10-34
 Table 5: Two STAGED ENERGY BALANCE

PREP ARA TION ZONE PROCESSING ZONE

HEAT SUPPLY HEAT SUPPLY
(MEGACALITHM) (MEGACAL/THM)
Gases from Procssing Zone 550.4 Blast Enthalpy 388.0
Heat from Burden -22 Solids from Preparation Zone 381.
Heat from Coke -0.6 Injected Fuel Oil 0.9
TOTAL 547.6 Combustion at Tuyeres 594.8

PREPARA TION ZONE TOTAL 1364.8

HEAT DEMAND
PROCESSING ZONE
(MEGACALITHM)
REA T DEMAND
Heating of Solids 381. (MEGACAL/THM)
Reduction of Ore Oxides 11.9 Hot Meta Heating and Melting 2812,
Flue Dust 0.4 Slag Formation and Heating 72.1
Decomposition of Fe and Mg -0.4 Si, Mn, & P in Slag
Carbonates
303
Wustite Reduction 221.8
Heat of Vaporization of Moii.1ure 352
Decmposition of Carbonates -0.5
Top Gas Humidity 4.3
Si and P in Hot metal 5.0
Reduction ofCO¡and CaCO~ -0.2
Nees iit Tuyeres 323
Top Gas Enthalpy 61.2
Reduction ofCO¡ and CaCO" -02
Heat Losses 53.9
Limestone Calcination 5.8
TOTAL 547.6
Gases to Preparation Zone 550.4

Heat Losses 172.4

TOTAL 1364.8

TABLE 6: MATERIAL BALANCE DATA

VARIABLE COKE BLAST CO¡ CO H¡ ET ACO ETAH

Input Data 784.2 35386.3 21.8 22.5 2.8 49.2 52.0

,¡
I
ETA+CKE 784.2 34692.0 22.6 23.4 2.7 49.2 52.4

CKE + BLT 784,2 35032.5 22,8 22.9 2.7 49.9 52.0

BLT+ETA 788.1 35032.5 22.5 23.3 2,7 49.2 52,0

10-35
Figure 1- Mechanical Limit Switch

Figure 2 - Electrical Proximity Switch

10-36
 '" magnetic
~coi

Figure3 -Principle
" f0 Magnetic
. FI0 w Meter

,.. l ..fe, ..jì

-
-- _.
-:-.: _:l'
¡it -:.._..~......,
--.;:......9
.... ....~
-: -.: - - :: - ::::1) 1;:..- , FC1

~i
Fe.

Figure
. 4 - . f 0the Corio. i"
Principle iS Meter

10-37
Figure 9 - Mechanical Stockrod Installation

LOAO BUTTON

LOAO SUPPORt
COLUMN WITH
BONO€.O STRAtN
GAUG E 5

HERMtTlCALLY
SEAtED
GAUGE CAeL£: CONtECTlON
C~AliaER P'1p, fOR POWER
SOUflGE AMO
OUTPUT CONNE C T I Of

800Y

1f
;l
8ASE-..,
Figure 10 - Load Cell Construction

10-40
 Figure i 1- Above or In-Burden Retractable Probe - Paul Wurth Design.

~
Magnetomet e~
Col í ng vvter

Thormocooplc and Gas sampl jog pipe

61: "Trmocple,. k: Gas samplíng piPß
i
~ : Mågntómc to,-

Figure 12 - Fixed In-Burden Probe

i I

I
10-41

i I

i
 Optimum SF Perormance Requirements

Incing Co of impleg
~nnce Steg
--~~'!~~-~- - -- -- - -- -- - -- --- -- - - - -- -- - - ----
DM-i Oper-ng Range
-:--:ontl,:-: -.~

- --- - - - - - -- - - - -- - ---- :-:::':~7.:~õdelS:';.::. ~;::::

MInium
Reulnt fo
Stbl Openin

111lng Co
of Unatabk
Openron ~1Í~;~5~~~~g~tr~~¡
~
Fact lnfncng Owraa Furnce

Figure 17: Blast furnace performance and its contributing factors

The fron Blat Furnace

foiD, + C TOP GAS (298K)
(298K) CO, COi ,Ni

AIR (296 Kl--

(Oi,Ni) L

\
Far, C
(IBOOK)

Figure 18: Simplified Representation of Material and Energy Balance Concepts (9)

10-44
 Material Balance
Raw Materials i Top Gas
Flue Dust
Iron Bearg
Coke
Fluxes
Ivsc.
Blast Conditions
\
l
l
Wind
Hot Blast
Molten Material Moistue
Hot Metal Fuel
O:x-ygen
Slag

Figure 19: More Detailed Material Balance Concepts (10)

o
y= Fe

2 x=Q
u-r c

T
-l
Pi 1000

800

Fe
600

t E 400
COz
CO + COz

Figure 20: Operating Line (Rist Diagram) of the Blast Furnace (11)

10-45
 \"~'--i
j L ¡\
I 1,0",,1, I

~~O!~~I: Or°lSKJ
/
i c..oc:w:.:: iU.HlI..
r-" ()~.I,')~O..¡il aci)l

nlt'li-- "UI('Æ' i
f UtotNt
\ I
./
\. J (
jI ¡-'-",=~
¡i
\ ".:~~~:.
I ._", " a..~

l, ,,:::~..':~:~ l
\ (
i
~\- I
:K liJ J~ ~Ò
I -- ! iiw 210:i
Te:t.P'R~rUR€. I( Oft RAn:; IN GAS

Figure 21: Conceptual Division of the Blast Furnace - Physical Position (9)

20 iap G"~: 100 - !sa.c

.. 1"- -l(1 '4. f;ai" ~o ~l.:"J. C:J .. ~s.r hz

11 ii
=_~ ,,i
äs ê;:'à; " ..
16
_ _ ~ ,_ _ _~:; 'lJ:,'~o~._(;~'_¡:"~': :0,
~ i F~J()\ 1" CO: JFoi(l" ::1,1
14
- -:~-E:l~A-L - - M . :I::~~~; - i:ll~ - - - _. -

~- -,'w
., --¡ I,., ii
ilI Zi Š
I¡ QI..t
- 1______ r------
12

Dútae .. io ~:t I r'fO .. cO = F~ -+ CO i

Abov 10
TUN... ~ 2 i ~ ~ UlIflECUCEO F'tO
a: -i 0..
~ ~ ~ ~
8

6
~ ~ i
4
~-iL_m_
¡- ~ ;".-.;:l-.;;;u--
~; :
OIRECT _ AEDUc.lICM '\ ctC-9! ., t~Q 4. tOi
..0 MElfi NÇ lONE, ~I. '\ C(\ . e, . ''0
ÅND lP~ 1r1',~O .. ç"" N,'I . CO
2 , "
P1C$ ~ 5Ç . it,. s~.
~i(l:: . 2; :I ii . -ieo .. ..
5.C~O"C 5"CO
T UN..
500 1000 1500 2000
Tenmeratu °C

Figure 22: Furnace Zones - temperature and reactions (12)

10-46
 (l flAuHi:OUllo(KI :)y
A.""!;h.,IW.,i-:
BLAST FURNACE f
..'eek 4 f 199 r

(~ ' DRR DII;GRAM CO~E: .~a,2 ï

or.J ..~ II HlJ ,ii
4~) OiL ~5.1i Gi/IHI1
FUR MlE '-~:i k9/11111
opn1lUM ci7.2 i-g/I K.. I
""- ..-----,' -.'.'..'. -'.---
:E
1: flUS' VOLUME " 1. i(" )1 h !
::4C
"
'"
..
",~'.. Sìi: 1th
02- EIII1CHM.EIlT ....., km'/h ¡

n:WPEAnlRi; ioa~' C
~ /w OISTUni::U3 lj.l rn J
c
" HEATiOSS£S itJl GJI¡"
e 3S
-:J

3C
ZS .J M
Ojrc-cl;o rcd.JdiOlFak- ~

liar r.ET.\L 2112 iId

,Si-% 0,44 'I
Sl.~.il!c". 0.12 ~
H
., t
Hm""..tu.. 14 'C

Figure 23: CDRR (Carbon Direct Reduction Rate) Diagram (13)

-.i E
..
""
-.: C)

-lQ)co
c:
i:
a
..'-
u co t
opmum operating po1nt
(minImum fuel roe)

% Direct Reduction
I (1 oper.ting point ...uaing top g.. uti UzaUon and
coke r.te ere correct
I
(2) operating point ...uaing cok. rate end bl..t
conauaptlon .ra correct
(3) operating point ...uaing top ga. ut i 1 ization .nd
bl..t coneu-øtion ar. correct

Figure 24: Ilustration of CDRR model to detect Measurement Error (15)

10-47
 Wu
/------
1lrmrron Fe Q

f -\Ç
ic _.\/'-,/-_.V... ,- '...'\..v
%Si r~')
¡-30
L,c,
~-~._~--._~./
-, - ,.,...,
%S (: /' \. ",'".. ~. --../
"~,_/'..__.-i _.-..__.__._~
L:o)
Fuel Oil Rate t 2 ~

gIN
in 3 :i iS

Blat Moistu
3 20
glNin
r '0
'3
0" ~"1 ~. 5"
12'Ó.67 1:ó'ó7 . ',-\'ó7

Figure 25: Blast Furace Control Results using "Wu" Heat Index as Control Variable and
Steam Injection as Control Strategy (11)

0;

l-,
i
)~
-Ê..
!i
GI

-..
D
=-

Ii
II
~
..C.. -l,

..:¡.. R
Gmler Wil;

-t- so!- Sloifr +- l'i.kis

=-
Loo!

i
Cl:
"Eo
:t-
M~
·0
Cc.
.lo
8&: o
Ceiier WaJl
Radial Position
in Furnace

Figure 26: Burden Profile from Mathematical Model (19)

10-48
 o

C 2ui,~5
o ni; ~f7
-u
~
"

c~
'"
~ 2
'"

I
,
I
L-Obs-e r-ied
i,
¿.Est;mCl1f:d

..
Q

Dis1Qni;~ from ient~r (rnl

Figure 27: Comparison of Estimated and Actual Surface Profie (20)

a b c:

Inti mesh Clunologial Mesh Generated

lis an fl by use of
lis of soli c:lunologial
lis of soli

Figure 28: Mesh Generation for Gas Flow Models (24)

10-49
 t-556h
60

,,$
40 --_~_.-il' - .....
~..' .

A.
20
.r .. . ./
./~:.l++'- /"
00
....--

~'O__.f;
. IV""
0

-40
-20

0
a)

50 100
B 150 200 250
Ac
t-570h
60

40 .,,.+'" t,~'

... "
t,~". . _.~
~'
.+ '
~ 6"

20 /,fil.;:.B.._.Aè
A.
0

-20

-40
0
b)

50 100
B 150 200 250
Ac
Figure 29: Evolution of the Indices after Changes in the Charging Program (26)

;l~i~i.i l/l ~~ojM:n1 .~ì;lllrl ",'q~i:i~ liii~' fri\

l5

2( IIII I\
mr--'
::1 d!l ::; iil ::1 ~íi¡l
ID; r)ì. i! \ ,i... ',1'. .... "'.-d !Cr. i" I . \~
'.': ~-~
~i ',.~ ~ _~"'I 5. '~;~
)1 Jr" i i ,-1-.. .. 1--' Jr.-H
s '

.iI
:'1 I!i~"i!:'!
r-"
tJ. tN'~¡1
- ii ¡'~-~
l ! -l,-.'
i"(lO!', I
L'~J
~
-"-l :,

,i~glilnl
H~ig\1 hil

l5Î
i
HI'l~1 :r
;5¡-
.

./ '
r,t ,
HF;tfhi 1n.1

i~r .l~f
,

1( f

~~ ~

ffi
ID

l,
fl 2D tr
¡ 'ff4
· !II l ¡ ii
20!
,

,\\ æl
I.S

" I

1\
ii ; ,i
,. ~". .._~.'

I.n;il :
I

~
i
10

c'
~
"H=: 'Ir 'di=~
1--'
, I
..-' .1

\. .i-¡!
'
_'_w''-
'-l~'"
i
'...
., 1Ul1
i '\ . -' ~

'~
;
:1

:1
tt
' - ~' \

/B. i
10

I
1t

Figure 30: Comparison of calculated and measured cohesive zone profies (25)

10-50
 burden _r--

Tc¿t;-
/LJ
1/-'- /
/' COk//
/
////
Ce ntre Wall
Figure 31 a: Burden Distribution Profile for Predictions shown in 31 b (27)

l\
Ì' 1-)1,
¡§,
i 'i
liA~. ~.~~3

'I
'.... . \o.

! -~
i
i
0.9 ~)i
l- if
!
! ..1
!.. Î 1.0
~L1~
ì ..,~I\O'5

LJ
!
LJ
r r
gas solid degree of relative
temperature temperature reduction gas
pressure

Figure 31 b: Predicted impact of burden distribution profile on furnace operating

conditions (27)

10-51
 100
o,ô
800
'-'-.',' -ij-- ,-
co
"V" Profie 0-4 600
TJ
co
.., ',- Tern!)
Gas
Temp
400 ~:C)
(1,2

2ú
(a~
o o
Centre Wal
Radi Location
100
., _ .. J -""".
C.6
"w" Profie
0.4
- ~'- 7)
co
,-" ~ 8C

00:)
Gas
nco ~. -'-, T"mp Temp
400 tC)
0.2
20
lbJ
o o
Centre Wal
Radi Location
Figure 32a: Predicted below burden gas and temperature profies for both a "V" and a
"W" shaped cohesive zone (27)

O'6~ 100 ,

80
0-4 r- / \ _
~\,'/ ç:; --.- -----
;( CO

::2 ~
600
T) r// '\\ 1 Gas
T ernp
40 (C)

200

0' o
C~r."TRE WALL
RADIAL LOCATiON

Figure 32b: Actual below burden gas and temperature measurements. Actual burden
distribution indicated a cohesive zone between a "V" and a "W" shape (27)

1 0-52
 Top
Furce Model
"'
o._o_~
.. o cenl(fr \0.
i '. centre
o
Relative L o \, "'.
o ,
Distance o '.
o "
i 0 \
o i
o I

..... 0 \i

wall
waii\
'''''~
.. ..
'I

T uye re II.
zoo 400 '00 *00 1000 1Z00 zoo 400 '00 *00 1000 1Z00

Temperature DC
Figure 33: Comparison of vertical temperature readings from cohesive zone model and
actual furnace readings (27)

70

60

-~
50 '0/
. ~;Z 0 PrdÎchid
. li,All

40 tp il eE " :i
-0
t;: .ii
on
'l
0
0)
E

CJ
30 fUANACE
. - B
" - C
0- D
20 .Ð - E

AVG. DCV'AllO",' 3.5'

10 20 30 4Q 50 60 70

0= ( 38.9 a'¿ T h
A Pr W H N (in)
)"5

Figure 34: Measured tuyere penetration versus predicted (31)

10-53
 ..Õ 10600
~
~inoo
vi;
~ 980
A
~
~400

0.2

0.0
~ 6 7 9 10 I 2
1.3-6~ l-4-óó
CON$((UT IVE CASTS

Figure 35: Mb (heat) index versus silicon on a cast basis (35)

1.0

0.8 o
~.,
Si%
0.6 ...~
" .n
"i 0",
l-cP
" ~4l
~~ -
L ~,' .;~ ". ',.p \ 'r
0.4
- ~ 00~ J!4ú ,,0 ..
0.2

0.0 L

1000 1100 1200 1300 1400 1500

1.0
"
Si %
0.8
.,
., .A\ .~.. .~ "
0.6
.:~..
iC li
" DD
'iW ~D ..
~ ~lIo
l. .. _ ,.-

0.4

0.2

0.0
y a 0

-1_
Q
"' .

3aoo 3300 3400 3500 3600 3100

Figure 36: ARV modeling using 25 step ahead prediction of silicon (38)

10-54
 5.00
c 4.2\1
, . N~~~"''ic~
~fi-cr;DI '.
3.40 CmeQ5
4.00
~
Si 3.20
Sr r;ol
2.40
150
O.SO

() ~
Mn tOO

0.50
(ì ·

C),50
~ "" ), \ rl~,l'V\
Mn meas

S 0.12.0 S IlDS
0.080
0.040
0
601 627 653 61a 704 120 Tal' 11

Figure 37: Predicted versus actual readings using a thermodynamic model (41)

10-55
 Head ofn\'C

JQun1iiJ
JlSF

. iii1 magnCÜr
nill~C furn¡ict' plarLl ì C..mr::1 complllCf ~
ri S.io. t.t:f.plam ¡ ~. . I.,. Trans.mission Reception-
Coke plan¡ J ,d:i:i datu
00'1 i.i1la ì := QDT '
lronmaking. d1illl bas P f;PilaC¡O,n and ~r)i~T . PL
L ~ L pro"cssItS.- '-, I --
~-
-- (,:::onl~n\. or c!.:\.) ¡,~~ :.
o.. ._,.,. ,...'" ......-~---
i 1 r.'umta or dau ~inLSf
Vi Pi.~! c.oll;.l., J,i.
I ""uil'"",,t
0' .....
ll:~'" rurn~r~ 'O¡xn~,nK ~ii¡
~ ,79 it'
~" ~Scìl:li: Iti.I.l
~Š~~;i';;;iiii ll!~ or, .\'idd :ril.:J llll1l .....1_:)
ii,::I111,¡ OI~!11
S~ iumpi'"," '~~¡,1
· Çuifii ~ ,1"1,,
CClh 1I1'"l J 610 "R~'. m¡lc,iil ~III!
~-,..,"~,,,,-----, r'Jrl ,f~\i I)U~.I'
. l'rr.J'mlCn ,al¡~_ i
J ~iR --__--:.~.K ~~::__
lw~::I."I_,_._.., (I" I

---- (i~/ (~J. ~-=--::-----,

('-_._-_."..._.._..,...._.................:... -""'._- --_... t:~J C:~,J r~ :_.:,..:._~~~...-~_I,.-_~q i t.~'J E.J (~:.:J
"._-_..._._~........~"._...._.,.,. ....... .._-_......'.
\':i..~I,l Weida IH),lllilihin\l ::utKenL~r x 1 S,ik:i \YQrl(,1 U!;ii furnace ~tllll:r xI R & l) L~boHllOrie~ rr
Wiikiin;;il~\. $¡Iliering 41,ld ~"ni(~u \\'orh n1;i1 furnai;t: ~'i:,'~er )( )
Irmllllìik iii~ 1,.,-bm1I"lIY 1;lti, ,'. i
ór~ ireiiijng plil\ x f Nippmi ~iC'~) ciicmi.al ~ulXenicr X)
.\uror:\n Works lr~)omilkini; office x I Hnd 0 ffcc JrunmiLk ¡lit i(d,:ii,'al Iliv, -; I
Oi(iI \Vmb Jronm.:king cc:nitr )( I Y,i w.'11 a li'ulimJ.bli~ í'1~Jl cn~l:lCl':iii~ (liv,
K:ii:ii~lii Worl~ lrotliiiaking offi,e x I ft & D l.atioriiivrk$ I l( J
Himbn~1\ Work$ Iri"nmaking. 1.~f1cc ~ I 'l.iwaia Works Cøkc~U¡"~c;11~r ;.'1
lmnm:ikil1~ !t"'hn,,~logy Iiib. xI
N~SÇlY~ Wc"k~ llhi~i fiirn¡¡:t:c~nlcr )( 1

Figure 38: Overview Computer Control Sysem "NICE" (49)

 Control Chart Areas - Upper Half

Determne Hot Meta 2 Std Deviation

Temperatue
Zone "8"
1 Std Deviation

Zone "A"
Plot Hot Meta Mean
Temperatue on the
Control Cha

Determine Exreme
Yes
Conditon and Apply
-, appropriate Control
Strategy

No

Apply Appropriate
Statistical Tests

"~/ Do Test Fail I

No
+
Yes
Take Appropriate
Control Action

r- Take Appropriate
Control Action
,i
i

Figure 39: Hot Metal Temperature Conceptual Control Strategy

10-57
 M

II
2 3 4 B "j 10 11 12

Figure 40: A schematic representation of a "Fuzzy Logic" membership function. A value

of 1 = full membership (69)

Slip warnìrHJ

Probabllly of occurrence Is 63% (It 12 : 10

RO$l.lt$ of dloQnQ~I:; of slip ore dl)é to thé

foRowlnQ reasons:

Raa:5cn Probability
From burden descent judqement,
sUp tends to occur 22%
From prc::zura 10:5:; judi¡amcnL
slip tonds to QCc:ur
13%

From tomperature judgement :slip

li:l1ds 10 occur 22%
POSSlbllltV Of 0 JorQe vOlume of
1'e~ldlJoJ pia iron ond sroq is tiioh
o~
Slip tends f,) oi;cur (Sen"o.. 47% Gro:::
Coefficient of effect 0 f
ProviQUi slip l'orninQ probability 30%
Probability of Qçcurrlioce of
sUp this time 6~% Gre:;:;

End of indiciifloii of how

diiii;nosis was fecchad

Figure 41: Slip Index model- evaluation results (71)

10-58
 Xd ~'_ w.
. """,.:~, ' /-~--"
".¡' '):.." \,
l2..~~
./"'\
I"-tJ../i / L
x, --~ J- )' (f) ~_,___~ y. x)__"

/' \\" --
JU -
X. .r in
Mi
Layer

Figure 42: Schematic representation of a neuron and neural network (73)

( ~

Ûp:..illf¡ l)t)
Scccnu S,iep
~
T'op Gu Tc;tpcmu
Top Gas Compaoc'J
S h~rll'reH~rc TI.lrcSiep
4
Brj~k Temprature
Ouida. ft1
(:;""lI"~ CGn:Ji.O)r, ) IIIl:táiiC'i
c£ Brll.i... r.llem
C.lÚ R.i~ F-tCCIS! cf(ìlÖ fbw
Bh~i Volu r:t, o1nr: .Mt(".JtiI)R
Maierial Cc-i,i"r, ~'L-, ~, ¡,f Be II-Icsi f'a!llni
i ".\J.Ll:alÎ(tn or
Be 1I-le,S5 Pallern
Ore NotclJi
Co\( Nos
VulJlî10( Ve.i\ic.
C1la!iil,S Cl)~.c

Figure 43: A neural network structure used for predicting impact of burden distribution
changes on gas flow conditions in the furnace (74)

10-59
 LECTURE #11

MAINTENANCE RELIABILITY STRATEGIES

IN AN IRONMAKING FACILITY

Gary De Grow

Dofasco Inc.
P.O. Box 2460, Hamilton, Ontario, Canada L8N 3J5

ABSTRACT

The approach to maintenance in the 1970's was to "Fix it when it Breaks" or

commonly known as Reactive Maintenance. One of the main reasons for this was due to
the focus on productivity. This approach was costly due to ineffective use of
maintenance resources and high inventory of spares. In most integrated plants, this could
be tolerated because there were multiple facilities or process streams so production losses
would be minimaL. At the same time, maintenance resources became great "Fire
Fighters".

In the 1980's the focus was now moved to quality aspects, but maintenance costs
could no longer be tolerated. Maintenance costs were escalating exponentially with no
end in sight. Work began to examine maintenance costs and develop plans to control
costs by improving equipment availability. Planning and scheduling maintenance
activities were thought to be the ultimate solution, however this alone was not the answer.

The focus in the 1990's moved toward the total equipment concept, targeting
reliability as the key to long-term success. This paper is an example of an equipment
reliability program developed to meet the challenges in an ironmakng facility.

11-1
 INTRODUCTION

In the 1970-80' s, Dofasco Hamilton fit the typical mould of the integrated steel mill
with two generations of facilities, i.e., four blast furnaces, two steelmaking facilities and
two hot mills. In the early 1990's, a decision was made to shut down our older Stream 1
facilities and optimize our newer Stream 2 facilities. This meant shutting down two blast
furnaces, No.1 Steelmaking consisting of three basic oxygen vessels, and direct coupling
our two remaining blast furnaces to one oxygen KOBM vessel at our No.2 Steelmaking
facility. The Ironmaking facility must now provide a continuous supply of iron without
interruption to Steelmaking. This meant that the blast furnace/hot metal areas in
Ironmaking must extend shutdown intervals to coincide with steelmaking vessel bottom
changes/relines. In other words, there would be 12-14 week intervals between alternate
blast furnace shutdowns.

To meet this challenge, maintenance had to move from a Reactive to a "Pro Active"
organization anticipating what wil happen in the future and planning and scheduling
corrective action ahead of time.

Activities were aimed at three key elements:

· Maintenance Process
· People
· Organizational Structure

11-2
 1. MAINTENANCE PROCESS

There are many processes in a steel plant, i.e., Chemical, Metallurgical, etc. In
any good maintenance program, there is a maintenance process. This process consists
of six basic critical elements:
. Work Identification
. Planning
. Scheduling
. Execution
. Follow up
. Analysis

a) Work Identification

One of the keys to a successful equipment maintenance program is knowing the

condition of your equipment at any time, anticipating what wil happen in the future
and planning/scheduling corrective actions in advance. The work identification stage
is critical to the program.

There are a number of tools to be used in the work identification. These tools are:

. Preventative Maintenance Activities

. Predictive Maintenance Activities
. Reliability Centred Maintenance
. Root Cause Failure Analysis
. Process Information Systems
. Computerized Maintenance ManagemenUIntelligent Condition Monitoring
System
. Regular Communication Meetings

11-3
Preventative Maintenance Activities

Preventative Maintenance inspection wil involve the senses, i.e., visual, touch,
sound. Information gathered through inspections is recorded via check sheets or hand
held data loggers (HHL). The HHL information is preferred as the data can be
downloaded directly into the computerized database.

The PM Program will also include:

· Routine Lubrication
· Cleaning
· Minor Adjustments, servicing
· Varous assessments/audits (environmental) etc.

Predictive Maintenance (PdM)

Predictive Maintenance itself does not prevent anything. It does however give
information on condition changes that wil result in a breakdown.
Predictive Maintenance consists of:
· Non Destructive Testing
. Liquid Penetrant
. Magnetic Particle
. Ultra Sonic
. Radiography
. Eddy Current
. Acoustics
. Fibre Optics
. Vibration/Alignment
. Stress Analysis
. Pump Performance/Flow Monitoring

· Lubrication Analysis
. Trace Metals
. Particle Counting
. Wear Particle Analysis/Ferrography

11-4
 . Infrared Thermography
. Electric Circuit Monitoring
. Refractory Insulation Monitoring

· Motor Circuit Analysis

. Insulation Testing
· Merger Testing
· Dielectric AbsorptionlPolarization Index
· DC High Pot Test

. Stator WindinglPower Circuit Testing

· Surge Testing
· Surge Comparison Testing
· Rotor Fault Diagnosis

Reliabiltv Centred Maintenance (RCM)

Reliability Centred Maintenance is a process used to determine the maintenance

requirements of any physical asset in its operating context.

This process involves addressing seven key questions about the asset selected:

· What are the functions and associated performance standards of the asset in its
present operating context?
· In what ways does it fail to fulfill its functions?
· What causes each functional failure?
· What happens when each failure occurs?
. In what way does each failure matter?
· What can be done to prevent each failure?
· What should be done if a suitable preventative task cannot be found?

Utilization of this process wil deliver the benefits of:

. Improved safety and environmental protection
· Improved operating performance including output, product quality and service
· Increased maintenance cost effectiveness
· Longer equipment life
· A comprehensive equipment maintenance database
· Motivation and ownership by the team
· Improved teamwork between Operation, Maintenance and Technology

11-5
 Root Cause Failure Analysis (RCFA)

Root cause failure analysis is designed to scrutinize every component of the failed
system. Based on scientific/engineering data, and the RCFA Group expertise,
possible failure modes are systematically eliminated leaving the remaining modes as
the causes of the failure. RCFA is initiated:

· After a costly breakdown

· For an in-depth look at a specific equipment failure. Events leading up to and
surrounding equipment failure are known
· To determine root cause failure and revise existing Standard Operating Practices,
create PM Tasks, or re-design equipment

A typical RCFA will require approximately 12 hours of group meetings to

complete.

Process Information System

Many facilities utilize on-line data collection systems to capture process

information. This process information, in many instances, is valuable equipment
condition data that can be used to monitor and trend equipment deterioration and plan
correcti ve acti on.

Computerized Maintenance Management (CMMS)

Intellgent Condition Monitoring System (lCMS)

A Computerized Maintenance Management System (CMMS) is a key tool

required moving to world class maintenance in any facility, but it alone will not
achieve the desired result. What is needed is a system that wil support a planned
maintenance process and have the ability to input, track, and trend equipment specific
data.

11-6
 Dofasco uses a CMMS system with a fully integrated component called the
Intelligent Condition Monitoring System (ICMS). The major components of this
system are:

· Reliability Centred Maintenance

. Equipment Maintenance Program
· Mathematical calculation
· Rules based failure model
. Route planning
. Data collection
· Graphical data

ICMS in conjunction with maintenance components of CMMS, i.e.,

planning/scheduling and spare parts inventory support Dofasco's planned
maintenance process.

The Equipment Maintenance Program (EMP) is the critical element of the system.
The EMP is the list of work activities to be performed to maintain a piece of
equipment at its required level of performance. These activities wil be a combination
of PM, PdM, RCM or RCFA Tasks.

Data points can then be used as condition indicators to give information on

equipment conditions. The indicators can be:
· Numeric
. Alphanumeric
. Boolean

Alarms can be established based on severity limits and rules set up to identify
when equipment is progressing toward failure. Data collection can be accomplished
in three modes:
· Operator check sheet
. Plant System Signals (PI)
· Hand Held Data Loggers (IlL)

11-7
 Once input into ICMS, searches can be run on non-normal condition and move
down to the data points in question for action. The system will allow a spin-off work
request, but action/follow-up must occur, as the alarm does not disappear until
corrected.

Re2ular Communication Meetin2s

One of the simplest tools used for work identification is regular communication
meetings. A daily manufacturing meeting allows operators, maintenance, trades and
technical personnel to review the past 24 hours and identify potential trends in
equipment condition. This is, in part, total productive maintenance activities
peiformed during the course of their shift. This information is gathered during their
inspection rounds and through the PI system monitoring the process.

Information reviewed in these meetings may then either require immediate

follow-up or allow for work requests to be issued to plan and schedule corrective
work.
Work Prioritization

With these various methods of work identification, a strategy must be deployed to

prioritize the facility assets, and then the best tools required must be identified to
determine equipment condition.

In the Ironmaking facility, two key steps were utilized to get started:
· Equipment Criticality Assessment
· Predictive Maintenance Needs Assessment

11-8
 · Equipment Criticality Assessment

The equipment criticality assessment tool was developed to focus on equipment

reliability improvement as a means to improve manufacturing results. The
consequences of equipment failure are assessed in key areas of business performance,
i.e.,
. Safety
. Environmental impact
. Product quality
. Throughput
. Customer Service
. Operating Costs

For each consequence area, a consistent criteria has been defined and a weighting
factor assigned for performing the assessment. The criticality risk number is
determined by multiplying the total consequences of failure by the
frequency/probability of failure.

This assessment was performed on all major equipment within the Ironmaking
Business Unit including the supporting hot metal facilities. The resulting risk
numbers provided this prioritization of assets to focus on the critical equipment that
has the highest impact on performance in the Iron Business Unit.

· Predicitive Maintenance Needs Assessment

In any facility a maintenance program of sorts has been established to minimize

unplanned outages. In many instances, a time-based program was set up to plan and
schedule equipment rebuilds/replacements and inspections. This method, however
may not be the most cost-effective approach to equipment reliability.

11-9

¡
 Once the equipment criticality has been established a predictive maintenance
needs assessment should be applied. This involves a team of PdM Technology
experts to conduct an audit of the existing maintenance program. This audit is
conducted at the operating plant level and involves interviews with maintenance and a
review of inspections on critical equipment.

An in-depth evaluation looking at applicable predictive technologies is applied to

the equipment taking into consideration the existing program, process requirements
and common cases of failures.

The resulting recommendations are documented and submitted to the Business

Unit. The end result will reduce maintenance costs by utilizing the PdM activities
and redirecting efforts of Tradespeople toward more value added maintenance
activities.

b) Plannin2

Planning is identifying a road map of where you are going with your equipment
maintenance program in order to achieve equipment reliability. There must be a short
term and long term plan for a successful reliability plan. The long term plan is vital,
as reliability cannot be reached in one year.

The short term consists of tasks that are required to perform the work now on the
critical equipment. The short-term tasks include:
· Equipment Hierarchy
· Bills of Materials
· Equipment Spare Parts Inventory
· Procedures
· Backlog of Work
· Type of Work
· Repair/Rebuild Programs

11-10
 · Equipment Hierarchy

In the Computerized Maintenance Management System, an equipment hierarchy

is needed to break down the equipment to the level where maintenance is performed.
The hierarchy is also based on physical/geographical locations within the plant. This
hierarchy will allow for accurate cost allocation for equipment maintenance.

· Bil of Materials

The Bil of Materials is of importance to ensure that the Master Parts Catalogue in
the Computerized Maintenance Management System accurately reflects the
equipment in the plant. It is important to ensure parts inventory is correct.

· Equipment Spare Parts Inventory

The Spare Parts Inventory is tracked and monitored via the CMM System. Spare
parts must be identified, stocked and available when needed for planned work. Just-
in-time delivery plays a larger part in reducing parts inventory and relies on the
suppliers/manufacturers to provide the parts where and when needed. Single sourcing
of certain commodities is also playing a larger part in the purchasing /partnering
strategy to further reduce inventory.

· Procedures

Procedures are the instructions of the "what and how" to complete the work.
These instructions may take the form of Job ProcedureslPurge Procedures and must
identify all safety requirements including such items as confined space entry. A
database of procedures has been developed for all the major work to be performed.
Before the procedures are used, they must be reviewed, updated and revised to ensure
content accuracy.

11-11
 . Backlog of Work

The CMMS is required to develop and continually update the outstanding work.
This work will come through the various methods listed in the work identification.
The backlog is used by the Planners to develop packages of work to be ready for
scheduling and execution.

. Type of Work

The type of work can be categorized into three areas:

· Daily
· PM
· Shutdown

The daily work requests may require little planning preparation and can be
planned by the First Line Supervisor. Other work requests may involve considerable
detail involving parts, procedures, etc. and would be passed on to the Trades
Planners.

PM work requests for the most part wil require minimal planning, as they
become pait of regular inspection routes.

Shutdown work requests are generated from the CMMS work backlog. Because
shutdown planning is the most in-depth and detailed planning function in our
Ironmaking facility, a guideline document was developed for shutdown planning and
scheduling. This document identifies a standardized approach to planned and
scheduled outages or shutdowns to:

· Eliminate or minimize start-up delays.

· Improve safety and efficiency of the workforce.
· Improve understanding of the amount and scope of work.
· Provide a basis for monitoring and controlling tasks during shutdown.

11-12
 The resulting document defines the roles and responsibilities of the shutdown
team, requirements of a good CMMS work plan and scope, schedule-building
techniques, downloading requirements for scheduling various critical meetings and
follow-up audit.

· Repair/Rebuild Program

A repair/rebuild program must be an integral part of the planning process in order

to improve maintenance. Approximately 70% of all failures can be classified as
"Maintenance Induced". These failures are in part caused by:
· Lack of skills to do the work,
· No job procedures
. No documented specifications
· Lack of job planning

The repair/rebuild program wil require equipment experts within your

organization working with equipment builders to develop accurate rebuild
procedures. These procedures will contain all pertinent specifications, materials and
measurements to repair the pars or equipment back to original equipment
manufacturer standards.

The benefits of having such a program will reduce maintenance cost by

eliminating re-work, improving equipment life and minimizing unplanned/emergency
maintenance.

c) Scheduling

The scheduling for the routine PM and minor type repair/corrective work is
performed by the First Line Supervisor.

The more complicated jobs are planned by the Planners and scheduled by the First
Line Supervisors with input from the Planner.

11-13
 Shutdown Planning requires a much larger project management tool to accurately
define the planning/scheduling needs for our blast furnaces. A scheduling tool "Open
Plan" is utilized in Ironmaking. All backlog work orders are to have accurate task
durations, estimated labour requirements (trades, contractors, and operations) and the
work centre/area. This information, as well as the available Tradespeople is
downloaded in this scheduling software tool.

The Scheduler will add the logic to the plan for each job and the logic/pert
diagram developed for the shutdown.

The First Line Supervisors and Coaches of both Maintenance and Operations
review the logic to develop the final approved logic diagrams.

From this point on, meetings will be conducted to develop crew assignments for
the work using their own internal resources, field service resources and contractors.

d) Work Execution

In the work execution step the resources required to complete the assigned work
are analyzed. In order to optimize our trades resources needed to perform the work,
several coordination groups have been established:

· Field Coordination Team

· Shop Coordination Team
· Manpower Sharing Team

In the Shop and Field Coordination Teams, Planners from the Primary Business
Units meet with Operating Services Supervision. Together, these people prioritize
the work sent to Operating Services and develop a plan for the work. Any work that
is entered on a rush basis is usually sent to a contractor or an outside shop.

11-14
 The Manpower Sharing Coordination Team was established to provide a resource
sharing pool of Tradespeople to accommodate peak loads, i.e., shutdown, in the
primary facilities of Cokemaking, Ironmaking, Steelmaking and Hotmili. The
purpose of this group is to maximize the use of internal plant personnel and minimize
contractor requirements for major shutdowns.

In a similar way, the Business Unit also allows for flexibility within their
manufacturing facilities. Manpower sharing wil occur between facility teams to
accommodate variation in workloads and work schedules.

Shift maintenance staffing has been reduced to two Tradesmen. This is possible
due to the maturity of the Equipment Reliability Program. Focus should be to
maximize Tradespeople on days to perform planned/scheduled work and reduce shift
size and possibly eliminate supervision.

e) Follow-up

The follow-up step in the maintenance process consists of a number of activities

that immediately take place once maintenance has been executed. These activities
are to:

· Record historical data

· Record future work identified during execution
· Revise parts and procedures
· Upgrade equipment

· Record Historical Data

The work order completion comments must contain key points if they are to be
¡
useful for analysis and improvements in the Equipment Maintenance Program. The
i

information must be thorough, accurate and identify the problem, explanation of

repairs performed, determine the cause of failure and time required to complete the
work.

11-15
 · Record Future Work Identified During Execution

This future work could pertain to other work identified during job execution. It
could also refer to follow-up work - PM, PdM activities that need to be performed to
base line the as installed condition of the equipment. Examples of these PMldM
activities could be vibration testing on rotating equipment, lubrication sample
analysis and on-line motor testing. These activities would be performed to confirm
acceptable limits and base line as installed record data for future monitoring and
trending. Visual inspections may also be part of the follow-up to ensure the
equipment is operating at its desired level of performance.

· Revise Parts and Procedures

During the executions, problems may be encountered with the parts used or
procedures followed to complete the repair. It is imperative that these issues are
corrected to ensure future repairs will not encounter the same pitfalls.

· Equipment Upgrades/Redesign

Follow-up may also include equipment upgrades/redesign to improve equipment

reliability and reduce maintenance and, ultimately costs. This may involve
Technology/Maintenance/Operations personnel in conjunction with outside suppliers
and manufacturers to redesign equipment. A recent example in lronmaking would be
the replacement of the high maintenance cost double drum stockrod winches with the
microwave systems to measure blast furnace burden stockline levels.

It may require a higher initial cost, however, when considering the return on the
investment, the cost of maintenance would far outweigh the initial capital outlay.

11-16
 In the planned and scheduled shutdown approach to performing maintenance, the
post audit meetings are a fundamental step for continuous improvement in the
maintenance process. The post audit meeting closes the gap by identifying the
problems encountered during the shutdown, lists solutions, identifies the persons
responsible for the actions, expected completion date and any additional comments.
A database is regularly updated until all actions are completed.

f) Analvsis

Performance measures are the means to identify gaps from your target or
benchmarks and trigger actions to close the gap. These measures should be as close
to the action as possible to motivate people to trigger corrective action in order to get
back on track.

The performance measures for maintenance can be separated into three

categories:

· Functional Maintenance Metrics

· Functional Operations Metrics
· Business Metrics

· Functional Maintenance Metrics

Functional Maintenance Metrics are associated with measuring the best practices
of sound maintenance programs/processes. These best practices are the activities
related to the maintenance process. As a result, a number of projects will be
implemented and measured on the maintenance process steps.

These measures should keep everyone thinking about how well they are executing
best practices and what impact they will have on your business.

11-17
 · Functional Operations Metrics

The Functional Operations Metrics are performance measures surrounding the

operating process. These measures involve quality availability and operating rate. At
this level, pareto analysis may be performed to identify the various key factors, i.e.,
equipment, process, raw material, and/or utilities that caused the unscheduled
shutdown, operating rate, or quality defect.

· Business Metrics

The Business Metrics consists of a core set of measures that result from the
functional measures or best practices. These measures are:

· Maintenance Costs
· Equipment Availability
· Failure Rate
· Overall Equipment Effectiveness

· Failure Rate

Failure rate is the measure of all equipment failures in the facility. The failures
can be categorized from pareto diagrams to identify areas of opportunity over time.
This number is not as important as the trending of the failures.

11-18
 · Overall Equipment Effectiveness

Overall Equipment Effectiveness (OEE) is a measure that reflects how the

equipment is performing overall while it is being operated. This measure takes into
account equipment availability, equipment efficiency and product quality or, in other
words is it operating at its desired rate. The product of these three factors yields a
number that quantifies equipment performance relative to design performance for the
net available time a machine was scheduled to run. The calculated number has no
, i
meaning when comparing to dissimilar processes. It is important however, to
evaluate the trend in GEE measure over time to monitor improvement.

This measure is important to communicate to all the individuals in the Business

Unit who can impact the equipment/facility performance. The GEE measure is a
common measure that includes both the production and maintenance components of a
facility.

In Iron, the measures consist of:

OEE% = Q% xPE% xA%

%Q The appraisal of quality which is the measure of the % of product in

control on hot metal temperatures

%PE The performance efficiency that is the measure of the rate of operation of
the equipment or the percent of time the blast furnace is operating at full wind.

A % The percentage of time that the furnace was not shutdown excluding
inventory stops.
, I

11-19
· External Maintenance Assessment

In addition to the internal performance analysis, an external consultant was

contracted to benchmark Dofasco's capabilities in the areas of maintenance and
reliability.

The consultant has benchmarked numerous companies in North America and

developed a good baseline comparison. The team of consultants worked closely with
our own personnel to evaluate our strengths and improvement areas, and compared
them to world class organizations. From this, the consultant identified opportunities
and assisted in the development of a plan to improve our practices, systems and
procedures. Each Business Unit in Dofasco was evaluated in the areas of:
. Management Commitment . Quality
. Work Identification . Coordination
. Safety . Planning
. Information . People
. Scheduling

11-20
 2. PEOPLE

The key to moving the equipment reliability program ahead is the people. Our
Tradesmen are the experts in the trade and equipment knowledge so we must harness
the knowledge and focus everyone on continuous improvement. Everyone must
clearly understand the goals and be committed to continuous improvement toward
them. This requires a change in the culture from a reactive approach to a pro-active
approach in maintenance. It must be supported by Operations, Technology and
Maintenance people within all levels of the Business Unit and the Company.

. Team Development

To demonstrate the commitment to changing the culture, all employees were senI
to various sessions on team development and empowerment. Training sessions such
as Pecos River and Athenium training were used to help execute the change. This
provided opportunities to breakdown the traditional trade/department bamers and
refocus on equipment reliability using a team approach.

Manufacturing trios or teams in the Ironmaking Business Unit were developed

consisting of Maintenance, Operations and Technical Supervisors for each operation
facility including No.3 Blast Furnace, No.4 Blast Furnace and the Hot Metal areas.
This developed a sense of ownership by the cross-functional groups down to and
including the floor leveL.

· Trade/Technology Training

The Tradespeople's skills must be continually upgraded to keep in tune with

technological changes. It requires attendance at Ironmaking conferences to
benchmark other steel plants. Similarly, the Tradesmen must also keep ahead of the
predictive maintenance developments in technology, equipment and techniques.

11-21
 . Communications

Communication is an important factor in order to develop and maintain

commitment to change. Regular lunchroom meetings, equipment reliability
newsletters, performance measurement graphs are tools used to identify the
improvements and the achievements. Remember, celebrate the successes as they
were accomplished by your people. This will help to heighten awareness, increase
participation and the sense of ownership.

3. STRUCTUR

The last piece of the puzzle for equipment reliability is the Business Unit
maintenance organizational structure. Equipment reliability must be a separate
function within maintenance, divorced from the day-to-day planned corrective work.
The Equipment Reliability Group is a headed by a Maintenance Coach and together
the group's mandate is:

· PM Inspection
· Lubrication
· Data Analysis
· PreventativelPredictive/Technology Application and Analysis
· Shutdown Planning and Scheduling

. PM Inspection

A group of Tradespeople of the various trade disciplines perform pre-identified

inspections and record findings on check sheets or input to Hand Held Data Loggers.
They wil make minor adjustments and repairs. However, their primary responsibility
is the PM Program. Anything found during the inspection wil be put on the work
requests and sent to the appropriate Trade Planner/Supervisor to perform the
necessary repairs. Each Tradesperson wil have a facility responsibility to assist in
the development of ownership.

11-22
. Lubrication

The lubrication team is also assigned to individual facilities. They are responsible
for daily, weekly, monthly time-based lubrication. In addition to lubrication, they
perform sampling on critical hydraulic and oil lubrication systems for analysis. A
First Line Supervisor/Teamleader will work with the team to identify, interpret and
trend lubrication laboratory analysis and develop action plans to identify root causes
and eliminate the problem. He wil also work closely with the various trades and the
vibration analysis to monitor unusual vibration and oil related issues.

· Data Analysis

The data analysis is involved with reviewing the various information collected
and working with the numerous reliability specialists to identify areas for opportunity
to improve the equipment reliability program. From the various information collected
through inspection, the reliability specialists can analyze and trend critical equipment
measurements and work with Tradespeople/Planners to develop corrective action
plans before failures occur. They wil use the different tools, i.e., reliability centre
maintenance, and root cause failure analysis to develop an appropriate equipment
maintenance program.

11-23
· Preventative/Predictive Technology Application Analysis

This group consists of specialized Tradespeople/Technicians in the various trade

areas whose primary purpose are to investigate analyze and implement Predictive
Maintenance Technologies and to continuously improve the Equipment Reliability
Program. This group consists of:

. Vibration Specialist
. Electrical Specialist
. Alignment Specialist
. Instrumentation Specialist
. Hydraulic Team
. Repair/Rebuild Team
. ICMS Development Team

The ICMS Development Team consists of a full time equipment specialist

working with various Tradespeople from the Floor and Reliability Teams. This team
is responsible for implementing the ICMS Program to the various equipment in the
facilities.

· Shutdown Planning and Scheduling

Two Schedulers work with the Equipment Reliability Group, daily Planners and
Supervisors to plan and schedule all the work on backlog. For a shutdown, a logic
diagram and critical path items are identified to determine the duration of the
shutdown. The group monitors work at the furnace during the shutdown execution
and continually update the logic diagram as work is completed. This ensures all work
is completed and assists in identifying the importance of adherence to the scheduled
sequencing.

11-24
 CONCLUSION

The implementation of an Equipment Reliability Program wil positively impact

the bottom line at your facilities. You must, however view maintenance as a process
and ensure that you have the dedicated personnel and structure in place to support the
program.

The end result wil change the perception of maintenance from an expense to
being recognized as an integral part of the manufacturing organization. Results will
be:
· Throughput increase caused by equipment reliability
· Safety Program Improvement
. Personnel development and improved efficiency
· Quality improvements
· Cost reductions on parts inventory, contracts and maintenance

Effective implementation of an Equipment Reliability Program has resulted in an

increase in intervals between blast furnace maintenance stoppages from two weeks to
eight weeks to align with steelmaking KOBM bottom change/vessel relines.
Opportunities for further increasing shutdown intervals are realistic as the Equipment
Reliability Program continues to mature and the steelmaking shop continues to
implement new technology to increase vessel life.

11-25
.J
i
G ;,

o
o
o
o
o -::

o ,,'

o .' .

o
,0 a

o ,( "':

o
o II

o
a

o
.'"~

o " a

o
o - (

;"Q J" .
.: . . ~I