Professional Documents
Culture Documents
KEYWORDS: Rock Mechanics, Hard-rock mine empirical methods. The Confinement Formula developed during the mining of coal seams. In
pillars, Pillar strength, Underground mining. utilizes a mine pillar friction term calculated hard-rock mining, research has been much more
from the average minor/major stress ratio with- limited. The most notable works are that of
Paper reviewed and approved for publication in the pillar core. Statistically, the new formula Hedley and Grant (1972) and more recently, the
by the Rock Mechanics Committee of CIM. provides better results at predicting pillar work of Hudyma (1988). The limited amount of
strength for the combined database than the work on pillar strength determination for hard-
best empirically fit pillar methods that currently rock mining has given an insight into pillar
ABSTRACT exist. The Confinement Formula allows for strength behaviour, however no methods that
the determination of the strength of mine pillars can be used with a high degree of confidence
This paper presents the results of research that with an increased level of confidence over pre- have been developed.
has been undertaken in order to develop an viously applied methods. The empirical strength formulae developed
improved method of determining the strength to date have generally taken one of two forms:
of hard-rock mine pillars. Pillars are found in (1) the Shape Effect Formula; or (2) the Size
various shapes and sizes in all underground Introduction Effect Formula. Equation 1 is the general form of
mining operations. Prudent engineering design the equation that these strength formulae follow.
requires that if pillars are to perform as desired, This paper presents a new method, The
wa
both the pillar strength and the pillar stress
must be determined. A detailed pillar stability
Confinement Formula, to be used for determin-
ing the strength of hard-rock mine pillars. The
Ps = K[A + B*
( )hb
] . . . . . . . . . . . . . . . (1)
study has been combined with an extensive method was developed through a combination where,
database of published pillar case histories (178) of detailed research at Westmin Resources Ltd. Ps = Pillar strength (Mpa)
resulting in a new hybrid pillar strength for- and the assimilation with all available published K = Strength constant related to pillar
mula, The Confinement Formula, that utilizes hard-rock pillar case histories. The total com- material (Mpa)
classic strength of rock methods combined with bined database contains 178 case histories that w = Pillar width (m)
represent pillars that are classified as stable, h = Pillar height (m)
unstable or failed. A, B = Empirically derived constants which
Per J. Lunder Traditional pillar strength formulations were when added equal 1. In the case of the
graduated with a B.A.Sc. in geological
engineering from The University of reviewed and The Confinement Formula repre- Size Effect Formula, A is equal to 0
British Columbia in 1983. He worked sents an advancement in pillar design methodol- and B is equal to 1
for eight years as a production
geologist with Giant Yellowknife Mines ogy that takes into account the factors more a, b = Empirically derived power constants. In
and Blackdome Mining Corporation. In common when dealing with the strength of rock. the case of the Shape Effect Formula,
1994 he received a M.A.Sc. in rock
mechanics from The University of Empirical strength formulae developed to date a and b are equal
British Columbia. He spent three years
working in the Geomechanics Group at
have generally used the pillar width/height ratio The Shape Effect Formula infers that for a
the Noranda Technology Centre, Pointe as a primary input factor for pillar strength deter- given rock type, a pillar of a given shape (pillar
Claire, Quebec and is currently
employed as senior ground control mination. This is in contrast to conventional width/height ratio) will have a constant strength,
engineer, Timmins Division for Royal Oak Mines. He is a member of CIM strength of rock methods, whereby the primary independent of change in size of the pillar. There
and APEO.
input factors are the major and minor principal are two variations of the Shape Effect Formula.
stresses on a sample. This paper presents the The first utilizes a linear relationship between pil-
Rimas Pakalnis most common historical methods, a summary of lar stress and pillar width/height ratio. The second
is presently assistant professor in the
Department of Mining and Mineral the databases used, and the methodology used utilizes a power relationship between pillar stress
Process Engineering atThe University of to develop The Confinement Formula. and pillar width/height ratio.
British Columbia. He is a graduate
mining engineer from McGill University The Size Effect Formula infers that for a
(1978) and The University of British
Columbia, where he received his
given rock type, a pillar of a given shape will have
M.A.Sc. and Ph.D. (1986). Dr. Pakalnis Historical Methods reduced strength as the size of the pillar increases.
is the author of several technical
papers relating to various aspects of This formula is a modified power formula where
underground rock mechanics. Research The strength of mine pillars has been the the pillar width and the pillar height are subject to
areas include applied mine design,
fibreglass cable bolts and modelling. subject of extensive research work in the past. differing power terms. The use of a Size Effect
He has consulted over the past ten years to most major mining companies The primary focus has been for coal mining oper- Formula was adopted due to the belief that sam-
across Canada. He is a Registered Professional Engineer in British Columbia
and Ontario. ations, where large regular arrays of pillars are ples of increasing size would have a lower rock
September 1997 51
Determination of the strength of hard-rock mine pillars
Fig. 1. Schematic illustration of pillar stability classification method used at Westmin Resources Ltd. and the common
mass strength due to the increased number of pillar stability classification.
structural features within a sample. It has been
shown, however, that above a sample size of a side
length of 1.0 m to 1.5 m, the resultant decrease in
sample strength due to increasing sample size
becomes negligible (Hoek and Brown, 1980).
Empirical strength methods for hard-rock
pillars have been developed and can be fit to
the existing pillar case histories. To some
degree, this is accomplished with ease due to
the limited number of case histories available,
and the narrow range of pillar shapes in each
database. The detailed compilation of the pillar
databases presented here allows for the devel-
opment of a more detailed strength formula
covering a broad range of pillar shapes.
Table 3. Distribution of pillar stability classifications for pillar case histories in the combined and individual
databases drops below one. A two-dimensional boundary
Pillar stability Combined Westmin Hudyma Von Kimmelman Hedley and Others element modelling exercise was undertaken to
classification database Resources (1988) et al. (1984) Grant (1972) determine the relationship between pillar
Failed 68 18 12 29 3 6 width/height ratio and the average pillar con-
Unstable 52 11 9 11 2 19 finement. The results of this modelling exer-
Stable 58 2 26 7 23 0
cise showed that a relationship between pillar
width/height and the average pillar confine-
ment does exist. Equation 4 was found to
interpretation is the unstable classification. Ps = (K*UCS)(C1+C2 * kappa) . . . . . . . . (3) relate pillar width/height ratio and the aver-
This classification can also been referred to as age pillar confinement with a good degree of
the transition zone from stable condition to where, accuracy at a modelled mining extraction ratio
failed condition. Ps = Pillar strength (Mpa) of 75 %.
Table 3 shows the distribution of the case K = Rock mass strength size factor 1.4
[ ]
___
histories according to the common pillar stabil- UCS = Unconfined compressive strength of w (w/h)
Cpav = 0.46 log( +0.75) . . . . . . . . . (4)
ity classification for each of the databases that intact pillar material (MPa) h
make up the combined database. There is a good C1, C2 = Empirically derived constants
distribution of pillars in the three pillar stability kappa = Mine pillar friction term where,
classes from each of the databases with the Cpav = Average pillar confinement
exception of Hedley and Grant (1972) which had Pillar Size Strength Term w = Pillar width (m)
only three failed and two unstable pillars. h = Pillar height (m)
The only common strength data available The advantages of using average pillar
for the combined database were the unconfined confinement in the place of pillar width/height
Pillar Strength Determination compressive strength (UCS) of intact pillar mate- ratio for strength determination may not be
rial. It was determined that the size term in immediately obvious. Where pillars are of regular
As shown in the previous sections, pillar Equation 2 could be replaced by product of the rectangular or square shape, the pillar width and
strength has been assessed using empirical rela- rock mass strength size factor, K, and the height can be acquired readily and will accurate-
tionships that relate the pillar width/height ratio unconfined compressive strength of intact pillar ly represent the shape of the pillar. However,
and a rock mass strength term. The computed material. Detailed analysis of the individual data- where pillars are of irregular shape, or are con-
strength is then compared to the predicted pillar bases was performed and the ranges of the K fined on one or two sides, the effective pillar
stress in order to assess actual or predicted pillar values for each of the databases was deter- dimensions are not so easy to assess, Usually a
performance. Conventional rock strength meth- mined. These values ranged between 30% and best guess is made on what to use for pillar
ods (Mohr-Coulomb, Hoek Brown), however, 51%. The results of this analysis led to the defi- width/height ratio. Using the average pillar con-
make use of the applied and confining stresses nition of the average rock mass strength size finement as determined from numerical model-
on a sample when determining sample strength. factor to be 44% of the unconfined compres- ling allows for a correct assessment of the
The Confinement Formula combines these sive strength of the intact pillar material. shape term in pillar strength.
two approaches to develop a hybrid strength
formula that utilizes a mine pillar friction term Pillar Shape Term Mine Pillar Friction Term
and empirical strength constants. The empirical
constants were determined in order to best fit The shape term in The Confinement The Confinement Formula utilizes a term
the strength curves for the case histories in the Formula makes use of a new term to represent that resembles the effect of increasing the fric-
combined database. pillar shape called the mine pillar friction tion angle of a material. This frictional effect of
The method presented here, like those that term, kappa. Kappa is determined from what mine pillars is determined from the average pil-
have preceded it, represents pillar strength with we have termed the average pillar confine- lar confinement. For a given average pillar
two multiplicative terms, one representing the ment which is used in place of the pillar confinement value, Mohr circle diagrams can
in-situ rock mass strength and the other that width/height ratio for the purposes of assessing be constructed and an effective friction term
accounts for the variation of pillar strength as a pillar shape. The average pillar confinement determined. As the average pillar confinement
result in the change in pillar shape. This is gen- is defined as the ratio of the average minor and (and pillar width/height ratio) increases the
eralized by Equation 2. average major principal stresses at the mid- slope of subsequent Mohr circle plots result in
height of a pillar. Empirical constants have been what would appear to be a decreased value of
Ps = Size Shape . . . . . . . . . . . . . . . . . . . . . (2) applied to Equation 3 after a detailed analysis of the friction term. Using the complementary value
the combined database. The method used to of the slope gives us the mine pillar friction
where, develop the average pillar confinement and term used in The Confinement Formula.
Ps = Pillar strength (MPa) the mine pillar friction term is discussed in the Equation 2 is the formula for the mine pillar
Size = Strength term that incorporates the following sections. friction term. Figure 2 shows the relationship
size effect and strength of intact between the mine pillar friction term and the
pillar material Average Pillar Confinement average pillar confinement.
Shape = Geometric term that incorporates the
1 Cpav
shape effect of the pillar A means of utilizing the average pillar kappa = tan[cos1()] . . . . . . . . . . (5)
The development of each of the terms in The confinement in a pillar strength formula has 1 + Cpav
Confinement Formula is discussed in the fol- been investigated. Numerical modelling using
lowing sections. The general form of The different rock mass failure criteria show that where,
Confinement Formula is represented by the mid-height of a pillar is the first point at kappa = Mine pillar friction term
Equation 3. which the factor of safety against pillar failure Cpav = Average pillar confinement
September 1997 53
Determination of the strength of hard-rock mine pillars
Fig. 2. Relationship between the mine pillar friction term and the average pillar confinement.
The Confinement Formula
where,
Fig. 3. Pillar stability graph for The Confinement Formula with average pillar confinement plotted on the x-axis. All Ps = Pillar strength (MPa)
of the case histories in the combined database are plotted according to their pillar stability classification.
UCS = Unconfined compressive strength of
intact pillar material (MPa)
kappa = Mine pillar friction term
Discussion
September 1997 55