Pressure Drop in an

Interconnected Pressurised Fluidised Bed Reactor for

Chemical Looping Combustion

Research Report

Jens Wolf

February 2004

Energy Processes
Department of Chemical Engineering and Technology
Royal Institute of Technology
TRITA – KET R189
ISSN 1104–3466
ISRN KTH/KET/R--189--SE
Abstract
Chemical Looping Combustion (CLC) is a new technique for CO2 capture in power
generation systems. In CLC, pure CO2 is obtained by applying a two-step combustion
with inherent CO2-separation. A subsequent CO2 separation process is not necessary.
Therefore, CLC has the potential to capture CO2 with lower penalties in efficiency.
The two-step combustion may be realised in an interconnected pressurised fluidised
bed reactor (IPFBR). This report presents a mathematical model for a rough
calculation of the pressure drop of such an IPFBR for CLC.
Table of contents

1 BACKGROUND 5

2 THE CLC REACTOR – HYPOTHETICAL DESIGN 7

3 CALCULATION OF THE PRESSURE LOSS IN THE IPFBR 9

3.1 The freeboard-entrainment model 9

3.2 Pressure drop and booster power 9

3.3 Terminal velocity 10

3.4 Height of the fluidised bed in the fuel reactor 11

3.5 Mean value of solid fraction in the freeboard 11

4 CALCULATION OF THE PRESSURE DROP IN A CYCLONE 13

5 CALCULATION OF THE PRESSURE SHELL 15

5.1 The cylinder 15

5.2 The ellipsoidal head 15

6 CONCLUSION 17

7 NOMENCLATURE 19

7.1 Greek letters 19

7.2 Indices 20

7.3 Abbreviation 20

8 REFERENCES 21

4
1 BACKGROUND
Figure 1 illustrates the principles of CLC. A solid oxygen carrier circulates between
two fluidized bed reactors and transports oxygen from the combustion air to the fuel;
thus, the fuel is not mixed with air. The oxygen carrier is a metal oxide that is reduced
in a fuel reactor thereby oxidizing the fuel. The oxygen carrier is then transported into
an air reactor where it is oxidized by air and after passing a cyclone it is recycled back
into the fuel reactor. The reactions occur in two separate reactors.
The connection between the oxygen carriers and the system is the reactor. We assume
a reactor that we call the interconnected pressurised fluidised bed reactor (IPFBR)
which is described in Chapter2. Compared to a commercial gas turbine combustor the
IPFBR will cause a higher pressure drop when integrated in a power generation
process such as, for example, a natural gas fired combined cycle.

N2 (+ O2)
Cyclone
Oxygen carrier
Metal
Oxygen
Air-
reactor CO2
H2O

Fuel reactor

N2+O2 (air) CH4 (Fuel)

Fig. 1. Principles of the CLC

5
2 THE CLC REACTOR – HYPOTHETICAL DESIGN
The dimension of the air reactor is mainly determined by the size of the cyclones, the
fuel reactor, and a particle seal between the cyclones and the fuel reactor. The
cyclones may be about 16 m high and 4 m in diameter for the capacity of the 800
MWth (thermal input) of the CLC system. The input velocity is 15 m/s. For the fuel
reactor, we assumed a height of 5 m and a diameter of 12 m. This height should lead
to a sufficient separation of the particles from the gas stream without any cyclones. In

To turbine
From compressor

H2O/CO2 ca 25 m

4–5m

Fuel gas/ methane

ca 22 m
ca 5 m ca 15 m

H2O/CO2
To turbine

From compressor

Fig. 2. Example for a ca. 800 MW CLC reactor system.

7
order to leave enough space for the downcomer and the particle seal, we assume a
height of 25 m for the air reactor (the riser) (Figure 4). The diameter of the riser is 5
to 6 m. For this dimension assumption, a pressure vessel will be needed that has a
diameter of at least 18 m and a height of about 30 m.
Figure 2 shows the assumed design of the IPFBR with the fuel reactor of the bubble
bed type and the air reactor being essentially a pneumatic transport reactor. The
oxygen carrier particles are separated from the hot gases by a cyclone system similar
to that in a pressurised fluidised bed combustion (PFBC) system.

8
3 CALCULATION OF THE PRESSURE LOSS IN THE IPFBR

3.1 The freeboard-entrainment model

The distribution of the fraction of solids over the freeboard is calculated using a
freeboard-entrainment model described by Kunii and Levenspiel2. According to this
model, the reactor may be divided into four fluidisation regions:
• At the bottom is a relatively short entry zone. Because the contribution of this zone

to the total mass of solids in the reactor is of minor importance, the entry zone has
been neglected in the following calculations.
• Following the entry zone, there is a portion of the vessel of almost constant solid

fraction. These lower portions may be called the dense region. The solid fraction
εsd in this region was assumed to be 0.11. This figure was obtained by
extrapolating experimental results presented by Kunii and Levenspiel2. The section
of the vessel between the surface of the dense phase and the exit of the reactor is
called the freeboard, and its height is called the freeboard height (Hf).
• Above the dense region is an upper entrained region where the solid fraction
decreases progressively to about εs = 0.01-0.02. When increasing the freeboard
height, eventually, a solid fraction of 0.01 is reached. This may be called the
transport disengaging height (TDH). When the freeboard height exceeds the TDH,
the entrainment rate does not change significantly.
• At the TDH, the fast fluidised bed may turn into a saturated pneumatic transport
with a particle fraction of εsp = 0.01.
If the freeboard is higher than the TDH, the maximum flowrate of solids between the
air reactor and the fuel reactor (also called carryover) is limited by the saturated
pneumatic transport flow.

3.2 Pressure drop and booster power

The pressure drops of the fuel reactor and air rector are calculated using Equations 1
and 2:
∆Pf = ∆PB,f + ∆PD,f (1)
∆Pa = ∆PB,a + ∆PD,a + ∆PC (2)
where ∆PB,f is the pressure drop in the fluidised bed and ∆PD,f the pressure drop
caused by the gas distribution. ∆PC is the pressure drop over the cyclone. The pressure
drop over the fluidised bed ∆PB is calculated using Equation (3), which is derived
form the hydrodynamics described by Carberry and Varma1 and Kunii and
Levenspiel2.

9
 M ⋅g
∆PB = 1.2 ⋅ (3)
A
According to Equation (3), the pressure drop is the total fluidised mass (M) per area
(A) of the reactor multiplied by the gravitation constant. Because of the large diameter
of the reactors, the pressure loss caused by friction with the wall is neglected.
The fluidised mass in the fuel reactor Mf is calculated as the product of the carryover
(m
& s ) and the residence time (τf) of the particles in the fuel reactor (Equation 4).

Mf = m
& s ⋅ τf (4)
The fluidised mass in the air reactor Ma is calculated using Equation (5),
M a = (H d ⋅ ε sd + H f ⋅ ε sF ) ⋅ A a ⋅ ρ s (5)
where Hd is the height of the dense region, εsd its solid fraction and Hf is the height of
the freeboard. ε sF is the mean value of the solid fraction over the freeboard. The
detailed calculation of ε sF is shown in Section 3.5.
The pressure drop across the gas distributors is calculated from Equation (6)2
∆PD = 0.4 ⋅ ∆PB (6)
For the cyclones, an overall pressure drop (∆PC) has been calculated according to a
model described by Sinnott3 (Chapter 4). In this study four pairs of cyclones are used
(Figure 4).
The gas turbine compressor or an additional booster fan has to overcome the pressure
& was calculated by the flow equation
loss in the fluidised beds. This booster power W F

for a reversible adiabatic process and the isentropic efficiency of the fan (Equation 7).

W& = m &
⋅ (h 2 − h1 ) =
m
&
⋅ c p ⋅ (T2 − T1 ) (7)
F
ηfan ηfan
By assuming a perfect gas, T2 is calculated for the reversible adiabatic process where
n is equal to the isentropic coefficient (Equation 8).
 1− n 
 
 P1   n 
T2 = T1   (8)
 P1 + ∆P 
For the power loss in the air reactor, T1 is the temperature of the air after compression.
For the fuel reactor T1 was 180°C, because the natural gas was preheated to 180°C.
The heat capacity (cp) was assumed to be constant.

3.3 Terminal velocity

The terminal velocity exists when the velocity drag force equals the gravitational
force. At this state, each particle is individually supported and they no longer rest
upon one another (Equation 8).

10
 12
4⋅g ⋅dp ρ 
uT =  ⋅  s − 1 (9)
ρ 
 3 ⋅ C D  g 
The drag coefficient, CD, was calculated as a function of the Reynolds number for
particles at terminal velocity (ReT).
a1
CD = b
(10)
Re T
u T ⋅ d p ⋅ ρg
Re T = (11)
µ
The constants a and b were approximated as Howard5 (Table 1).
Table 1 Constants a and b for calculation of the drag coefficient.
Range of ReP Region a1 b
0 < ReP < 0.4 Stoke’s law 24 1
0.4 < ReP < 500 Intermediate law 10 0.5
500 < ReP Newton’s law 0.43 0

3.4 Height of the fluidised bed in the fuel reactor

The height of the fluidised bed in the fuel reactor was calculated using Equation 12.
M
H Bf ⋅ (1 − ε f ) = (12)
Af ⋅ ρ s
The overall voidage εf of the bubble bed (fuel reactor) was assumed to be 0.62. This is
in the range presented by Kunii and Levenspiel2 and Basu and Fraser4.

3.5 Mean value of solid fraction in the freeboard

The coefficient a in Equation 13 is estimated from Equation 14.
ε sF − ε sp
= e − a 2 ⋅z F (13)
ε sd − ε sp
a 2 ⋅ u 0 = constant (for constant dp) (14)
The constant is estimated to five based on experimental results presented by Kunii
and Levenspiel2. However, the constant has to be determined experimentally for the
particles of oxygen carrier and the relatively large diameter of the air reactor for a fast
fluidisation.
The mean solid fraction over the freeboard is calculated with Equation 15.

εsF =
1
HF ∫0
HF
ε sF ⋅ dz F = ε sp +
ε sd − ε sp
a ⋅ HF
(
⋅ 1− e
−a 2 ⋅H F
) (15)

11
4 CALCULATION OF THE PRESSURE DROP IN A CYCLONE
According to Sinnott3, Stairmand developed two standard designs for gas-solid
cyclones: a high efficiency cyclone and a high throughput design. The high
throughput design, Figure 3, is suitable for high gas rates which we have in the
IPFBR.

Fig. 3. Standard cyclone dimension of a high rate cyclone (Picture from Sinnott)

The pressure drop in the cyclone will be due the entry and exit losses, and friction
kinetic energy losses in the cyclone. The empirical equation given by Stairman can be
used to estimate the pressure drop (∆PC):
ρ f  2  2  2rt  
∆PC = u 1 1 + 2φ  − 1 + 2u 22  (17)
203    re  
Here the inlet duct velocity is u1 and the exit duct velocity is u2. rt is the radius of the
circle to which the center line of the inlet is tangential and re is the radius of exit pipe.
The factor φ can be taken from Figure 4, where the parameter Ψ is proportional the
ratio of As, which is the surface area of the cyclone exposed to the spinning fluid, and
A1 , which is the area of the inlet duct:
As
Ψ = fC
A1
The friction factor (fC) is 0.005 for gases. For design purpose As can be taken as equal
to the surface area of a cylinder with the same diameter as the cyclone and length
equal to the total height of the cyclinder (barrel pus cone).

13
Fig. 4. Cyclone pressure drop factor (Sinnott)

14
5 CALCULATION OF THE PRESSURE SHELL
The dimensions of the pressure shell can be calculated according to the Swedish
standards for pressure vessels (Tryckkärlskommissionen6). In order to simplify the
calculation no holes and welding seams are considered.

5.1 The cylinder

The minimum thickness of the wall (Smin) for the cylinder was calculated with
Equation (18).
D⋅P
Smin = (18)
σ
20 des ⋅ z
Sf
D is the inner diameter of the shell and P is the design pressure (overpressure). The
security factor was taken as Sf = 1.5. Z is a strength factor, which depends on holes
and welding seams in the shell. Here, we neglect the impact of holes and welding
seams and set the strength factor to one.
σdes is the design stress for the steal and can be found, for example, in the Swedish
standards for pressure vessels (Tryckkärlskommissionen7). These calculations are only
valid for Smin/D ≤ 0.05.

5.2 The ellipsoidal head

The minimum wall thickness of the head can be calculated according to Equation
(19).
Dy ⋅ P ⋅ y
Smin = (19)
σ
20 ⋅ des ⋅ z
Sf
Here Dy is the outer diameter of the ellipsoidal head and y is a form factor, depending
on the shape of the head. If the form for the head is determined by Equations (20) to
(22), the form factor is y = 1.3.

h = height of the ellipsoidal head = 0.25 ⋅ D y [mm] (20)

R = crown radius = 0.8 ⋅ D y [mm] (21)
r = knuckle radius = 0.154 ⋅ D y [mm] (22)

For the ellipsoidal head, the security factor is Sf = 1.1 mm

15
6 CONCLUSION
The presented mathematical model for calculating the pressure drop of an
interconnected pressurised fluidised bed reactor (IPFBR) for chemical looping
combustion (CLC) gives an idea about the pressure loss in such a reactor depending
on its dimensions. However, the size of the cyclone system is very important for the
overall size of the reactor. For this reason a more detailed model for the cyclone
system is required.

17
7 NOMENCLATURE
A = area [m2]
A1 = area of the inlet duct (Eq. 17) [m2]
As = surface area of cyclone exposed to the spinning fluid (Eq. 17) [m2]
CD = drag coefficient [-]
cp = heat capacity [kJ/kg,K]
D = inner diameter of the shell (Eq. 18) [mm]
Dy = outer diameter of the ellipsoidal head (Eq. 19) [mm]
dp = particle diameter [mm]
fC = friction factor, taken as 0.005 for gases (Eq. 17) [-]
g = gravitation constant [m/s2]
HBf = height of the fluidised bed in the fuel reactor [m]
Hd = height of the dense region [m]
HF = height of the freeboard [m]
Hf = height of the fuel reactor [m]
hi = specific enthalpy [kJ/kg]
M = mass of fluidised bed [kg]
m& = mass flowrate [kg/s]
n = isentropic coefficient (Eq.8) [-]
P = pressure [bar]
P = design pressure (overpressure) (Eq. 18 + 19) [bar]
∆PC = cyclone pressure drop (Eq. 17) [mbar]
ReT = Reynolds number for particles at uT [-]
rt = radius of circle to which the center line of the inlet is tangential [m]
re = radius of exit pipe (Eq. 17) [m]
Sf = security factor (Eq. 18 + 19) [-]
T = temperature [°C]
uT = terminal velocity [m/s]
u1 = inlet duct velocity (Eq. 17) [m/s]
u2 = exit duct velocity (Eq. 17) [m/s]
W& = fan power [MW]
F

z = variable of height [m]

z = strength factor (Eq. 18 + 19) [-]

7.1 Greek letters

ε = voidage = (Vg-Vs)/Vg [-]
εs = solid fraction = 1- ε [-]

19
εs = mean solid fraction [-]
εsd = solid fraction in the dense region [-]
εsp = solid fraction at saturated pneumatic conditions [-]
ηfan = isentropic efficiency of the fan [-]
µ = dynamic viscosity [N,s/m2]
ρf = gas density (Eq. 17) [kg/m3]
ρg = fluid density [kg/m3]
ρs = particle density [kg/m3]
φ = factor from Figure 4 (Eq. 17) [-]
Ψ = Parameter in Figure 4 (Eq. 17) [-]
τ = residence time/ reaction time [s]
σdes = design stress (Eq. 18 + 19) [N/mm2]

7.2 Indices
a = air reactor
B = fluidised bed
C = cyclone
d = dense region
D = distributor
F = freeboard
f = fuel reactor
g = gas
s = solid

7.3 Abbreviation
CLC = chemical-looping combustion
PFBC = pressurized fluidised bed combustion
IPFBR = two interconnected pressurised fluidised bed reactors
TDH = transport disengaging height
TIT = turbine inlet temperature (temperature of the gas when it
enters the first expander step

20
8 REFERENCES

1. Carberry, J. and Varma, A., Chemical Reaction and Reactor Engineering,

Dekker, New York, USA 1986, ISBN: 0-8247-7543-0.
2. Kuni, D. and Levenspiel, O., Fluidization Engineering, 2nd Edition,
Butterworth-Heinemann, USA 1991, ISBN 0-409-90233-0.
3. Sinnott, R.K, Coulson and Richardson’s, Chemical Engineering, Vol. 6,
Butterworth-Heinemann, Great Britain 1996, ISBN 0-7506-2558-9
4. Basu, P. and Fraser, S. A., Circulating Fluidized Bed Boilers – Design and
Operations, Butterworth-Heinemann, USA 1991, ISBN 0-7506-9226-X.
5. Howard, J.R., Fluidized Bed Technology - Principles and Application, Adam
Hilger, Bristol, UK, 1989
6. Tryckkärlskommissionen, Tryckkärlsnormer – Normer för
hållfastighetsberäkning av tryckkärl, fifth edition, Lagerblads Tryckeri AB,
Stockholm, Sweden 1987, ISBN 91-85254-00-2.
7. Tryckkärlskommissionen, Tryckkärlsnormer Kapitel 4 - Material, Lagerblads,
Karlshamn, Sweden 1987, ISBN 91-85254-00-2.

21