-
This article was downloaded by: [Johns Hopkins University]On: 13
January 2015, At: 08:24Publisher: Taylor & FrancisInforma Ltd
Registered in England and Wales Registered Number: 1072954
Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T
3JH, UK
Click for updates
Petroleum Science and TechnologyPublication details, including
instructions for authors andsubscription
information:http://www.tandfonline.com/loi/lpet20
A Comparative Study of Different KineticLumps Models in the
Fluid CatalyticCracking Unit Using COMSOL MultiphysicsD. Yousuoa
& S. E. Ogbeideaa Department of Chemical Engineering, Faculty
of Engineering,University of Benin, Benin City, NigeriaPublished
online: 22 Dec 2014.
To cite this article: D. Yousuo & S. E. Ogbeide (2015) A
Comparative Study of Different KineticLumps Models in the Fluid
Catalytic Cracking Unit Using COMSOL Multiphysics, Petroleum
Science andTechnology, 33:2, 159-169, DOI:
10.1080/10916466.2014.958237
To link to this article:
http://dx.doi.org/10.1080/10916466.2014.958237
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy
of all the information (theContent) contained in the publications
on our platform. However, Taylor & Francis,our agents, and our
licensors make no representations or warranties whatsoever as tothe
accuracy, completeness, or suitability for any purpose of the
Content. Any opinionsand views expressed in this publication are
the opinions and views of the authors,and are not the views of or
endorsed by Taylor & Francis. The accuracy of the Contentshould
not be relied upon and should be independently verified with
primary sourcesof information. Taylor and Francis shall not be
liable for any losses, actions, claims,proceedings, demands, costs,
expenses, damages, and other liabilities whatsoever orhowsoever
caused arising directly or indirectly in connection with, in
relation to or arisingout of the use of the Content.
This article may be used for research, teaching, and private
study purposes. Anysubstantial or systematic reproduction,
redistribution, reselling, loan, sub-licensing,systematic supply,
or distribution in any form to anyone is expressly forbidden. Terms
&
-
Conditions of access and use can be found at
http://www.tandfonline.com/page/terms-and-conditions
Dow
nloa
ded
by [J
ohns
Hop
kins U
nivers
ity] a
t 08:2
4 13 J
anua
ry 20
15
-
Petroleum Science and Technology, 33:159169, 2015Copyright C
Taylor & Francis Group, LLCISSN: 1091-6466 print / 1532-2459
onlineDOI: 10.1080/10916466.2014.958237
A Comparative Study of Different Kinetic Lumps Modelsin the
Fluid Catalytic Cracking Unit Using
COMSOL Multiphysics
D. Yousuo1 and S. E. Ogbeide11Department of Chemical
Engineering, Faculty of Engineering, University of Benin,
Benin City, Nigeria
The COMSOL Multiphysics computational fluid dynamics software
was used to simulate the fluidcatalytic cracking (FCC) riser
reactor of the FCC unit. The 4-, 5-, 10-, 20-, and 35-lump kinetic
modelswere used to describe the kinetics of the cracking reactions
in the riser reactor. The results of the kineticlump models of the
distributions of pressure, velocity, temperature, and yields of
products were comparedusing practical values from the Port Harcourt
Refinery Company plant. The results showed that the higherthe
kinetic lumps the better the accuracy of the prediction of the
product yields and that the 10-, 20-, and35-lump kinetic models
could be used to predict the yield of various fractions of the
riser reactor usingCOMSOL Multiphysics.
Keywords: Fluid catalytic cracking, FCC, computational fluid
dynamics, CFD, Riser reactor, lumpingschemes, product yields
1. INTRODUCTION
The kinetics modeling of fluid catalytic cracking (FCC) has been
traditionally based on using alumping strategy: chemical species
with similar behaviors are grouped together forming a similarnumber
of pseudo species. The lumping of species is important to make the
kinetic modeling atraceable exercise. Detail work on kinetic
lumping has been reported previously (Weekman and Nace,1970;
Pitault et al, 1994; Gao et al., 1999; Gupta et al., 2005; Ahari et
al., 2008; Hernandez-Barajaset al., 2009; Heydari et al., 2010;
Jiang et al., 2013).
In this study, the COMSOL Multiphysics computational fluid
dynamics (CFD) software was usedto simulate the FCC riser reactor
of the FCC unit (FCCU). The 4-, 5-, 10-, 20- and 35-lump
kineticmodels were used to describe the kinetics of the cracking
reactions in the riser reactor to investigatethe kinetics of
lumping scheme in the riser and to ascertain the comparative
advantage of one lumpover the other.
Address correspondence to D. Yousuo, Department of Chemical
Engineering, Faculty of Engineering, University ofBenin, P.M.B.
1154, Benin City, Nigeria. E-mail: [email protected]
Color versions of one or more of the figures in the article can
be found online at www.tandfonline.com/lpet.
159
Dow
nloa
ded
by [J
ohns
Hop
kins U
nivers
ity] a
t 08:2
4 13 J
anua
ry 20
15
-
160 D. YOUSUO AND S. E. OGBEIDE
FIGURE 1 The FCC reactor.
2. METHODOLOGY
2.1 The Riser Kinetic Model
The modeling was based on the schematic flow diagrams of the
riser reactor in the FCCU reactoras presented in Figure 1. The FCCU
reactor consists of the riser reactor, reactor catalyst
stripper,reactor separator or disengager, reactor cyclones, and
other auxiliary parts. The riser reactor is 33 m
Dow
nloa
ded
by [J
ohns
Hop
kins U
nivers
ity] a
t 08:2
4 13 J
anua
ry 20
15
-
DIFFERENT KINETIC LUMPS MODELS IN THE FCC UNIT 161
HFOPh
HFONh
HFOAh
HFORhLFONh
G
LFOAh
LFORh
LFOPh
C
K9
K6
K4
K12
K10
K7 K3
K5
K19
K11
K16
FIGURE 2 Ten-lump kinetic scheme.
long and the diameter is 0.8 m. The diagram for the 10-lump
kinetic scheme of the FCCU is shown inFigure 2 while that of the
other lumps have been shown elsewhere (Yousuo, 2014). The lumps for
the10-lump kinetic scheme are the heavy fuel oils from the
paraffins (HFOph), the heavy fuel oils fromthe naphtenes (HFONh),
the heavy fuel oils from the aromatic substituent groups (HFOAh),
the heavyfuel oils of the carbons among the aromatic rings (HFORh),
the light fuel oils from the paraffins(LFOph), the light fuel oils
from the naphthenes (LFONh), the light fuel oils from the
aromaticsubstituent groups (LFOAh), the light fuel oils of the
carbons among the aromatic rings (LFORh),gasoline (G), and COKE
(C). In this model COKE (C) represents 50% coke and 50% C1-C4
gases.The rate expressions of the 10-lump kinetic scheme and other
details are shown elsewhere (Jacobet al., 1976).
2.2 Plug-flow Reactor Equations
The reactor model is an ideal plug-flow reactor, described by
the mass balance in Eq. (1). Assumingconstant reactor cross section
and flow velocity, the species concentration gradient as fraction
ofresidence time ( ) is given in Eq. (2). The reaction rates are
given by rf = KjCi and to account forthe different time scales, two
different activity functions are used. For the non-coking reactions
theactivity function is given in Eq. (3).
(1)
(2)(3)
Dow
nloa
ded
by [J
ohns
Hop
kins U
nivers
ity] a
t 08:2
4 13 J
anua
ry 20
15
-
162 D. YOUSUO AND S. E. OGBEIDE
The reaction rates are modified by the activity according to Eq.
(4). For the coking reactions, theactivity function is given by Eq.
(5) where is a deactivation constant depending on the
residencetime. The modified reaction rates are given by Eq. (6).
The coke content is given by Eqs. (7) and (8).The values of a, b, ,
and are obtained from Gupta et al. (2005), Ahari et al. (2008), and
Yousuo(2014) as shown in Eqs. (9) and (10).
(4)(5)
(6)
(7)
(8)(9)
(10)For the mass transport, the inlet and outlet concentrations
are obtained from Eq. (11) and the velocityand pressure for ideal
gases are obtained from Eqs. (12) and (13), respectively. The
static head ofthe catalyst in the riser can be calculated using Eq.
(14). The details on choosing the void fractionvariable, assumed
gas velocity, slip factor, and the vaporization heat of the feed in
the riser inlet havebeen shown elsewhere (Gupta et al., 2005;
Yousuo, 2014).
Inlet : c = cin, Outlet : c = cout (11)
(12)
(13)
(14)For momentum transport, the inlet and outlet pressure are
obtained from Eq. (15)
(15)For energy balance, neglecting pressure drop, the energy
balance for an ideal reacting gas, as well asan incompressible
reacting liquid is given by Eqs. (16) and (17). The inlet
temperature is calculatedputting into consideration the energy
balance of the components. Equation (18) is used in calculatingthe
inlet temperature while Eq. (19) is used for calculating the outlet
temperature.
(16)
Dow
nloa
ded
by [J
ohns
Hop
kins U
nivers
ity] a
t 08:2
4 13 J
anua
ry 20
15
-
DIFFERENT KINETIC LUMPS MODELS IN THE FCC UNIT 163
(17)
At z = z0 = 0, ws = 0, Qext = 0, Eq. (16) and (17) becomes
i MiCp,idTdV
Q = 0
This implies that
That is
(18)
At z = h or z, ws = 0, Qext = 0, Eqs. (16) and (17) become
i MiCp,idTdV
= Q
That is,
This implies that
That is,
By our correlation
is
hence
Outlet: (19)
Dow
nloa
ded
by [J
ohns
Hop
kins U
nivers
ity] a
t 08:2
4 13 J
anua
ry 20
15
-
164 D. YOUSUO AND S. E. OGBEIDE
TABLE 1Boundary Conditions
SETTING BOUNDARY 3 BOUNDARY 4 BOUNDARIES 1 and 2
TemperatureBoundary type Inlet outlet wallBoundary condition
Temperature Temperature Thermal insulationValue T 0 T n
ConcentrationBoundary type Inlet outlet wallBoundary condition
Concentration Concentration Insulation/SymmetryValue cin for all
species cout for all species
Velocity and pressureBoundary type Inlet outlet wallBoundary
condition Velocity Pressurre, no viscous stress No slipValue w0 =
vs, uo = v0 = 0 P0 = Pn
2.3 Boundary Conditions
The boundary conditions for the riser reactor are shown in Table
1.
3. MATERIALS, MESH GENERATION AND SIMULATION
3.1 Materials
The average molecular weight, the thermodynamic properties of
the feed, the plant operating condi-tions and the properties of the
catalyst used in this study, the specific heat of different lumps,
and thekinetic parameters for cracking reactions can be found
elsewhere (Port Harcourt Refinery CompanyProject, 1987; Gupta et
al., 2005; Ahari et al., 2008).
3.2 Mesh Generation and Simulation
The extra fine mesh generator of the COMSOL Multiphysics
software was used to produce gridrefinement in the riser reactor.
The riser reactor was meshed into 77,358 triangular elements.
Figure 3shows the computational grid used to represent the
computational domain of the riser reactor. Thesimulations in this
work used the three-dimensional model of the COMSOL multiphysics
CFDsoftware in a Windows Vista Home Premium HP Pavilion dv 6500
Notebook PC (processor: IntelCore 2 Duo CPU T5450 @ 1.661.67 GHz;
memory [RAM]: 250 GB; and type: 32-bit operatingsystem).
4. RESULTS AND DISCUSSION
Figure 4 shows the temperature in the reactor riser. Gas
oil/heavy diesel oil, medium pressure steam,and fresh catalyst
enter the reactor riser at a temperature of 505, 464, and 1004 K,
respectively. Themedium pressure steam atomizes the gas oil/heavy
diesel oil as they travel up along the reactor riserincreasing
catalysis and the rate of reaction. The hydrocarbons and catalyst
mixture travel upward
Dow
nloa
ded
by [J
ohns
Hop
kins U
nivers
ity] a
t 08:2
4 13 J
anua
ry 20
15
-
DIFFERENT KINETIC LUMPS MODELS IN THE FCC UNIT 165
FIGURE 3 Computational domain and grid.
Dow
nloa
ded
by [J
ohns
Hop
kins U
nivers
ity] a
t 08:2
4 13 J
anua
ry 20
15
-
166 D. YOUSUO AND S. E. OGBEIDE
FIGURE 4 The temperature in the riser reactor.
and the temperature inside the FCC riser decreases because of
the endothermic cracking reactions.The mixture temperature of the
riser falls sharply to 803 K for the Port Harcourt Refinery
Company(PHRC) plant because sensible heat of catalyst coming from
the regenerator is utilized in providingheat for raising the
sensible heat of feed, for vaporizing the feed, and for further
heating of thevaporized feed.
Figure 5 shows the yield in the reactor riser of the PHRC plant
of the 10-lump kinetic model.HFOPh, HFONh, HFOAh, and HFORh are
broken down and as a result their weight fraction de-creased along
the riser from the inlet to the outlet. LFOP i, LFONi, LFOAi, and
LFORi lumps wereformed and later broken down to G and C lumps. The
gasoline (G) yield was 51% and coke (C)
FIGURE 5 The yield in the riser reactor.
Dow
nloa
ded
by [J
ohns
Hop
kins U
nivers
ity] a
t 08:2
4 13 J
anua
ry 20
15
-
DIFFERENT KINETIC LUMPS MODELS IN THE FCC UNIT 167
TABLE 2Kinetic Lumps, Predicted Values, and Deviation of
Predicted Values From the Practical Values
Predicted Value
PHRC Plant Data Practical Value 4-lump 5-lump 10-lump
20-lump
Gasoline yield, wt% 49.50 67 60 51 50.5Coke yield, wt% 5.90 6 6
6.25 7Outlet temperature, K 805 829 821 803 803
Deviation of Predicted Values From the Practical Values
Gasoline yield, wt% 49.50 17.5 10.5 1.5 1Coke yield, wt% 5.90
0.1 0.1 0.35 1.1Outlet temperature, K 805 24 16 2 2
yield was 12.50%. For the 10-lump kinetic model, coke (C)
represents 50% coke and 50% C1C4gases and therefore the actual coke
yield was 6.25%.
Table 2 shows the predicted values of the 4-, 5-, 10-, 20-, and
35-lump kinetic models, whichwere compared with the PHRC plant
practical values and the deviations of the predicted valuesfrom the
PHRC plant values. From the tables it is observed that with four
lumps, the deviation ofthe predicted values from the PHRC plant
values is 17.5% for gasoline yield and 0.1% for cokeyield, and
outlet temperature is 24 K; with five lumps, the deviation is 10.5%
gasoline yield and0.1% for coke yield, and the outlet temperature
is 16 K; with 10 lumps, the deviation is 1.5% forgasoline yield and
0.35% for coke yield, and outlet temperature is 2 K; with 20 and 35
lumps, thedeviation is 1% for gasoline yield and 1.1% for coke
yield, and outlet temperature is 2 K. Thisimplies that the
predicted values and the practical values from the PHRC plant
become closer asthe lumps increases and because the difference in
the deviation of the predicted values of gasoline,coke, and the
temperature for 10, 20, and 35 lumps are very small, and any from
the 10-lump kineticscheme could be used to predict the yield of
gasoline, coke, and temperature of the riser reactorusing COMSOL
Multiphysics.
5. CONCLUSION
The 4-, 5-, 10-, 20-, and 35- kinetic schemes were effectively
used to describe the kinetics of thecracking reactions in the FCCU.
The results showed that the predicted values and the
practicalvalues from the PHRC plant become closer as the lumps
increased. The results also showed that forgasoline, coke, and the
temperature the difference in the deviation of the predicted values
for 10, 20,and 35 lumps are very small and this could mean that any
from the 10-lump kinetic scheme couldbe used to predict the yield
of gasoline, coke, and temperature of the riser reactor using
COMSOLMultiphysics.
REFERENCES
Ahari, J. S., Farshi, A., and Forsat, K. (2008). A mathematical
modeling of the riser reactor in industrial FCC unit. Pet.
Coal50:1524.
Gao, J., Xu, C., Lin, S., and Yang, G. (1999). Advanced model
for turbulent gas-solid flow and reaction in FCC riser
reactors.AIChE J. 45:1095.
Dow
nloa
ded
by [J
ohns
Hop
kins U
nivers
ity] a
t 08:2
4 13 J
anua
ry 20
15
-
168 D. YOUSUO AND S. E. OGBEIDE
Gupta, R., Kumar, V., and Srivastava, V. K. (2005). Modeling and
simulation of fluid catalytic cracking unit. Rev. Chem.
Eng.21:95131.
Jacob, S. H., Gross, B., Voltz, S. E., and Weekman, V. M.
(1976). A lumping and reaction scheme for catalytic cracking.AICHE
J. 22:701713.
Jiang, L., Zheng-Hong, L., Xing-Ying, L., Chun-Ming, X., and
Jin-Sen, G. (2013). Numerical simulation of the turbulentgas-solid
flow and reaction in a polydisperse FCC riser reactor. Power
Technol. 237:569580.
Hernandez-Barajas, J. R., Vazquez-Roman, R., and Felix-Flores,
M. G. (2009). A comprehensive estimation of kineticparameters in
lumped catalytic cracking reaction models. Fuel 88:169178.
Heydari, M., AleEbrahim, H., and Dabir, B. (2010). Study of
seven-lump kinetic model in the fluid catalytic cracking unit.Am.
J. Appl. Sci. 7:7176.
Pitault, I., Nevicato, D., Foressier, M., and Bernard, J.-R.
(1994). Kinetic model based on a molecular description for
catalyticcracking of vacuum gas oil. Chem. Eng. Sci.
49:42494262.
Port Harcourt Refinery Company Project. (1987). Nigerian
National Petroleum Corporation Process. Project 9465A: Area 3FCCU
16.
Weekman, V. W., and Nace, D. M. (1970). Kinetics of catalytic
cracking selectivity in fixed, moving and fluid-bed reactors.AICHE
J. 16:397404.
Yousuo, D. (2014). Application of COMSOL multiphysics in the
simulation of the fluid catalytic cracking riser reactor
andcyclones. PhD thesis, Benin City, Nigeria: Department of
Chemical Engineering, University of Benin.
NOMENCLATURE
c: Concentration, mol/m3E: Activation energy for
rate constant, J/molg: Acceleration due to
gravity, m/sec2P: The pressure of gases, paR, r: Rate expression
valueT: Tempersature, Kt, : Residence time, secv: Volume, m3z:
Axial distance from the
inlet, mCP cat (Cpcat): Specific heat of
catalyst, J/kgKCp ds(Cpds): Specific heat of steam,
J/kgKCpL GO (CPLgo): Specific heat of liquid
gas oil, J/kgKCpV GO (CPVgo): Specific heat of gaseous
gas oil, J/kgKCi: Species molar concen-
trations, mol/m3cin: Inlet concentration,
mol/m3cout: Outlet concentration,
mol/m3Kd: Deactivation constant
M go (Mgo): Mass flow rate of gas oil,kg/sec
M ds (Mds): Mass flow rate of steam,kg/sec
M ca (Mcat): M cat (Mcat): Mass flowrate of catalyst, kg/sec
Pin: Inlet pressure, paRg (Ru): Gas constant, J/(mol.K)Tcat:
Temperature of the cata-
lyst, K: Void fractionTgo: Temperature of gas oil,
KTvap: Gas oil vapourization
temperature, Kv0: Outlet velocity, m/secTds: Temperature of
the
steam, KV R, , V: Reactor volume, m3Ws: Additional work termQ:
Heat due to chemical re-
action, J/m3.secQext: Heat added to the sys-
tem, J/m3.sec: Viscosity, N.S/m2: Density, kg/m3: Slip fact
Dow
nloa
ded
by [J
ohns
Hop
kins U
nivers
ity] a
t 08:2
4 13 J
anua
ry 20
15
-
DIFFERENT KINETIC LUMPS MODELS IN THE FCC UNIT 169
Subscripts
j: Refers to lump j that is crackedi: Refers to lump i that is
formedp (or s): Particle/solid
a (or f): Air/fluidcat: Catalystc: Coke content
Dow
nloa
ded
by [J
ohns
Hop
kins U
nivers
ity] a
t 08:2
4 13 J
anua
ry 20
15