1531

A publication of

CHEMICAL ENGINEERING TRANSACTIONS

VOL. 57, 2017 The Italian Association
of Chemical Engineering
Online at www.aidic.it/cet
Guest Editors: Sauro Pierucci, Jiří Jaromír Klemeš, Laura Piazza, Serafim Bakalis
Copyright © 2017, AIDIC Servizi S.r.l.
ISBN 978-88-95608- 48-8; ISSN 2283-9216 DOI: 10.3303/CET1757256

Optimization of Fluidized Catalytic Cracking Unit Regenerator

to Minimize CO2 Emissions
John M. Yakubua, Raj Patela, Iqbal M. Mujtaba*a
Chemical Engineering Division, School of Engineering, University of Bradford, Bradford, BD7 1DP, UK
[email protected]

The Fluidized Catalytic Cracking (FCC) is known for its ability to convert products considered as wastes from
various refinery units into useful fuels such as gasoline and diesel. It is considered as one of the most
important units of the refinery. In spite of its importance, it is also considered as the major contributor of
carbon dioxide. This pollutant, CO2 comes mainly from the regeneration of spent catalyst in the regenerator. In
this study, minimization of carbon dioxide exiting the dense bed of the regenerator is considered using model
based techniques. The model for the regenerator dense bed was adopted from the literature for simulation.
From the simulation, the exit mole fraction of carbon dioxide was found to be 16.21%. The minimized mole
fraction was obtained as 15.36%. This is a 5.24% reduction on the yield of carbon dioxide which in turn will
reduce the overall mass of carbon dioxide produced and released into the air or captured and stored. Any
reduction achieved on carbon dioxide emission is progress made on solving the problems of global warming.

1. Introduction
The FCC unit converts heavy petroleum products such as vacuum gas oils and residues into more valuable,
lower molecular-weight fuels such as gasoline and light products. The FCC unit has two major reactors: the
riser where cracking reactions take place and the regenerator where the burning of coke is accomplished.
During this process referred to as regeneration, large amount of flue gases; CO, CO 2, SO2, SO3, NO, N2O and
N2 are generated (Wauquier, 1994). The flue gases are mostly considered as pollutants to the environment,
hence, they are required to be found in little quantity in the air. The amount of CO 2 emitted from the FCC unit
is about 30% of the total CO2 emitted from the refinery and it is considered the highest in oil refineries (de
Mello, Gobbo, Moure, & Miracca, 2013). Hence, the refinery is a major contributor to the Green House Gas
(GHG), a culprit of global warming. To stop the use of fossil fuels may not be practicable because various
projections make clear that fossil fuels will continue to be needed while renewable energy sources are not
sufficient.
A recent report from the Nobel Prize-winning Intergovernmental Panel on Climate Change (IPCC) concluded
that global CO2 emissions must be cut by 50-80% by 2050 in order to elude the most destructive effects of
climate change (CCP, 2016). To cut down on the CO2 emission, Carbon Capture and Storage (CCS) is
playing a vital role (de Mello et al., 2013; Metz, Davidson, Coninck, Loos, & Meyer, 2005), however, the
approach has been proposed for more than 30 years, little is achieved with respect to commercial success of
CCS projects. The principal concern is where to stockpile the immense volume of captured pure CO 2 every
year (Peng & Zhuang, 2012). Therefore, an approach capable of mitigating the emission is required. To
achieve this goal, the use of operational changes to bring about emissions reduction can be carried out in the
FCC unit to reduce the extent of emission before it is being captured and stored (Moore, 2005).
This work will focus on minimizing the yield of CO2 from FCC regenerator flue-gas as an important step in
mitigating CO2 emission of the refinery. Simulation and optimization of the regenerator can identify the scope
for reducing the emission. The FCC regenerator is divided into dense bed and freeboard. The dense bed is
modelled as a mixed-tank model for energy and coke balances but a plug flow reactor model for gas
component balances. The freeboard is modelled as a plug flow reactor. In this work, only the dense bed is
considered because most of the solids and gases are in the dense bed where almost all reactions take place

Please cite this article as: Yakubu J., Patel R., Mujtaba I.M., 2017, Optimization of fluidized catalytic cracking unit regenerator to minimize co2
emissions, Chemical Engineering Transactions, 57, 1531-1536 DOI: 10.3303/CET1757256
1532

(Pinheiro et al., 2012), and the fact that the dense bed model can be used for the overall regenerator
dynamics (Bollas et al., 2007).
To carry out the optimization studies, a single objective function was developed and implemented in gPROMS
software which uses a successive reduced quadratic programing (SRQPD) optimization technique, a
Sequential Quadratic Programming based solver to minimize the yield of CO2. The optimization is done using
the mathematical models (Han & Chung, 2001a, 2001b) of the regenerator and results obtained will be
compared with CO2 emissions from literature data.

2. Regenerator
The FCC regenerator involves very strong exothermic coke burning reaction which takes place in fluidized bed
reactor with composite hydrodynamics (Pinheiro et al., 2012). Different FCC units have different regenerator
configurations. Some have single stage and others have two-stage regenerators like the Orthoflow F unit
(Chiyoda, 1980), however, all regenerator have similar coke burning kinetics. Most FCC regenerators have
two sections; dense bed and a freeboard. The dense bed is divided into two, that is, the emulsion containing
much of catalyst where coke burning reactions takes place to produce regenerated catalyst, and the bubble
phase with some entrained catalyst but having much of gaseous reactions converting CO to CO 2. The
regenerator dynamics follow the well-known two-phase theory of fluidization (Kunii & Levenspiel, 1991). In the
regenerator, several coke burning reactions take place; the homogenous and heterogeneous. The
heterogeneous reactions happen in the phases where catalyst is present that is the emulsion phase and the
freeboard, while the homogeneous reactions happen in gaseous phase. The regenerator bed in this work is
modelled as a perfectly mixed reactor, hence no temperature gradient is considered in the bed (Cuadros,
Melo, Filho, & Maciel, 2012).
The coke burning kinetics are as follows:
𝐶𝐻𝑞 + (0.5 + 0.25𝑞)𝑂2 → 𝐶𝑂2 + 0.25𝑞𝐻2 𝑂 (1)
𝑟1𝑖 = 𝑘1𝑅𝐺 𝐶𝑐𝑘𝑖 𝐶𝑂2𝑖 , i = emulsion, freeboard (2)
𝐶𝐻𝑞 + (1 + 0.25𝑞)𝑂2 → 𝐶𝑂2 + 0.25𝑞𝐻2 𝑂 (3)
𝑟2𝑖 = 𝑘2𝑅𝐺 𝐶𝑐𝑘𝑖 𝐶𝑂2𝑖 , i = emulsion, freeboard (4)
Equations 1 and 3 are coke burning reactions, while Equations 2 and 4 are their rates of reaction respectively
𝑘 𝑘 𝜎
(Weisz, 1966). The constants of reaction𝑘1𝑅𝐺 = 1∗𝑅𝐺, and 𝑘2𝑅𝐺 = 2∗𝑅𝐺 .
1+ 𝜎 1+ 𝜎

The gaseous reactions that take place in the regenerator is obtained from Han and Chung (2001a) and given
in Equation (5):
2𝐶𝑂 + 𝑂2 → 2𝐶𝑂2 (5)
When Equation (5) is homogeneous, it occurs in all the phases of the reactor, and its rate of reaction is given
in Equation (6).
𝑟3𝑖 = 𝑘3∗𝑅𝐺 𝐶𝐶𝑂𝑖 𝐶𝑂0.5 𝐶 0.5 ,
2 𝑖 𝐻2 𝑂𝑖
i=bubble, emulsion, freeboard (6)

When Equation (5) is heterogeneous, it occurs only in the emulsion and freeboard phases of the reactor, and
its rate of reaction is given in Equation (7).
𝑟4𝑖 = 𝑘4∗𝑅𝐺 𝐶𝐶𝑂𝑖 𝐶𝑂0.5
2𝑖
, i = emulsion, freeboard (7)

2.1 The regenerator mathematical model

The mathematical model of the regenerator was adopted from the work of Han and Chung (2001a) and the
initial conditions and some process variables were taken from Han and Chung (2001b). The emulsion phase is
modelled as a CSTR for energy and coke balances but a plug flow reactor model for gas component balances
in the bubble phase. The component balance equations for the gaseous phases in emulsion and bubble are
defined by the partial differential equations:
𝜕𝐶𝑖𝐸 𝜕𝐶𝑖𝐸 𝐶𝑖𝐸 𝑆𝑔𝐸 𝐾𝐼
= −𝑣𝑔𝐸 − + (𝐶𝑖𝐵 − 𝐶𝑖𝐸 ) + 𝑅𝑖𝐸 (8)
𝜕𝑡 𝜕𝑧 𝜀𝑔𝐸 𝜀𝑔𝐸

𝑖 = 𝑂2 , 𝐶𝑂, 𝐶𝑂2 , 𝐻2 𝑂, 𝑁2 ,
𝜌𝑐 𝜀𝑐𝐷 (0.5+0.25𝑞)𝑟1𝐸 (1+0.25𝑞)𝑟2𝐸
𝑅𝑂2𝐸 = − [ + + 0.5𝑟4𝐸 ] − 0.5𝑟3𝐸 (9)
𝜀𝑔𝐸 𝑀𝑤𝑐𝑘 𝑀𝑤𝑐𝑘
 1533

𝜌𝑐 𝜀𝑐𝐷 𝑟1𝐸
𝑅𝐶𝑂𝐸 = [ − 𝑟4𝐸 ] − 𝑟3𝐸 (10)
𝜀𝑔𝐸 𝑀𝑤𝑐𝑘

𝜌𝑐 𝜀𝑐𝐷 𝑟2𝐸
𝑅𝐶𝑂2𝐸 = [ + 𝑟4𝐸 ] + 𝑟3𝐸 (11)
𝜀𝑔𝐸 𝑀𝑤𝑐𝑘
𝜌𝑐 𝜀𝑐𝐷𝑞
𝑅𝐻2𝑂𝐸 = [0.5𝑟1𝐸 + 0.5𝑟2𝐸 ] (12)
𝜀𝑔𝐸 𝑀𝑤𝑐𝑘

𝑅𝑁2𝐸 = 0 (13)
𝜕𝜀𝑔𝐸 𝜕(𝑣𝑔𝐸 𝜀𝑔𝐸 )
𝑆𝑔𝐸 = + (14)
𝜕𝑡 𝜕𝑧
(0,𝑧) (𝑧)
I.C 𝐶𝑖𝐸 = 𝐶𝑖𝐸 𝑖 = 𝑂2 , 𝐶𝑂, 𝐶𝑂2 , 𝐻2 𝑂, 𝑁2 , (15)
(0)
(𝑡,0) 𝑓𝑖𝐸 𝜌𝑔𝑅𝐺
B.C 𝐶𝑖𝐸 = (0) 𝑖 = 𝑂2 , 𝐶𝑂, 𝐶𝑂2 , 𝐻2 𝑂, 𝑁2 , (16)
𝑀𝑤𝑔𝑅𝐺

𝜕𝐶𝑖𝐵 𝜕𝐶𝑖𝐵 𝐶𝑖𝐵 𝑆𝑔𝐵 𝐾𝐼

= −𝑣𝑔𝐵 − + (𝐶𝑖𝐸 − 𝐶𝑖𝐵 ) + 𝑅𝑖𝐵 (17)
𝜕𝑡 𝜕𝑧 𝜀𝑔𝐵 𝜀𝑔𝐵

𝑖 = 𝑂2 , 𝐶𝑂, 𝐶𝑂2 , 𝐻2 𝑂, 𝑁2 ,
𝑅𝑂2𝐵 = − 0.5𝑟3𝐸 (18)

Figure 1 shows the schematic diagram of the regenerator as modelled in this work.

Catalyst to
Riser
Gases

Dense bed

Bubble Emulsion

Air Catalyst from

Riser

Figure 1: Schematic diagram of the regenerator dense bed model

Table 1 shows the regenerator parameters and initial conditions used in this work.

Table 1: Regenerator parameters and initial conditions

Parameter Value
Mass flowrate of Air (kg/s) 66.09
3
Density of catalyst (kg/m ) 1410
Holdup of catalyst (kg) 182000
Temperature of dense bed (K) 991
3
O2 in emulsion and bubble at t = 0 (kg mol/m ) 0.0005
3
CO in emulsion and bubble at t = 0 (kg mol/m ) 0.0003
3
CO2 in emulsion and bubble at t = 0 (kg mol/m ) 0.004
3
H2O in emulsion and bubble at t = 0 (kg mol/m ) 0.003
3
N2 in emulsion and bubble at t = 0 (kg mol/m ) 0.02

3. Optimization problem formulation

Different modelling and optimization platform/software such as Matlab and Hysys were used for FCC
regenerator simulations/optimization but not gPROMS, in spite of its robustness. gPROMS uses the
successive reduced quadratic programing (SRQPD), it is a nonlinear programming optimization technique
capable of handling the nonlinearity of the partial differential and algebraic equations that described the
regenerator. In this work gPROMS is used for the regenerator dense bed optimization to minimize the yield of
CO2 from the dense bed of the FCC unit regenerator.
1534

The optimization problem can be described as:

Given the fixed volume of the dense bed regenerator
Optimize the mass flowrate of air
So as to minimize the yield of CO2
Subject to constraints on the yield of CO

Mathematically, the optimization problem can be written as;

min 𝑇(𝑥)𝑜𝑟 𝐹𝐽 (𝑥) 𝑍 (26)

𝑠. 𝑡.
𝑓(𝑥, 𝑧 ′ (𝑥), 𝑧(𝑥), 𝑢(𝑥), 𝑣) = 0 (model equations) (27)
𝑥𝑓 = 𝑥𝑓∗ (28)
𝐹𝐿 ≤ 𝐹𝑎𝑖𝑟 ≤ 𝐹𝑈 (29)
∗
𝑌𝐶𝑂 < 𝑌𝐶𝑂 (30)
Where 𝑍 is the yield of carbon-dioxide, 𝐹𝑎𝑖𝑟 the mass flow rate of air into the regenerator, 𝑥𝑓 the height of the
regenerator, 𝑌𝐶𝑂 the yield of carbon monoxide, 𝐹𝐿 and 𝐹𝑈 the lower and upper bounds of the mass flowrate of
𝑘𝑔
air (60 ≤ 𝐹𝑎𝑖𝑟 ≤ 80 ), 𝑥𝑓∗ the fixed height of the regenerator and 𝑌𝐶𝐷 ∗ the maximum allowable limit for carbon
𝑠
monoxide 𝑌𝐶𝑂 < 0.0002 .

4. Results and discussions

This section shows both simulation and optimization results. The results are presented to show the capability
of gPROMS in solving complex nonlinear PDAEs by validating the results against those predicted by the same
model but using different solution software as DSim-FCC (Han & Chung, 2001b).

4.1 Simulation
When air comes in contact with coke on the surface of the catalyst, the coke gets burned and the catalyst is
regenerated under high temperature which provides enough energy that is required for the endothermic
cracking of gas oil in the riser. Figures 2 shows the yields of carbon dioxide from the regenerator during the
coke burning reactions.

0,13-0,14 0,14-0,15 0,15-0,16 0,16-0,17

0,17
0,16
Mole fraction

0,15
0,14
0,13
0
2000
4000
6000

100…
8000

120…
140…
160…

180…

200…

220…

Time (s)
240…

6
3
0
Height (m)

Figure 2: Concentration of carbon dioxide from dense bed - simulation

 1535

At 66.09 kg/s gas flowrate and constant temperature of 991 K, the mole fraction of carbon dioxide at the exit of
the dense bed of the regenerator is 0.1621. This is 16.21 % carbon dioxide, 0.23 % carbon monoxide, 10.95
% water, 72.24 % nitrogen and 0.36 % oxygen. These results are very much closer to gases mole fractions
obtained by Han and Chung (2001b) where the model of the regenerator in this work was taken. Han and
Chung (2001b) obtained 14.80 % carbon dioxide, 0.60 % carbon monoxide, 9.20 % water and 0.20 % oxygen.
The CO2 yield from this simulation (16.21%) is higher than what was obtained by Han and Chung (2001b)
(14.80%), this is because the mass flowrate of air used in this simulation is 66.09 kg/s while the mass flowrate
of the simulation of Han and Chung (2001b) is 34 kg/s. Again, only the regenerator dense bed was considered
in this work while Han and Chung (2001a, b) considered the entire FCC unit (riser, disengage, stripper and the
regenerator – including the freeboard which was not considered in this work). This could account for the
difference in the CO2 yield for both simulations, however, the simulation results in this work agrees with other
literature results with little margin of errors of less than 5% (Zheng, 1994).
Figure 3 shows the results of the minimization of carbon dioxide using the optimized process condition.

0,135-0,14 0,14-0,145 0,145-0,15 0,15-0,155

0,155
Mole fraction

0,15

0,145

0,14

0,135

Time (s)
6
3
0
Height (m)

Figure 3: Concentration of carbon dioxide from dense bed – optimization

The mass flowrate of air for the simulation is 66.09 kg/s, while the optimized mass flowrate of air is 83.09 kg/s.
This is a 20 kg/s increase in the mass flowrate of air to the regenerator bringing about a slight reduction on the
mole fraction of carbon dioxide, which is 0.1536 at the exit of the dense bed. Compared with the mass
fraction of 0.1621 of the simulation result, it shows a decrease of 5.24 % of carbon dioxide emitted at the exit
of the reactor. Though, the optimization result in this work could not compare favourably with Han and Chung
(2001b) simulation result, it shows the simulation result of this work was minimized by 5.24%. As stated
earlier, the regenerator model used in this model was for the dense bed only, further work would be required
to include the freeboard of the regenerator and all other units of the FCC as was done by Han and Chung
(2001a, b). It is expected that with the increase in air mass flowrate, more carbon dioxide should be produced,
due to availability of oxygen to burn more coke. However, it was observed that the catalyst holdup decreased
slightly, and that could reduce the amount of coke available for the exothermic reaction. This observation is
consistent with what was presented by Han and Chung (2001b).

5. Conclusions
The regenerator of FCC unit was simulated and optimized to minimize the carbon dioxide exit concentration
so as to cut down on emission of the greenhouse gas. With an increase of 20 kg/s mass flowrate of air, 5.24
1536

% of carbon dioxide was reduced. On carbon dioxide emission, 5.24 % reduction is good step in cutting down
the effect of CO2 emission from the FCC unit on global warming.
The regenerator model considered is not exhaustive, because it did not consider the freeboard and other
section of the FCC unit like the riser, disengage and stripper, hence, further work is ongoing to include all the
units as simulated concurrently by Han and Chung (2001a, b). This will capture the entire hydrodynamics of
the regenerator which will provide detailed insight into the optimization of the unit. Nevertheless, with 5.24 %
reduction on this simulation, it shows that using operational changes in process variables of the regenerator
can bring about great reduction in CO2 emission.

Acknowledgments

Thanks to Petroleum Technology Development Fund, Nigeria, for the sponsorship.

Reference
BOLLAS, G. M., VASALOS, I. A., LAPPAS, A. A., IATRIDIS, D. K., VOUTETAKIS, S. S. & PAPADOPOULOU,
S. A. 2007. Integrated FCC riser—regenerator dynamics studied in a fluid catalytic cracking pilot plant.
Chemical Engineering Science, 62, 1887-1904.
CCP. 2016. What is CO2 Capture & Storage? [Online]. Available:
http://www.co2captureproject.org/what_is_co2_capture_storage.html [Accessed 27/09/2016 2016].
CHIYODA 1980. Operating and Unit Manual for the Nigerian National Petroleum Corporation.
CUADROS, J. F., MELO, D. C., FILHO, R. M. & MACIEL, M. R. W. 2012. Fluid Catalytic Cracking
Environmental Impact: Factorial Design Coupled with Genetic Algorithms to Minimize Carbon Monoxide
Pollution. Chemical Engineering Transactions; The Italian Association of Chemical Engineering, 26.
DE MELLO, L. F., GOBBO, R., MOURE, G. T. & MIRACCA, I. 2013. Oxy-combustion Technology
Development for Fluid Catalytic Crackers (FCC) – Large Pilot Scale Demonstration. Energy Procedia, 37,
7815-7824.
HAN, I.-S. & CHUNG, C.-B. 2001a. Dynamic modeling and simulation of a fluidized catalytic cracking process.
Part I: Process modeling. Chemical Engineering Science, 56, 1951-1971.
HAN, I.-S. & CHUNG, C.-B. 2001b. Dynamic modeling and simulation of a fluidized catalytic cracking process.
Part II: Property estimation and simulation. Chemical Engineering Science, 56, 1973-1990.
KUNII, D. & LEVENSPIEL, O. 1991. Fluidization Engineering, London, Butterworth Heinemann Seris in
Chemical Engineering.
METZ, B., DAVIDSON, O., CONINCK, H. D., LOOS, M. & MEYER, L. (eds.) 2005. CARBON DIOXIDE
CAPTURE AND STORAGE, Cambridge University Press. 40 West 20th Street, New York, NY 10011–
4211, USA: Cambridge University Press.
MOORE, I. 2005. Reducing CO2 emissions. Refining [Online]. Available:
http://www.eptq.com/view_article.aspx?intAID=211 [Accessed 28/09/2016].
PENG, P. & ZHUANG, Y. 2012. The Evaluation and Comparison of Carbon Dioxide Capture Technologies
Applied to FCC Flue Gas. Renewable and Sustainable Energy, Pts 1-7, 347-353, 1479-1482.
PINHEIRO, C. I. C., FERNANDES, J. L., DOMINGUES, L., CHAMBEL, A. J. S., GRACA, I., OLIVEIRA, N. M.
C., CERQUEIRA, H. S. & RIBEIRO, F. R. 2012. Fluid Catalytic Cracking (FCC) Process Modeling,
Simulation, and Control. Industrial & Engineering Chemistry Research, 51, 1-29.
WAUQUIER, J.-P. 1994. Petroleum Refining. Crude oil. Petroleum Products. Process Flosheets, Paris,
France., IFP Publications.
WEISZ, P. B. 1966. Combustion of Carbonaceous Deposits within Porous Catalyst Particles III. The CO 2/CO
Product Ratio. Journal of Catalysis, 6, 425-430.
ZHENG, Y.-Y. 1994. Dynamic Modeling and Simulation of a Catalytic Cracking Unit. Computers & Chemical
Engineering, 18, 39-44.