Chemical Engineering Research and Design 1 3 2 ( 2 0 1 8 ) 235–251

Contents lists available at ScienceDirect

Chemical Engineering Research and Design

journal homepage: www.elsevier.com/locate/cherd

Mathematical modeling, simulation, and analysis

for predicting improvement opportunities in the
continuous catalytic regeneration reforming
process

Badiea S. Babaqi a,b,d , Mohd S. Takriff a,b,∗ , Siti K. Kamarudin a,c ,

Nur Tantiyani A. Othman a
a Department of Chemical and Process Engineering, Faculty of Engineering and Built Environment, Universiti
Kebangsaan Malaysia, 43600 Bangi, Selangor, Malaysia
b Research Centre for Sustainable Process Technology, Faculty of Engineering and Built Environment, Universiti

Kebangsaan Malaysia, 43600 Bangi, Selangor, Malaysia

c Fuel Cell Institute, Universiti Kebangsaan Malaysia, 43600 Bangi, Selangor, Malaysia
d Department of Chemical Engineering, Faculty of Engineering and Petroleum, Hadhramout University, Mukalla,

Hadhramout, Yemen

a r t i c l e i n f o a b s t r a c t

Article history: A mathematical model and simulation of the continuous catalytic regeneration reforming
Received 5 October 2017 process (CCRRP) were developed to identify the key opportunities for predicting the out-
Received in revised form 1 January put parameters and improving the process performance. This proposed model was used
2018 to monitor the profiles of reformate yield, temperatures and pressures of reactors, octane
Accepted 11 January 2018 number, hydrogen yield, and light gases. It includes a description of reforming reactions
Available online 1 March 2018 by using the lumping technique to reduce the complexity of the reactions that occur dur-
ing the CCRRP. The new network model of various reactions containing 36 lumps and 55
Keywords: reactions was investigated. The primary reactions included dehydrogenation, dehydrocy-
Kinetic-reactor model clization, isomerization, hydrocracking, and hydrodealkylation of the reforming process.
New reaction network The simulation results of the model have been validated by comparison with plant data.
Mathematical modeling Average absolute deviation (AAD%) of reformate yield, temperatures and pressures of reac-
Simulation tors, octane number, hydrogen yield, and light gases reached 2.5%, 1.03%, 2.6%, 1.3%, 0.43%,
Analysis and 0.93% respectively. The evaluation of the output parameters was within the acceptable
CCRRP limit and a fair agreement.
© 2018 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

1. Introduction products, and the operating conditions and energy consump-

tion. The process simulation is described by mathematical
Modeling and development of processes for improving design equations of the process parameters such as feed composi-
efficiency represent the main challenges in realizing strate- tion, pressure, temperature, and geometrical configurations,
gic economic benefits. Simulation and analysis of industrial which form an integrated approach to the system. In any
processes are often carried out to study the behavior of the oil refinery sector, simulation is used to monitor the process

∗
Corresponding author at: Department of Chemical and Process Engineering, Faculty of Engineering and Built Environment, Universiti
Kebangsaan Malaysia, 43600 Bangi, Selangor, Malaysia.
E-mail address: sobritakriff@ukm.edu.my (M.S. Takriff).
https://doi.org/10.1016/j.cherd.2018.01.025
0263-8762/© 2018 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
236 Chemical Engineering Research and Design 1 3 2 ( 2 0 1 8 ) 235–251

Nomenclature
Abbreviations
PnPx partial pressure of component nPx , kPa nPx component of paraffinic hydrocarbons (P)
PiPx partial pressure of component iPx , kPa iPx component of paraffinic hydrocarbons bran-
PNx partial pressure of component Nx (AKCH or shed (IP)
AKCP), kPa Nx component of naphthenes, alkylcyclohexane
PAx partial pressure of component Ax , kPa (AKCH) or alkylcyclopentane (AKCP)
PH2 partial pressure of component H2 , kPa Ax component of aromatic, alkyl benzene (A)
Pt total pressure of components, kPa RON research octane number
PAx+n partial pressure of component, n = l, kPa
Px , iPx light gases respect to breaking probability
rj rate of reaction performance and as a tool for guiding and controlling the oper-
Kj constant of rate, kmol/kg cat h kPa ations to achieve optimum production.
Kc equilibrium constant Catalytic reforming is one of the most important processes
Kj rate constant of reaction in oil refineries that produce high-octane-number gasoline.
ko frequency factor or pre-exponential factor Catalytic reforming processes are commonly classified into
E activation energy, kJ/kmol three types based on the regeneration systems of the cata-
R gas constant (8.314 kJ/kmol K) lyst: (i) semi-regenerative catalytic reforming process (SRCRP),
T temperature, K (ii) cyclic regenerative catalytic reforming process (CRCRP),
To temperature standard usually 298 K and (iii) continuous catalytic regeneration reforming process
HR heat of reaction, kJ/kmol (CCRRP). These processes vary according to the need to shut
H◦ f enthalpy of formation for the reactants or the down the reactor for catalyst regeneration. The mechanism
products at T for the regeneration steps could be classified into fixed-bed
◦
Hf enthalpy of formation of component at T catalyst system; fixed-bed catalyst combined a swing reactor
◦
Hf 0 standard enthalpy of formation at standard and a move-bed catalyst with special regenerator of SRCRP,
temperature (298 K) CRCRP or CCRRP type respectively (Babaqi et al., 2016).
T0 temperature at standard (298 K) Several reactions occur in the catalytic reforming process
T temperature at the reaction, K that increase the gasoline octane number: (i) dehydrogenation
Fxm mole flow rate of component (x) in reactor (m) of naphthenes, (ii) dehydrocyclization of paraffins, (iii) iso-
by kmol/h merization of normal paraffins, (iv) hydrocracking of paraffins
rjm rate of reaction (j) of component (x) in reactor and conversion into lower-molecular-weight paraffins, and
(m) by kmol/kg h (v) hydrodealkylation of aromatics. Dehydrogenation, dehy-
wm weight of the catalyst in reactor (m) by kg drocyclization, and isomerization are the desired reactions
Txm temperature of component (x) in reactor (m) by because they control the octane number and hydrogen purity.
K In contrast, hydrocracking is undesirable because it cracks
Cpxm specific heat capacity of component (x) by paraffins into smaller paraffins that produce light gases (lower
mol kJ/kmol K octane, LPG). Additionally, hydrocracking consumes hydrogen
Pm pressure of fluid by (kPa) in reactor (m) that decreases the reformate yield (George and Abdullah, 2004;
wm weight of the packed bed of catalyst by m in Sadighi and Mohaddecy, 2013; Zahedi et al., 2008).
reactor (m) The catalytic reforming process plays a significant role in
Gum mass flux through the system by kg/m2 s in the transformation of low-octane naphtha into higher-octane-
reactor (m) number reformate for gasoline blending and aromatic-rich
Dpm particle diameter in the bed by m in reactor (m) reformate for petrochemical production. It also produces
As area external curved surface of catalyst (m2 ) high-purity hydrogen gas as a by-product. Because of the sig-
nificance of the desirable gasoline for increasing the quantity
Greek letters and quality, process improvement is necessary. Most previ-
∅m void fraction, dimensionless in reactor (m) ous research focused on the catalytic reforming process with
m viscosity of fluid by kg/m s in reactor (m) different types of units via the simulation and modeling of
fm density of fluid by kg/m3 in reactor (m) reactors for improving the process design. For example, in
c density of catalyst by kg/m3 1997, Vathi and Chaudhuri carried out the modeling and simu-
HjmR heat of reaction by kJ/kmol lating of a commercial catalytic reforming process. This study
focused on the concentration and temperature profiles of
Subscript each reactor to evaluate the process performance (Vathi and
x number of components 6, 7, 8, 9, 10 Chaudhuri, 1997). In 2005, Yongyou et al. simulated a com-
j number of reactions 1, 2, 3, . . ., n mercial catalytic reforming process and focused on dynamic
1
m fraction respect to sum of breaking probability simulation. The dynamic simulation provides information on
sp stoichiometric coefficient of the product from the control system analysis and process design with dynamic
the balanced reaction process optimization. They used a numerical technique for
sr stoichiometric coefficient of the reactants from dealing with partial differential equations, and linked the
the balanced reaction reaction model with catalyst deactivation, the furnace model,
and the separator model (Yongyou et al., 2005). Weifeng et al.
(2006) simulated and optimized a whole industrial catalytic
reforming process and focused on the process profit. They
 Chemical Engineering Research and Design 1 3 2 ( 2 0 1 8 ) 235–251 237

used an 18-lump kinetic model of reactions simulated on the that represented the feed and the products. The simulation
Aspen plus platform. The obtained process profit was about results obtained with the model were compared with actual
1000 Yuan/h (Weifeng et al., 2006). plant data (Mohaddecy, 2016).
Alves et al. (2008) focused on the simulation and analysis This work focused on modeling and simulation for pre-
of fixed-bed reactors for the catalytic reforming of naphtha dicting the output parameters and monitoring the profiles to
and analyzed the reformate composition. The simulation and improve the process performance. The proposed kinetics and
analysis results were in fair agreement with the plant data reactor models for a new network of reactions were based
(Alves et al., 2008). Stijepovic et al. (2009) developed a kinetic on 36 lumps and 55 reactions. Mathematical models were
model for the catalytic reforming of naphtha and focused implemented for mass and heat integration by the particle
on reformate composition and temperature profiles that pro- swarm optimization (PSO) technique using MATLAB Program
vide information about the conversion. The simulation results Software.
were in very close agreement with the plant data (Stijepovic
et al., 2009). In 2010, Stijepovic et al. proposed a new optimiza-
2. Process description
tion approach for the simulation of the catalytic reforming
process and focused on the temperature profiles and residence
The CCRRP is the most modern process used for producing
time in the reactor by applying a quasi-steady state. The refor-
high-octane-number gasoline and aromatic-rich compounds.
mate produced provided more detailed quality specifications
This process includes three sections: (i) a reaction section
(Stijepovic et al., 2010). Saidi et al. (2011) simulated a contin-
with the regenerator catalyst, (ii) compressor section with
uous catalytic reforming process based on a process model
recontact, and (iii) product separation section. The products
and used MATLAB software for improving the process perfor-
obtained in this process are reformate, LPG, off-gas, and
mance. Good agreement was observed between the simulated
hydrogen-rich gas. The main purpose of the CCRRP is to
compositions and the actual compositions observed at the
upgrade low-octane heavy naphtha to a high-octane refor-
reactor outlet in a CCR plant (Saidi et al., 2011). In the same
mate for motor fuel mixing. Fig. 1 shows a simplified process
year, Gyngazova et al. simulated a catalytic reforming pro-
flow diagram of the CCRRP (Babaqi et al., 2017).
cess and focused on the reactivity of a moving-bed catalyst
by using a mathematical model. This process model was used
for predicting and optimizing the operation parameters such 3. Methodology
as octane number, reactor outlet temperature, and yield. The
reformate product composition was in perfect agreement with
The existing data for CCRRP were extracted, and then,
the experimental data (Gyngazova et al., 2011).
mathematical modeling, including the kinetic reactions and
Weifeng et al. (2013) carried out the simulation and
reactors, was performed for predicting the output parameters
optimization of a commercial continuous catalytic naphtha
and monitoring the profiles to improve the process perfor-
reforming and focused on the variables affecting the yield
mance. The particle swarm optimization (PSO) technique was
of aromatics, in addition to carrying out a detailed analysis
implemented to simulate the existing process. The MATLAB
for selecting the operating conditions. The aromatics yield
Program Software is used to achieve that task.
increased by about 0.49% in weight, which was equivalent to
a process profit of process approximately 5 million Yuan/year
(Weifeng et al., 2013). Aizeng et al. (2013) developed an ultra- 3.1. Reaction kinetics model
low-pressure naphtha reforming process and focused on the
process conditions and parameters. They found that the The new model includes a description of reforming reactions
research octane number (RON) of the C + 5 product reached by using the lumping technique to reduce the complexity of
104 when the capacity of the process was 100%, with 115% of the reactions that occur during this process. The new lumped
the design value (Aizeng et al., 2013). Sadighi and Mohaddecy reaction network of the CCRRP is illustrated in Fig. 2. The
(2013) simulated and optimized the continuous naphtha model enclosed lumps in the range of carbon number C6
reforming process and parameters such as octane number, to C10 according to the feedstock of paraffinic hydrocarbons,
reactor outlet temperature, and yield. They used a Res-sim branched paraffinic hydrocarbons, naphthenic hydrocarbons,
module and Petro-sim simulator for simulation. The simu- and aromatics. In addition, C8 aromatics are subdivided into
lation results indicated that the outlet temperature of the ethylbenzene, o-xylene, m-xylene, and p-xylene. Furthermore,
reactors was about 515 ◦ C, the yield of the product increased light hydrocarbons include methane, ethane, propane, butane,
from 99.55% to 99.67%, and RON increased from 82.04% to pentane, isobutane, and isopentane. The new network of var-
82.4% (Sadighi and Mohaddecy, 2013). ious reactions contains 36 lumps and 55 reactions. The main
More recently, Ivanchina et al. (2014) optimized the oper- reactions that occur during reforming include dehydrogena-
ating parameters such as octane number, reactor outlet tion, dehydrocyclization, isomerization, hydrocracking, and
temperature, and yield via mathematical modeling. The hydrodealkylation. The reaction kinetics are expressed by the
results indicated that operation under the optimal parameters rate of the reaction. The reaction is classified as reversible
leads to an increase in the product output, at a constant pro- or irreversible. Reversible reactions take place during the
duction cost (Ivanchina et al., 2014). Kavousi and Mokhtarian dehydrogenation of naphthenic hydrocarbons, as well as dur-
(2015) simulated continuous catalytic reforming using Petro- ing the dehydrocyclization and isomerization of paraffinic
sim software and focused on the operational variables. The hydrocarbons, branched paraffinic hydrocarbons, and naph-
obtained result of production of octane gasoline was about thenic hydrocarbons. Irreversible reactions take place during
98.6 than 97.4 (Kavousi and Mokhtarian, 2015). Mohaddecy the hydrocracking of paraffinic hydrocarbons and branched
(2016) studied a catalytic fixed-bed naphtha plant and used paraffinic hydrocarbons, and during the hydrodealkylation of
the discrete lumping approach for simulation. The proposed aromatics. The rate equations are expressed according to the
model consisted of 21 components, with 21 reaction pathways type of reaction.
238 Chemical Engineering Research and Design 1 3 2 ( 2 0 1 8 ) 235–251

Fig. 1 – Simplified process flow diagram of CCRRP.

All reactions of the naphtha feedstock in the reactors can naphthenic hydrocarbons (AKCH and AKCP) and composites,
be expressed in terms of the proposed network for a complete is shown by the formula below:
understanding of the model. These reactions are classified as
nPx ↔ Nx + H2
described below.
where nPx represents normal paraffinic hydrocarbons, Nx is
3.1.1. Dehydrogenation naphthenic hydrocarbons (AKCP), and x is the carbon num-
Dehydrogenation is significant due to the formation of aro- ber (6–10). The dehydrocyclization rate can be expressed by
matics that increase the octane number of gasoline. It is the applying Eq. (2):
fastest among all the reactions that occur in the reactors and is  
PAKCPx PH2
endothermic, thus leading to a decrease in temperature (Antos −rj(nPx ) = KjnPx PnPx − (2)
KcjnPx
et al., 1995). The dehydrogenation of naphthenic hydrocar-
bons (cyclo-alkanes: AKCH) into aromatics (alkyl benzene: A)
is expressed as per the following formula: Branched paraffinic hydrocarbons are also converted into
naphthenic hydrocarbons (AKCP) according to the reaction
shown in the formula below:
Nx ↔ Ax + 3H2
iPx ↔ Nx + H2

where Nx denotes cyclo-alkanes (AKCH), Ax is alkyl benzene where iPx donates iso-paraffinic hydrocarbons, Nx is naph-
(A), and x is the carbon number (from 6 to 10). The dehydro- thenic hydrocarbons (AKCP), and x represents the carbon
genation rate is expressed by applying Eq. (1): number (6–10). The reaction rate can be expressed by Eq. (3):
 
  PAKCPx PH2
PAx P3H −rj(iPx ) = K3iPx PiPx − (3)
−rj(AKCHx ) = KjAKCHx PAKCHx − 2
(1) KcjiPx
KcjAKCHx

Dehydrocyclization of naphthenic hydrocarbons (AKCH)

3.1.2. Dehydrocyclization produces normal paraffinic hydrocarbons and isoparaffinic
Dehydrocyclization is also important, as it increases the hydrocarbons according to the following formula:
octane number of gasoline. It is an endothermic reaction that
Nx + H2 ↔ nPx
slowly causes a decrease in temperatures and increase in
pressure (Antos et al., 1995). Dehydrocyclization of paraffinic where Nx denotes naphthenic hydrocarbons (AKCH), nPx
hydrocarbons (n-alkanes: nP and branched alkanes: IP) into donates normal paraffinic hydrocarbons, and x represents the
 Chemical Engineering Research and Design 1 3 2 ( 2 0 1 8 ) 235–251 239

Fig. 2 – Scheme of reaction network in four reactors of the catalytic reforming process.

carbon number (6–10). The reaction rate can be expressed by The reaction rate can be expressed by applying Eq. (6):
applying Eq. (4):
 
PiPx
  −rj(nPx ) = KjnPx PnPx − (6)
PnPx KcjnPx
−rj(AKCHx ) = KjAKCHx PAKCHx PH2 − (4)
KcjAKCHx
In addition, isomerization converts AKCP into AKCH
For branched paraffinic hydrocarbons (iPx ), the following according to the following formula:
reaction is used:
AKCPx ↔ AKCHx

Nx + H2 ↔ iPx The reaction rate is expressed in Eq. (7):

 
The reaction rate is expressed by applying Eq. (5): PAKCHx
−rj(AKCPx ) = KjAKCPx PAKCPx − (7)
KcjAKCPx
 
PiPx
−rj(AKCHx ) = KjAKCHx PAKCHx PH2 − (5)
KcjAKCHx 3.1.4. Hydrocracking
Hydrocracking reaction is an irreversible reaction taking the
3.1.3. Isomerization place of paraffinic hydrocarbons that produces light gases due
Isomerization is important as it increases the octane num- to the breaking of carbon atom bonds. Hydrocracking is an
ber of gasoline due to branched hydrocarbons. This reaction exothermic reaction that gradually leads to an increase in
leads to rearrangement of the molecules of the component temperature, thereby contributing to heat energy. This reac-
without any change in volume. Isomerization is similar to as tion occurs at high temperature and pressure and is promoted
fast exothermic reaction, which results in an increase in tem- by the acidity function of catalyst (Matar and Hatch, 2001).
perature (Matar and Hatch, 2001). Isomerization is promoted The hydrocracking reaction involves paraffinic hydrocarbons
by the acidity function of catalyst (Antos et al., 1995), and that are converted into lighter components according to the
it converts paraffinic hydrocarbons into branched paraffinic breaking probability of chemical bonds, as shown below:
hydrocarbons, as shown in the formula below:  x 
1 
nPx + H2 → Px + Px
nPx ↔ iPx m
n=1
240 Chemical Engineering Research and Design 1 3 2 ( 2 0 1 8 ) 235–251

where nPx donates normal paraffinic hydrocarbons, P̄x donates 3.1.6.2. Heat of reaction. The heat of reaction HR can be cal-
1 culated using Eq. (12) (Oxtoby et al., 2011).
light components, m is a fraction with respect to the sum of
the breaking probability of chemical bonds, and x represents
the carbon number (1–10). The reaction rate can be expressed    
◦ ◦
HR = sp Hf − sr Hf (12)
by Eq. (8): products reactants

P 
nPx
−rj(nPx ) = KjnPx (8) Eq. (13) is applied for calculating the enthalpy of formation
Pt
of each component at temperature:
Furthermore, isoparaffinic hydrocarbons are converted
into light components according to the breaking probability T
◦ ◦
via hydrocracking, as follows: Hf = Hf 0 + Cp dT (13)
T0
 x 
1  
x
iPx + H2 → iPx + Px The heat capacity can be calculated using the polynomial
m
n=1 n=1 shown in Eq. (14):
where iPx denotes iso-paraffinic hydrocarbons, P̄x denotes
1
light components, iPx denotes iso-light components, m is a Cp = A + BT + CT 2 + DT 3 (14)
fraction with respect to the sum of the breaking probability of
chemical bonds, and x represents the carbon number (1–10).
where A, B, C, and D constants are taken from the literature
The reaction rate can be expressed using Eq. (9):
(Reid et al., 1977). The heat of reaction has a positive signal
P  in endothermic reactions, which indicates that the enthalpy
iPx
−rj(iPx ) = KjiPx (9) of the products is higher than that of the reactants. On the
Pt
other hand, the heat of reaction has a negative signal in the
3.1.5. Hydrodealkylation exothermic reactions, which implies that the enthalpy of the
Hydrodealkylation is an irreversible reaction that is often products is lower than that of the reactants.
related to aromatic hydrocarbons, and is exothermic. This
reaction proceeds at elevated temperatures and pressures, 3.2. Reactor model
and is supported by the acidity function of the catalyst.
Hydrodealkylation is the slowest reaction compared with A plug flow reactor is used to describe chemical reactions in
other reactions (Tasker and Riggs, 1997). Hydrodealkylation of continuous, flowing systems, and has been used recently in
aromatics affords simpler aromatics with light gases accord- the reforming process. The development of the reactor model
ing to the breaking probability, as shown by the following requires derivation of appropriate equations for material bal-
formula: ance, energy balance, and drop pressure, to provide a full
 x  understanding of the process design and reactions. The struc-
1  
x

Ax+n + H2 → Ax + Px ture and control volume of the reactor model used herein is
m
x=6 x=1 illustrated in Fig. 3. The mathematical model was developed by
deriving governing equations of control volume, with several
1
where Ax+n , Ax are aromatics, Px denotes light components, m assumptions, namely (Fogler, 2006):
is the fraction with respect to the sum of the breaking proba-
bility of chemical bonds, and x represents the carbon number
from 1 to 10. The reaction rate can be expressed in Eq. (10): 1. One-dimensional plug flow model.
2. Steady state.
−rj(Ax ) = KjAx (PAx PH2 ) (10) 3. Changing radial flow direction.
4. Uniform and unvarying axial flow direction.
3.1.6. Thermodynamics of the reaction 5. Axially moving homogeneous catalyst.
The thermodynamics of the reaction in the kinetic model 6. Ideal gas law considered at high temperature.
including the rate of the chemical reaction and the heat of 7. Adiabatic system.
reaction are explained in the following subsections. 8. Radial and axial dispersion are ignored.

3.1.6.1. Rate of a chemical reaction. The rate of a chemical

3.2.1. Material balance
reaction depends on several factors such as the concentra-
The general equation for material balance is used to develop
tion, temperature, and activation energy, which affects the
the model in which four reactors are integrated as a stack in
speed of the reaction to reach thermodynamic equilibrium.
the reforming process, as:
The Arrhenius equation was proposed by Svante Arrhenius in
1889, and it describes the rate of a chemical reaction, as in Eq. ⎡ ⎤
(11) (Missen et al., 1999): 
4

rn
⎣ d(Fxm ) = rjm ⎦ , x = 1, 2, . . ., xn (15)
E dwm
Kj = ko e− RT (11) m=1 j=1

where ko represents a frequency factor, E represents the 3.2.2. Energy balance

activation energy (kJ/kmol), R represents the gas constant For the energy balance equation with respect to the control
(kJ/kmol K), and T is the temperature (K). volume, the general equation of energy balance is used to
 Chemical Engineering Research and Design 1 3 2 ( 2 0 1 8 ) 235–251 241

Fig. 3 – Structure and control volume of the reactor model in catalytic reforming.

develop the model for four reactors integrated as a stack in 3.3. RON model
the reforming process, and is calculated using Eq. (16):
The estimation of the RON of hydrocarbons for the feed and
rn products can be accomplished using many methods to pre-

4
dTxm j=1
R r ]
[Hjm jm dict the task. RON can be calculated for each pure component
[ = xn ] (16)
dwm [Fxm Cpxm ] by using a polynomial equation that correlates to the normal
m=1 x=1
boiling point as shown in Eq. (18) (Albahri et al., 2002):

3.2.3. Pressure drop RON = A + BT + CT 2 + DT 3 + ET 4 (18)

From a mechanical perspective, the pressure drop is a key fac-
tor in chemical processes due to the flow of fluid through a where T = TBP ∗ 0.01, TBP represents the normal boiling point
packed bed or column. A packed bed is used widely in cat- (◦ C), and A, B, C, D, and E are coefficients.
alytic reactors and separation processes. The pressure drop in The RON of a hydrocarbon mixture is calculated by assum-
catalytic reactors or pipes occurs because of the resistance of ing that the mixture consists of paraffinic hydrocarbons,
a fluid arising from changes in kinetic energy, friction, vertical isoparaffinic hydrocarbons, naphthenic hydrocarbons, and
pipe difference, or elevation (Bird et al., 2002). For creating a aromatics families. The equation is expressed as the sum of
pressure drop model, a uniform cross section of the reactor is the RON for each pure component multiplied by the volume
assumed, which is illustrated in Fig. 4. For the pressure drop fractions of the components, as shown in Eq. (19) (Nikolaou
of the four reactors (m) with multicomponent (x), the general et al., 2004):
Eq. (17) is applied:

n

4    RON = RONx yx (19)

 dPm G2um 150 × 10−3 (1 − ∅m )m x=1
=−
dwm Dpm fm ∅3m Asm c Dpm Gum
m=1
7  where RONx denotes the research octane number of the pure
+ × 10−3 (17) component and yx is the mole fraction of the components.
4

3.4. Estimation of model parameters based PSO

All equations describe the mathematical model for inte- To validate the proposed model, the simulation results of
grating mass balance, energy balance, and the pressure drop. this model were compared with the plant data. Conditions
More details on the model are provided in Appendix A. The identical to the operating conditions of the actual plant were
equations are simultaneously coupled ordinary differential taken into account in the simulation. The estimation of kinetic
equations, and involve thirty-eight dependent variables and parameters for the rate constants (k) in every reaction in the
one independent variable. The 4th order Runge–Kutta method reaction network is a very important step to simulate this
can be used to solve this problem using MATLAB. process. The frequency factor (ko ) and activation energy (E)
242 Chemical Engineering Research and Design 1 3 2 ( 2 0 1 8 ) 235–251

Fig. 4 – Structure and creation of pressure drop model of the reactor in catalytic reforming.

as initial values for guesswork were taken from the literature calculation of the objective function. If the objective func-
(Vathi and Chaudhuri, 1997). tion is lesser than “global best,” it is saved as the best value,
The kinetic parameters can be estimated by heuristic opti- which then updates the position and velocity. If it is not lesser
mization to determine the minimum error between the actual than “global best,” it is rejected and updates the position and
plant data and output model results. The equation of the velocity accordingly. The objective function will search for the
chemical reaction rate of each component in the network optimum value with new random values of the input param-
is a function of the component molar flow rate and kinetic eters. The iterations will be continued until a maximum is
constant (Kj ). From the Arrhenius Law, each rate constant reached and will then end.
(Kj ) has a frequency factor and activation energy; these are
considered as two unknown variables according to kinetic
models. Kinetic parameters have 110 unknowns for every reac- 4. Results and discussion
tor; these parameters can be determined by using an objective
function to minimize the error between the predicted and To validate the proposed model, the simulated prediction
plant results, i.e. we achieve the desired convergence. The results of this model were compared with plant data. The
sum of the average absolute deviation is used as a function for kinetic parameters of the proposed model for dehydro-
obtaining the desired convergence by applying particle swarm genation, dehydrocyclization reactions and for isomerization,
optimization (PSO), as follows: hydrocracking, and hydrodealkylation reactions are reported
  in Tables A.1 and A.2, respectively in Appendix A These param-
1  
r n
eters are the values predicted by considering the minimum
OF = × |(Fx,actual − Fx,model )|
n error between the actual plant data and predicted model out-
i=1 x=1
puts within acceptable limits.
where Fx,actual denotes the molar flow rate of the component The simulated results of the model exhibit issues in terms
(x) from the actual plant; Fx,model represents the molar flow rate of the behavior of the output parameters. Nevertheless, it
of component (x) from model; n is number of components, and provides a complete understanding and knowledge of the pro-
r is number of reactors. cess performance. These parameters include the reformate
The source of the algorithm to determine the kinetic yield, temperatures and pressures of reactors, octane number,
parameters of CCRRP via PSO is presented in Fig. 5. The algo- hydrogen yield, and light gases. One of the most important
rithm starts with initial guesses, and then generates random issues is the reformate yield (output composition) profiles for
values of the input parameters. After first setting up the four reactors, which are shown in Figs. 6–10. These figures
iteration, the fitness and running of the process model are compare the reformate composition between the model (indi-
evaluated; this is followed by thermodynamic calculations and cated by lines) and plant data (expressed by symbols). The total
the integration of mass and energy of the four reactors, and error average of the comparison was about 2.5%.
 Chemical Engineering Research and Design 1 3 2 ( 2 0 1 8 ) 235–251 243

Fig. 5 – Algorithm structure for the estimation of kinetic parameters for rate constants.

According to Fig. 6, the composition of paraffinic hydrocar- The AKCH naphthenic hydrocarbons show a sharp decline
bons in the reformate from FP6 to FP10 decreases on account on account of their consumption (Fig. 8). Moreover, the
of its consumption. The main reactions that control this are concentration of these naphthenic hydrocarbons decreases
dehydrocyclization, isomerization, and hydrocracking. The rapidly throughout the reactions, particularly in the first
composition presents a decrease in the first, second, and third reactor and slowly at the other reactors. This is attributed to
reactor, and declines steadily in the fourth reactor due to the dehydrogenation reaction, which occurs quickly and is
material conversion to products and the endothermic reac- endothermic, and therefore leads to a fall in the concentra-
tion of dehydrocyclization. Fig. 7 shows the composition of the tion of the naphthenic hydrocarbons and the temperature of
paraffinic hydrocarbons branched from FIP6 to FIP10. A slight the reactor. On the other hand, as seen in Fig. 9, the AKCP
increase in the value of the branched paraffinic hydrocarbons naphthenic hydrocarbons show a considerable increase with
can be observed for increasing octane number. The concentra- rising octane number. The concentration of these naphthenic
tion of paraffinic hydrocarbons branched (IP) grows steadily hydrocarbons increases gradually due to dehydrocyclization
because of dehydrocyclization and isomerization reactions, and isomerization reactions.
especially in the first, second, and third reactors, but drops Finally, Fig. 10 shows the composition of the aromatic pro-
in the fourth reactor due to the hydrocracking reaction. files from FA6 to FA10. A sharp increase in the amount of
Figs. 8 and 9 show the composition of naphthenic hydro- aromatics can be observed, as a result of the conversion of
carbons i.e., alkylcyclohexane and alkylcyclopentane profiles naphthenic hydrocarbons (AKCH) into aromatics by the dehy-
from AKCH6 to AKCH10 and AKCP6 to AKCP10, respectively. drogenation reaction. Especially, there is a sharp increase in
244 Chemical Engineering Research and Design 1 3 2 ( 2 0 1 8 ) 235–251

Fig. 6 – Comparison of the molar flow rate composition of paraffinic hydrocarbons between the model and plant (solid lines
and symbols represent the predicted model and plant data, respectively).

Fig. 7 – Comparison of molar flow rate composition of paraffinic hydrocarbons branched between the model and plant (solid
lines and symbols represent the predicted model and plant data, respectively).

the amount of benzene (FA6) and toluene (FA7), as compared compared to that in other reactors, as a result of dehydro-
to other aromatics. The increase in conversion levels leads to genation reactions that are very fast and mostly endothermic.
the raising of the octane number. The concentration of the aro- Therefore, to maintain the reaction temperature under the
matics (A) increases quickly throughout the reactions, mostly operating conditions of commercial catalytic reformers, the
in the first reactor, and gradually at other reactors. This is reactors are always designed such that they are adjacent to
because the dehydrogenation reaction takes place very quickly the heaters. The hydrocarbon mixture heats via the heater and
and is endothermic; this leads to an increase in the concentra- enters the second reactor, thus maintaining the reaction tem-
tion of the aromatics and decline in the temperature of reactor. perature at 822 K. The difference between the temperature of
The next issue with the process is the operating condi- the model and that of the plant is about 9.19 K. In the second
tions that are established from important parameters. Fig. 11 reactor, the temperature drops due to the dehydrogenation
compares the temperature profile of the four reactors for the reaction of the remaining naphthenic hydrocarbons (AKCH)
predicted model and the actual plant in the reforming process. and dehydrocyclization reactions of paraffinic hydrocarbons
Clearly, the temperature in the first reactor rapidly declines as (P) and paraffinic hydrocarbons branched (IP). The tempera-
 Chemical Engineering Research and Design 1 3 2 ( 2 0 1 8 ) 235–251 245

Fig. 8 – Comparison of molar flow rate composition of naphthenic hydrocarbons (AKCH) between the model and plant (solid
lines and symbols represent the predicted model and plant data, respectively).

Fig. 9 – Comparison of molar flow rate composition of naphthenic hydrocarbons (AKCP) between the model and plant (solid
lines and symbols represent the predicted model and plant data, respectively).

ture of the model and the plant differ by about 3.23 K. Further- sure drop in the reactor is related to the accumulated catalyst
more, the temperature drops in the third and fourth reactors weight or length of the reactor. Hence, the pressure drop for
due to exothermic hydrocracking reactions, although this drop the model is economical, due to the reduced flow force of fluid
is lesser than that in the first and second reactor. The differ- from compressors, which is useful from a mechanical perspec-
ence in temperature is about 6.53 and 4.2 K for the third and tive. The total error average for the pressure profiles of the four
fourth reactor, respectively. On comparing the temperature reactors between the model and plant was roughly 2.62%.
profiles of the four reactors for the predicted model and actual The other important parameter in the catalytic reforming
plant, the total error average was about 1.03%. As the process process is the RON. Table 1 compares the octane number of
requires continuous energy supply in order to maintain the the predicted model and the actual plant, obtained from cal-
temperature of the reactor at 822 K, an integration of four reac- culated equations. Feed naphtha, with RON = 49.4, is the input
tors with four furnaces can be used as a means to identify to the process for raising the octane number to RON = 75.32,
opportunities for design improvement, and thus save energy. according to the predicted model. The output of the octane
Fig. 12 compares the pressure drop between the model and number from the actual plant reached RON = 76.32, indicating
plant. The pressure drop of the plant is lower than that of the that the behavior of the predicted model is close to that of the
model due to the effect of a plug of the catalyst. The pres-
246 Chemical Engineering Research and Design 1 3 2 ( 2 0 1 8 ) 235–251

Fig. 10 – Comparison of molar flow rate composition of aromatics between the model and plant (solid lines and symbols
represent the predicted model and plant data, respectively).

Fig. 11 – Comparison of temperature profile of four reactors between the model and plant (solid lines and symbols represent
the predicted model and plant data, respectively).

actual plant. The error in the RON between the model and the ratio of the amount of hydrogen as a product to the amount of
plant was around 1.3%. hydrogen fed. The error for hydrogen yield between the model
The other parameter, namely hydrogen production, is a and plant was around 0.43%.
byproduct of the catalytic reforming process, and is very Finally, the parameter for the production of light gases
important in hydro-treating and cracking processes. Fig. 13 should be considered in the catalytic reforming process. Fig. 14
compares the molar flow rate of hydrogen between the model shows the composition of light gases from FC1 to FiC5; it agrees
and plant, and shows fairly good agreement. Dehydrogena- well with the plant data. The increase in the amount of light
tion reactions, seen in the first reactor, are very important for gases in the third and fourth reactors, as compared to other
hydrogen production. In the second, third, and fourth reac- reactors, is as a result of hydrocracking reactions of paraffinic
tors, the amount of hydrogen only slightly increases due to and iso-paraffinic hydrocarbons and hydrodealkylation reac-
hydrocracking and hydrodealkylation reactions that consume tions of aromatics. The error for hydrogen yield between the
hydrogen. The yield of pure hydrogen can be estimated as the model and plant was around 0.93%.
 Chemical Engineering Research and Design 1 3 2 ( 2 0 1 8 ) 235–251 247

Fig. 12 – Comparison of pressure drop between the model and plant (solid lines and symbols represent the predicted model
and plant data, respectively).

Table 1 – Comparison between the octane number of the predicted model and actual plant.
Components Naphtha feed fresh Mole fraction Mole fraction RON com- RON of naphtha RON of plant RON of
composition plant output model output ponent fresh model

n-Hexane 0.08225 0.08153 0.08536 24.602 2.024 2.006 2.100

n-Heptane 0.11677 0.06358 0.06417 −0.265 −0.031 −0.017 −0.017
n-Octane 0.07695 0.01195 0.01040 −15.801 −1.216 −0.189 −0.164
n-Nonane 0.05121 0.01365 0.01273 −14.809 −0.758 −0.202 −0.189
n-Decane 0.03842 0.00229 0.00369 −12.772 −0.491 −0.029 −0.047
2-Methyl pentane 0.08366 0.08745 0.08717 93.195 7.797 8.150 8.124
2-Methylhexane 0.03850 0.04957 0.05090 75.857 2.921 3.760 3.861
2-Methylheptane 0.01381 0.03215 0.03361 64.281 0.887 2.067 2.160
2-Methyloctane 0.03096 0.04462 0.04528 36.627 1.134 1.634 1.659
2-Methylnonane 0.00484 0.02937 0.03041 34.343 0.166 1.009 1.044
Cyclohexane 0.11154 0.00003 0.00001 83.019 9.260 0.002 0.001
Methyl cyclohexane 0.06993 0.00001 0.00000 53.289 3.727 0.000 0.000
Ethyl cyclohexane 0.07031 0.00001 0.00001 46.690 3.283 0.000 0.000
Propyl cyclohexane 0.00749 0.00001 0.00005 16.699 0.125 0.000 0.001
Butyl cyclohexane 0.00590 0.00001 0.00001 71.113 0.420 0.001 0.001
Methyl cyclopentane 0.01627 0.04403 0.04571 91.341 1.486 4.021 4.176
Ethyl cyclopentane 0.01315 0.05495 0.05638 79.528 1.046 4.370 4.484
Propyl cyclopentane 0.01003 0.03664 0.03770 51.722 0.519 1.895 1.950
Butyl cyclopentane 0.00474 0.02212 0.02318 24.724 0.117 0.547 0.573
Pentyl cyclopentane 0.00370 0.00712 0.00911 −7.523 −0.028 −0.054 −0.069
Benzene 0.04193 0.13143 0.13061 107.463 4.506 14.123 14.036
Toluene 0.06101 0.14027 0.13629 120.208 7.334 16.862 16.383
Ethylbenzene 0.00944 0.03046 0.02794 107.593 1.015 3.277 3.007
o-Xylene 0.00752 0.03542 0.03573 115.222 0.866 4.081 4.116
m-Xylene 0.00835 0.02010 0.01657 117.277 0.979 2.357 1.943
p-Xylene 0.00723 0.02644 0.02460 118.177 0.855 3.124 2.908
Propylbenzene 0.00422 0.01494 0.01573 101.983 0.430 1.523 1.604
Butylbenzene 0.00986 0.01987 0.01664 100.416 0.990 1.995 1.671

Total 1.00000 1.00000 1.00000 49.363 76.317 75.316

5. Conclusion light gases, and their profiles, were monitored by the proposed
model. The proposed model is a new network of various reac-
A mathematical model, simulation and analysis in the con- tions, and contains 36 lumps and 55 reactions. Results for the
tinuous catalytic regeneration reforming process (CCRRP) was simulation of the model were validated by comparison with
developed to identify the major areas for improvement of plant data. The average absolute deviation (AAD%) of the refor-
the process performance via predictions of output param- mate yield, temperatures and pressures of reactors, octane
eters. These parameters, i.e., reformate yield, temperatures number, hydrogen yield, and light gases was 2.5%, 1.03%, 2.6%,
and pressures of reactors, octane number, hydrogen yield, and 1.3%, 0.43%, and 0.93% respectively, which was within accept-
248 Chemical Engineering Research and Design 1 3 2 ( 2 0 1 8 ) 235–251

Fig. 13 – Comparison of molar flow rate of hydrogen between the model and plant (solid lines and symbols represent the
predicted model and plant data, respectively).

Fig. 14 – Comparison of molar flow rate composition of light gases between the model and plant (solid lines and symbols
represent the predicted model and plant data, respectively).

able limits. This simulation of the output parameters of the dFnP6

= −r2(nP6 ) + r4(nP6 ) − r6(nP6 ) − r8(nP6 )
CCRRP process will enable the selection of suitable operating dW
conditions for improving process design. 
10 
9,10
+ r8(C6 H14 )x + r9(C6 H14 )x
x=7 x=7
Acknowledgement

The authors would like to acknowledge the Universiti dFnP7

= −r2(nP7 ) + r4(nP7 ) − r6(nP7 ) − r8(nP7 )
Kebangsaan Malaysia (UKM) for funding this work the follow- dW
ing grants DIP-2017-007 and Ggpm-2014-039. 
10 
8,10
+ r8(C7 H16 )x + r9(C7 H16 )x
n=8 x=8
Appendix A.

For mass balance of all components can be expressed as the

dFnP8
net reaction for one reactor by these equations: = −r2(nP8 ) + r4(nP8 ) − r6(nP8 ) − r8(nP8 )
dW
 Chemical Engineering Research and Design 1 3 2 ( 2 0 1 8 ) 235–251 249


10

9
+ r8(C8 H18 )x + r9(C8 H18 )x dFA8EB  10

x=9 x=9
= r1(A8EB ) − r10(A8EB ) + r10(A8EB )x
dW
x=9

dFnP9  10
dFA8OX
= −r2(nP9 ) + r4(nP9 ) − r6(nP9 ) − r8(nP9 ) + r8(C9 H20 )x = r1(A8OX ) − r10(A8OX )
dW dW
x=10

dFA8MX
dFnP10 = r1(A8MX ) − r10(A8MX )
= −r2(nP10 ) + r4(nP10 ) − r6(nP10 ) − r8(nP10 ) dW
dW
dFA8PX
= r1(A8PX ) − r10(A8PX )
dFiP6  10
dW
= −r3(iP6 ) + r5(iP6 ) + r6(iP6 ) − r9(iP6 ) + r9(iC6 H14 )x
dW
x=7
dFA9  10
= r1(A9 ) − r10(A9 ) + r10(A9 )x
dW
dFiP7  10
x=10
= −r3(iP7 ) + r5(iP7 ) + r6(iP7 ) − r9(iP7 ) + r9(iC7 H16 )x
dW
x=8 dFA10
= r1(A10 ) − r10(A10 )
dW
dFiP8  10
= −r3(iP8 ) + r5(iP8 ) + r6(iP8 ) − r9(iP8 ) + r9(iC8 H18 )x
dW
x=9 dFH2  10  10  10 10

= r1(H2 )x + r2(H2 )x + r3(H2 )x − r4(H2 )x

dW
dFiP9  10 x=6 x=6 x=6 x=6

= −r3(iP9 ) + r5(iP9 ) + r6(iP9 ) − r9(iP9 ) + r9(iC9 H20 )x 

10

10 
10 
10
dW
x=10 − r5(H2 )x − r8(H2 )x − r9(H2 )x − r10(H2 )x
x=6 x=6 x=6 x=6
dFiP10
= −r3(iP10 ) + r5(iP10 ) + r6(iP10 ) − r9(iP10 )
dW
dFC1  10  
10 10

dFAKCH6 = r8(CH4 )x + r9(CH4 )x + r10(CH4 )x

= −r1(AKCH6 ) − r4(AKCH6 ) − r5(AKCH6 ) + r7(AKCH6 ) dW
dW x=6 x=6 x=7

dFAKCH7
= −r1(AKCH7 ) − r4(AKCH7 ) − r5(AKCH7 ) + r7(AKCH7 ) dFC2  10
 10
 10
dW = r8(C2 H6 )x + r9(C2 H6 )x + r10(C2 H6 )x
dW
x=6 x=6 x=8
dFAKCH8
= −r1(AKCH8 ) − r4(AKCH8 ) − r5(AKCH8 ) + r7(AKCH8 )
dW
dFC3  10
 10
 10

= r8(C3 H8 )x + r9(C3 H8 )x + r10(C3 H8 )x

dFAKCH9 dW
= −r1(AKCH9 ) − r4(AKCH9 ) − r5(AKCH9 ) + r7(AKCH6 ) x=9 x=6 x=9
dW

dFAKCHN10 dFC4  10
 10
 10
= −r1(AKCH10 ) − r4(AKCH10 ) − r5(AKCH10 ) + r7(AKCH10 ) = r8(C4 H10 )x + r9(C4 H10 )x + r10(C4 H10 )x
dW dW
x=6 x=7 x=10

dFAKCP6
= r2(AKCP6 ) + r3(AKCP6 ) − r7(AKCP6 )
dW
dFC5  10 
8,9,10
= r8(C5 H12 )x + r9(C5 H12 )x
dW
dFAKCP7 x=6 x=6
= r2(AKCP7 ) + r3(AKCP7 ) − r7(AKCP7 )
dW

dFiC4  10
dFAKCP8 = r9(iC4 H10 )x
= r2(AKCP8 ) + r3(AKCP8 ) − r7(AKCP8 ) dW
dW x=6

dFAKCP9
= r2(AKCP9 ) + r3(AKCP9 ) − r7(AKCP9 )  10
dW dFiC5
= r9(iC5 H12 )x
dW
x=6
dFAKCP10
= r2(AKCP10 ) + r3(AKCP10 ) − r7(AKCP6 )
dW where C1 , C2 , C3 , C4 , C5 , iC4 and iC5 represent methane, ethane,
propane, butane, pentane, isobutane and isopentane respec-
dFA6  10
tively.
= r1(A6 ) + r10(A6 )x
dW For energy balance of all components can be expressed by
x=7
these equations:
rn  
dFA7  10 R r
Hjm jm
= r1(A7 ) − r10(A7 ) + dTx j=1
dW
r10(A7 )x = xn
x=8
dW [Fxm Cpxm ]
x=1
250 Chemical Engineering Research and Design 1 3 2 ( 2 0 1 8 ) 235–251

For pressure drop of reactors can be expressed by these

rn  equations:
j=1
H1R (−r16 ) + H2R (−r26 ) + H3R (−r36 )
     
dTx +H4R (−r46 ) + H5R (−r56 ) dPx G2u 150 × 10−3 (1 − ∅) 7
= xn =− + × 10−3
dW
x=1
[Fx Cpx ] dW Dp f ∅3 [As c ] Dp Gu 4

Table A.1 – Kinetic parameters of the proposed model for dehydrogenation and dehydrocyclization reactions.
Reactions Items ko E Kj

Dehydrogenation K1akch6 16.83E−03 430.24 15.80E−03

reaction K1akch7 92.03E−03 429.77 86.42E−03
K1akch8EB 22.02E−03 436.97 20.65E−03
K1akch8OX 28.48E−03 436.80 26.714811
K1akch8MX 5.30E−03 431.02 4.98E−03
K1akch8PX 16.87E−03 434.53 15.84E−03
K1akch9 66.72E−03 431.56 62.64E−03
K1akch10 44.86E−03 420.71 42.18E−03

Dehydrocyclization K2nP6 8.95E−06 715.74 8.06E−06

reaction K2nP7 3.93E−03 717.18 3.54E−03
K2nP8 3.91E−03 720.65 3.52E−03
K2nP9 13.83E−03 715.31 12.46E−03
K2nP10 40.53E−03 717.26 36.50E−03
K3iP6 24.93E−03 718.46 22.45E−03
K3iP7 41.82E−03 712.50 37.68E−03
K3iP8 4.50E−03 717.25 4.05E−03
K3iP9 24.32E−03 715.48 21.90E−03
K3iP10 1.87E−03 714.48 1.68E−03
K4akch6 4.03E−10 715.50 3.63E−10
K4akch7 4.81E−08 713.72 4.33E−08
K4akch8 5.81E−09 716.46 5.23E−09
K4akch9 8.49E−08 717.91 7.64E−08
K4akch10 1.55E−07 713.50 1.40E−07
K5akch6 2.44E−09 712.61 2.20E−09
K5akch7 9.92E−10 712.78 8.94E−10
K5akch8 3.91E−08 718.84 3.52E−08
K5akch9 8.22E−08 715.51 7.40E−08
K5akch10 8.53E−08 711.52 7.68E−08

Table A.2 – Kinetic parameters of the proposed model for isomerization, hydrocracking and hydrodealkylation reactions.
Reactions Items ko E Kj

Isomerization K6nP6 5.06E−10 562.32 4.66E−10

reaction K6nP7 3.71E−10 565.71 3.41E−10
K6nP8 1.24E−09 567.50 1.14E−09
K6nP9 4.24E−10 564.67 3.90E−10
K6nP10 4.60E−10 564.60 4.24E−10
K7akcp6 4.55E−03 478.88 4.24E−04
K7akcp7 4.95E−05 471.31 4.62E−05
K7akcp8 6.90E−05 472.11 6.44E−05
K7akcp9 28.86E−05 463.45 2.70E−04
K7akcp10 1.75E−09 466.39 1.63E−09

Hydrocracking K8nP6 27.80E−03 751.72 24.90E−03

reaction K8nP7 27.75E−03 751.08 24.86E−03
K8nP8 4.12E−03 753.03 3.69E−03
K8nP9 119.27E−03 748.99 10.69E−02
K8nP10 148.82E−03 749.23 13.34E−02
K9iP6 3.48E−03 750.41 3.12E−03
K9iP7 5.96E−03 744.27 5.35E−03
K9iP8 4.68E−03 745.28 4.20E−03
K9iP9 41.46E−03 746.08 37.17E−03
K9iP10 47.52E−03 743.71 42.6E−03

Hydrodealkylation K10a7 2.31E−17 382.16 2.19E−17

reaction K10a8EB 2.80E−18 384.64 2.65E−18
K10a8OX 3.16E−18 383.69 2.99E−18
K10a8MX 3.43E−18 382.13 3.24E−18
K10a8PX 3.84E−18 391.11 3.62E−18
K10a9 5.62E−18 389.73 5.31E−18
K10a10 1.36E−17 391.56 1.28E−17
 Chemical Engineering Research and Design 1 3 2 ( 2 0 1 8 ) 235–251 251

References Missen, R.W., Mims, C.A., Saville, B.A., 1999. Introduction to

Chemical Reaction Engineering and Kinetics. J. Wiley.
Mohaddecy, S., 2016. Modeling catalytic naphtha reforming
Aizeng, M., Youchun, X., Dong, Y., Xinkuan, Z., Jieguang, W., 2013.
process using discrete lumping approach. Indian J. Chem.
Development and commercial application of ultra-low
Technol. 21 (5–6), 309–320.
pressure naphtha reforming technology with continuous
Nikolaou, N., Papadopoulos, C., Gaglias, I., Pitarakis, K., 2004. A
catalyst regeneration. China Pet. Process. Petrochem. Technol.
new non-linear calculation method of isomerisation gasoline
15 (4), 1–8.
research octane number based on gas chromatographic data.
Albahri, T.A., Riazi, M.R., Alqattan, A.A., 2002. Octane number
Fuel 83 (4), 517–523.
and aniline point of petroleum fuels. Fuel Chem. Div. Prepr. 47
Oxtoby, D.W., Gillis, H.P., Butler, L.J., 2011. Principles of Modern
(2), 710–711.
Chemistry, 7th ed. Cengage Learning, The United States of
Alves, R.M., Menten, F., Maejima, W.S., Guardani, R., Nascimento,
America.
C., 2008. A study on naphtha catalytic reforming reactor
Reid, R.C., Prausnitz, J.M.A., TK, S., 1977. The Properties of Gases
simulation and analysis. In: Braunschweig, B., Jou-lia, X.
and Liquids, 3rd ed. McGraw-Hill, New York.
(Eds.), Paper presented at the 18th European Symposium on
Sadighi, S., Mohaddecy, R.S., 2013. Simulation, sensitivity analysis
Computer Aided Process Engineering.
and optimization of an industrial continuous catalytic
Antos, G.J., Aitani, A.M., Parera, J.M., 1995. Catalytic Naphtha
naphtha reforming process. Eur. Chem. Bull. 2 (10), 777–781.
Reforming. Science and Technology/Marcel Dekker Inc., New
Saidi, M., Mostoufi, N., SotudehGharebagh, R., 2011. Modeling and
York/New York.
simulation of continuous catalytic regeneration (CCR) process.
Babaqi, B.S., Takriff, M.S., Kamarudin, S.K., Othman, N.T.A.,
Int. J. Appl. Eng. Res. (Dindigul) 2 (1).
Ba-Abbad, M.M., 2016. Comparison of catalytic reforming
Stijepovic, M.Z., Linke, P., Kijevcanin, M., 2010. Optimization
processes for process integration opportunities. Brief Rev. Int.
approach for continuous catalytic regenerative reformer
J. Appl. Eng. Res. 11 (19), 9984–9989.
processes. Energy Fuels 24 (3), 1908–1916.
Babaqi, B.S., Takriff, M.S., Kamarudin, S.K., Othman, N.T.A.,
Stijepovic, M.Z., Vojvodic-Ostojic, A., Milenkovic, I., Linke, P.,
Ba-Abbad, M.M., 2017. Energy optimization for maximum
2009. Development of a kinetic model for catalytic reforming
energy saving with optimal modification in Continuous
of naphtha and parameter estimation using industrial plant
Catalytic Regeneration Reformer Process. Energy 120, 774–784.
data. Energy Fuels 23 (2), 979–983.
Bird, R.B., Stewart, W.E.A., Lightfoot, E.N., 2002. Transport
Tasker, U.A., Riggs, J.B., 1997. Modeling and optimization of a semi
Phenomena, 2nd ed. John Wiley & Sons, Inc., The United
regenerative catalytic naphtha reformer. AIChE 43, 740–753.
States of America.
Vathi, G.P., Chaudhuri, K.K., 1997. Modelling and simulation of
Fogler, H.S., 2006. Elements of Chemical Reaction Engineering,
commercial catalytic naphtha reformers. Can. J. Chem. Eng.
4th ed. Pearson Education, Inc.
75 (5), 930–937.
George, J.A., Abdullah, M.A., 2004. Catalytic Naphtha Reforming.
Weifeng, H., Hongye, S., Yongyou, H., Jian, C., 2006. Modeling,
Marcel Dekker, New York.
simulation and optimization of a whole industrial catalytic
Gyngazova, M.S., Kravtsov, A.V., Ivanchina, E.D., Korolenko, M.V.,
naphtha reforming process on Aspen Plus platform. Chin. J.
Chekantsev, N.V., 2011. Reactor modeling and simulation of
Chem. Eng. 14 (5), 584–591.
moving-bed catalytic reforming process. Chem. Eng. J. 176,
Weifeng, H., Xiuqiang, Z., Yanping, H., Fengyue, J., 2013.
134–143.
Simulation, sensitivity analysis and optimization of a
Ivanchina, E.D., Sharova, E.S., Yakupova, I.V., 2014. Mathematical
commercial continuous catalytic naphtha reformer. Inf.
modelling method application for optimisation of catalytic
Technol. J. 12 (19), 4953–4958.
reforming process. Procedia Chem. 10, 197–202.
Yongyou, H., Weihua, X., Weifeng, H., 2005. Dynamic modeling
Kavousi, K., Mokhtarian, N., 2015. Simulation the continuous
and simulation of a commercial naphtha catalytic reforming
catalytic reforming (octanizer) unit of Isfahan Refinery
process. Chin. J. Chem. Eng. 13 (1), 74–80.
Gasoline Production Complex with PETROSIM Software. Int.
Zahedi, G., Mohammadzadeh, S., Moradi, G., 2008. Enhancing
Sci. Investig. J. 4 (3), 14–33.
gasoline production in an industrial catalytic-reforming unit
Matar, S., Hatch, L.F., 2001. Chemistry of Petrochemical Processes,
using artificial neural networks. Energy Fuels 22 (4), 2671–2677.
2nd ed. Gulf Professional Publishing.