Applied Thermal Engineering 112 (2017) 1040–1056

Contents lists available at ScienceDirect

Applied Thermal Engineering

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

Research Paper

A novel integrated thermally coupled moving bed reactors for naphtha

reforming process with hydrodealkylation of toluene
Davood Iranshahi a,⇑, Reza Saeedi a, Kolsoom Azizi a, Mahshid Nategh b
a
Department of Chemical Engineering, Amirkabir University of Technology (Tehran Polytechnic), No. 424, Hafez Avenue, Tehran 15914, Iran
b
Department of Chemical Engineering, School of Chemical and Petroleum Engineering, Shiraz University, Shiraz 71345, Iran

h i g h l i g h t s

A novel thermally coupled reactor in CCR naphtha reforming process is modeled.

The required heat of Naphtha process is attained with toluene hydrodealkylation.

A new kinetic model involving 32 pseudo-component and 84 reactions is proposed.

The aromatics and hydrogen production increase 19% and 23%, respectively.

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

Article history: Due to the importance of catalytic naphtha reforming process in refineries, development of this process
Received 21 June 2016 to attain the highest yield of desired products is crucial. In this study, continuous catalyst regeneration
Revised 21 September 2016 naphtha reforming process with radial flow is coupled with hydrodealkylation of toluene to prevent
Accepted 19 October 2016
energy loss while enhancing aromatics and hydrogen yields. In this coupled process, heat is transferred
Available online 20 October 2016
between hot and cold sections (from hydrodealkylation of toluene to catalytic naphtha reforming pro-
cess) using the process integration method. A steady-state two-dimensional model, which considers coke
Keywords:
formation on the catalyst pellets, is developed and 32 pseudo-components with 84 reactions are inves-
Naphtha reforming
Continuous catalyst regeneration
tigated. Kinetic model utilized for HDA process is homogeneous and non-catalytic. The modeling results
Thermal coupling reveal an approximate increase of 19% and 23% in aromatics and hydrogen molar flow rates, respectively,
Hydrodealkylation in comparison with conventional naphtha reforming process. The improvement in aromatics production
Optimization evidently indicates that HDA is a suitable process to be coupled with naphtha reforming.
Differential evolution (DE) Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction Naphtha constitutes typically 15–30 wt% of the crude oil and is
obtained from the middle part of the atmospheric distillation
1.1. Naphtha reforming tower. Among the products of this reforming process are aromatic
components containing benzene, toluene and xylene. The other
Even though fossil fuels have disadvantages such as environ- product is hydrogen, which is served as feed in some refining pro-
mental pollution, they are considered as the most substantial cesses such as hydrocracking [8,9]. Besides, hydrogen is a critical
source of energy [1–5]. Hence, it is of great importance to apply feedstock and raw material in petroleum units. Refineries require
processes with the purpose of increasing their quality along with a large amount of hydrogen to remove sulfur and nitrogen compo-
reducing their pollution. Catalytic naphtha reforming, used for nents and to produce lighter fuels [10].
upgrading octane number of hydrocarbons, is regarded as one of Pressure drop is one the main issues in industrial processes.
the most critical processes in refineries. The catalytic reformer Remedy of this problem is presenting new configurations. Tubular
plays a very significant role in refineries. It provides high value- reactors with radial flow are one of these proposed configurations.
added reformate for gasoline pool; hydrogen for improving feed- In comparison with axial flow reactors, radial flow reactors have
stock by hydrogen-consuming hydrotreatment processes [6,7]. advantages such as low pressure drop, high flux capacity, using
smaller catalyst and easier regeneration process [11,12]. Catalysts
used in this process are Pt-Sn/Al2O3. Tin is used to prevent coke
⇑ Corresponding author. formation on the palatine side of catalysts, to enhance selectivity
E-mail address: [email protected] (D. Iranshahi).

http://dx.doi.org/10.1016/j.applthermaleng.2016.10.118
1359-4311/Ó 2016 Elsevier Ltd. All rights reserved.
 D. Iranshahi et al. / Applied Thermal Engineering 112 (2017) 1040–1056 1041

Nomenclature

ai catalyst activity in endothermic reaction side n0 number of components in exothermic reaction side
aA acidic function activity in endothermic reaction side Nj molar flux of component j, kmol m2 h1
aC A acidic function activity for coke formation in endother- P total pressure, kPa
mic reaction side P An partial pressure of n carbon aromatic, kPa
aC M metallic function activity for coke formation in PACHn partial pressure of n carbon alkyl-cyclohexane, kPa
endothermic reaction side PACPn partial pressure of n carbon alkyl-cyclopentane, kPa
aM metallic function activity in endothermic reaction side PBen partial pressure of benzene, kPa
a0 catalyst activity in exothermic reaction side P H2 partial pressure of hydrogen in endothermic reaction
Ar cross section area of reactor in radial direction, m2 side, kPa
Ap available side area for heat transfer, m2 PIPn partial pressure of n carbon iso-paraffin, kPa
C concentration in endothermic reaction side, kmol m3 PD partial pressure of diphenyl, kPa
C Exo concentration in exothermic reaction side, kmol m3 PNPn partial pressure of n carbon normal-paraffin, kPa
C ACP alkyl-cyclopentane concentration, kmol m3 PT partial pressure of toluene, kPa
C CA coke weight fraction on acidic function of catalyst in r radius, m
endothermic reaction side, kg kgcat1 ri rate of ith reaction, kmol kgcat1 h1
C CM coke weight fraction on metallic function of catalyst in r in rate of inth reaction, kmol kgcat1 h1
endothermic reaction side, kg kgcat1 r oC rate of coke formation on fresh catalyst in endothermic
Cp specific heat capacity at constant pressure in endother- reaction side, kg kgcat1 h1
mic reaction side, kJ kmol1 K1 rCA rate of coke formation on acidic function of catalyst in
C Exo
P specific heat capacity at constant pressure in exother- endothermic reaction side, kg kgcat1 h1
mic reaction side, kJ kmol1 K1 rCM rate of coke formation on metallic function of catalyst in
CT total concentration in endothermic reaction side, endothermic reaction side, kg kgcat1 h1
kmol m3 r 0Ca rate of coke formation on active sites in exothermic
C Exo
T total concentration in exothermic reaction side, reaction side, kg kgcat1 s1
kmol m3 R gas constant, J mol1 K1
dp particle diameter, m Ri inner diameter, m
dr control volume thickness in radial direction, m Ro outer diameter, m
dz control volume length, m Sa specific surface, m2 g1
De effective diffusivity, m2 s1 T temperature of endothermic reaction side, K
Ec coke formation activation energy in endothermic reac- T Exo temperature of exothermic reaction side, K
tion side, J mol1 U overall heat transfer coefficient, W m2 K1
E0 activation energy of the hydrogenation of nitrobenzene, ur radial velocity in endothermic reaction side, m s1
kJ mol1 uExo
r radial velocity in exothermic reaction side, m s1
Hj enthalpy of component j in endothermic reaction side, Yj yield of component j
J mol1
HExo
j enthalpy of component j in exothermic reaction side, Greek letters
J mol1 e void fraction of catalyst bed in endothermic reaction
k Thermal conductivity, W m1K1 side
kC A constant of deactivation equation for acidic function in l viscosity, kg m1 s1
endothermic reaction side, kg kgcat1 kPan1 m1.5 v ij stoichiometric coefficient of component j in reaction i in
kmol1.5 endothermic reaction side
kC M constant of deactivation equation for metallic function vj stoichiometric coefficient of component j in exothermic
in endothermic reaction side, kg kgcat1 kPan1 m1.5 - reaction side
kmol1.5 qb reactor bulk density in endothermic reaction side,
kin reaction rate constant for endothermic reaction (in) kg m3
K in equilibrium constant for endothermic reaction (in) /s sphericity
0
k reaction rate constant for exothermic reaction, h1 endothermic reaction side angle
mol kg1 s1 h2 exothermic reaction side angle
L length of reactor, m DH heat of reaction in endothermic reaction side, kJ mol1
m number of reactions in endothermic reaction side DHExo heat of reaction in exothermic reaction side, kJ mol1
Mj molecular weight of component j, kg kmol1 aA constant of deactivation, m3 kmol1
M mean molecular weight in the flow, kg kmol1 aC A constant of deactivation, m3 kmol1
n number of components in endothermic reaction side aC M constant of deactivation, m3 kmol1
n1 constant of deactivation equation in endothermic reac- p Pi number
tion side aM constant of deactivation, m3 kmol1
n2 constant of deactivation equation in endothermic reac-
tion side Superscript
nA acidic function activity power number in endothermic Exo exothermic reaction side
reaction side
nCA acidic function activity power number in endothermic
reaction side Subscript
e exit condition (in inner radius of reactor condition)
nM metallic function activity power number in endother-
mic reaction side i numerator for reaction in endothermic reaction side
nC M metallic function activity power number in endother- j numerator for component in endothermic and exother-
mic reaction sides
mic reaction side
1042 D. Iranshahi et al. / Applied Thermal Engineering 112 (2017) 1040–1056

n number of carbon atom IBP initial boiling pint, °C

T total IP iso paraffin
LOD length over diameter
Abbreviations M metallic function of catalyst
A acidic function of catalyst NOS number of subsections
ACH alkyl-cyclohexane NP normal paraffin
ACP alkyl-cyclopentane OTCR optimized thermally coupled reactor
CCR continuous catalyst regeneration reformer PI process integration
CR conventional reactor Pt platinum
HDA hydrodealkylation of toluene Sn tin
FBP final boiling point
HC hydrocarbon

and stabilize of catalysts [13–17]. Pt-Sn catalysts regenerate easily [23]. Khademi et al. optimized a thermally coupled process to
and can be used in continuously regenerative process [18,19]. enhance methanol, benzene and hydrogen production [24].
The design of a pure heat exchangers is based on choosing an
appropriate DTmin [25]. Basically in heat exchangers there is
1.2. Thermally coupled reactor trade–off between capital and energy costs. Knowing the details
of this costs, one can obtain an optimum DTmin. For the usual heat
In order to reduce energy consumption and undesired products exchangers, the range of optimum DTmin is narrow and falls in
formation in industrial processes, process integration (PI) concept range of 5–20 °C [25,26]. Therefore, the temperature of hot and
has a great potential [20]. Among the integration methods, cou- cold streams are adjusted by capital and energy costs. As rule of
pling of exothermic and endothermic processes in multifunctional thumb, abovementioned range may guarantee an optimum design
reactors has been reported widely in the literature. It is worth without any detail information about energy, utility and capital
mentioning that the difference between hot and cold sources is a costs.
major key [21]. Hunter and McGuire investigated the concept of This rule does not satisfy the thermally coupled reactors,
indirect thermal coupling of endothermic and exothermic reac- because of following reasons. First, in thermally coupled reactors,
tions experimentally and deduced that exothermic reactions can the range of temperature for hot and cold streams is restricted
be employed as a heat source for endothermic ones [22]. Abo- and can’t be changed arbitrary. For example the temperature of
Ghander et al. modeled the thermal coupling of dehydrogenation fluid stream in a reactor can’t be decreased too much, because in
of ethylbenzene to styrene with hydrogenation of nitrobenzene lower temperature catalyst doesn’t work. Also, an increase in the
to aniline in a novel membrane reactor for co-current and temperature of reactor may cause hot spot, temperature run away
counter-current operations using a pseudo-homogeneous model and explosion [27].

Hydrogen Recycle Gas

Net Hydrogen

Comperessor
Off Gas

Heater Heater Heater Heater

Preheated Naphtha & Hydrogen

Recycle Gas Heat

Exchanger
Gas
Feed
Separator
Pump

Fig. 1a. A schematic diagram of conventional naphtha reforming process with moving bed reactors.
 D. Iranshahi et al. / Applied Thermal Engineering 112 (2017) 1040–1056 1043

Secondly, the mole fraction of desired product in effluent mix- Naphtha reforming reactions are greatly endothermic, temper-
ture of reactor is strongly under the influence of temperature pro- ature and subsequently the reaction rate decrease along the reac-
file through the reactor. The reactor effluent always experiences a tor. In order to remove this negative effect and benefits of
sequential post-reactor treatments in separation apparatus such as thermally coupled reactor, including higher thermal efficiency
flash drum and distillation column. The mole fraction of desired and lower operational costs, combination of these concepts has
products in reactor effluent not only limits its price, but also affects been proposed. Jafari et al. suggested a novel configuration, the
the capital and operating costs of the separation processes. For generated heat in nitrobenzene hydrogenation reaction is trans-
example the number of trays and boiler duty of a distillation col- ferred to naphtha reforming section. Yield enhancement is
umn is under the influence of the feed composition. This limita- observed in both of the reactions, and aromatics, hydrogen and
tions does not exist in ordinary heat exchangers. aniline production increase in the thermally coupled configuration
As mentioned, the optimum DTmin for a pure heat exchanger [28]. Also Iranshahi et al. proposed a new configuration for naph-
could be obtained by energy and capital cost only, while in novel tha reforming reactors to enhance the production of hydrogen
heat exchanger reactors, the price of products and separation units and aromatics, which are indicators of process efficiency [29].
must be considered too. In this regard, one can introduce the above Karimi et al. optimized the parameters and conditions of a novel
costs to capital and energy costs trade–off. reactor in which naphtha reforming is thermally coupled with
In this study the optimization is applied, just from the process nitrobenzene hydrogenation [30].
viewpoint and due to lack of information, cost is not considered. Hydrogenation of nitrobenzene to aniline, which was coupled
Our objective is to maximize yield and production of products. with naphtha reforming in the above-mentioned study, may result
Therefore the temperature profile of both hot and cold streams in some undesirable issues, e.g. occurrence of hot spots in the reac-
as well as other parameters will be obtained in such way that, tors, due to its highly exothermic nature. Additionally, a-alumina-
we have maximum value of the objective function. supported Palladium catalysts are exposed to deactivation by coke

Naphtha Reforming Naphtha Reforming

process catalyst process catalyst

Top View

Naphtha Reforming Naphtha Reforming

process catalyst process catalyst

Fig. 1b. A schematic diagram of optimized moving bed thermally coupled naphtha reactor.
1044 D. Iranshahi et al. / Applied Thermal Engineering 112 (2017) 1040–1056

formation, which is slow at temperatures below 598 K but experi- 1.4. Objective
ences an extreme growth at higher temperatures [17]. Therefore,
finding a more suitable exothermic process for thermal coupling In the present work, attempts have been made to present a
with naphtha reforming can address these operating and financial novel configuration with higher performance and lower energy
problems. consumption for conventional catalytic naphtha reforming and
hydrodealkylation of toluene processes. To achieve this goal, the
concept of continuous catalytic regeneration process is combined
1.3. Hydrodealkylation of toluene with thermally coupled reactor. The attained conventional reactor
model has been validated by comparing with industrial data. DE
According to the aforementioned statements, the exothermic method has been used to optimize the thermally coupled reactor
hydrodealkylation of toluene (HDA) is suggested to be coupled characteristics and operating conditions via optimization of 26
with the naphtha reforming process. Hydrodealkylation of toluene decision variables. Finally, the performance of the optimized ther-
to benzene is one of the most important processes in Eastern Eur- mally coupled reactor (OTCR) has been compared with that of con-
ope owing to the fact that benzene is more practical than toluene. ventional reactor (CR) to assess OTCR superiority. The modeling
HDA process contains nine unit operations: a reactor, a furnace, a results show enhancement in the aromatics and hydrogen produc-
recycle compressor, a vapor-liquid separator, two heat exchangers tion, respectively.
and three distillation columns. Hydrogen and toluene are usually
employed as feed. Benzene is the main product and methane and
diphenyl are the byproducts of this process [31–33]. This process 2. Process description
can be carried out in the presence or absence of catalysts. Catalytic
process is based on the hydrogenation reaction in the presence of Fig. 1(a) demonstrates a simple flow diagram for catalytic naph-
catalysts such as nickel or clinoptilolite [34,35]. tha reforming process. Four radial-flow moving-bed reactors in

Fig. 2. Reaction network of naphtha reforming process.

 D. Iranshahi et al. / Applied Thermal Engineering 112 (2017) 1040–1056 1045

Table 1
The considered reactions in the network kinetic model of endothermic side.

Reactions Reaction rate Carbon Heat of Reaction

number reaction speed


Dehydrogenation ACHn $ An þ 3H2 ð1Þ P P3
r 1n ¼ k1n P ACHn  AnK 1nH2 ð2Þ 6–9+ DH > 0 Very rapid


Dehydrocyclization of paraffin to ACHn þ H2 $ NPn ð3Þ r 2n ¼ k2n P ACHn P H2  K 2n PNP
ð4Þ 6–9+ DH > 0 Rapid
naphthene
 +
ACHn þ H2 $ IPn ð5Þ r 3n ¼ k3n P ACHn P H2  PKIPn ð6Þ 6–9


3n

NP n $ ACP n þ H2 ð7Þ P ACPn PH2

r 4n ¼ k4n P NPn  K 4n ð8Þ 5–9+


IP n $ ACP n þ H2 ð9Þ r 5n ¼ k5n P IPn  PACPn PH2
ð10Þ 5–9+
K 5n


Dehydrocyclization of paraffin to NP n $ An þ 4H2 ð11Þ P P4
r 6n ¼ k6n P Npn  AnK 6nH2 ð12Þ 6–9+ DH > 0 Rapid
aromatic

IP n $ An þ 4H2 ð13Þ PAn PH2
r 7n ¼ k7n P Ipn  K 7n
4
ð14Þ 6–9+


Isomerization of naphthenes ACP n $ ACHn ð14Þ r 8n ¼ k8n P ACPn  PKACHn ð15Þ 6–9+ DH < 0 Rapid


8n

Isomerization of paraffins NP n $ IP n ð16Þ P IPn

r 9n ¼ k9n P NPn  K 9n ð17Þ 4–9+


Isomerization of aromatics R$P ð18Þ r 10n ¼ k10n P R  KP10nP
ð19Þ 8


Transalkylation 2R $ P þ C ð20Þ r 11n ¼ k11n P 2R  PKP11n
PC
ð21Þ 7, 8 DH > 0 Rapid


Cracking of paraffins P 2 þ H2 ! 2P 1 ð22Þ PNPn
r 12n ¼ k12n Pt ð26Þ 2 DH < 0 Very slow
P 3 þ H2 ! P 1 þ P 2 ð23Þ 3
NP 4 þ H2 ! 23 ðP 3 þ P 2 þ P 1 Þ ð24Þ 4
NP 5 þ H2 ! 12 ðNP4 þ P 3 þ P 2 þ P 1 Þ ð25Þ 5
P5
NP n þ n3 6–9+
3 H 2 ! 15 ð27Þ
n
i¼1 P i


IP 4 þ 32 H2 ! 12 ð3P 1 þ P 2 þ P 3 Þ ð28Þ r 13n ¼ k13n PPIPnt ð31Þ 4 DH < 0 Veryslow
IP 5 þ H2 ! 12 ðIP 4 þ P 3 þ P 2 þ P 1 Þ ð29Þ 5
P5
IP n þ n3 6–9+
3 H2 ! 15 ð30Þ
n
i¼1 P i
P5

Cracking of Naphthenes ACHn þ n3 H2 ! 15 n
ð32Þ r 14n ¼ k14n PACHn
ð33Þ 6–9+ DH < 0 Very slow
i¼1 P i Pt


ACP 5 þ 2H2 ! 12 ðNP 4 þ P 3 þ P 2 þ P 1 Þ ð34Þ r 15n ¼ k15n PACPn ð36Þ 5
P5 Pt
ACP n þ n3 H2 ! 15
n
ð35Þ 6–9+
i¼1 P i
Hydrodealkylations Anþ1 þ H2 ! Ak þ C m H2mþ2 ð37Þ r 16n ¼ k16n P Anþ1 P 0:5
H2 ð38Þ 7–9+ DH < 0 Very slow

Table 2
Catalyst deactivations model for endothermic reaction side.
8
>
> Reaction occurs on acid function ak ¼ aA
<
Reaction occurs on metal function ak ¼ aM
r i ¼ ak  r oi
>
> Reaction occurs on acid or metal function ak ¼ meanðaA ; aM Þ
:
Reaction occurs on acid as well as metal functions ak ¼ meanðaA ; aM Þ
Reaction occurs on metal and acid function
ri ¼ ak  r oi ð40Þ r C k ¼ aC k  r oC k ð41Þ
dak
 dC Ck
¼ ak  ank k ð42Þ daC nC
 dC Ck ¼ aC k  aC kk ð43Þ
8 8 k
> n ¼1 ak ¼ expðak  C C k Þ > nC k ¼ 1 aC k ¼ expðaC k  C C k Þ
< k >
< n –1
nk –1 ak ¼ 1
 a ¼ 1  
ð44Þ Ck Ck
ð45Þ
>
:
1
> 1
ð1þðnk 1Þak C C k Þ
nk 1
>
: nC 1
ð1þðnC k 1ÞaC k C C k Þ k

kC k exp ðRT
Ec
Þ
roC k ¼ H n2  C 0:5
ACP ð46Þ
P n1 2
HC


Reaction accours on Acid function ! k ¼ A
k:
Reaction accours on Metal function ! k ¼ M

Table 3 series, associated with catalyst regenerator are employed in this

Required constants for calculating catalyst deactivation rate and activity in endother- process. Regenerated catalysts enter the top of the first reactor
mic reaction side.
and move axially through the reactor. Finally, they exit from the
Parameter Value Unit bottom of the reactor and move to the top of the next reactor. This
nM 1 – procedure continues for all the reactors. The formation of coke on
aM 26 kgcat: the surface of catalysts is an inevitable process. Thus, the catalysts
kgcoke
nC M 1 – are sent to the regenerator and then enter the first reactor again. As
aC M 12.34 kgcat: it is observed from Fig. 1(b), catalysts enter the reactor in the z
kgcoke
nA 1 – direction, while the feed flows radially into the catalyst bed after
aA 14.5 kgcat: entering the reactor. Typically, before entering the feed to the first
kgcoke
nC A 1 – reactor, the ratio of H2/HC is set by mixing it with a hydrogen
aC A 10.18 kgcat: stream coming out of the separator. The feed advances down the
kgcoke
kC M 0.384 n
kgcoke ðkPaÞ 1 m1:5 reactor and then moves through the catalytic bed radially. Then,
0:5
kgcat: ðkmolÞ :h
n the obtained stream enters an axial collector, moves to the follow-
kC A 1.386 kgcoke ðkPaÞ 1 m1:5
kgcat: ðkmolÞ0:5 h ing reactor and this process is repeated for all the reactors. At the
Ec 4055 J
mol end, the exit stream from the fourth reactor is sent to a separator in
n1 0.94 – order to separate the hydrogen-rich gas from hydrocarbons and
n2 1.33 –
the liquid hydrocarbons is sent to the next reactor. The effluent
1046 D. Iranshahi et al. / Applied Thermal Engineering 112 (2017) 1040–1056

Table 4
Specific properties and operating conditions for the conventional and optimized moving bed thermally coupled configurations.

Parameter Numerical value Unit

H2/HC mole ratio 2.193 –
Mole percent of hydrogen in recycle 0.83 –

Distillation fraction of naphtha feed

ASTM D86 Naphtha feed (°C)
IBP 81
5% 91.2
10% 93.2
20% 96.9
30% 101.1
40% 105.7
50% 111.4
60% 117.6
70% 124.5
80% 132.7
90% 143.1
95% 150.5
FBP 159

Specific properties and operating conditions for the conventional configuration

Parameter Numerical value Unit
Naphtha feed stock 233637.01 kg/h
Reformate 216488.63 kg/h

Parameter 1st Reactor 2nd Reactor 3rd Reactor 4th Reactor

Inlet temperature (K) 798 798 798 798
Inlet pressure (kPa) 595 550 505 460
Inner and outer Diameter(m) 1.25, 2.19 1.25, 2.35 1.30, 2.53 1.30, 2.89
Length (m) 8.50 10.35 12.1 15.39
Catalyst distribution (wt%) 12 18 25 45

Specific properties and operating conditions for the optimized thermally coupled configuration
Parameter 1st Reactor 2nd Reactor 3rd Reactor 4th Reactor
Inner and outer diameter (m) 1.25, 3.12 1.25, 2.95 1.30, 2.73 1.30, 2.77
Length (m) 15.57 15.51 15.14 16.09
NOS (Number Of Subsection) 18.97 17.61 14.78 13.48
Rh ¼ h1 =ðh1 þ h2 Þ 0.42 0.49 0.64 0.64

Endothermic side
Parameter Numerical Value Unit
Naphtha feed stock 233637.01 kg/h
Reformate 216488.63 kg/h

Parameter 1st Reactor 2nd Reactor 3rd Reactor 4th Reactor

Inlet temperature (K) 798 798 798 798
Inlet pressure (kPa) 595 550 505 460
Catalyst distribution (wt%) 0.2379 0.2416 0.2460 0.2745

Exothermic side
Composition Numerical value Unit
Toluene 0.5 –
Hydrogen 0.5 –
Benzene 0.0 –
Methane 0.0 –
Parameter 1st Reactor 2nd Reactor 3rd Reactor 4th Reactor
Feed molar flow rate (kmol/h) 4394.8 3619.3 3139.9 2902.3
Inlet temperature (K) 946.6 942.4 937.791 931.741
Inlet pressure (kPa) 3400 3400 3400 3400

Table 5
Typical properties of catalyst and reactor’s bed.
stream is removed through a tube and sent to the next reactor. This
Endothermic side of CR, TCR and OTCR configurations
procedure continues for all reactors.
Parameter Numerical Value Unit Since catalytic naphtha reforming is an endothermic process, an
dp 1.8 Mm exothermic process is coupled with it in the proposed model. Thus,
Pt 0.3 wt% it not only prevents waste of energy, but also increases the perfor-
Sn 0.3 wt%
mance of the endothermic side. In this thermally coupled configu-
sa 220 m2/g
qB 680 kg/m3 ration, the cross sectional area of reactor is divided into some
e 0.36 – subsections in some of which exothermic reaction takes place
 D. Iranshahi et al. / Applied Thermal Engineering 112 (2017) 1040–1056 1047

Table 7
Endothermic The optimized parameters for moving bed thermally coupled reactor in the naphtha
side Catalyst reforming process.
particle Decision variable Min. Max. Optimized
value
Number of subsections (NOS) for the 1st 10 25 18.9719
reactor
Number of subsections (NOS) for the 2nd 10 25 17.6158
reactor
Number of subsections (NOS) for the 3rd 10 25 14.7825
reactor
Number of subsections (NOS) for the 4th 10 25 13.4835
reactor
Total fresh naphtha feed to the plant 2000 3000 2530.0487
(kmol/h)

Δz
Exothermic Exothermic side molar flow rate of 1st 4000 4500 4394.777
side reactor (kmol/h)
Exothermic side molar flow rate of 2nd 3500 4000 3619.323
reactor (kmol/h)
Exothermic side molar flow rate of 3rd 3000 3500 3139.919
reactor (kmol/h)
Exothermic side molar flow rate of 4th 2500 3000 2902.264
reactor (kmol/h)
The ratio of Rh for the 1st reactor 0.30 0.80 0.42657
The ratio of Rh for the 2nd reactor 0.30 0.80 0.49730
The ratio of Rh for the 3rd reactor 0.30 0.80 0.64233
The ratio of Rh for the 4th reactor 0.30 0.80 0.64828
Fig. 3. A differential element for mass and energy balances.
Inlet temperature of exothermic reaction 850 950 946.614a
side of first reactor (K)
Inlet temperature of exothermic reaction 850 950 942.463a
side of second reactor (K)
Table 6 Inlet temperature of exothermic reaction 850 950 937.791a
Mass & energy balances and auxiliary relations for the optimized thermally coupled side of third reactor (K)
configuration. Inlet temperature of exothermic reaction 850 950 931.741a
side of fourth reactor (K)
Mass balance (Endothermic Reaction side)
LOD of first reactor 3 6 4.9948
@C j qb Pm j ¼ 1; 2; . . . ; n (59)
i¼1 ai  mij r i
Cj C j @ur
@r ¼  r  ur @r þ ur
LOD of second reactor 3 6 5.2554
i ¼ 1; 2; . . . ; m
LOD of third reactor 3 6 5.5461
Mass balance (Exothermic Reaction side) LOD of fourth reactor 3 6 5.8055
@C Exo C Exo C Exo @uExo (60) Hydrogen mole fraction in the recycle gas 0.55 95 0.8991
@r
j
¼ j
r  uExo
j r
@r þ uExo
1
ðmj r 0 Þ j ¼ 1; 2; . . . ; n0
r r
Catalyst distribution for the 1st reactor 0.1 0.5 0.237948
Energy balance (Endothermic Reaction side) (weight fraction)
q P Catalyst distribution for the 2nd reactor 0.1 0.5 0.241564
@T
@r¼  ur C TbC P m
i¼1 ðDHi  ai  r i Þ þ prRh ur C T C P  U  ðT
NOS Exo
 TÞ (61)
(weight fraction)
Energy balance (Exothermic Reaction side) Catalyst distribution for the 3rd reactor 0.1 0.5 0.245979
@T Exo
¼  uExo C Exo
1
ðDHExo  r 0 Þ þ prð1R NOS
 U  ðT  T Exo Þ (62) (weight fraction)
@r r C Exo Exo Exo Exo
h Þur C T C P
T P
Catalyst distribution for the 4th reactor 0.1 0.5 0.274480
Velocity distribution (Endothermic Reaction side) (weight fraction)
qb Pn Pm
i¼1 ai  mij r i
@ur ur ur @C T (63)
@r ¼  r  C T  @r þ C T j¼1 a
Conditional.
Velocity distribution (Exothermic Reaction side)
@uExo @C Exo Pn0 (64)
j¼1 ðmj r Þ
uExo uExo 0
@r ¼  r  C Exo  @r þ C Exo
r r r T 1
T T

Mass and energy balance for Collector (Endothermic Reaction side)

@F j 2pRi
whilst in the others endothermic reaction is carried out. Here, HDA
(65)
@z ¼ NOS ure C je  Rh j ¼ 1; 2; . . . ; n
process is considered to be a heat source. The feed which contains
2pRi T 2pRi
@T
¼ F T C P  NOS
1
 C Te ure Rh C pe ðT e  TÞ  F T ð NOS ure  Rh  C Te Þ  CTP @C P (66)
@z @z toluene and hydrogen enter the reactor radially and after releasing
Mass and energy balance for Collector (Exothermic Reaction side) heat due to the reaction, enters to the collector and exits. In the
@F Exo pRi (67) proposed model, the effluent stream does not enter the next reac-
@z
j
¼ 2NOS  uExo Exo
re C je ð1  Rh Þ j ¼ 1; 2; . . . ; n0
(68) tor; rather fresh feed enters each reactor to increase the production
@T Exo C Exo Exo
Te ure 2pRi
¼ Exo   ð1  Rh Þ  C Exo Exo
Pe ðT e  T Exo Þ and heat transfer.
@z F T C Exo
P
NOS
 
T Exo 2pRi Exo T Exo @C Exo
 Exo  uExo
re ð1  Rh ÞC Te  Exo P 3. Kinetic model
FT NOS CP @z

Ergun equation (Pressure drop)

3.1. Naphtha reforming process
l ð1eÞ 1:75q ð1eÞ 2
2
(69)
dP
dr
¼ 150
2 2 e3 ur þ / dp e3 ur
/s dp s

Additional relation Naphtha is a complex mixture of hydrocarbons containing 300

Rh ¼ h1hþh
1 (70) different components that are similar inherently [36] and different
2

Nu ¼ 4:21  ðRep Þ0:365 (71) reactions might take place among them. The most determining
l ¼ expðC 1 þ C 2 =T þ C 3 lnT þ C 4 T c5 ), (72) reactions are as follows [37,38]:
d3

2 d5

2
C P ¼ d1 þ d2 T d  þ d4 T d 
sinh 3
T sinh 5
T
Dehydrogenation and isomerization of naphtha to aromatics.

Dehydrogenation of paraffins to olefins.
1048 D. Iranshahi et al. / Applied Thermal Engineering 112 (2017) 1040–1056

Table 8
Comparison between predicted molar flow rate in conventional configuration and plant data (steady-state model validation).

Pseudo components Molecular weight Input (mole fraction) Output plant (kmole/h) Output model (kmole/h) Deviation (kmole/h)
n-P6 (C6H14) 86.178 0.0229 71.84 70.98 0.86
n-P7 (C7H16) 100.205 0.0292 46.65 46 0.65
n-P8 (C8H18) 114.232 0.0239 8.1 7.63 0.47
n-P9 (C9H20) 128.259 0.0156 1.15 1.25 0.1
i-P6 (C6H14) 86.178 0.0232 216.06 216.06 0
i-P7 (C7H16) 100.25 0.0314 110.18 109.69 0.49
i-P8 (C8H18) 114.232 0.0338 22.75 22.67 0.08
i-P9 (C9H20) 128.259 0.0244 1.79 1.95 0.16
ACH6 (C6H12) 84.162 0.0077 0.51 0.92 0.41
ACH7 (C7H14) 98.189 0.0084 0.92 1.43 0.51
ACH8 (C8H16) 112.216 0.0115 2.33 2.34 0.01
ACH9 (C9H18) 126.243 0.0018 0.04 0.05 0.01
ACP5 (C5H10) 70.135 0.0001 2.14 2.11 0.03
ACP6 (C6H12) 84.162 0.003 26 26.02 0.02
ACP7 (C7H14) 98.189 0.0065 2.33 3.12 0.79
ACP8 (C8H16) 112.216 0.0084 0.6 0.6 0
ACP9 (C9H18) 126.243 0.0012 0.01 0.02 0.01
A6 (C6H6) 78.114 0.0086 205.84 206.39 0.55
A7 (C7H8) 92.141 0.0109 453.93 454.33 0.4
A8 (C8H10) 106.168 0.0021 163.02 163.47 0.45
A9 (C9H12) 120.195 0.0026 323.66 324.46 0.8
A8 (C8H10) 106.168 0.0015 113.62 113.4 0.22
A8 (C8H10) 106.168 0.0016 120.84 121.55 0.71
A8 (C8H10) 106.168 0.0036 276.56 276.8 0.24
H2 2.016 0.6226 10071.31 10090.82 19.51
P1 (CH4) 16.043 0.0211 398.53 396.64 1.89
P2 (C2H6) 30.07 0.0231 399.16 397.72 1.44
P3 (C3H8) 44.097 0.0202 352.36 351.69 0.67
P4 (C4H10) 58.124 0.0106 189.21 189.93 0.72
P5 (C5H12) 72.151 0.0035 71.06 71.69 0.63
i-P4 58.124 0.0073 140.75 139.38 1.37
i-P5 72.151 0.0076 149.21 149.49 0.28
Total – – 13942.46 13960.6 18.14

Table 9
Comparison between operating conditions in conventional configuration and plant data.

Parameter 1st Reactor 2nd Reactor 3rd Reactor 4th Reactor

Outlet temperature of plant (K) 707 725 743 761
Outlet temperature of model (K) 708.9 725.1 742.2 758.1
Outlet pressure of plant (kPa) 581 535 490 446
Outlet pressure of model (kPa) 585.1 540.3 495.7 451.4
Catalyst feed flow rate of plant (kg/s) 0.18 0.22 0.26 0.37
Catalyst feed flow rate of model (kg/s) 0.18 0.22 0.26 0.37

Fig. 4. Temperature profile of endothermic side of reactor. Fig. 5. Temperature profile of exothermic side of reactor.
 D. Iranshahi et al. / Applied Thermal Engineering 112 (2017) 1040–1056 1049

Dehydrocrystalization of paraffins to olefins and aromatics. model these omissions have been taken into account to increase

Isomerization to isoparaffins. the model accuracy. Fig. 2 indicates suggested reaction network

Hydrocracking of paraffins and naphthenes to light ends. for naphtha reforming process [44]. The components and reactions
are illustrated in Table 1. Also the catalyst deactivations model for
Lumped models have been presented to reduce the complexity endothermic reaction side and the constants are listed in Tables 2
of this intricate reaction system. The first attempt to model naph- and 3, respectively.
tha reforming process using this approach was made by Smith
et al. in 1959 [39]. In their model, the feedstock is subdivided to 3.2. Hydrodealkylation of toluene
naphthenes, paraffins and aromatics. Another valuable study
reported in the literature was carried out by Kmak, Krane et al., In this process, toluene is hydrodealkylated in a homogeneous
and Ramage et al. [40–43]. In these models, chemically similar and gaseous non-catalytic bed where benzene and methane are
components are merged in one group and marked as a pseudo- produced as products. In another equilibrium reaction, benzene
component. Then, the reactions among these pseudo-components loses hydrogen and converts to diphenyl. Benzene is the main pro-
are considered. In this study, a kinetic model similar to Vathi and duct whilst hydrogen and diphenyl are by-products. These reac-
Chaudhuri’s model [44] is used. It should be mentioned that they tions and their rate laws are listed below [31,33].
ignored some of the main lumps and reactions. However, in our

Toluene þ H2 ! Benzene þ CH4 DH ¼ 50 kJ=mole ð47Þ

2Benzene $ Diphenyl þ H2 DH ¼ 0 kJ=mole ð48Þ

 
25616 1=2
R1 ¼ 1:9580  108 exp pT pH2 ð49Þ
T
 
25616 2
R2 ¼ 1:2115  106 exp pB  5:1662  106
T
 
25616
 exp pD pH2 ð50Þ
T
The specifications of the conventional and proposed reactors
and catalysts are tabulated in Tables 4 and 5, respectively.

4. Mathematical model

In recent years modeling and simulation has been widely used

for reactors to predict and improve their performance. Therefore
many types of reactors like fixed-beds, fluidized beds and moving
beds reactors are modeled, such as Xiong et al. studies [45–48] for
Fig. 6. Total molar flow rate of naphtha reforming process. biomass fast pyrolysis in fluidized bed.

5
OTCR

4.5

4 CR
H 2 /HC molar ratio

3.5

2.5

2 1
1
0.8 0.5
0.6
0.4
0.2
Radius of reactor 0 0 Length of reactor
(Dimensionless) (Dimensionless)

Fig. 7. Comparison of H2/HC molar ratio between CR and OTCR.

1050 D. Iranshahi et al. / Applied Thermal Engineering 112 (2017) 1040–1056

A two-dimensional homogeneous model is proposed to model
Steady state condition is considered.

the catalytic naphtha reforming process. Since catalysts move axi-
The ideal gas law is applied to the components.
ally (in z direction) and feed moves radially (in r direction), the
Gradient along perimeter is neglected.
resultant equations will be partial differential equations (PDEs).
Catalyst bed is assumed to be homogeneous.
To develop a model, heat and material balances and continuity
Diffusion of mass and heat in catalyst pellets is neglected.
equation must be written. Fig. 3 shows a differential element for
Heat loss is considered negligible.
mass and energy balances. Since there are several components in
the feed and material balance is written for each of them, a group Because of the coke generation on the catalyst surface and cat-
of differential equation is obtained. For each exothermic and alyst deactivation, its deactivation equation must be solved simul-
endothermic side, equations should be written separately. There taneously. The equations in mathematical modeling are
is no mass transfer between these two sides. However, heat trans- summarized in Table 6 [49–51].
fer between two sides must be taken into account in the calcula-
tions. Because of pressure drop in the catalyst bed, momentum
equation must be considered too. Ergun equation is applied to 5. Numerical solution
reflect the pressure drop. Physical properties such as viscosity, heat
capacity and etc. are determined based on dependent variables and As it was mentioned in the previous section, governing equa-
replaced in the main equations. tions are in the form of partial differential equations; therefore, a
The following assumptions are taken into account in the proper method should be used to solve these equations. In this
modeling: study, finite difference method is applied. In this method, the reac-

(a)
0.06
Total coke
)
cat

0.05
/kg
coke

0.04
Coke weight function (kg

0.03

0.02
Acidic function

0.01

0
1
1
Metalic function 0.8
0.5 0.6
0.4
0.2
0 0
Length of reactor Radius of reactor
(Dimensionless) (Dimensionless)

(b)
1

0.9
Mean catalyst activity

0.8

0.7

0.6
1 0.5
0.8
0.6 Radius of reactor
0.4
0.2 (Dimensionless)
Length of reactor 0 0
(Dimensionless)

Fig. 8. (a) Coke weight fraction, (b) Mean activity of catalyst in endothermic side of OTCR.
 D. Iranshahi et al. / Applied Thermal Engineering 112 (2017) 1040–1056 1051

tor is divided into a series of nodes and heat and mass equations in a way to have the same order of magnitude. Some constraints
are solved for each node according to the boundary conditions. should be defined to maintain the other condition which affects
As the number of node increases, the solution accuracy and time the reactions in their ideal state. The constraints of the current
is increased [52,53]. In this study 400 and 1200 nodes are consid- optimization are as follows:
ered in directions of z and r directions, respectively.
DPi 6 35 kPa ð52Þ

6. Optimization T endoi 6 800 K ð53Þ

In order to achieve the highest production rate of the desired C ci 6 0:05% ð54Þ
product, Differential Evolution (DE) method has been applied. In
this technique, an objective function must be defined first of all. X
4
The following objective function is considered to maximize the wi ¼ 1 ð55Þ
yield and production of aromatics and hydrogen. Also, decision i¼0

variables which are related to the formation of products are

required. Decision variables and their values are illustrated in Y OTCR
i > Y CR
i ð56Þ
Table 7.
  DF OTCR > DF CR ð57Þ
DF Ar DF H2 i i
OF ¼ ðY Ar þ Y H2 Þ þ þ ð51Þ
2000 2500 Finally, the optimization process is completed by minimizing
the penalty function while maximizing the objective function. This
In this equation, Y and DF are yield and production, respec-
optimization problem is carried out in Matlab. The governing equa-
tively. The above terms are divided by appropriate numbers
tions are as follows:
(2000 for aromatics production and 2500 for hydrogen production)

(a) (a)

(b) (b)

Fig. 9. Comparison of (a) Iso paraffin consumption, (b) Normal paraffin consump- Fig. 10. (a) Alkylcyclohexane and (b) Alkylcycopatane consumption in CR and
tion between CR and OTCR. OTCR.
1052 D. Iranshahi et al. / Applied Thermal Engineering 112 (2017) 1040–1056

X
n G2 ¼ Maxð0; ðT i  800ÞÞ ð60Þ
Result ¼ OF þ 107 G2i ð58Þ
i¼1
G3 ¼ Maxð0; ðC ci  0:05ÞÞ ð61Þ
G1 ¼ Maxð0; ðDPi  30ÞÞ ð59Þ
G4 ¼ Maxð0; ðW 1 þ W 2 þ W 3 þ W 4 Þ  1:0001Þ ð62Þ

G5 ¼ Maxð0; 0:9999  ðW 1 þ W 2 þ W 3 þ W 4 ÞÞ ð63Þ

(a) G6 ¼ Maxð0; Y CR OTCR

i  Yi Þ ð64Þ

G7 ¼ Maxð0; DF CR OTCR
i  DF i Þ ð65Þ

7. Validation of model

The accuracy of the presented model has been investigated by

comparing the obtained results with the industrial data from
domestic plant [54]. The modeling and industrial results for com-
ponents’ exit molar flow rates are tabulated in Table 8; and it is
obvious that they are in satisfactory agreement. Also, the outlet
temperatures and pressure for each reactor, as well as molecular
weight of the process exit stream are illustrated in Table 9. The
minor deviation between plant data and modeling results can be
justified with the applied simplifying assumptions in addition to

(b) (a)

(c) (b)

Fig. 11. A comparison between, (a) Aromatic production, (b) Hydrogen production Fig. 12. (a) Rate of cracking of IP6 to light ends and (b) Rate of dehydrogenation
and (c) Light ends production of CR and OTCR along radius of each reactor. ACH6 in CR and OTCR.
 D. Iranshahi et al. / Applied Thermal Engineering 112 (2017) 1040–1056 1053

the fact that the utilized kinetic model for the reaction system 8.3. Variation of molar flow rate of pseudo-components
might undervalue the true reaction rates.
The aim of the catalytic naphtha reforming process is conver-
sion of paraffinic and naphthenic components to the aromatics,
hydrogen and lighter components. So, paraffinic and naphthenic
8. Results and discussion
pseudo-components are reactants and their mole fraction decrease
along the reactor. Iso paraffins and normal paraffins consist of var-
In this study, the catalytic naphtha reforming process is coupled
ious components which convert to aromatics and naphthenes.
with hydrodealkylation i.e. benzene production process. The
Paraffins are converted to aromatics during dehydrocyclization.
obtained results are presented in two and three dimensional
Besides, normal paraffins are converted to iso paraffins during
curves according to axial and radial variation of the following
the isomerization reactions. Iso-paraffin consumption is illustrated
parameters in the reactor.
in Fig. 9(a). A part of isoparaffin is consumed in each reactor and
the residual is sent to the next reactor. Temperature rise in each

Operating conditions for each section (exothermic and
reactor (because of the coupled exothermic reaction) leads to an
endothermic section).
increment in isoparaffin consumption besides reducing its molar

Physical properties, catalyst activity and coke formation.
flow rate.

Production and consumption of pseudo-components and their
Fig. 9(b) displays consumption of normal paraffins. According to
conversion.
the reactions, normal and isoparaffins are converted to aromatics
in separated reactions. It should be mentioned that conversion of
normal paraffin is more than conversion of isoparaffins and it
8.1. Variations of operating conditions in reactors increases with temperature increment in the reactor.
Naphthenic components are one of the other main ingredients
The temperature profile in the catalytic conversion side is illus- in the feed. This pseudo-component is divided into two compo-
trated in Fig. 4. Due to the endothermic nature of the process, nents; Alkylcyclopentane, and Alkylcyclohexane, each of them
higher amount of products can be achieved with an increase in
temperature of each reactor. By comparing the profiles obtained
for each reactor, it can be observed that the exothermic reaction
has a desirable effect on the catalytic naphtha reforming process (a)
and results in temperature rise of 10–20 °C in each reactor. Fig. 5
reveals the temperature profile in the exothermic side. The tem-
perature increases initially in this section as a result of high initial
concentration of reactants; and then it decreases because of heat
transfer to the endothermic side. In order to increase benzene pro-
duction and heat transfer between two sides, the inlet temperature
of the second and third reactor is higher than the first and second
reactor, respectively.
Total molar flow rate of each reactor in endothermic side is
demonstrated in Fig. 6. The total molar flow rate in each reactor
is increased as a positive impact of temperature rise on products
formation in endothermic section. It should be mentioned that
the concentration of reactants and their molar flow rate decreases.
But generation of products is much more than consumption of
reactants, which compensates the reactants reduction. H2/HC ratio
is another operating factor affecting coke formation. In proposed
OCTR reactor, hydrogen formation and reactants consumption lead
to H2/HC increment and reduction of coke formation consequently
(see Fig. 7).
(b)

8.2. Variations of catalyst activity and coke formation

As it was discussed, catalyst activity decreases because of

unavoidable coke formation during the process which has a nega-
tive effect on the yield of products. Coke weight fraction on acidic
and metallic functions along the length and radius of the reactor is
exhibited in Fig. 8(a). As it is noticed, coke formation on acidic
functions is lower than metallic ones. Average formation of coke
on the catalyst surface is lower than 0.05. If coke percentage
increases to more than 0.05, it will not be deposited in the regen-
erator properly.
Catalyst activity, which decreases with increasing coke forma-
tion, in all the reactors is illustrated in Fig. 8(b). Fresh catalysts
enter the first reactor. Hence, activity of catalyst in this reactor is
higher than the other reactors. On the other hand, the activity of
catalysts in the third reactor is the lowest due to the fact that deac- Fig. 13. (a) Methane and (b) benzene production in exothermic side along the
tivated catalysts enter this reactor. reactors.
1054 D. Iranshahi et al. / Applied Thermal Engineering 112 (2017) 1040–1056

has its various ingredients. Naphthenic pseudo-components are increases due to the progress in isomerization and dehydrocycliza-
converted to aromatics and paraffinic components during dehy- tion of naphthenes and paraffins reactions. Temperature increment
drogenation and dehydrocyclization processes. Also, naphthenic in each reactor increases aromatics production in comparison with
compounds are converted to each other in isomerization reactions. conventional reactors.
Based on these reactions, they are transferred to the next reactor to Hydrogen is one of the other main products, which is produced
have more conversion as reactants. The proposed model has a during dehydrogenation and dehydrocyclization reactions; while
favorable impact on these pseudo-components according to consumed during cracking of naphthenes and paraffins. Fig. 11(b)
Fig. 10(a) and (b) in comparison with conventional reactors. represents that hydrogen production in thermally coupled reactor
According to the rate law of coke formation, the amount of Alkyl- is more than that conventional reactor. By coupling HDA with
cyclopentane (ACP), in addition to H2/HC ratio, has a direct effect naphtha reforming, more heat is available for the hydrogen
on this reaction. It is clearly spotted in Fig. 10(b) that consumption production in endothermic dehydrogenation of naphthenes. There-
of ACP in the proposed reactor increases, which leads to reduction fore, its yield increases.
in coke formation along with an increase in yield of the desired Light ends, other products of the catalytic naphtha reforming
products. process, are produced during cracking reactions, which take place
As it was discussed, aromatics are the main products of the cat- at high temperatures. They are used as feedstock for LPG
alytic naphtha reforming process. ortho-xylene, para-xylene, ethyl production in refineries. Fig. 11(c) describe that the temperature
benzene and benzene are considered as aromatics in this study. As in each reactor of the suggested configuration is higher than the
it is illustrated in Fig. 11(a), production of aromatic in each reactor conventional one. Subsequently, the yield of light ends is higher.

35
CR OTCR
(a)
Mean molecular weight (kg/kmol)

30

25

20
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Radius of reactor (Dimensionless)

10
-5 (b)
1.8
Viscosity of endothermic side (Pa.s)

1.75

1.7

1.65

1.6
1 1
0.8 0.8
0.6 0.6
0.4 0.4
Radius of reactor 0.2 0.2 Length of reactor
(Dimensionless) 0 0
(Dimensionless)

Fig. 14. (a) Mean molecular weight of endothermic side and (b) Endothermic side viscosity.
 D. Iranshahi et al. / Applied Thermal Engineering 112 (2017) 1040–1056 1055

The reason for choosing thermally coupled reactor is the effect of [5] D.J. Roddy, Development of a CO2 network for industrial emissions, Appl.
Energy 91 (2012) 459–465.
temperature on the reactions. With heat transfer from the heat
[6] F. Ciapetta, D. Wallace, Catalytic naphtha reforming, Catal. Rev. 5 (1972) 67–
source, temperature of endothermic side and consequently reac- 158.
tion rate increase. Higher reaction rates increase reactants con- [7] J.A. George, M.A. Abdullah, Catalytic Naphtha Reforming, Marcel Dekker, New
sumption and products formation. Since the reactions are York, 2004.
[8] T. Namioka, A. Saito, Y. Inoue, Y. Park, T.-J. Min, S.-A. Roh, K. Yoshikawa,
numerous, two reactions are chosen and shown in Fig. 12 Hydrogen-rich gas production from waste plastics by pyrolysis and low-
(a) and (b) for instance. It is evident that cracking of IP6 and dehy- temperature steam reforming over a ruthenium catalyst, Appl. Energy 88
drogenation of ACH6 increase as the temperature increase. It (2011) 2019–2026.
[9] H. Taghvaei, M.M. Shirazi, N. Hooshmand, M.R. Rahimpour, A. Jahanmiri,
should be notified that cracking reaction is more temperature- Experimental investigation of hydrogen production through heavy naphtha
sensitive than other reactions. cracking in pulsed DBD reactor, Appl. Energy 98 (2012) 3–10.
Fig. 13(a) and (b) shows methane and benzene production. in [10] Z. Liao, J. Wang, Y. Yang, G. Rong, Integrating purifiers in refinery hydrogen
networks: a retrofit case study, J. Cleaner Prod. 18 (2010) 233–241.
HDA process, respectively. Hydrodealkylation of toluene is a [11] M.R. Rahimpour, D. Iranshahi, E. Pourazadi, K. Paymooni, A.M. Bahmanpour,
equimolar reaction, so the molar flow rate of benzene and methane The aromatic enhancement in the axial-flow spherical packed-bed membrane
must be equal, but benzene involves in another reaction which is naphtha reformers in the presence of catalyst deactivation, AIChE J. 57 (2011)
3182–3198.
consumed. Therefore molar flow rate of benzene is slightly less [12] X. Song, Z. Wang, Y. Jin, M. Gong, Investigations on hydrodynamics of radial
than methane. flow moving bed reactors, Chem. Eng. Technol. 16 (1993) 383–388.
[13] O.A. Bariås, A. Holmen, E.A. Blekkan, Propane dehydrogenation over supported
Pt and Pt–Sn catalysts: catalyst preparation, characterization, and activity
8.4. Variations of physical properties and other parameters measurements, J. Catal. 158 (1996) 1–12.
[14] R. Burch, L. Garla, Platinum-tin reforming catalysts: II. Activity and selectivity
in hydrocarbon reactions, J. Catal. 71 (1981) 360–372.
The average molecular weight of components in the reactor is [15] S. de Miguel, A. Castro, O. Scelza, J.L.G. Fierro, J. Soria, FTIR and XPS study of
demonstrated in Fig. 14(a). According to this figure, molecular supported PtSn catalysts used for light paraffins dehydrogenation, Catal. Lett.
weight inside the reactor is decreased. Also, the total molecular 36 (1996) 201–206.
[16] Z. Huang, J. Fryer, C. Park, D. Stirling, G. Webb, Transmission electron
weight of the stream is declined due to conversion of heavy frac- microscopy and energy dispersive X-ray spectroscopy studies of Pt–Sn/c-
tions to lighter products such as hydrogen. Production of hydrogen Al2O3 catalysts, J. Catal. 159 (1996) 340–352.
and lighter ends in suggested model is more than conventional [17] J. Völter, U. Kürschner, Deactivation of supported pt and pt-sn catalysts in the
conversion of methylcyclopentane, Appl. Catal. 8 (1983) 167–176.
reactors. Thus, reduction of molecular weight in the proposed con- [18] M. González-Marcos, B. Inarra, J. Guil, M. Gutierrez-Ortiz, Development of an
figuration is slightly more than conventional reactors. industrial characterisation method for naphtha reforming bimetallic Pt-Sn/
Variation of viscosity along radius and length of reactors is illus- Al2O3 catalysts through n-heptane reforming test reactions, Catal. Today 107
(2005) 685–692.
trated in Fig. 14(b). Temperature has a direct effect on the viscos- [19] J.M. Hill, R. Cortright, J. Dumesic, Silica-and L-zeolite-supported Pt, Pt/Sn and Pt/
ity. Average temperature is the lowest and the highest in the first Sn/K catalysts for isobutane dehydrogenation, Appl. Catal. A: Gen. 168 (1998)
and last reactor, respectively. So, gradual increment of viscosity 9–21.
[20] F. Friedler, Process integration, modelling and optimisation for energy saving
from the first to the last reactor is approved. It should be men-
and pollution reduction, Appl. Therm. Eng. 30 (2010) 2270–2280.
tioned that viscosity increment affects the pressure drop. Accord- [21] A.P. Rossiter, Succeeding in process integration, Process Industries Expo User
ingly, the increase in the reactor temperature should not be too Conference, vol. 31, 2003.
[22] J.B. Hunter, G. McGuire, Method and apparatus for catalytic heat exchange,
high that causes a high pressure drop inside the reactor.
Google Patents (1980).
[23] N.S. Abo-Ghander, J.R. Grace, S.S. Elnashaie, C.J. Lim, Modeling of a novel
membrane reactor to integrate dehydrogenation of ethylbenzene to styrene
9. Conclusion with hydrogenation of nitrobenzene to aniline, Chem. Eng. Sci. 63 (2008) 1817–
1826.
[24] M. Khademi, P. Setoodeh, M. Rahimpour, A. Jahanmiri, Optimization of
In the proposed configuration, conventional reactor for the cat- methanol synthesis and cyclohexane dehydrogenation in a thermally
alytic naphtha reforming process is replaced by a thermally cou- coupled reactor using differential evolution (DE) method, Inte. J. Hydrogen
pled reactor to prevent heat loss and enhance yield of aromatics Energy 34 (2009) 6930–6944.
[25] R. Smith, Chemical Process Design and Integration, Wiley, 2005.
and desired products. Each reactor is divided into several subsec-
[26] I.C. Kemp, Pinch Analysis and Process Integration: A User Guide on Process
tions in which exothermic and endothermic reactions occur simul- Integration for the Efficient Use of Energy, Elsevier Science, 2011.
taneously. Since the catalytic naphtha reforming process is one of [27] H.S. Fogler, Elements of Chemical Reaction Engineering, fourth ed., Prentice
Hall, New York, 2006.
the most important industrial processes, the suggested model
[28] M. Jafari, R. Rafiei, S. Amiri, M. Karimi, D. Iranshahi, M.R. Rahimpour, H.
can be an acceptable alternative instead of conventional reactors. Mahdiyar, Combining continuous catalytic regenerative naphtha reformer
To optimize and maximize inlet parameters and products yield, with thermally coupled concept for improving the process yield, Int. J.
differential evolution method (DE) has been employed in this Hydrogen Energy 38 (2013) 10327–10344.
[29] D. Iranshahi, R. Rafiei, M. Jafari, S. Amiri, M. Karimi, M.R. Rahimpour, Applying
study. In the suggested model, aromatics and hydrogen production new kinetic and deactivation models in simulation of a novel thermally
are increased up to 19% and 23%, respectively, in comparison with coupled reactor in continuous catalytic regenerative naphtha process, Chem.
those in the conventional reactor. Moreover, another advantage of Eng. J. 229 (2013) 153–176.
[30] M. Karimi, M.R. Rahimpour, R. Rafiei, A. Shariati, D. Iranshahi, Improving
this new configuration is that the temperature of reactors can be thermal efficiency and increasing production rate in the double moving beds
adjusted by changing the inlet flow rates of exothermic reaction. thermally coupled reactors by using differential evolution (DE) technique,
Appl. Therm. Eng. 94 (2016) 543–558.
[31] J.M. Douglas, Conceptual Design of Chemical Processes, McGraw-Hill, New
References York, 1988.
[32] S. Emets, K. Hoo, U. Mann, A modified hierarchy for designing chemical
processes, Ind. Eng. Chem. Res. 45 (2006) 5037–5043.
[1] Z.A. Aboosadi, A. Jahanmiri, M. Rahimpour, Optimization of tri-reformer
[33] W.L. Luyben, B.D. Tyreus, M.L. Luyben, Plantwide Process Control, 1998.
reactor to produce synthesis gas for methanol production using differential
[34] J. Papp, D. Kalló, G. Schay, Hydrodemethylation of toluene on clinoptilolite, J.
evolution (DE) method, Appl. Energy 88 (2011) 2691–2701.
Catal. 23 (1971) 168–182.
[2] A. Demirbas, Competitive liquid biofuels from biomass, Appl. Energy 88 (2011)
[35] C.C. Zimmerman, R. York, Thermal demethylation of toluene, Ind. Eng. Chem.
17–28.
Process Des. Dev. 3 (1964) 254–258.
[3] M. Ding, T. Hayakawa, C. Zeng, Y. Jin, Q. Zhang, T. Wang, L. Ma, Y. Yoneyama, N.
[36] M.Z. Stijepovic, A. Vojvodic-Ostojic, I. Milenkovic, P. Linke, Development of a
Tsubaki, Direct conversion of liquid natural gas (LNG) to syngas and ethylene
kinetic model for catalytic reforming of naphtha and parameter estimation
using non-equilibrium pulsed discharge, Appl. Energy 104 (2013) 777–782.
using industrial plant data, Energy Fuels 23 (2009) 979–983.
[4] M. Khobragade, S. Majhi, K. Pant, Effect of K and CeO2 promoters on the activity
[37] J. Ancheyta-Juárez, E. Villafuerte-Macías, Kinetic modeling of naphtha catalytic
of Co/SiO2 catalyst for liquid fuel production from syngas, Appl. Energy 94
reforming reactions, Energy Fuels 14 (2000) 1032–1037.
(2012) 385–394.
1056 D. Iranshahi et al. / Applied Thermal Engineering 112 (2017) 1040–1056

[38] M.A. Rodríguez, J. Ancheyta, Detailed description of kinetic and reactor [47] Q. Xiong, S. Aramideh, A. Passalacqua, S.-C. Kong, BIOTC: an open-source CFD
modeling for naphtha catalytic reforming, Fuel 90 (2011) 3492–3508. code for simulating biomass fast pyrolysis, Comput. Phys. Commun. 185
[39] R. Smith, Kinetic analysis of naphtha reforming with platinum catalyst, Chem. (2014) 1739–1746.
Eng. Prog. 55 (1959) 76–80. [48] Q. Xiong, S.-C. Kong, A. Passalacqua, Development of a generalized numerical
[40] W. Kmak, A kinetic simulation model of the powerforming process, in: AIChE framework for simulating biomass fast pyrolysis in fluidized-bed reactors,
Meeting, Houston, TX, 1972. Chem. Eng. Sci. 99 (2013) 305–313.
[41] H. Krane, A. Groh, B. Schulman, J. Sinfelt, 4. Reactions in catalytic reforming of [49] R.B. Bird, Transport phenomena, Appl. Mech. Rev. 55 (2002) R1–R4.
naphthas, in: 5th World Petroleum Congress, World Petroleum Congress, [50] H. Perry, Perry’s Chemical Engineering Handbook, eighth ed., McGraw-Hill,
1959. New York, 2007.
[42] M.P. Ramage, K.R. Graziani, F. Krambeck, 6 Development of mobil’s kinetic [51] D. Wen, Y. Ding, Heat transfer of gas flow through a packed bed, Chem. Eng.
reforming model, Chem. Eng. Sci. 35 (1980) 41–48. Sci. 61 (2006) 3532–3542.
[43] M.P. Ramage, K.R. Graziani, P.H. Schipper, F.J. Krambeck, B.C. Choi, KINPTR [52] E.B. Nauman, Chemical Reactor Design, Optimization, and Scaleup, John Wiley
(Mobil’s kinetic reforming model): a review of Mobil’s industrial process & Sons, 2008.
modeling philosophy, Adv. Chem. Eng. 13 (1987) 193–266. [53] G.I. Marchuk, J. Ruzicka, Methods of Numerical Mathematics, Springer-Verlag,
[44] G.P. Vathi, K.K. Chaudhuri, Modelling and simulation of commercial catalytic New York, 1975.
naphtha reformers, Can. J. Chem. Eng. 75 (1997) 930–937. [54] D. Iranshahi, S. Amiri, M. Karimi, R. Rafiei, M. Jafari, M.R. Rahimpour, Modeling
[45] Q. Xiong, S. Aramideh, A. Passalacqua, S.-C. Kong, Characterizing effects of the and simulation of a novel membrane reactor in a continuous catalytic
shape of screw conveyors in gas–solid fluidized beds using advanced regenerative naphtha reformer accompanied with a detailed description of
numerical models, J. Heat Transfer 137 (2015) 061008. kinetics, Energy Fuels 27 (2013) 4048–4070.
[46] Q. Xiong, S.-C. Kong, Modeling effects of interphase transport coefficients on
biomass pyrolysis in fluidized beds, Powder Technol. 262 (2014) 96–105.