Journal of the Taiwan Institute of Chemical Engineers 76 (2017) 19

Contents lists available at ScienceDirect

Journal of the Taiwan Institute of Chemical Engineers

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

Modeling and optimization of an industrial Claus process: Thermal

and catalytic section
H. Ghahraloud, M. Farsi, M.R. Rahimpour
Department of Chemical Engineering, School of Chemical and Petroleum Engineering, Shiraz University, Shiraz, Iran

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

Article history: The main goal of this research is modeling and optimization of an industrial modied Claus process to
Received 28 September 2016 achieve maximum sulfur recovery. The modied Claus process consists thermal conversion in furnace
Revised 20 January 2017
and catalytic conversion in two series xed bed reactors. The furnace and catalytic reactors are modeled
Accepted 1 March 2017
based on the mass and energy conservation laws at steady state condition. To prove the accuracy of the
Available online 29 April 2017
developed mathematical model, simulation results of conventional process are compared with the avail-
Keyword: able plant data. Then, the optimal condition of Claus process is calculated considering sulfur recovery as
Modied Claus process the objective function using Genetic algorithm as a useful method in global optimization. The attainable
Process modeling decision variables are inlet temperature of furnace and xed bed reactors, feed distribution along the
Optimization furnace and ow rate of air in the furnace. The simulation results show that sulfur recovery is improved
Sulfur recovery about 4.63% in the optimized process compared to the conventional process. In addition, performance of
Auto-thermal reactor
auto-thermal reactor as a substitute for conventional adiabatic reactors is investigated to enhance sulfur
recovery and reducing sulfur contaminants emission in the Claus process.
2017 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

1. Introduction hydrogen sulde as one of the main environmental pollutants,

Elemental sulfur is a bright yellow crystalline solid at room
temperature. Under normal conditions, sulfur atoms form cyclic oc-
tatomic molecules with chemical formula S8 . It is a useful element
that is used to manufacture of fertilizers with other principal users
including rubber industries, cosmetics, and pharmaceuticals. The
modied Claus process is the most popular method for recovery of
elemental sulfur from hydrogen sulde in oil and natural gas re-
nery units. The modied Claus process is suitable when hydrogen
sulde concentration and feed capacity is high.
Sulfur contaminants emission and sulfur management is an im-
portant challenges from environmental and economic viewpoints.
So, a small improvement in eciency, operational cost and en-
ergy management in sulfur recovery units is worthy. Generally, the
Claus process comprises two main stages including thermal and
catalytic sections. In the rst stage, a part of inlet H2 S is burned
in a furnace at high temperature and elemental sulfur, sulfur diox-
ide and water vapor are produced. After sulfur condensation, fur-
nace euent is fed to the catalytic stage. In the catalytic stage,
un-reacted hydrogen sulde and produced sulde dioxide react
to form elemental sulfur and vapor [1,2]. Due to importance of
Corresponding author.
E-mail addresses: rahimpor@shirazu.ac.ir, Rahimpor230 0 0 0@gmail.com
(M.R. Rahimpour).
many researchers have focused on sulfur recovery from hydrogen
sulde and other sulfur contaminants. Some scholars studied ki-
netics of thermal conversion of hydrogen sulde at different tem-
peratures and residence time. Several researches have focused on
identifying catalyst, mechanism of reactions, rates, and catalyst de-
activation in the catalytic section. Generally, various catalysts such
as metal oxides were proposed for Claus reaction and decomposi-
tion of H2 S. Between metal oxides, vanadium oxide has presented
better performance and attracted more attention [3,4]. Monnery et
al. modeled the thermal section and waste heat boiler of Claus pro-
cess based on an equilibrium model [5]. Nasato et al. presented
a mathematical model to evaluate the kinetic and thermal char-
acteristics of waste heat boiler in modied Claus Sulfur Recovery
Unit [6]. The results showed that, hydrogen production could be
increased over 20% by reconguration of waste heat boiler tubes.
Pierucci et al. developed a detail kinetic scheme based on the
free radical approach to predict ame temperature and conversion
of hydrogen sulde [7]. Zarenezhad and Hosseinpour investigated
some applicable techniques to increase furnace temperature [8].
The results showed that acid gas enrichment is a reliable method
to create the desired furnace temperature, when a high ow of
lean acid gas is processed. Manenti et al. modeled furnace and
waste heat boiler based on a detailed kinetic considering 2400 re-
actions and 140 species. They optimized the process condition to
achieve maximum elemental sulfur recovery, acid gas conversion

http://dx.doi.org/10.1016/j.jtice.2017.03.005
1876-1070/ 2017 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
2 H. Ghahraloud et al. / Journal of the Taiwan Institute of Chemical Engineers 76 (2017) 19

and steam generation [9,10]. Manenti et al. revised a conventional

kinetic model of thermal section of sulfur recovery units to account Nomenclature
light hydrocarbons, ammonia and other species that there are in
the feedstock of industrial units [11]. Zarei et al. selected a detail av specic surface area of catalyst pellet (m2 /m3 )
kinetic model for thermal conversion of hydrogen sulde and ob- Ac cross section area of each tube (m2 )
tained the kinetic parameters of model [12]. Comparison between Cp specic heat of the gas at constant pressure
simulation results and plant data shows that the mean absolute (J/mol/K)
error is about 7.6%. Elsner et al. applied the mixture of catalyst Ct total concentration (kmol/m3 )
and water adsorbent in the catalytic section of Claus process as D tube diameter (m)
a reliable method to shift the reaction and overcome equilibrium dp catalyst particle diameter (m)
limitations [13]. Nabikandi and Fatemi modeled the catalytic sec- Ei activation energy for ith reaction (J/mol)
tion of Claus process based on the both equilibrium and kinetic Dij binary diffusion coecient of component i in j
approaches [14]. The simulation results showed that although both (m2 /s)
equilibrium and kinetic models present reliable results, the kinetic Dim diffusion coecient of component i in the mixture
model reveals more accuracy compared to the equilibrium model. (m2 /s)
Modeling of uidized bed reactor for Claus process showed that F total molar ow rate (mol/s)
replacing the xed bed reactor with a uidized bed decreases cat- hf gassolid heat transfer coecient in reactor
alyst loading and pressure drop [15,16]. Abdel Fattah et al. used (W/m2 /K)
computational uid dynamic tool to develop a three dimensional hi tube side heat transfer coecient in reactor
nite element model for catalytic section of Claus process [17]. (W/m2 /K)
They investigated the effect of feed temperature and composition ho shell side heat transfer coecient in reactor
on the specication of outlet product. (W/m2 /K)
Due to importance of environmental legislation and limitations, Hf,i heat of reaction for reaction i (kJ/kmol)
any decrease in sulfur contaminants emission is attractive. In this H i heat of adsorption of water on catalyst surface
study, the thermal and catalytic sections of Claus Sulfur Recov- (kJ/kmol)
ery Unit is modeled and optimal operating condition of considered ki reaction rate constant (mol/gcat /h)
process is determined to achieve maximum sulfur recovery. Then, KH2 O i water adsorption equilibrium constant in the rate
the simulation results of optimized process are compared with the equation (k/Pa)
conventional system. In the recent decades, auto-thermal reactor KE equilibrium constant (1/kPa0.5 )
concept for weakly endothermic or endothermic xed-bed reac- L reactor length (m)
tions was developed. Auto-thermal reactors consists a combination P total pressure (bar)
of a chemical reactor with a heat exchanger, where the generated Pi partial pressure of component i (kPa)
heat by the chemical reaction(s) in the reactor is used to heat up ri rate of reaction (mol/gcat /h)
the feed. In this research as the second goal, the eciency of auto- R Universal gas constant (J/mol/K)
thermal reactor is investigated as a substitute for conventional adi- Tg bulk gas phase temperature in reaction side (K)
abatic reactors in Claus process. Tc bulk gas phase temperature in shell side (K)
Ts solid phase temperature (K)
ug velocity of uid phase (m/s)
U overall heat transfer coecient (W/m2 /K)
2. Process description yi mole fraction of component i in the uid phase
Z axial reactor coordinate (m)
The Claus process was invented in 1883 by English scientist,
Carl Friedrich Claus. The Claus is the most popular sulfur recov- Greek letters
ery process to produce elemental sulfur from hydrogen sulde. The viscosity of gas phase (Pa s)
considered Claus process in this research is Split-Flow type and catalyst effectiveness factor
consists a furnace reactor and two series adiabatic reactors. The gas density (kg/m3 )
feed is divided in two streams. The rst stream is mixed with b catalytic bed density (kg/m3 )
air and fed to the furnace, while the second part of feed stream porosity factor
is injected at the middle part of furnace. In the furnace, a part
of hydrogen sulde is converted to the sulfur and sulfur dioxide.
To prevent sulfur condensation in the pipelines and heat recovery
from furnace euent, the output stream from furnace is cooled in
a boiler, and the elemental sulfur is condensed and separated from 3. Auto-thermal conguration
furnace euent. In addition, since the hydrogen sulde conver-
sion is characterized by a large number of radical reactions; there- It is appeared that removing of the generated heat from reac-
fore, it is mandatory to freeze the reactive situation to avoid re- tion zone and decreasing reactor temperature shift thermodynamic
combination reactions. Usually, about 6070% of the total amount equilibrium limitation of the reversible reactions toward the com-
of elemental sulfur is produced in the thermal section. Then, the pletion. Therefore, it concluded that a structure in form of a shell
stream is heated and fed to the rst catalytic reactor. The outlet and tube exchanger is satiable to enhance capacity of sulfur recov-
stream from rst reactor is cooled in the boiler to separate ele- ery plant. The proposed reactor is a shell and tube heat exchanger
mental sulfur and after heating is fed to the second reactor. Gener- that the tube side is packed with the catalyst slice. The feed stream
ally, the catalytic recovery of sulfur consists three sub steps includ- is fed to the shell side of the reactor, receives the generated heat
ing heating, catalytic reaction, cooling and condensation. The outlet of reaction in tube side. Thus, the feed temperature increases along
stream from second reactor is fed to the boiler to separate elemen- the shell side and reaction temperature decreases. Then, the pre-
tal sulfur and uncondensed part is sent to the separation section. heated feed is entered to the tube side and reactions occur over
Fig. 1 depicts the schematic of Claus process. the catalyst surface.
 H. Ghahraloud et al. / Journal of the Taiwan Institute of Chemical Engineers 76 (2017) 19 3

4. Reaction kinetic Table 1

(a) Furnace and reactor design specications, (b) feed specication of Claus
process.
In the thermal conversion, hydrogen sulde reacts in a sub-
stoichiometric combustion at temperatures above 850 C and ele- (a)
Parameter Value
mental sulfur is produced. Apart of un-reacted H2 S is converted
to SO2 . The following main and side reactions take place in the Furnace
thermal section: Length (m) 3.86
Injection point (m) 2.56
H2 S H2 + S2 (1)
Diameter (m) 1.1

Reactor
H2 S + 1.5O2 SO2 + H2 O (2)
Catalyst bed length (m) 0.7
First reactor catalyst volume (m3 ) 3.3
Second reactor catalyst volume (m3 ) 2.7
H2 S + 0.5SO2 H2 O + 0.75S2 (3) Bed void fraction 0.45
Inlet pressure (bar) 0.4
Inlet temperature to the rst reactor (C) 235
CH4 + 1.5O2 CO + 2H2 O (4) Inlet temperature to the second reactor (C) 210

(b)
Stream 1 Stream 2 Air
C2 H6 + 2.5O2 2CO + 3H2 O (5)
Feed composition Mole% Mole% Mole%

H2 0.287 0.161
C3 H8 + 3.5O2 3CO + 4H2 O (6) H2 S 85.234 43.029
NH3 0.018 39.775
H2 O 14.305 17.035 9.028
C4 H10 + 4.5O2 4CO + 5H2 O (7) N2 0.005 71.867
O2 19.104
CO2 0.013
CO + 0.5S2 COS (8) C1 0.079
C2 0.031
C3 0.015
CO2 + H2 CO + H2 O (9) C4+ 0.013
Total mole ow rate (kmol/h) 39.091 2.489 86.739
Total mass ow rate (kg/h) 1238.5 61.0 2417.8
Temperature (C) 68 80 114
CO + H2 S COS + H2 (10)
Pressure (kg/cm2 g) 0.714

NH3 0.5N2 + 1.5H2 (11)

CH4 + 2S2 CS2 + 2H2 S (12)

In the catalytic reactors, produced CS2 and COS in the thermal
section are hydrolyzed. The following main and side reactions take
place on the catalyst surface:
2H2 S + SO2 1/2S6 + 2H2 O (13)

COS + H2 O CO2 + H2 S (14)

CS2 + 2H2 O CO2 + 2H2 S (15)

The rst reaction is reversible and exothermic, however sec-
ond and third one are irreversible and endothermic. Generally, the
process is exothermic and temperature increases along the reactor
[14]. Industrially, the catalytic reactions take place over the surface
of alumina catalyst. The reaction scheme and rate of reactions in
the furnace and catalytic reactors are presented in Tables 1 and 2,
respectively.

5. Process modeling Fig. 1. Schematic of an industrial Claus SRU process.

5.1. Model assumptions

(r ) D UD
Since | UC B p |  | Dep |, the axial dispersion is negligible com-
In this study, a one-dimensional model is developed to simulate A
pared to mass transfer by convection in axial direction.
thermal and catalytic sections of Claus process. In the considered
Since Biot number of catalyst is smaller than 0.1, temperature
model, the following assumptions are adopted:
distribution in catalyst is negligible.
Since Re > 2100, plug ow pattern is considered. Due to low pressure and high temperature condition, the gas
Since Per  1, radial diffusion of mass and energy is negligible. mixture is an ideal gas.
4 H. Ghahraloud et al. / Journal of the Taiwan Institute of Chemical Engineers 76 (2017) 19

Table 2
Rate of thermal reactions.

Reaction no. Rate of reaction Reference

1 rH2 S = 4.3 106 exp( RT 26

)CH2 CS2 3.6 108 exp( RT 48
)CH2 S [22]
2 rH2 S = 14 exp( RT
11
)PH2 S PO12.5 [23]
3 rH2 S = 1590 0 exp( 49 RT
.9
)PH2 S PS0O.52 50 0 exp( 49RT.9 )PH2 O PS02.75 [24]
4 rCH4 = 1013.2 exp( 48RT
.4
)CC0H.74 CO0.28 [25]
5 rC2 H6 = 1.3 1012 exp( RT 30
)CC02.1H6 CO1.265 [26]
6 rC3 H8 = 1012 exp( RT
30
)CC03.1H8 CO1.265 [26]
7 rC4 H10 = 8.8 1011 exp( RT 30
)CC04.1H10 CO1.265 [26]
8 rCO = 3.78 105 exp( 6700 T
)CCOCS2 2.05 109 exp( 21630 T
)CCOSCt [27]
9 rCO2 = 3.95 1010 exp( 31220 T
)CCO2 CH0.25 [28]
10 rCO = 1.59 105 exp( 13340 T
)CCOCH0.25S [28]
11 rNH3 = 0.0042 exp( 16 RT
.5
)PN1H.253 [29]
12 rCH4 = 5.53 1010 exp( 19320 T
)CCH4 CS2 [30]

Table 3
Rate of catalytic reactions. 
N  
Reaction no. Rate of reaction Reference
a v h f ( T g T s ) + b ri H f,i = 0 (21)
i=1
PH O P 0.25
2 S6
ks (PH2 S PS0O.5 )
rH2 S =
KE
13 2
[15] Pressure drop along the bed is calculated by Ergun momentum
(1+KH2 O PH2 O )2
kCOS PCOS PH2 O
14 rCOS = (1+KH2 O PH2 O ) [31] balance:
(1 )u2g
kCS2 PCS2 PH2 O
15 rCS2 = (1+KH2 O PH2 O ) [32] dP (1 )2 ug
ki = k0i exp( RT Ei
) = 150 + 1 . 75 (22)
KH2 O i = KH2 O 0i exp( RTHi )
dz 3 d2p 3 d p
4
KE = 9.502 107 exp( 1.11T10 )
5.4. Auto thermal reactor model
No. k0i KH2 O 0i Ei Hi
13 6.91 0.338 30.77 The mass and energy balances for uid phase in the reaction
14 19.75 3.43 40.41 98.10 zone of auto thermal reactor are as follows:
15 2.30 1.25 25.27 83.22
F dyi
g
 
+ avCt kgi ysi ygi = 0 (23)
Ac dz
The specication of furnace, catalytic reactors and feed of an
F g dT g Di
industrial case are tabulated in Table 3(a) and (b). C + av h f ( T s T g ) U (T g T c ) = 0 (24)
Ac p dz Ac
5.2. Furnace model The mass and energy balance equations in the shell side are:
g
F dyi
To develop the mass and energy balance equations, a differen- =0 (25)
tial element along the axial direction of reactor is considered and Ac dz
governing equations are written over the element. The mass and
F g dT c Di
energy balance equations in the furnace are as follows: C + U (T g T c ) = 0 (26)
Ac p dz Ac
F dygi 
N
+ ri = 0 = 0 (16) It is mentioned the mass and energy balance over the catalyst
Ac dz particle for adiabatic and auto thermal reactors are the same.
i=1

F g dT g   
N 5.5. Auxiliary equations
Cp + ri H f,i = 0 (17)
Ac dz
i=1 In the heterogeneous models, the physical properties of chemi-
cal species and overall mass and heat transfer coecients between
5.3. Adiabatic reactor model phases must be estimated. The overall mass transfer coecient be-
tween solid and gas phase in the catalytic section has been ob-
In the heterogeneous models, to account the external mass and tained from the proposed correlation by Smith [18].
energy resistances, concentration and temperature difference be-
tween gas and solid phases are considered. The mass and energy kgi = 1.17Re0.42 Sci0.67 ug 103 (27)
balances for uid phase are as follows: Mass transfer diffusion coecient for each component in the
F dygi  g
 mixture is estimated by Wilke [19]:
+ avCt kgi ysi yi = 0 (18) 1 yi
Ac dz Dim =  (28)
yi
i= j Di j
F g dT g
Di
C + av h f ( T s T g ) U (T g T c ) = 0 (19) Binary mass diffusion coecient of components is calculated by
Ac p dz Ac
the following equation [20]:
Over the solid phase the mass and energy balance could be 
written as: 107 T 3/2 1 /Mi + 1 /M j
  Di j =  2 (29)
3/2
avCt kgi ygi yi + ri b = 0
s
(20) P v ci
+ v3c /j 2
 H. Ghahraloud et al. / Journal of the Taiwan Institute of Chemical Engineers 76 (2017) 19 5

The overall heat transfer coecient between solid and uid Table 4
Comparison between simulation results and plant data.
phase (hf ) is predicted from Cussler [21]:
 C 2 / 3
0.407 Simulation results Plant data Relative error
hf P 0.458 ud p
= (30) Thermal section
CP K B
Temperature (C) 1266 1227 3.17
The overall heat capacity in Eq. (30) is calculated as following: H2 S 7.64 7.664 0.31
SO2 3.32 3.308 0.36

N Component ow S2 11.7 11.697 0.02
CP = yi CPi (31) rate (kmol/h) H2 2.033 2.06 1.31
i=1 CO 0.0731 0.075 2.53
COS 0.0228 0.023 0.87
6. Process optimization Catalytic section

Temperature (C) 229.7 228 0.74

6.1. Genetic algorithm
H2 S 1.437 1.43 0.48
Component ow SO2 0.1839 0.18 2.12
The gradient based optimization techniques suffer from trap- rate (kmol/h) CS2 0.004063 0.004 1.55
ping at local optimum and considered initial guess has a signicant COS 0.0 0 0148 0.0 0 015 1.07
effect on the quality of obtained optimal solution. These techniques
do not ensure global optimum in case of nonlinear and complex
problems. Genetic algorithm is a particular class of evolutionary Table 5
optimization algorithms that has been based on biological rules in Comparison between simulation results of optimized and conventional pro-
cess.
the world such as mutation, selection and crossover. It is started
by generating an initial population set, which contains a set of ran- Decision variables Optimal Conventional
domized strings of genes and covers the problem domain. The size Inlet temperature of furnace (C) 481 Acid gas 70
of this population remains constant along the procedure. At every Air 114
generation, the individuals in the considered population are eval- Inlet temperature of rst reactor (C) 198.1 235
Inlet temperature of second reactor (C) 188.8 210
uated according to tness function. To develop a new population
Feed distribution ratio 0 0.3
in the next generation, some of the best individuals are selected Air ow rate (kmol/h) 82.45 86.74
according to their tness. To create new different individuals, the
previous population is updated by crossover and mutation mech-
anisms. Crossover concerns two selected individuals (parents) to
form two new individuals (children). Mutation consists in ipping
software. In this section, the reliability and accuracy of developed
bits of individuals strings at random. In practice, over successive
mathematical model of Claus process is validated by comparing
generations, the population moves toward an optimal solution in
the simulation results with the available plant data in a domes-
the considered feasible domain.
tic plant. As represented in Table 4, the simulation results have a
6.2. Optimization problem formulation good agreement with the available plant data at the same process
condition.
From practical viewpoint, process optimization is dened as In this section, simulation results of the conventional and opti-
nding the best operational condition of a process considering mized processes are analyzed and the predicted molar ow rates
bounds and constraints. Process optimization is the discipline of and temperature proles are presented. Table 5 presents values of
adjusting a process to operate at optimal condition by a specied decision variables at optimized and conventional condition. It is
set of input parameters without violating constraints. In this re- appeared that, although a part feed stream is injected to the mid-
search, the optimal condition of Claus process is calculated consid- dle part of furnace at the conventional condition; the total feed
ering sulfur recovery as the objective function using Genetic algo- stream should be injected to the inlet of furnace in the optimized
system. The presented data show that injection a part of feed
rithm. The considered decision variables are inlet temperature of
furnace and reactors, feed distribution and ow rate of air in the stream at middle part of furnace has a negative effect on the sul-
furnace. To ensure that the feed temperature of catalytic section fur recovery. Indeed, the simulation results prove the performance
of Straight-Through Process over Split Flow in the considered in-
is not too low for the hydrolysis and Claus reaction to occur, the
lower bound on temperature of feed is set at 150 C. dustrial unit. Generally, applying the obtained optimal values for
decision variables on the conventional Claus process increases the
7. Numerical solution sulfur recovery about 1.07%. It is appeared that sulfur recovery at
outlet of furnace and catalytic sections are 68.03% and 95.3% at
A set of nonlinear ordinary differential equations are obtained conventional condition, respectively. While, sulfur recovery at out-
by combination of mass and energy governing equations with the let of furnace and catalytic sections are 72.65% and 96.32% at op-
kinetic expressions and auxiliary correlations. This set of equations timized system, respectively. The conversion of H2 S in the opti-
is solved with 4thorder RungeKutta method. mized system increases about 1.1% compared to the conventional
process. Fig. 2 shows the sulfur recovery prole along the furnace
8. Results and discussions and catalytic reactors at conventional and optimized condition. It
is appeared that about 70% hydrogen sulde is converted to the
8.1. Optimization of conventional process elemental sulfur in the thermal section. In optimized system, high
feed temperature increases rate of reactions and the equilibrium
Generally, the process simulators are applicable tools in design, is reached immediately at the inlet of the furnace. While, in the
development, analysis, and optimization of chemical processes, en- conventional unit due to lower feed temperature, equilibrium con-
vironmental systems, power stations, biological processes, and sim- dition is established in a longer length after the inlet. Sulfur re-
ilar technical functions. Indeed, process simulation is a model- covery increases along the catalytic beds and approach toward the
based representation of chemical processes and unit operations in equilibrium at the last part of the catalytic section. Condensation
6 H. Ghahraloud et al. / Journal of the Taiwan Institute of Chemical Engineers 76 (2017) 19
Fig. 2. Sulfur recovery prole along the furnace and catalytic beds.
of elemental sulfur in the outlet stream from furnace and rst re-
actor shifts reactions toward the higher sulfur production in the
next steps.
Fig. 3 shows the proles of hydrogen sulde and sulfur diox-
ide along the furnace and xed bed reactors at optimal and con-
ventional condition. In the furnace, hydrogen sulde and oxygen
reacts and sulfur dioxide is produced. Thus, hydrogen sulde con-
centration decreases, while sulfur dioxide increases along the fur-
nace. Since a part of hydrogen sulde is injected at middle part
of furnace in the conventional process, hydrogen sulde decreases
at the rst part of furnace, and increases sharply at injection point
and then decreases smoothly. The stoichiometry coecient of reac-
tants in the main catalytic reaction shows that the optimum value
of sulfur dioxide to hydrogen sulde is two. It is appeared that sul-
Fig. 3. (a) H2 S and (b) SO2 prole along the furnace and catalytic beds.
fur dioxide to hydrogen sulde ratio at the inlet of catalytic section
is 2.3 and 2.02 at conventional and optimal condition, respectively.
Fig. 4 (a) and (b) depicts the temperature prole along the
furnace and catalytic reactors. Since the net of reactions heat is
exothermic, thus temperature increases along the furnace and cat-
alytic reactors. Since hydrogen sulde conversion as the main re-
action is reversible and exothermic, temperature has a signicant
effect on the equilibrium conversion in the catalytic section. Gen-
erally, decreasing temperature can shift equilibrium conversion of
main reaction toward the higher H2 S and SO2 conversion. On the
other hand, it decreases rate of reaction and applying the larger
reactor to convert reactant is necessary. In addition, since hydroly-
sis reactions are non-reversible, decreasing temperature drops rate
of CS2 and COS conversion. Thus, there is an optimal temperature
prole to improve sulfur recovery in the Claus process. It is ap-
peared that applying the obtained optimal temperatures on the
system increase sulfur recovery signicantly.
Fig. 5 shows the COS prole along the conventional and opti-
mized systems. In the Claus Sulfur Recovery Unit, carbon monox-
ide, methane and sulfurous components are converted to COS and
CS2 at high temperature. It is appeared that one of the main prob-
lems in the conventional system is COS production at the second
half of furnace after injection point. Based on the simulation
results, although COS concentration increases along the furnace
length, the produced COS and CS2 in the thermal section are hy-
drolyzed to the carbon dioxide and hydrogen sulde in the cat-
alytic section. It is appeared that applying the obtained optimal
operating condition on the system decreases COS content in the
outlet of second catalytic reactor from 0.004 to 0.002.
Fig. 6 shows prole of hydrogen molar ow rate along the con-
ventional and optimized systems. Generally, higher temperature
and lower oxygen injection in the furnace increase rate of hydro-
gen sulde decomposition, so the higher hydrogen is produced at
the optimal condition. The hydrogen production capacity at the
optimal and conventional condition is about 4.66 and 2.03, re-
spectively. The results show that hydrogen production is improved
about 129.5% in the optimized system. Thus, operating at higher
temperature increases elemental sulfur and hydrogen production
from hydrogen sulde decomposition.
 H. Ghahraloud et al. / Journal of the Taiwan Institute of Chemical Engineers 76 (2017) 19
Fig. 4. (a) Temperature prole along the furnace, (b) and catalytic reactors.
Fig. 5. COS prole along the furnace and catalytic beds.
8.2. Auto-thermal reactor
In this section, the obtained results of the auto-thermal reactor
are presented at optimal condition. Temperature prole, as one of
the most important parameters, is illustrated along the adiabatic
and auto thermal reactors in Fig. 7. In the auto-thermal reactor,
the feed is entered to the shell side and ows along the shell. It
receives heat from the tube side and feed temperature increases.
After that feed stream is entered to the tube side and sulfur con-
taminants are converted to the elemental sulfur over the catalyst.
In the tube side before maximum temperature point, due to higher
heat generation compared to the heat transfer between shell side
and reaction zone, temperature increases continuously. After the
maximum temperature point, the rate of heat transfer is domi-
nated over the heat generation, and temperature decreases along
the tube side. According to Le Chteliers principle, when an inde-
pendent variable of an equilibrium system changes the equilibrium
7
Fig. 6. H2 prole along the furnace and catalytic beds.
shifts in the direction that tends to reduce effect of the change.
Since the H2 S conversion is a reversible and exothermic reaction,
heat removal from reaction zone shifts H2 S conversion reaction to-
ward right side sand higher H2 S is converted to elemental sulfur.
Thus, higher temperature at the rst half and lower temperature
at the second half of reactor are the main advantages of auto ther-
mal reactor over the adiabatic reactor that has enhanced produc-
tion rate.
Table 6 presents a comparison between output product from
adiabatic and auto thermal reactors at same feed condition. The
lower molar ow rate of pollutants at the outlet of auto-thermal
reactor proves the performance of auto-thermal over the adiabatic
reactor in the Claus process. Hydrogen sulde mole ow rate in the
output of conventional and auto-thermal reactors is about 2.7 and
2.46 kmol/h, respectively. Generally, auto thermal reactor is more
effective to convert COS and CS2 compared to the adiabatic reactor.
Since COS and CS2 conversion reactions are non-reversible, higher
8 H. Ghahraloud et al. / Journal of the Taiwan Institute of Chemical Engineers 76 (2017) 19

9. Conclusion

In this research, the Claus Sulfur Recovery Unit was modeled

based on the mass and energy conservation laws at steady state
condition. The process consists two main parts, thermal conversion
in furnace and catalytic conversion in two series adiabatic xed
bed reactor. Comparison between simulation results and plant data
proved the accuracy of the developed mathematical model. In or-
der to maximize sulfur recovery, a single objective optimization
problem was formulated considering inlet temperature of furnace
and reactors, feed distribution and ow rate of air in the furnace as
decision variables. The simulation results showed that sulfur recov-
ery is improved about 1.07% in the optimized process compared to
the conventional. In addition, higher temperature and lower oxy-
gen injection in the furnace increase rate of hydrogen sulde de-
composition, so hydrogen production is improved about 129.5% at
the optimal system. The simulation results proved the performance
of Straight-Through Process over Split Flow Process in the consid-
ered domestic unit. Generally, applying obtained optimal values on
the system increases production capacity and decreases sulfur con-
taminant emission. In addition, the results showed that produced
Fig. 7. Temperature prole along the conventional and autothermal reactor. elemental sulfur in the auto-thermal reactor is about 1.54 kmol/h
that is improved about 5.18% compared to the conventional reactor.
Thus, the auto-thermal reactor is a good candidate as substitute for
Table 6
conventional adiabatic reactor in the Claus process to decrease sul-
Comparison between output product from adiabatic and auto thermal
reactors. fur contaminants emission.

Capacity Conventional Auto thermal References

(kmol/h) reactor reactor Improvement %

H2 S 2.76 2.46 10.87 [1] Goar B. Sulfur recovery technology. New York, NY: American Institute of Chem-
ical Engineers; 1986.
SO2 0.84 0.69 17.86
[2] Clark PD. Sulfur and hydrogen sulde recovery. Kirk-Othmer Encyclopedia of
CS2 0.0 0 01807 0.0 0 01662 8.02
Chemical Technology. John Wiley & Sons, Inc.; 20 0 0.
COS 0.004064 0.003776 7.09
[3] Kaloidas VE, Papayannakos NG. Kinetic studies on the catalytic decomposition
of hydrogen sulde in a tubular reactor. Ind Eng Chem Res 1991;30(2):34551.
[4] Davydov AA, Marshneva VI, Shepotko ML. Metal oxides in hydrogen sulde ox-
idation by oxygen and sulfur dioxide: I. The comparison study of the catalytic
activity. Mechanism of the interactions between H2 S and SO2 on some oxides.
Appl Catal A Gen 20 03;244(1):9310 0.
[5] Monnery WD, Svrcek WY, Behie LA. Modelling the modied claus process re-
action furnace and the implications on plant design and recovery. Can J Chem
Eng 1993;71(5):71124.
[6] Nasato LV, Karan K, Mehrotra AK, Behie LA. Modeling reaction quench times in
the waste heat boiler of a Claus plant. Ind Eng Chem Res 1994;33(1):713.
[7] Pierucci S, Ranzi E, Molinari L. Modeling a Claus process reaction furnace via a
radical kinetic scheme. Comput Aided Chem Eng 2004;18:4638.
[8] Zarenezhad B, Hosseinpour N. Evaluation of different alternatives for increas-
ing the reaction furnace temperature of Claus SRU by chemical equilibrium
calculations. Appl Therm Eng 2008;28(7):73844.
[9] Manenti F, Papasidero D, Frassoldati A, Bozzano G, Pierucci S, Ranzi E. Multi-
-scale modeling of Claus thermal furnace and waste heat boiler using detailed
kinetics. Comput Chem Eng 2013;59:21925.
[10] Manenti F, Papasidero D, Bozzano G, Ranzi E. Model-based optimization of sul-
fur recovery units. Comput Chem Eng 2014;66:24451.
[11] Manenti F, Papasidero D, Ranzi E. Revised kinetic scheme for thermal furnace
of sulfur recovery units. Chem Eng Trans 2013;32:1185290.
[12] Zarei S, Ganji H, Sadi M, Rashidzadeh M. Kinetic modeling and optimization of
Claus reaction furnace. J Nat Gas Sci Eng 2016;31:74757.
[13] Elsner MP, Menge M, Mller C, Agar DW. The Claus process: teaching an old
dog new tricks. Catal Today 2003;79:48794.
[14] Nabikandi NJ, Fatemi S. Kinetic modelling of a commercial sulfur recovery unit
based on Claus straight through process: comparison with equilibrium model.
J Ind Eng Chem 2015;30:5063.
[15] Birkholz RK, Behie LA, Lana IGD. Kinetic modelling of a uidized bed Claus
plant. Can J Chem Eng 1987;65(5):77884.
Fig. 8. Mole ow rate of S6 along the conventional and autothermal reactor.
[16] Puchyr DM, Mehrotra AK, Behie LA, Kalogerakis N. Hydrodynamic and kinetic
modelling of circulating uidized bed reactors applied to a modied Claus
plant. Chem Eng Sci 1996;51(24):525162.
[17] Abdel-Fattah A-S, Fateen S-EK, Moustafa TM, Fouad MM. Three-dimensional
CFD simulation of industrial Claus reactors. Chem Eng Res Des 2016;112:7887.
temperature at the rst half part of auto-thermal results in the
[18] Smith JM. Chemical engineering kinetics. McGraw-Hill; 1981.
higher conversion compared to the adiabatic reactor. Fig. 8 shows [19] Wilke C. Estimation of liquid diffusion coecients. Chem Eng Progr
the molar ow rate of produced sulfur along the auto thermal and 1949;45(3):21824.
conventional adiabatic reactors. The produced elemental sulfur in [20] Robert CR, John MP, Bruce EP. The properties of gases and liquids. McGraw-
Hill, New York. 1987.
the auto thermal is about 1.54 kmol/h that is improved about 5.18% [21] Cussler EL. Diffusion: mass transfer in uid systems. Cambridge University
compared to the conventional reactor. Press; 2009.
 H. Ghahraloud et al. / Journal of the Taiwan Institute of Chemical Engineers 76 (2017) 19
[22] Dowling NI, Clark PD. Kinetic modeling of the reaction between hydrogen and
sulfur and opposing H2 S decomposition at high temperatures. Ind Eng Chem
Res 1999;38(4):136975.
[23] Hawboldt KA. Kinetic modelling of key reactions in the modied Claus plant
front end furnace. 1998.
[24] Monnery W, Hawboldt K, Pollock A, Svrcek W. New experimental data and ki-
netic rate expression for the Claus reaction. Chem Eng Sci 20 0 0;55(21):51418.
[25] Dryer F, Glassman I, editors. High-temperature oxidation of CO and CH4 . Sym-
posium (international) on combustion; 1973: Elsevier
[26] Westbrook CK, Dryer FL. Simplied reaction mechanisms for the oxidation of
hydrocarbon fuels in ames. Combust Sci Technol 1981;27(1-2):3143.
[27] Karan K, Mehrotra AK, Behie LA. COS-forming reaction between CO and
sulfur: a high-temperature intrinsic kinetics study. Ind Eng Chem Res
1998;37(12):460916.
9
[28] Karan K, Mehrotra AK, Behie LA. A high-temperature experimental and mod-
eling study of homogeneous gas-phase COS reactions applied to Claus plants.
Chem Eng Sci 1999;54(15):29993006.
[29] Monnery W, Hawboldt K, Pollock A, Svrcek W. Ammonia pyrolysis and oxida-
tion in the Claus furnace. Ind Eng Chem Res 2001;40(1):14451.
[30] Karan K, Behie LA. CS2 formation in the Claus reaction furnace: a kinetic study
of methanesulfur and methanehydrogen sulde reactions. Ind Eng Chem Res
2004;43(13):330413.
[31] Tong S, Dalla Lana I, Chuang K. Kinetic modelling of the hydrolysis of
carbonyl sulde catalyzed by either titania or alumina. Can J Chem Eng
1993;71(3):392400.
[32] Tong S, Dalla Lana I, Chuang K. Kinetic modeling of the hydrolysis of
carbon disulde catalyzed by either titania or alumina. Can J Chem Eng
1995;73(2):2207.