Journal of Natural Gas Science and Engineering 38 (2017) 235e244

Contents lists available at ScienceDirect

Journal of Natural Gas Science and Engineering

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

Modeling and multi-optimization of thermal section of Claus process

based on kinetic model
H. Kazempour, F. Pourfayaz*, M. Mehrpooya
Department of Renewable Energies and Environment, Faculty of New Sciences and Technologies, University of Tehran, POB 14395-1561, Tehran, Iran

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

Article history: The Claus process consists of two basic stages: thermal and catalytic sections. In this study, modeling by a
Received 17 April 2016 kinetic model and multi-objective optimization of the thermal section of Claus process were described.
Received in revised form The industrial data of the South Pars Renery in Asaluyeh, Iran was used to validate this model. In order
24 November 2016
to investigate the inuences of the inlet ow rates of fuel and air, inlet stream temperature, furnace
Accepted 25 December 2016
pressure and waste heat boiler (WHB) outlet temperature on the sulfur recovery efciency, the steam
Available online 29 December 2016
production and the H2S/SO2 ratio, a sensitivity analysis was done by simulator software. Three objects of
the sulfur recovery efciency, the steam production and the H2S/SO2 ratio were optimized by using the
Keywords:
Sulfur recovery
software and a multi-optimization approach based on the response surface methodology. The results
Claus showed that the decrease of the sulfur recovery efciency from 0.6129 to 0.6099 leads to the addition of
Furnace reactor 8.54 Kg mole/h to the medium pressure steam production capacity and more closeness of the H2S/SO2
Kinetic model ratio to number 2 for better performance of the catalytic section. However, the 66% improvement in the
H2S/SO2 ratio leads to increase the conversion of H2S in the catalytic section, compensating the decrease
of the sulfur recovery efciency in the thermal section. Moreover, the total fuel consumption was
reduced about 0.6843 kg mol/h.
2016 Elsevier B.V. All rights reserved.

1. Introduction other functions such as destroying of any contaminates that could

Natural gas may consist of toxic hydrogen sulde. It can be up to
30 percent (Dalrymple et al., 1991). Approximately 25% of the
natural gas being brought into production from new sources, re-
quires H2S removal and disposal (Kidnay et al., 2011). Consequently,
it would cause air pollutant near the hydrocarbon industries. Sulfur
recovery unit (SRU) is an important unit for environmental regu-
lations preventing the emission of acid gas in the air. This is a unit
for removing sulfur from natural gas to produce elemental sulfur.
Modied Claus process is the most common process that is used
around the world in hydrocarbon processing industry. It has two
sections: 1-thermal section 2-catalytic section. The thermal section
is more important because about 60% of the conversion takes places
in this part. Typically a thermal section of SRU is composed of a
burner, thermal reactor (TR) and waste heat boiler (WHB). These
three parts appear in the simulator in three plug reactors: ame
zone, anoxic zone and WHB (Zarei et al., 2016). This section has
* Corresponding author.
E-mail address: pourfayaz@ut.ac.ir (F. Pourfayaz).
damage or cause fouling in the equipment, as well as sulfur pro-
duction by thermal decomposition, and steam production in WHB.
In the catalytic section, the elemental sulfur is produced during
Claus reaction (Equation (3)) which includes the reaction between
SO2 produced in thermal section and unreacted H2S. The catalytic
section consists of two or three catalytic reactors placed down-
stream of the thermal section. The performance of this section
related to outlet H2S/SO2 ratio from the thermal section. Based on
Claus reaction the mentioned parameter should be equal to stoi-
chiometric value. More closeness of the H2S/SO2 ratio to number 2
leads to better performance and more sulfur recovery efciency for
the catalytic section.
In this work, the thermal section of Claus process was modeled
using a kinetic model. To study the effects of 5 parameters on three
objects, the sensitivity analysis was performed by the simulator
software. These parameters were: The inlet ow rates of fuel and
air, inlet stream temperature, furnace pressure and WHB outlet
temperature. Three objects also were: The sulfur recovery ef-
ciency, the steam production and the H2S/SO2 ratio. Finally, multi-
objectives optimization was done using the software and an opti-
mization approach based on the response surface methodology.

http://dx.doi.org/10.1016/j.jngse.2016.12.038
1875-5100/ 2016 Elsevier B.V. All rights reserved.
236 H. Kazempour et al. / Journal of Natural Gas Science and Engineering 38 (2017) 235e244

2. Process description
SRU operates based on reaction (1), but in modied Claus pro-
cess it is divided into two reactions and contains two sections, the
rst is the thermal section that has a furnace reactor and WHB that
oxidizes the 1/3 H2S and produce SO2 via the reaction (2) (GPSA,
2004). In this reaction, the oxygen is limiting component. More-
over, it is an important safety notice that ensuring no oxygen slips
to anoxic zone because of the possibility of a ashback in the
bypassed acid gas (Kohl and Nielsen, 1997). By dropping the tem-
perature, other reactions occurring in the thermal section such as
the formation of H2 and CO or COS and CS2 based on the reactions in
Table 1.
1 reaction in a plug ow reactor has been studied by Karan et al. in
H2 S O2 /S H2 O; DH@298 Kz  66672 kCal
2 2004 (Karan and Behie, 2004). Howbalt et al. in 2000 found that the
3 kinetic equation (Hawboldt et al., 2000). The furnace of Claus
H2 S O / SO2 H2 O; DH@298 Kz  56258 kCal
2 2
1 3
H2 S SO2 4H2 O Sx ; DH@298 Kz  10414 kCal
2 x
H2S burning in the rst stage occurs at high temperature.
Temperature is usually in the range of 1100e1400  C (Manenti
et al., 2014) that has potential for heat recovery and steam pro-
duction for any usage. Downstream of WHB is the second section
consisting of two or three converters (catalytic reactors). Reaction
number 3 is the main reaction in these converters. The rst con-
verter hydrolyzes COS and CS2 that may damage catalysts of this
section. Production of COS and CS2 takes place in the thermal
section, but hydrolysis of these components occurs in the catalytic
section of the process. The temperature of the outlet stream from
the furnace and each converter should be decreased down to below
the dew point of sulfur. In the furnace during burning H2S, other
reactions occurring in the thermal section are collected in Table 1.
Kinetics parameter mentioned in Table 2.
Fig. 1 shows a typical thermal section of Claus process. The exit
acid gas from gas treating unit is mixed with air in the burner and
oxidized. WHB coupled to furnace cools the outlet stream from the
furnace and then condenser should decrease the temperature to
form liquid elemental sulfur. The liquid sulfur collected in a tank
and the gasses stream is reheated and fed to the catalytic section.
The preheating prevents liquid sulfur formation and catalyst
deactivation. It also enhances the progression of the COS and CS2
hydrolysis in the catalytic section. Tong et al. in 1997 found that the
sulfur liquids may diffuse in pores of the catalyst bed and reduce
active sites on the catalyst surface (Tong et al., 1997).
Howbalt in 1998 proposed kinetic based on low power equation
for H2S oxidation (Hawboldt, 1998). Oxidation of methane leads to a
sustainable ame and the increase of temperature. Burning one-
third of H2S is controlled by inlet oxygen ow and in absence of
oxygen, the excess CH4 take part in CS2 formation. Kinetic of this
(1)
reverse reaction of H2S cracking has a low rate and presented the
(2)
process is an important source for hydrogen produced from H2S
pyrolysis and this component is stable in another part of the pro-
cess. But during the process, the reaction between H2 and CO2 leads
to produce CO. The carbon monoxide is a main carbon source for
the carbon and carbonyl sulde. Karan et al. in 1998 studied the
(3)
formation of COS in the range of 600e1150  C in a plug ow reactor
(Karan et al., 1998). Turns et al. in 2000 found the mechanism of the
carbon formation in the furnace (Turns et al., 2000). The presence
of CO and H2 causes carbon formation in a reverse reaction (Davis
et al., 2005). The efciency of sulfur recovery in the furnace is
based on two reactions: H2S pyrolysis and Claus reaction that takes
place in furnace. After SO2 production in ame zone, the reaction
between H2S and SO2 leads to produce water and elemental sulfur.
First, Kaloidas et al. in 1989 worked on the non-catalytic decom-
position of hydrogen sulde (Kaloidas and Papayannakos, 1989)
lastly Monnery et al. in 2000 by improvement this work established
a kinetic for the reaction (Monnery et al., 2000).
3. Modeling
Table 3 shows component molar ow and properties of SRU feed

Table 1
Reactions and kinetic equations for considered reactions in furnace.

No Reaction Kinetic Ref.


1 H2 S 32O2 /SO2 H2 O E (Hawboldt, 1998)
r 2 S Af exp RTaf PH2 S PO2 1:5


2 NH3 34O2 /32H2 O 12N2 (Monnery et al., 2001)
rN 3 A exp E RT PNH3
a 1:25


3 2CH4 2O2 /CO2 2H2 O dCH4  (Turns et al., 2000)
dt
A exp E 0:3 1:3
RT CCH4 CO2
a


4 H2 12O2 /H2 O (Peters, 1979)
r A exp E RT CH2 CO2
a


5 CO 12O2 /CO2 (Westbrook and Dryer, 1981)
r A exp E 0:25
RT CO2 CCO CH2 O
a 0:5



6 H2 S412S2 H2 E (Hawboldt et al., 2000)
rH2 S Af exp RTaf PH2 S PS2 0:5  Ar exp E RT
ar
PH2 PS2



7 2H2 S SO2 412S2 2H2 O E (Monnery et al., 2000)
rH2 S Af exp RTaf PH2 S PSO2 0:5  Ar exp E RT
ar
PH2 O PS2


8 CO2 H2 /CO H2 O Ea (Karan et al., 1999)
r A exp RT CH0:5 2
CCO2



9 CO H2 /C H2 O E (Davis et al., 2005)
rH2 Af exp RTaf CCO CH2 Ar exp E RT
ar
CC CH2 O


10 C O2 4CO2 Ea (Turns et al., 2000)
rC Af exp RT CC CO2



11 CO 12S2 4COS E (Karan et al., 1998)
rCOS Af exp RTaf CCO CS2  Ar exp E RT
ar
CCOS Ct


12 CH4 2S2 /CS2 2H2 S (Karan and Behie, 2004)
rCS2 A exp E RT CCH4 CCS2
a
 H. Kazempour et al. / Journal of Natural Gas Science and Engineering 38 (2017) 235e244 237

Table 2
Constants of kinetic equations.

No Af Eaf kcal=kgmole Ar Eaf kcal=kgmole Ref

1 14gmole=cm3 :s:atm1:5 11 (Hawboldt, 1998)

2 0.00421gmole=cm3 :s:atm1:5 16.5 (Monnery et al., 2001)
3 1.3108 1= 24,358 (Turns et al., 2000)
s
4 1.08106 mole=cm3 s 30 (Peters, 1979)
5 3.9810 gmole=cm3 s
14 40 (Westbrook and Dryer, 1981)
6 526 gmole=cm3 :s:atm1:5 45 14gmole=cm3 :s:atm2 23.4 (Hawboldt et al., 2000)
7 10 gmole=cm3 :s:atm1:5 10.6 0:5gmole=cm3 :s:atm 5.1 (Monnery et al., 2000)
8 3.951010 m3 =kgmole0:5 =s 7462 (Karan et al., 1999)
9 1900 m3 =kgmole:s 21.51 1  109 m3 =kgmole:s 2.39 1010 (Davis et al., 2005)
10 5109 m3 =kgmole:s 0.239 (Turns et al., 2000)
11 3.18 105 m3 =kgmole:s 13,314 2:18  109 m3 =kgmole:s 42,980 (Karan et al., 1998)
12 5.53 1010 m3 =Kgmole:s 0.038 (Karan and Behie, 2004)

Fig. 1. Typical thermal section of sulfur recovery unit.

and the air inlet to the furnace, reported from the South Pars Re- the waste heat boiler (WHB) (Monnery et al., 1993). Also Jones et al.
nery in Asaluyeh, Iran. in 2012 considered the reaction in WHB for a Claus process of an
The rst step of simulation contains the reactions occurring in integrated gasication combined cycle (Jones et al., 2012) and
all of the parts each one with its kinetics. Earlier reactions in the Manenti et al. in 2012 considered recombination reactions in the
ame zone have power low kinetics and very low residence time in boiler in modeling SRU thermal reactor including a detailed kinetic
ame zone, and then the reactions that have the equilibrium ki- scheme (Manenti et al., 2012). Generally, all the reactions have been
netics, take place in the anoxic zone. considered in the furnace and WHB.
Furnace modeled by the kinetic model as a plug ow reactor. In this study, a horizontal furnace was simulated with following
Monnery et al. in 1993 found that results of equilibrium calcula- assumptions:
tions do not match the industrial data taken both before and after
 Radiant dispersion is neglected
 The unit is in the steady state operation condition
 In adiabatic condition is assumed in this case insulation of the
Table 3
Compositions and properties of feed & air inlet to the furnace, reported from the furnace
South Pars Renery in Asaluyeh, Iran.  Unit is simulated by ideal gas rule because of high temperature
and low pressure in furnace
Acid gas Air
 The ow is assumed to be fully developed both in TR and WHB,
Molar ow (Kgmole/h) 732.6 728.1 with at proles across the sections
Temperature ( C) 240 240
Pressure (kPa) 177 1.82
 Axial dispersion is ignored in the turbulent ow. It is obtained
Composition (gmole fraction) from Pe number which is the ratio of heat or mass transfer in
H2S 0.4616 0.0000 bulk stream to diffusion. In this case, the range of Pe is shown in
CH4 0.0018 0.0000 the inequality (4)
SO2 0.0000 0.0000
N2 0.0001 0.7586
O2 0.0000 0.2011
CO 0.0000 0.0000
1
CO2 0.4989 0.0003 0:020Pe[500 (4)
COS 0.0000 0.0000 Pe
CS2 0.0000 0.0000
S2 0.0000 0.0000
S8 0.0000 0.0000  Fouling in the WHB tubes is not considered
H2O 0.0375 0.0400
H2 0.0000 0.0000 Sulfur recovery efciency is calculated based on following
NH3 0.0001 0.0000
equation
238 H. Kazempour et al. / Journal of Natural Gas Science and Engineering 38 (2017) 235e244

elemental sulfur Mole percent of S6 4:862  104 T 2  6:854  101 T  192

h1 (5)
The output of sulfur compounds of condensor
(7)
Operating pressure and pressure drop are about 170 and 7 kPa,
respectively. Reactions kinetics occurring in the furnace collected in
Mole percent of S8 3:431  105 T 2  2:086  101 T 172:1
Table 1. Reactions 1 to 5 with power low kinetics take place in the
ame zone that is 0.1e0.2 m of the rst furnace. Other reactions (8)
take place in the rest of the furnace reactor. Mass balance for each component j in the presence of a chemical
All the kinetics mentioned were presented elemental sulfur as reaction is as bellow:
S2. However, elemental sulfur has other allotropes such as S1 to 8.
Gargurevich in 2005 presented a plot of the concentration of sulfur Z
dNj
allotropes at various temperatures in the furnace (Gargurevich, Fj0  Fj rj dv (9)
dt
2005). At high temperatures, elemental sulfur exists in S2 form
while by decreasing the temperature at end of furnace and WHB Energy enters into the reactor through the bulk uid and leaves
other species appear. S6 and S8 are more stable as compared with through output bulk uid. Heat transfer from the walls of the
other species. Generally, S2, S6, and S8 are considered as dominant reactor, and heat released or absorbed by the reaction is considered.
Allotropes in the furnace. Distribution of selected sulfur allotropes
are showed in Fig. 2 that reported by Gamson et al. in 1953 b c d
A B/ C D (10)
(Gamson and Elkins, 1953). They tted the mole percents of S2, S6, a a a
and S8 in the following equations. R-square was 0.9894 and the Heat of the reaction per mole of the limiting component A at
validation range was 100e760  C and in this range, elemental sulfur reference temperature TR is derived from equation (11). Equation
is in S2 form. In these equations, T unit is  K: (12) shows the reaction heat capacity change.

Mole percent of S2 4:666  104 T 2  4:979  101 T 126:4 c d b

(6) DHrxn TR Hc TR HD TR  HB TR  HA TR (11)
a a a

Fig. 2. Distribution of sulfur vapor species (Gamson and Elkins, 1953).

 H. Kazempour et al. / Journal of Natural Gas Science and Engineering 38 (2017) 235e244 239

In this modeling, boiling feed water (BFW) is a source of water

c d b for plant use in WHB for cooling. Water stream property is at 115  C
DCp T Cpc T Cpd T  Cpb T  Cpa T (12)
a a a and 31 bars with liquid ow.
The enthalpy change per moles can be calculated by integration Fig. 3 shows temperature prole through the furnace and WHB.
of equation (12) from referenced temperature to the desired tem- The burner that named ame zone is beginning of furnace. In this
perature. Simultaneously, mass balance equation should be con- section, the temperature increases very fast and reaches about
structed in the range of reference temperature to the desired 1095  C. After this section and in anoxic zone, other power low
temperature. Equation (13) shows the enthalpy change of a specic kinetic reactions mentioned in Table 1 take place. Endothermic
component during the reaction. reactions reduce the temperature in this part. Then WHB is a heat
exchanger coupled to the furnace that its function is cooling the
DHrxn TR DHrxn T DCp T  Tc (13) outlet stream from the furnace. Two exothermic reactions of for-
mation of COS and CS2 have more conversion by decreasing the
According to the First Law of Thermodynamics and regardless of
temperature.
potential and kinetic energies changes, and energy losses from
Figs. 4e7 show the components molar ows along the furnace.
viscosity presented as equation (14).
Because of high temperature in the furnace, elemental sulfur was
Pq shown as S2 in the diagram. In Fig. 4, the molar ows of SO2, CO, S2
dT UaTa  T i1 rA  DHrxn T
Pm (14) and H2 components are presented. As observed in this gure, the
dV j1 Fj Cpj composition of SO2, CO and S2 change sharply at the initial length of
the furnace and then don't change considerably. It can be argued
that reaction between CO2 and H2 leads to produce CO and then
because of the reaction between H2 and CO, its concentration de-
4. Validation creases. SO2 is immediately produced at the ame zone (0.1e0.2 m
of the rst furnace) due to the oxidation reaction of H2S and then
In order to validate the kinetic model, the results are being consumed at the anoxic zone during Claus reaction. Production of
compared with the industrial data of the South Pars Renery in CO is started after the ame zone as a result of the reaction between
Asaluyeh, Iran. These results contain the molar ows of composi- CO2 and H2 and signicantly slowed equilibrium between CO and
tions of the outlet stream from the thermal section. CO2. S2 is also produced during Claus reaction. The molar ows of
The industrial data and kinetic modeling results have shown in H2S, CO2, H2O and N2 components are exhibited in Fig. 5. At the
Table 4, conrming the validity of the kinetic model. The most
challenging components are CO, COS, S2 and H2 as shown in this
Table. The amount of S2 in the outlet is negligible because of its very
low mole fraction. Deviations for COS, CO and H2 are high because
of the kinetically limited mechanism. However, since the mole
fractions of these components are low, the high deviation for them
did not affect on the accuracy of the simulation. Nabikandi et al. in
2015 had seen a large difference between the COS amounts ob-
tained from the simulation results and the industrial data
(Nabikandi and Fatemi, 2015). Davis et al. in 2005 proposed an
H2eCO kinetic model for high-temperature oxidation which
incorporated thermodynamic, kinetic, and species transport up-
dates (Davis et al., 2005).
By burning combustible substances in the beginning of furnace,
temperature rises to 1095  C. The exothermic reaction and
increasing temperature provide conditions for other reactions.
Concentrations of components show the reaction conversion in the
furnace. Fig. 3. Prole of temperature in the furnace and WHB.

Table 4
The kinetic modeling results compared with the industrial data of the South Pars Renery in Asaluyeh, Iran.

Composition (Kgmole/h) Data Kinetic model Absolute deviation Deviation%

H2S 78.97 78.94 0.03 0.04

CH4 0.95 0.92 0.03 3.16
SO2 40.49 41.81 1.32 3.26
N2 552.37 552.39 0.02 0.00
CO 27.51 24.44 3.07 11.16
CO2 334.23 327.94 6.29 1.88
COS 3.45 2.47 0.98 28.41
S2 0.11 0.08 0.03 27.27
S6 4.87 4.61 0.26 5.34
S8 23.16 22.88 0.28 1.21
H2O 300.20 311.94 11.74 3.91
H2 14.63 5.93 8.70 59.47
O] 0.00 0.00 0.00 0.00
CS2 1.02 0.99 0.03 2.94
C 10.51 10.76 0.25 2.38
NH3 0.03 0.03 0.00 3.70
240 H. Kazempour et al. / Journal of Natural Gas Science and Engineering 38 (2017) 235e244
Fig. 4. Molar ows of components in the furnace and WHB.
Fig. 5. Molar ows of components in the furnace and WHB.
Fig. 6. Molar ows of components in the furnace and WHB.
ame zone, H2S is immediately oxidized and after the ame zone, it
is consumed during Claus reaction. Burning combustible compo-
nents increase the H2O concentration and the oxidation reaction of
H2 cannot progress completely because of low availability of oxy-
gen. However, a few water amounts are generated from H2 oxida-
tion. H2O is also produced from H2S and CH4 oxidation and Claus
reaction. CO2 is consumed during the reaction with H2 and pro-
duction of CO. Fig. 6 shows the molar ows of CH4 and CS2. CH4 is
Fig. 7. Molar ows of components in the furnace and WHB.
immediately oxidized at the beginning of the furnace (ame zone)
and then because of low availability O2, it takes part in CS2 pro-
duction reaction in a linear trend. Finally, the molar ows of COS
and NH3 are presented at Fig. 7. At high temperature, S2 is a product
of pyrolysis reaction that reacts with CO and produces COS. Carbon
disulde and carbonyl sulde damage catalyst of the converter in
catalytic section. Half of NH3 is oxidized in beginning of the furnace
and the other half is transferred to other section as a stable
component. Because of limited resources of oxygen, a part of hy-
drocarbons are burned incompletely and produce more CO. Frac-
tion of the sulfur recovery for the thermal section of Claus process is
about 60%.
5. Sensitivity analysis
For three objects (sulfur recovery efciency, steam production,
and H2S/SO2 ratio) dened in the thermal section of Claus process, 5
parameters selected for analysis by the model. These parameters
are: Inlet molar ow fuel, inlet molar ow air, inlet stream tem-
perature, furnace pressure and WHB outlet temperature. These
parameters are reported in Table 5.
Methane acts as the fuel and the air-fuel ratio is calculated in
inlet ow to the burner. Fig. 8 shows the inuence of the fuel molar
ow on the sulfur recovery efciency and steam production. As
observed, increasing the fuel molar ow leads to decrease the
sulfur recovery efciency and increase the steam production. The
methane combustion increases the ame temperature so that the
presence of 10 Kgmole/h methane in inlet ow to the furnace rises
the temperature of the ame about 200  C. The increase of the
temperature leads to reduce the efciency of exothermic Claus
reaction. But, the temperature increase contrarily leads to improve
the performance in WHB and therefore increase the steam pro-
duction. However, the injection of an amount of the fuel to the
furnace is necessary for the stability of the ame.
In Fig. 9, the inuence of the fuel molar ow on H2S/SO2 ratio is
exhibited. As seen in this gure, H2S/SO2 ratio is decreased by the
fuel injection. H2S/SO2 ratio close to number 2 is a factor that
positively inuences on the catalytic part of the plant. Therefore, an
increase in the fuel molar ow negatively effects on the catalytic
part of the plant.
The stoichiometric ratio of air to H2S is not optimized ratio and
excess air is required. Effect of air injection to the furnace on H2S/
SO2 ratio is shown in Fig. 10. As above mentioned, more closeness
H2S/SO2 ratio to the value of 2 causes more sulfur recovery

Table 5
Parameters for sensitivity analysis.
Parameter
Inlet molar ow fuel
Inlet molar ow air
Inlet stream temperature
Furnace pressure
WHB outlet temperature
Fig. 8. Sulfur recovery efciency and steam production as a function of fuel molar ow.
Fig. 9. H2S/SO2 ratios at the end of thermal section as a function of the fuel molar ow.
Fig. 10. H2S/SO2 ratios at the end of thermal section as a function of the air molar ow.
efciency in catalytic section. According to Fig. 10, by increasing the
air molar ow, H2S/SO2 ratio is rstly close to the ideal value of 2,
H. Kazempour et al. / Journal of Natural Gas Science and Engineering 38 (2017) 235e244 241
Unit Default changes
Kgmole/h 0 0e43
Kgmole/h 728.1 400e1800

C 240 140e440
kPa 177 120e320

C 285 190e350
Fig. 11. Sulfur recovery efciency and steam production as a function of the air molar
ow.
reaches the number 2 and then becomes smaller than this value.
Fig. 11 presents the sulfur recovery efciency and the steam pro-
duction with respect to the air molar ow. The enhancement of the
inlet air ow to furnace leads to increase the sulfur recovery and
maximum amount of the recovery is about 750e800 Kgmole/h as
observed in Fig. 11. Moreover, the increase of the inlet air ow to
furnace increases temperature, improves the performance in WHB
and nally enhances the steam production. Similar results have
been shown by Zarei et al. (2016). The high temperature is desired
for pyrolysis of H2S but after an optimum point, the increase of the
temperature leads to reduce Claus reaction progress and therefore
the efciency of sulfur recovery due to the oxidation of H2S.
However, heat of the oxidation reaction causes to produce the more
steam.
Since the temperature of inlet stream signicantly inuences on
the reactions occurred in the furnace, the effect of this temperature
on the sulfur recovery efciency, steam production, and H2S/SO2
were investigated. In Fig. 12, the sulfur recovery efciency and
steam production versus temperature of inlet stream are shown. As
observed in this gure, the increase of the temperature leads to
increase the efciency of sulfur recovery and steam production. The
increase of temperature causes the progression of the endothermic
reactions such as pyrolysis of H2S and therefore the increase in the
sulfur recovery efciency. However, preheating of the inlet stream
to the furnace can cause the increase of costs. Fig. 13 exhibits the
inuence of the inlet stream temperature on the H2S/SO2 ratio.
According to this gure, by increasing this temperature, the H2S/
SO2 ratio decreases and reaches the optimized value of 2 at about
210  C.
According to the kinetics mentioned above (Table 1), the pres-
sure is an effective parameter on the reaction rate. The inuences of
pressure on the efciency of sulfur recovery, steam production, and
the H2S/SO2 ratio are presented in Figs. 14 and 15. Fig. 14 shows the
sulfur recovery efciency and steam production with respect to the
furnace pressure. The furnace pressure is increased by compressing
the inlet streams to the furnace. As seen in Fig. 14, the increase of
pressure can improve steam production. Moreover, by increasing
the furnace pressure, the sulfur recovery efciency rst increases,
then slightly decreases. The maximum efciency is observed at
about 260 kPa and about 60.5%. However, compression of the inlet
streams to furnace has high energy consumption. Fig. 15 exhibits
H2S/SO2 ratio versus the furnace pressure. Generally, the increase of
the furnace pressure undesirably inuences on outlet H2S/SO2 ratio
from the thermal section.
Fig. 16 illustrates the sulfur recovery efciency and steam pro-
duction with regard to WHB temperature. According to this gure,
242 H. Kazempour et al. / Journal of Natural Gas Science and Engineering 38 (2017) 235e244
Fig. 12. Sulfur recovery efciency and steam production as a function of the temper-
ature of inlet stream.
Fig. 13. H2S/SO2 ratios at the end of thermal section as a function of the temperature of
inlet stream.
Fig. 14. Sulfur recovery efciency and steam production as a function of the furnace
pressure.
the sulfur recovery efciency is approximately independent of
WHB outlet temperature, while by increasing the outlet tempera-
ture, the load of WHB and steam production decrease. However,
sulfur recovery efciency sways because of calculating errors for
the estimation of the nal response. In Fig. 17, the H2S/SO2 ratio
respecting the WHB outlet temperature is shown. As observed, by
increasing WHB outlet temperature from 190 to 340  C, H2S/SO2
ratio doesn't change while the increase of this temperature from
340 to 350  C leads to a slight decrease of H2S/SO2 ratio.
Fig. 15. H2S/SO2 ratio at the end of thermal section as a function of the furnace
pressure.
Fig. 16. Sulfur recovery efciency and steam production as a function of WHB outlet
temperature.
6. Optimization
Mentioned parameters are effective on three objects which
dened in the thermal section of Claus process. To determine the
importance of each parameter in optimizing, the weight factor for
each parameter is used. The main function of Claus process is
converting acid gas to elemental sulfur. Because of strict rules,
maximizing the efciency of sulfur recovery is the main object. For
this reason, the tail gas clean up unit added to the sulfur recovery
unit for increase efciency (Mahdipoor and Ashkezari, 2016). Cat-
alytic section based on Claus reaction can increase sulfur recovery
efciency to 98%. However, H2S/SO2 ratio (outlet from thermal
section) effects on performance of catalytic section.
Manenti et al. in 2014 modeled a sulfur recovery unit by a
complex combustion model with 3 variable parameters (Manenti
et al., 2014). But in this study, 5 variable parameters were consid-
ered for the kinetic model. These parameters consist of: air ow,
fuel ow, inlet temperature, furnace pressure and WHB outlet
temperature. In this multi-objective optimization problem, maxi-
mize the efciency of sulfur recovery for the thermal section,
maximize the steam production and the H2S/SO2 ratio proximity to
the number 2 are considered for optimization. Due to the very high
heating value of the fuel, the fuel ow rate was changed in the
range of 0e1.5 Kgmole/h. A response surface method used to obtain
the optimal values of these three objects. This method is based on
Central Composite Design with weight fraction for each variable
and objects and exhibits a collection of responses with desirability.
The results of the optimization are reported in Tables 6 and 7. The
 H. Kazempour et al. / Journal of Natural Gas Science and Engineering 38 (2017) 235e244
Fig. 17. H2S/SO2 ratios at the end of thermal section as a function of WHB outlet
temperature.
industrial data of the South Pars Renery in Asaluyeh, Iran have
been used as basic data shown in these tables.
As can be observed in these Tables, the sulfur recovery efciency
of the furnace thermal section decreases from 0.6129 to 0.6099 and
the steam production as an energy resource in the plant increases
to 1233.43 Kgmole/h. Moreover, this optimization results in the
2.14% decrease in air inlet to the furnace, the injection of
1.41 Kgmole/h fuel into the burner, the 5% decrease in inlet tem-
perature and the increase of furnace pressure to 235 kPa. More fuel
injection leads to better combustion, higher stability of the ame
and the enhancement of combustion product temperature. How-
ever, the inlet fuel rate is very low as compared with the total inlet
ow rate to the furnace (the sum of rates of acid gas, air, and fuel).
Therefore, the increase of the inlet fuel rate from 1.3 to 1.41 Kgmole/
h is neglectable. Moreover, the decrease of inlet temperature from
240 to 228  C leads to the reduction of the fuel consumed to pre-
heat the total inlet gas ow to the furnace, compensating the in-
crease in the inlet fuel rate. The average heat capacity (Cp, av) of the
inlet gas mixture (air acid gas) is about 36.1433 kJ/Kgmol. oK
(GPSA, 2004) and the rate of the inlet gas mixture is about
1445.1 kg mol/h, thus 626,762.9916 kJ per hour (kJ/h) is needed to
preheat the total inlet gas ow to the furnace from 228 to 240  C.
DH_ av mC
_ p; av T2  T1 1445:1  36:1433  240  228
626762:9916 kJ=h
(15)
The lower heating value (LHV) of the fuel (natural gas) is about
887,936 kJ/Kgmol (Sanaye et al., 2016). If the thermal efciency is
supposed to be about 90%, 0.7843 Kgmole/h of the fuel is approx-
imately required to preheat the inlet gas mixture from 228 to
240  C.

 
m_ Fuel; req DH_ av LHV 0:9 626762:9916=887936=0:9
0:7843 Kgmole=h
(16)
Thereupon, the reduction of the inlet temperature from 240 to
Table 6
The industrial data of the South Pars Renery in Asaluyeh, Iran and the optimized values.
WHB temp ( C) T ( C) P (kPa)
285 240 177
275 228 235
243
Table 7
The industrial data of the South Pars Renery in Asaluyeh, Iran and the optimization
results.
MP steam product (Kgmole/h) Sulfur recovery H2S/SO2 ratio Objects
1224.89 0.6129 1.95 Data
1233.43 0.6099 2.017 Optimized
228  C leads to the decrease of about 0.7843 Kgmole/h in the rate of
fuel consumed to preheat the total inlet gas ow to the furnace. It is
nally concluded that the total fuel consumption not only does not
increase but also decreases about 0.6843 Kgmole/h.
The reduction of the air inlet to the furnace also leads to enhance
the temperature. These increases in the furnace temperature
overcome the reduction of inlet temperature to the furnace. H2S
conversion takes effect from the mentioned variable parameters.
H2S Oxidation and followed by it Claus reaction rate decrease but
high temperature leads to increase the progress of the pyrolysis of
H2S. WHB temperature did not effect on the sulfur recovery ef-
ciency but the decrease of it to 275  C improved the load of WHB
and led to produce more steam. The reduction of the sulfur re-
covery efciency catches up with the improvement of H2S/SO2 ra-
tio. The proximity of this ratio to number 2 inuences on the
catalytic section operation and improves the conversion of H2S in
the catalyst bed. Diversion of this parameter from 0.05 decreases to
0.017, this improvement is about 66%. In addition, a steam with the
temperature of 214.90  C and the pressure of 21.1 bars (a medium
pressure steam) generated. The medium pressure steam produc-
tion increases to 1233.43 Kgmole/h from WHB. The desired
response has been selected with 92.01% desirability which ac-
cording to the numerous variables and objectives has good
accuracy.
7. Conclusions
In this study, modeling and multi-optimization of the furnace
for SRU plant based on Claus process were performed by a kinetic
model. The model results and the industrial data of the South Pars
Renery in Asaluyeh, Iran were compared, conrming the validity
of the kinetic model. Effects of the furnace pressure, inlet temper-
ature to the furnace, air and fuel ows to the furnace and WHB
outlet temperature on the sulfur recovery efciency, steam pro-
duction and outlet H2S/SO2 ratio were investigated. Maximizing
sulfur recovery efciency and medium pressure steam production
were the main aims for the system and H2S/SO2 molar ratio as
parameter should be close to number 2 for better performance of
the catalytic section. This optimization problem was solved using
the response surface method and the desirable response selected
with 92.01% desirability. The optimization results showed that the
steam production capacity can be improved about 8.54 kg mol/h by
decreasing the sulfur recovery efciency from 0.6129 to 0.6099.
However, the decrease in the sulfur recovery efciency leads to
more close the H2S/SO2 to number 2 and therefore the increase in
the conversion of H2S in the catalytic section, compensating this
decrease in the thermal section. Additionally, the total fuel con-
sumption was decreased about 0.6843 kg mol/h.
Air (Kgmole/h) Fuel (Kgmole/h) Variable
728.1 1.31 Data
712.5 1.41 Optimization
244 H. Kazempour et al. / Journal of Natural Gas Science and Engineering 38 (2017) 235e244

Acknowledgement Jones, D., Bhattacharyya, D., Turton, R., Zitney, S.E., 2012. Rigorous kinetic modeling
and optimization study of a modied claus unit for an integrated gasication
combined cycle (IGCC) power plant with CO2 capture. Industrial Eng. Chem.
The authors thank National Iranian Gas Company for its nancial Res. 51, 2362e2375.
support. Kaloidas, V., Papayannakos, N., 1989. Kinetics of thermal, non-catalytic decompo-
sition of hydrogen sulphide. Chem. Eng. Sci. 44, 2493e2500.
Karan, K., Behie, L.A., 2004. CS2 formation in the Claus reaction furnace: a kinetic
Nomenclature study of methane-sulfur and methane-hydrogen sulde reactions. Industrial
Eng. Chem. Res. 43, 3304e3313.
H Enthalpy (kCal/gmole, kj/gmole) Karan, K., Mehrotra, A.K., Behie, L.A., 1998. COS-forming reaction between CO and
sulfur: a high-temperature intrinsic kinetics study. Industrial Eng. Chem. Res.
Pi Partial pressure (Pa, atm) 37, 4609e4616.
R Reaction rate (gmole/m3.s) Karan, K., Mehrotra, A.K., Behie, L.A., 1999. A high-temperature experimental and
Gas constant (8.314 j/gmole.K) modeling study of homogeneous gas-phase COS reactions applied to Claus
R
plants. Chem. Eng. Sci. 54, 2999e3006.
C Concentration (gmole/m3) Kidnay, A.J., Parrish, W.R., McCartney, D.G., 2011. Fundamentals of Natural Gas
Cp Heat capacity (j/gmole.K) Processing. CRC Press.
F Molar ow (gmole/s) Kohl, A.L., Nielsen, R., 1997. Gas purication. Gulf Publishing Company, Houston, TX,
USA.
Af Arrhenius coefcient of forward reaction Mahdipoor, H.R., Ashkezari, A.D., 2016. Feasibility study of a sulfur recovery unit
V Volume (m3) containing mercaptans in lean acid gas feed. J. Nat. Gas Sci. Eng. 31, 585e588.
Ar Arrhenius coefcient of reverse reaction Manenti, F., Papasidero, D., Bozzano, G., Ranzi, E., 2014. Model-based optimization
of sulfur recovery units. Comput. Chem. Eng. 66, 244e251.
Ni Moles of component j
Manenti, G., Papasidero, D., Manenti, F., Bozzano, G., Pierucci, S., 2012. Design of SRU
Eaf Activation energy of forward reaction (j/gmole) thermal reactor and waste heat boiler considering recombination reactions.
T Temperature (K,  C) Procedia Eng. 42, 376e383.
Ear Activation energy of reverse reaction (j/gmole) Monnery, W., Hawboldt, K., Pollock, A., Svrcek, W., 2000. New experimental data
and kinetic rate expression for the Claus reaction. Chem. Eng. Sci. 55,
P Pressure (Pa, atm) 5141e5148.
Pe Peclet number Monnery, W., Hawboldt, K., Pollock, A., Svrcek, W., 2001. Ammonia pyrolysis and
Sulfur recovery efciency oxidation in the Claus furnace. Industrial Eng. Chem. Res. 40, 144e151.
Monnery, W.D., Svrcek, W.Y., Behie, L.A., 1993. Modelling the modied claus process
DHrxn or DH Enthalpy of reaction reaction furnace and the implications on plant design and recovery. Can. J.
Chem. Eng. 71, 711e724.
References Nabikandi, N.J., Fatemi, S., 2015. Kinetic modelling of a commercial sulfur recovery
unit based on Claus straight through process: comparison with equilibrium
model. J. Industrial Eng. Chem. 30, 50e63.
Dalrymple, A., Skinner, F., Maserole, N., 1991. Investigation of US Natural Gas
Peters, N., 1979. Premixed burning in diffusion amesdthe ame zone model of
Reserve Demographics and Gas Treatment Processes. Final Report to the Gas
Libby and Economos. Int. J. Heat Mass Transf. 22, 691e703.
Research Institute. Report No. GRI-91/0019 (January 1991).
Sanaye, S., Ghafurian, M.M., Tavakoli Dastjerd, F., 2016. Applying Relative Net Pre-
Davis, S.G., Joshi, A.V., Wang, H., Egolfopoulos, F., 2005. An optimized kinetic model
sent or Relative Net Future Worth Benet and exergy efciency for optimum
of H 2/CO combustion. Proc. Combust. Inst. 30, 1283e1292.
selection of a natural gas engine based CCHP system for a hotel building. J. Nat.
Gamson, B., Elkins, R., 1953. Sulfur from hydrogen sulde. Chem. Eng. Prog. 49,
Gas Sci. Eng. 34, 305e317.
203e215.
Tong, S., Dalla Lana, I.G., Chuang, K.T., 1997. Effect of catalyst shape on the hydrolysis
Gargurevich, I.A., 2005. Hydrogen sulde combustion: relevant issues under Claus
of cos and CS2 in a simulated claus converter. Industrial Eng. Chem. Res. 36,
furnace conditions. Industrial Eng. Chem. Res. 44, 7706e7729.
4087e4093.
GPSA, 2004. Engineering Data Book, vol. 2. Gas Processors Suppliers Association,
Turns, S., Holman, J.P., Lloyd, J., 2000. An Introduction to Combustion Concepts and
pp. 16e24.
Applications. McGraw Hill, Singapore.
Hawboldt, K., Monnery, W., Svrcek, W., 2000. New experimental data and kinetic
Westbrook, C.K., Dryer, F.L., 1981. Simplied reaction mechanisms for the oxidation
rate expression for H2S pyrolysis and re-association. Chem. Eng. Sci. 55,
of hydrocarbon fuels in ames. Combust. Sci. Technol. 27, 31e43.
957e966.
Zarei, S., Ganji, H., Sadi, M., Rashidzadeh, M., 2016. Kinetic modeling and optimi-
Hawboldt, K.A., 1998. Kinetic Modelling of Key Reactions in the Modied Claus Plant
zation of claus reaction furnace. J. Nat. Gas Sci. Eng. 31, 747e757.
Front End Furnace. Ph.D. Thesis. Department of Chemical and Petroleum En-
gineering, University of Calgary, Canada.