CHEMICAL ENGINEERING TRANSACTIONS  
 

VOL. 70, 2018 

A publication of 

 

The Italian Association 
of Chemical Engineering 
Online at www.aidic.it/cet 

Guest Editors: Timothy G. Walmsley, Petar S. Varbanov, Rongxin Su, Jiří J. Klemeš 
Copyright © 2018, AIDIC Servizi S.r.l. 

ISBN 978-88-95608-67-9; ISSN 2283-9216 

Optimization of the Pyrolysis Gasoline Hydrogenation Reactor 

Considering the Hydrogen Network Integration 

Lingjun Huang, Donghui Lv, Guilian Liu* 
School of Chemical Engineering and Technology, Xi’an Jiaotong University, Xi’an, Shaanxi Province, China, 710049 

guilianliui@mail.xjtu.edu.cn 

A mathematical model is developed considering the pyrolysis gasoline (pygas) hydrogenation kinetic, the 

temperature of the second-stage pygas reactor, hydrogen network integration, the variation of catalyst activity 

and product quality along time. The industrial data is employed to fit the linear equation about the average 

temperature of the second-stage reactor and operation time of catalyst. The relationship between catalyst 

activity and operation time is deduced, and the temperature of second-stage reactor is optimized within a certain 

range with the product quality requirements satisfied, as well as the economic benefit. Case study shows that 

this model can provide guidance for the design and improvement of pygas hydrogenation process. 

1. Introduction 

Pygas is an important by-product of ethylene industry with high content of aromatic hydrocarbon (benzene, 

toluene and xylene). Besides that, olefins, diolefins and organic compounds containing sulphur, nitrogen and 

oxygen also exist and these components make it unstable (Sun et al., 1999). Because of this, pygas needs to 

be hydrogenated before further processing. In industry, pygas is processed with a two-stage hydrogenation 

reactor. The diolefins are selectively hydrogenated in the first-stage, while olefins and compounds containing 

nitrogen, sulfur, and oxygen are hydrogenated in the second-stage. Both stages consume large amount of 

hydrogen. When the operating parameter of the reactor changes, the hydrogen consumption of the hydrogen 

network will change correspondingly. The reason is that the hydrogen separated from the effluent of this reactor 

is a source of the hydrogen network, while its inlet hydrogen is a sink. Therefore, the hydrogen consumption 

should be considered in the optimization of the pygas hydrogenation reactor. 

Navid et al. (2005) developed a two-stage model for the hydrogenation of pygas and illustrated that the 

prediction of this model is in satisfactory agreement with the industrial unit. Qian et al. (2011) evaluated the 

introduction of Zn and Mo into the Ni-based catalyst and proved that this catalyst has excellent catalytic activity 

for the hydrogenation of pygas. Zhou et al. (2007) proposed a Langmuir-Hinshelwood-type kinetic model for the 

selective hydrogenation of pygas, and the predicted results show that competitive adsorption of diolefins and 

mono-olefins exists over the active sites of the catalyst. Later, this model was applied in liquid-phase selective 

hydrogenation of pygas, and a diffusion-reaction mathematical model was developed (Zhou et al., 2010). 

Catalysts play a critical role in pygas hydrogenation, and its activity decreases along time. To compensate its 

deactivation and keep the reaction rate stable, the inlet temperature of the reactor is generally raised. This will 

have different effects on different reactions and should be considered in the optimization of pygas hydrogenation 

reactor. Castaño (2007) evaluated the effect of five different Pt-supported catalysts on conversion, selectivity 

and deactivation and provided a guidance for optimal catalyst design. Zhao (2006) analyzed the nickel-based 

catalyst applied in the first-stage pygas hydrogenation and reached a conclusion that CS2 leads to its 

deactivation. Similarly, for the palladium-based catalyst, the main cause of deactivation is also CS2 (Xue et al., 

2014). Deng et al. (2014) located the flowrate targets of interplant hydrogen conservation network via the 

improved problem table (IPT). Mao et al. (2015) integrated the hydrogen network of a refinery with the hydrogen 

consumption of Vacuum Gas Oil (VGO) hydro-cracking reactor considered. However, no research considered 

the hydrogen network integration together with the catalyst activity. 

In this work, the pygas hydrogenation reactor of a refinery is optimized with both the hydrogen network 

integration and the variation of the catalyst activity considered. The mathematical model will be built to analyze 

                                

 
 

 

 
   

                                                  
DOI: 10.3303/CET1870103 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Please cite this article as: Huang L., Lv D., Liu G., 2018, Optimization of the pyrolysis gasoline hydrogenation reactor considering the 
hydrogen network integration , Chemical Engineering Transactions, 70, 613-618  DOI:10.3303/CET1870103   

613



the influence of the second-stage reactor's temperature on the hydrogen consumption of whole hydrogen 

network and the product quality, and identify its optimal value. A case is studied to show the validity of the 

proposed model. 

2. Problem Statement 

In refinery, the hydrogen stream separated from the effluent of the second-stage pygas hydrogenation reactor 

is generally the source (SRp) of hydrogen network, while the inlet hydrogen of this reactor is a sink (SKq). For 

this reactor, the conversion of each component is Xi. There are Nm hydrogen sources and Nn hydrogen sinks in 

a hydrogen network, the flow rate and hydrogen concentration of each hydrogen source are FSRi and CSRi, 

respectively; those of each hydrogen sink are FSKj and CSKj, respectively. The purity of fresh hydrogen is 

99.99 %.  

The objective is to identify the optimal parameters of the second-stage pygas hydrogenation reactor considering 

the hydrogen network integration.  

3. Mathematical model 

In the pygas hydrogenation reactor, the catalyst activity will decrease along time. To compensate the 

deactivation of the catalyst and keep the reaction rate stable, the inlet temperature is generally raised. However, 

the upper limit of temperature exists. If the reactor is operated at the temperature greater than the upper limit, 

the catalyst might deactivate rapidly, and the reaction rate of side reactions increases significantly, or more side 

reactions occur. This will affect the product and properties of the desired product (Li et al., 2016). Because of 

this, the catalysts should be changed or regenerated once the temperature reaches the upper limit. For the 

second-stage reactor of the pygas hydrogenation, the average temperature of catalyst bed can be adjusted 

along the operation time according the fitting equation shown by Eq. (1) (Zhang and Wang, 1996).  

 
a

T te f  (1) 

Where Ta is the average temperature of catalyst bed, °C; t  is the operation time, day; e and f are constants. 

3.2 Relation between the conversion and operating temperature 

In the second-stage reactor, diolefins unreacted in the first-reactor will be hydrogenated to mono-olefines, and 

the latter can be further hydrogenated. For the second-stage reactor, Eq. (2) shows the relation between the 

conversion(X) and the operating temperature.  

         
   

  

i i i
a b a

i ai H i

i

i i

i

n

Y k exp E T p p a
X

F

2

1

1

,

/ R 273.15 1
1 1  (2) 

Where Xi is the conversion of component i, %; Y is the catalyst activity in the second-stage reactor; ω is the 

weight of catalyst, g; ki is the pre-exponential factor; Eai is the activation energy, J·mol-1; T is the reactor 

temperature, °C; pi is the pressure of component i , Pa; PH2 is the pressure of H2, Pa; ai and bi are reaction 

orders; Fi,in is inlet flow of component i, mol·h-1. 

When Xi is fixed, the relationship between Y and t can be deduced with Eq(1) substituted into Eq(2), as shown 

by to Eq(3).  

 
 



    
         

i

i i

a

a bi ai
i in i H

i

i

X E
Y F k p p

a t 2

1

,

1 1
exp

1 R e f 273.15
 (3) 

3.3 The effect of reactants conversion on the product price and hydrogen consumption 

If the reactor is operated at the temperature different from that identified according Eq(1), Xi will change, and 

this will affect the product purity, as well as the product price. The relationship between the product purity and 

conversion is shown by Eq(4), while that between the product price and conversion is shown by Eq(5).  



 
n

i i i

i

C g X h
1

 (4) 

 P Cm n  (5) 

614



Where P is the product price, RMB·t-1; C is the total mass fraction of benzene, methylbenzene and 

dimethylbenzene in the product, %; gi and hi are constants related to reactants' stoichiometric factors; m and n 

are constants.  

In the second-stage reactor, Hr is composed by two parts, the hydrogen consumption of diolefins hydrogenation 

( 
 1

n

i i i
F X ) and that of mono-olefines hydrogenation (Hm), as shown by Eq(6). When the reactor's parameters 

change, the hydrogen consumption of second-stage reactor changes to Hr’. And the variation of hydrogen 

consumption can be calculated by Eq(7) (Mao et al., 2016). 




 
1

r i m

n

i i

i

H F X H  (6) 

  
'

r r r
H H H  (7) 

Where φi is the stoichiometric ratio of H2 to reactant i in the second-stage reactor; Fi is the flow of reactant i. 

According to the work of Mao et al. (2015), the hydrogen utility adjustment (HUA), which denotes the variation 

of the hydrogen consumption of the hydrogen network, is also affected by Xi. When the pinch appears at the 

sink-tie-line lying above SKq, HUA can be calculated by Eq(8). And when it appears at the sink-tie-line lying 

between SKq and SRp, and below SRp, HUA can be calculated by Eq(9). 

  


i

i
u

u SR

H
F

c c
 (8) 


   

 

1
i p

i i

SR SR i
u r

u SR u SR

c c H
F H

c c c c
 (9) 

Where c is the hydrogen concentration of each stream. 

3.4 Objective of the optimization  

With the price of product calculated by Eq(5) and that of hydrogen given by Li et al. (2016), the economic benefit 

can be calculated according to Eq(10). Based on this, the reactor parameters, including the operation 

temperature, can be optimized. 

   
2H u

S PM P F  (10) 

Where ΔP is the variation of product price, RMBt-1; M is the weight of product, t; PH
2 
is the price of hydrogen, 

RMB·Nm-3.  

4. Case study 

The pygas hydrogenation reactor and hydrogen network of a refinery will be studied in this section. In the 

hydrogen network of this refinery, there are 26 sources and 14 sinks, and their data are shown in Table 1. SR8 

and SK4 are connected to the second-stage pygas hydrogenation reactor. The initial temperature of second-

stage reactor is 280 °C; the hydrogen pressure is 3 MPa; pygas flow rate is 29.9 t·h-1 and its density is 0.804 

g·cm-3; hydrogen to oil mole ratio is 3.6. In this reactor, there are the hydrogenation reactions of cyclopentene, 

styrene, hexene and thiophene, and the corresponding hydrogenation reaction kinetics are shown Table 2 (Mao 

et al., 2015), as well as the inlet flow rates and di, which is the amount of hydrogen required for hydrogenating 

per mole of component i.  

The catalyst activity will decline along operation time, and reaction temperature is generally raised to keep the 

reaction rate stable. The fitted relationship between average temperature and operation time is shown by 

Eq(11). And the upper limit on the operating temperature of this reactor is 300 °C (Zhang and Wang, 1996).  

=0.078 280T t  (11) 

  

615



Table 1: The data of sources and sinks 

Streams Purity/% Flowrate 

/Nm3h-1 

Streams Purity/%  Flowrate 

/Nm3h-1 

Streams Purity/%  Flowrate 

/Nm3h-1 

Sources 

SR0 97.60 29,800.00 SR9 90.00 8,650.00 SR18 91.97 5,314.00 

SR1 99.80 14,000.00 SR10 90.00 54,500.00 SR19 74.99 300.00 

SR2 95.00 1,000.00 SR11 88.00 63,000.00 SR20 69.67 170.00 

SR3 95.00 2,700.00 SR12 89.70 67,000.00 SR21 65.80 500.00 

SR4 93.50 640.00 SR13 88.00 31,000.00 SR22 63.96 474.00 

SR5 93.00 22,800.00 SR14 88.00 49,000.00 SR23 58.60 300.00 

SR6 92.00 26,500.00 SR15 87.00 6,500.00 SR24 50.00 3,000.00 

SR7 91.97 500.00 SR16 82.00 200.00 SR25 55.00 200.00 

SR8 91.00 18,000.00 SR17 82.00 2,300.00 SR26 55.00 50.00 

Sinks 

SK1 99.78 1,400.00 SK6 92.41 28,000.00 SK11 87.97 55,000.00 

SK2 99.60 4,000.00 SK7 91.70 9,100.00 SK12 89.40 67,000.00 

SK3 99.60 5,000.00 SK8 88.21 73,000.00 SK13 87.45 7,650.00 

SK4 96.00 25,000.00 SK9 90.26 56,000.00 SK14 92.10 5,954.00 

SK5 93.58 25,000.00 SK10 89.74 38,000.00    

Table 2: Reaction kinetics of all the hydrogen consumption component 

Components Hydrogenation reaction kinetics Inlet flow 

rate/mol·h-1 

di 

Cyclopentene 



 

2

6484

1 1

5
4 0.724 0.682

1.422 10 RT
H

r e p p   
54,810 1 

Styrene 



 

2

6554

2 2

5
5 0.775 0.475

6.022 10 RT
H

r e p p   
1,524.38 4 

Hexene 



 

2

5984

3 3

5
5 0.764 0.685

4.022 10 RT
H

r e p p   
18,090 1 

Thiophene 


 
2

71855

4

6 0.73 .8

4

0 4
7.251 10 RT

H
r e p p   

106.11 4 

Methylthiophene 



 

2

75955

5

6 0.83 0.8 5

5

3
8.311 10 RT

H
r e p p    

151.2 4 

Dromethyl-

thiophene 



 

2

64855

6

6 0.87 0.8 2

6

3
9.629 10 RT

H
r e p p    

31.86 4 

Table 3: The HUA of each sink-tie-lines 

Sink-tie-lines HUA / Nm3·h-1 Sink-tie-lines HUA / Nm3·h-1 

1   
u

F
,1

4,305.22  5    u rF H 2,5 0.181 2,756.2  

2   
u

F
,2

2,805.73  6   
2,6

0.343
u r

F H
 

3   
u

F
,3

2,348.94  7    
2,7

0.634 337.87
u r

F H
 

4    
u r

F H
2,4

0.145 2,181.45    

In this hydrogen network, seven sink-tie-lines can intersect the sources and are numbered from 1 to 7; the Pinch 

can only appear at these sink-tie-lines (Mao et al., 2015). By the hydrogen surplus method, it is identified that 

the initial Pinch purity is 88 % (volume fraction) and the hydrogen utility is 27,235.63 Nm3h-1. At the minimum 

utility consumption, the hydrogen surpluses of sink-tie-lines 1 to 7 are198.04 mol·h-1, 157.12 mol·h-1, 155.03 

mol·h-1, 165.79 mol·h-1, 217.74 mol·h-1, 0, and 52.71 mol·h-1, respectively. Sink-tie-lines 1, 2 and 3 lie above 

SK4, the corresponding T versus HUA relationship can be determined according to Eq(8). Sink-tie-lines 4, 5, 6 

and 7 lie between SK4 and SR8, the corresponding T versus HUA can be determined according to Eq(9). The 

results are shown in Table 3. The conversion of each components can be calculated based on Eq(2) and Table 

2, as shown in Table 4. Substituting the conversions into Eq(6), Eq. (12) is obtained. According to Eq(3), Eq(11), 

616



the data in Table 2 and the operating parameters, the relationship between catalyst activity about different 

reactants and operation time can be deduced respectively, as shown in Table 4. 

'
54,810 6,097.52 10,890 424.44 604.8 127.44 1,178.07

r cyc sty hex thi met dro
H X X X X X X        (12) 

Table 4: The conversion and catalyst activity of each components 

Components Conversion Catalyst Activity 

Cyclopentene 
  

      
   

cyc
X Y

T

3.623

2

7,799.49
1 1 445,000 exp

273.15
 

  
   

 
cyc

Y
t

7 7,799.49
7.533 10 exp

0.078 553.15
 

Styrene 
  

      
   

sty
X Y

T

4.44

2

7,883.69
1 1 30,082.64 exp

273.15
 

  
   

 
sty

Y
t

7 7,883.69
6.458 10 exp

0.078 553.15
 

Hexene 
  

      
   

hex
X Y

T

4.24

2

7,198.1
1 1 236,788.62 exp

273.15
 

  
   

 
hex

Y
t

6 7,198.1
2.23 10 exp

0.078 553.15
 

Thiophene 
  

      
   

thi
X Y

T

3.7

2

8,642.65
1 1 1,303,014 exp

273.15
 

  
   

 
thi

Y
t

7 8,642.651
1.57 10 exp

0.078 553.15
 

Methyl-

thiophene 
  

      
   

met
X Y

T

5.88

2

9,135.80
1 1 1, 422,164 exp

273.15
 

  
   

 
met

Y
t

7 9,135.795
1.225 10 exp

0.078 553.15
 

Dromethyl-

thiophene 
  

      
   

dro
X Y

T

7.69

2

7,800.70
1 1 1,860,975 exp

273.15
 

  
   

 
dro

Y
t

7 7,804.306
3.078 10 exp

0.078 553.15
 

 

If the product quality that affects product price is allowed to change within a given interval, based on the 

hydrogen network integration, the average temperature can be optimized with HUA and product price taken into 

consideration, as shown by Eq(13). The price of hydrogen is taken as 1.52 RMBNm-3 (Li et al., 2016), and the 

mass flow rate of the aromatics product is 28 th-1 (Hao, 2007). ΔFu is identified according to Table 3 and the 

method presented in Mao (2016). And P can be calculated by Eq(14) (Hao, 2007). According to the mass 

balance, the purity of benzene, methylbenzene and dimethyl in the benzene aromatic product, S can be 

calculated by Eq(15). Based on Tables 3 and 4, Eqs(12) and (13), the S~T~t surface is plotted in the three-

dimensional diagram, as shown by Figure 1a. This diagram clearly shows the variation of the economic benefit 

along time and operating temperature. 

  28 1.52
u

S P F  (13) 

 P C9, 400 5075.9  (14) 

   0.63 0.203 0.0057 0.067
cyc sty hex

S X X X  (15) 

 (a)           (b)  

Figure 1: (a) The S~T~t three-dimensional diagram. (b) The temperature against operation time with maximum 

economic benefit 

0 50 100 150 200 250 300 350
280

282

284

286

288

290

292

294

296

298

300

T
/℃

t/d

 

 

0 50 100 150 200 250 300 350
0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

E
c
o
n
o
m

ic
 B

e
n
e
fi
t 
D

if
fe

re
n
c
e
/R

M
B

Improved Temperature Curve

Initial Temperature Curve

Economic Benefit Difference

 

617



According to Eq(11), the initial temperature curve is drawn in Figure 1b. The optimal temperature at different 

operation time also can be calculated according to Table 4, as shown by the improved temperature curve. 

Furthermore, the variation of the economic benefit along operation time is illustrated by the economic benefit 

curve. According to the initial temperature curve, when t = 256d, the temperature reaches to the upper limit. If 

the reactor is operated according the improved temperature curve, the operation time of the catalyst can be 

raised to 337days, as shown in Figure 1b. 

5. Conclusion 

A model is presented for optimizing the temperature of the second-stage reactor of pyrolysis gasoline 

hydrogenation unit based on hydrogen network. This model can give a significant guidance for pyrolysis gasoline 

hydrogenation unit design and improvement, and the optimization results is more meaningful as the catalyst 

activity, product quality and hydrogen network integration are considered. The case study shows that this model 

clearly shows the effect of operation temperature variation on economic benefit at different time, optimal 

temperature can be identified according to the initial and improved temperature curves, and the service life of 

catalyst can be prolonged effectively and the economic benefit is increased by adjusting the temperature. 

In the proposed model, the numerical relationship between reactor temperature and operation time is 

represented by a linear fit, this will affect the accuracy of the optimization results. In the future, more accuracy 

relation should be deduced and applied. Furthermore, more aspects should be taken into consideration to 

calculate the economic benefit over reaction temperature. This will be studied in the future. 

Acknowledgement 

Financial supports provided by the National Natural Science Foundation of China (21476180) and (U1662126) 

are gratefully acknowledged. 

References 

Castaño P., Pawelec B., Fierro J.L.G., Arandes J.M., Bilbao J., 2007, Enhancement of Pyrolysis Gasoline 

Hydrogenation over Pd-promoted Ni/Sio2–Al2O3 Catalysts, Fuel, 86, 2262-2274. 

Deng C., Zhou Y.H., Li Y.T., Feng X., 2014, Flowrate Targeting for Interplant Hydrogen Networks, Chemical 

Engineering Transactions, 39, 19-24. 

Hao Z., 2007, The Industrial Application and Optimization of Pyrolysis Gasoline Hydrogenation Plant, MEng 

Thesis, Tianjin University, Tianjin, China. 

Li D.D., Nie H., Sun L.L., 2016, Hydrotreating Engineering and Technology, Beijing, China, ISBN: 978-7-5114-

3668-9, 1662. 

Li H.K., Deng C., 1998, Qualitative Analysis of Light Component of Splitting Decomposition Petrol Using Kovats 

Retention Indices, Journal of Instrumental Analysis, 17, 67-69. 

Mao J. B., 2016, Integration of Hydrogen Networks Considering the Variation of the Coupled Sink and Source, 

MEng Thesis, Xian Jiaotong University, Xian, China. 

Mao J.B., Liu G.L., Wang Y.J., Zhang D., 2015, Integration of a Hydrogen Network with the Vacuum Gas Oil 

Hydrocracking Reaction, Chemical Engineering Transactions, 45, 85-90. 

Mao J.B., Shen R.J., Wang Y.J., Liu G.L., 2015, An Integration Method for the Refinery Hydrogen Network with 

Coupling Sink and Source, International Journal of Hydrogen Energy, 40, 8989-9005. 

Navid M., Rahmat S.G., Mohammad A., Javad E., 2005, Simulation of An Industrial Pyrolysis Gasoline 

Hydrogenation Unit, Chemical Engineering & Technology, 28, 174-181. 

Qian Y., Liang S.Q., Wang T.H., Wang Z.B., Xie W., Xu X.L., 2011, Enhancement of Pyrolysis Gasoline 

Hydrogenation over Zn- and Mo-Promoted Ni/γ-Al2O3 Catalysts, Catalysis Communications, 12, 851-853. 

Sun L.X., He Q.Y., Hu Y.X., 1999, Evaluation Method for Selective Hydrogenation Catalysts, Petroleum 

Processing and Petrochem Icals, 30, 44-47. 

Xue H.F., Geng Z.J., Wang F., Qin P., Ma H.W., Fan G.N., 2014, Cause of Poisoning of Palladium-Based 

Catalyst for Hydrogenation of Pyrolysis Gasoline, Petrochemical Technology, 43, 1076-1081. 

Zhang X.J., Wang L.Q., 1996, Prediction of Pyrolysis Gasoline Hydrogenation Catalyst Life with Linear 

Regression Analysis, Heilongjiang Petrochemical Technology, 4, 1-4. 

Zhou Z.M., Cheng Z.M., Cao Y.N., Zhang J.C., Yang D., Yuan W.K., 2007, Kinetics of the Selective 

Hydrogenation of Pyrolysis Gasoline, Chemical Engineering & Technology, 30, 105-111. 

Zhou Z.M., Zeng T.Y., Cheng Z.M., Yuan W.K., 2010, Kinetics of Selective Hydrogenation of Pyrolysis Gasoline 

over an Egg-Shell Catalyst, Chemical Engineering Science, 65, 1832-1839. 

618