Developing a Steady-state Kinetic Model for Industrial Scale Semi-Regenerative Catalytic Naphtha Reforming Process

The catalytic reforming of heavy naphtha (heavy straight run gasoline or HSRG) is a favourite process in petroleum refineries due to producing high-octane gasoline.1 The semi-regenerative naphtha reformer is the oldest type where reactions are carried out in three or four adiabatic fixed-bed reactors in series, each of which is equipped with a pre-heater. This plant usually operates at temperatures between 450 °C and 520 °C, total pressure between 25 and 35 atm, and hydrogen-to-hydrocarbon amount ratios between 3 and 7.1,2,3


Introduction
The catalytic reforming of heavy naphtha (heavy straight run gasoline or HSRG) is a favourite process in petroleum refineries due to producing high-octane gasoline. 1The semi-regenerative naphtha reformer is the oldest type where reactions are carried out in three or four adiabatic fixed-bed reactors in series, each of which is equipped with a pre-heater.This plant usually operates at temperatures between 450 °C and 520 °C, total pressure between 25 and 35 atm, and hydrogen-to-hydrocarbon amount ratios between 3 and 7. 1,2,3 Usually, the feed of the catalytic reforming process is HSRG, including four hydrocarbon groups i.e. paraffins, olefins, naphthenes and aromatics (PONA) with number of carbon atoms between 5 and 10.
The main reforming reactions occurring through the catalytic beds are dehydrocyclization, hydrocracking, isomerization, dehydrogenation and cyclization.Some of these reactions are desired for increasing octane number of gasoline, whereas others are undesired because they decrease it.For paraffins, increment of octane number is the result of reactions increasing the number of branches, such as cyclization and aromatization.Therefore, normal paraffins conversion to isoparaffins, naphthenes, and aromatics can increase the octane number. 3e catalytic reforming process is often modelled based on: 1) the number of reactive species, and 2) the type of the used kinetic model.However, the presence of many com-ponents as reactants or products causes numerous reactions.Therefore, the situation is extremely sophisticated for process modelling.To decrease these complications, reactants in the mixture are classified into limited groups called pseudo-components or lumps.The number of selected lumps in the mixture is a determinant factor for designing the reforming model.Arrhenius and Langmuir-Hinshelwood kinetics are widely used for catalytic reforming models.
In the field of catalytic naphtha process modelling, a simple and first model was suggested by Smith 4 in which naphtha reforming was considered as a combination of only four reactions.Then, in 1997, Taskar suggested a model for the catalytic reforming reaction that consisted of 35 pseudo-components in the reaction network and 36 reactions. 5ollowing the use of Arrhenius kinetics, a well-known model was proposed by Padmavathi 6 in 1997 in which 26 pseudo-components, such as alkyl cyclohexane (ACH), alkyl cyclopentane (ACP), normal paraffins (NP), isoparaffins (IP), aromatics (A), hydrogen (H 2 ) and light hydrocarbons (C 1 to C 5 ) were used in the network.Ancheyta et al. 7 developed a kinetic model for the naphtha catalytic reforming process.This model utilized a lumped mathematical model, presenting the reactions ranging from 1 to 11 carbon atoms for paraffins, and from 6 to 11 carbon atoms for naphthenes and aromatics.In 2003 Rahimpour et al. presented a kinetic model for industrial scale catalytic naphtha reformers, including deactivation of the catalyst.The impact of inlet temperature, operating pressure and catalyst mass distribution on the performance of the reactors was examined.The results indicated an increase in aromatic yield with increasing inlet temperature.However, manipulating the operating pressure had no appreciable effect on the gasoline yield.Additionally, this model estimated catalyst deactiva-tion where the corresponding parameters were estimated using the plant data. 8In 2006 Hou et al. presented a new 18-lump kinetic-based mathematical model for an industrial continuous catalytic reforming plant.In this model, reaction temperature and concentration profiles of all reactors, heater duties, catalyst deactivation, recycle gas composition and octane number for different feedstock or operating conditions could be predicted. 9In the next effort in 2009 Arani et al. developed a lumping procedure to obtain kinetic and thermodynamic parameters of catalytic naphtha reformer using a model consisting of 17 lumps ranging from C 6 to C 8 + hydrocarbon, and including 15 reaction pathways. 10In the same year Fatemi et al. developed a mathematical model for a commercial naphtha catalytic reformer which included three sequencing catalytic fixed-bed reactors at steady-state condition.They used a detailed kinetic scheme involving 26 pseudo-components connected by a network of 47 reactions.The output variables of the reformer, such as RON (research octane number) and yield of gasoline showed good agreement with actual data obtained from the under study reforming unit. 11Recently Ziaoon presented a detailed kinetic model including 24 components, 1 to 11 carbon atoms for paraffins, and 6 to 11 carbon atoms for naphthenes and aromatics, forming 71 reactions. 12 this paper, in order to model an industrial scale catalytic naphtha reforming process, a steady-state kinetic model including 9 pseudo-components and 4 reactions was developed.In comparison to the complex models proposed for this process, this model can also show low average absolute deviation against actual data.

Process description
A commercial fixed-bed catalytic naphtha reforming unit, called Platformer, licensed by Chevron research cooperation was chosen as a case study.The feed of the plant prior to entering the catalytic reformer should undergo hydrodesulphurization (HDS) reaction in the hydrotreatment unit.Then, the produced naphtha, called Platcharge, is introduced to the reforming process.The most commonly used types of catalytic reforming units have three or four reactors each having a fixed catalytic bed. 3,4 shown in Fig. 1, Platcharge is first preheated by the first furnace (H-1), and then it enters the first reactor (R-1) where the naphthenes are dehydrogenated to aromatics.The product stream from the first reactor then passes through the second reactor (R-2), and enters the third reactor (R-3).The overall reforming reactions are endothermic; therefore, a preheater (H-1, H-2, and H-3) should essentially be provided before each reforming reactor.
Next, the product stream from the third reactor enters the flash separator (V-1) wherein the produced hydrogen is separated, and recycled to the beginning of the process.This recycled stream is then mixed with the fresh naphtha feed (Platcharge).Finally, the liquid product leaving the separator is introduced to the gasoline stabilizer in which the LPG and light gases are separated from the gasoline.So, the vapour pressure of the gasoline can be set according to the market requirement.
The specifications of feed and catalyst distribution of the studied catalytic naphtha reforming plant are given in Tables 1 and 2. The normal operating conditions of the unit are presented in Table 3.
The design pressure of the studied catalytic reforming plant was 34 bar.However, depending on the feed specification and deactivation of the catalyst, the operating pressure of the plant could be varied between 27 and 32 bar.So, the effect of the pressure on reforming reactions was included in the model, which is discussed later.

Development of kinetic model 4
In the present study, in order to simulate a catalytic reforming unit, the naphtha feed is classified into three general lumps i.e. aromatics, naphthenes and paraffins.Also considered were hydrogen, methane, propane, butane, and pentane.The reactions within this model are classified into four groups, as follows:

Naphthenes to paraffins
Rate constants concerning this reaction are as follows: 4 (3)

Hydrocracking of paraffins
The rate of paraffins cracking and rate constants concerning this reaction are as follows: 4 (5)

Hydrocracking of naphthenes
The rate of naphthenes cracking and rate constants concerning this reaction are as follows: 4 Using the presented rate equations, the mass and energy conservations can be written as follows: where n is the number of each presumed carbon of pseudo-components 4 which is between 7 and 6 for the feed in the model.

Results and discussions
After developing the kinetic model using data obtained from a pilot-scale catalytic naphtha reforming unit, it should be scaled up to the industrial scale.An optimization subroutine was used to determine the coefficients until reaching a suitable consistency between the industrial data and the model results.In this subroutine, the Levenburg-Marquardt optimization algorithm was used, and the following target function was optimized: (13) The magnitudes of constants are presented in Table 4 for the studied reforming plant.
T a b l e 4 -Reaction constants calculated using optimized approach T a b l i c a 4 -Konstante optimiziranim pristupom  Other significant operating parameters in catalytic reforming are product volume yield and RON.Comparisons of the predicted product volume yield and RON with the actual data are shown in Figs. 5 and 6.From these figures, close mappings between the measured and simulated product volume yield and RON can be observed.Moreover, it is found that the presented model can simulate the RON and volume yield of gasoline with AAD of 0.32 % and 4.8 %, respectively.These results confirm that the presented approach can reliably be applied by refineries to monitor the operation of the catalytic reforming plant.The other significant parameter in the catalytic reforming plant is the hydrogen purity of the recycle stream.Fig. 7 presents the comparison between the simulated hydrogen purity and the actual data.These results show that the kinetic model can predict the purity of the produced hydrogen with AAD of 3.19 %.It is concluded that the presented kinetic model is also good for predicting the purity of hydrogen produced in the catalytic naphtha reforming unit.

Conclusions
The catalytic reforming of heavy naphtha (heavy straight run gasoline or HSRG) is a favourite process in petroleum refineries for producing high-octane gasoline.In this research, significant process variables of a commercial naphtha catalytic reforming plant were modelled using a heterogeneous kinetic model.These variables included the outlet temperatures of the first, second and third reactors, RON of gasoline, product volume yield and hydrogen purity.To evaluate the proposed model, the results were compared against data obtained from a commercial catalytic naphtha reformer.It was found that the mean relative absolute deviation (AAD) of the mentioned parameters were 0.38 %, 0.52 %, 0.54 %, 0.32 %, 4.8 % and 3.2 %, respectively.Therefore, a close mapping was confirmed between the simulated variables and data obtained from an industrial-scale reforming plant.These results show that the presented kinetic model can reliably be utilized to monitor the operation of the catalytic reforming plant.

List of symbols and abbreviations
Popis simbola i kratica

F i g . 1 -
Catalytic reforming flowchart (semi-regenerative) S l i k a 1 -Dijagram toka poluobnavljajućeg katalitičkog reformiranja T a b l e 1 -Specification of feed of catalytic naphtha reforming process T a b l i c a 1 -Sirovina za katalitičko reformiranje benzina

F i g . 2 -F i g . 3 -F i g . 4 -
Simulated outlet temperature of 1 st reactor against actual values S l i k a 2 -Simulirana ulazna temperatura 1. reaktora u usporedbi sa vrijednostima Simulated outlet temperature of 2 nd reactor against actual values S l i k a 3 -Simulirana ulazna temperatura 2. reaktora u usporedbi sa stvarnim vrijednostima Simulated outlet temperature of 3 rd reactor against actual values S l i k a 4 -Simulirana ulazna temperatura 3. reaktora u usporedbi sa stvarnim vrijednostima

F i g . 5 -
Simulated volume yield of gasoline against actual values S l i k a 5 -Simulirano obujamsko iskorištenje benzina u usporedbi sa stvarnim vrijednostima F i g .6 -Simulated RON of product against actual values S l i k a 6 -Simulirani RON produkta u usporedbi sa stvarnim vrijednostima