Kinetic study and Modeling of heavy Naphtha Catalytic Reforming process in AL-Daura Refinery

In the present work, kinetics and modeling of heavy naphtha catalytic reforming process in AlDaura refinery-Midland refineries Company were studied. A proposed reaction scheme involving (15 pseudo components) connected together by a network of 30 reactions for components in the C6-C8+ range have been modeled. In the present work, kinetics and modeling of heavy naphtha catalytic reforming process in AL-Daura refinery-Midland refineries Company were studied. A proposed reaction scheme involving (15 pseudo components) connected together by a network of 30 reactions for components in the C6-C8+ range have been modeled. The proposed model has been solved numerically using the 4th order Runge–Kutta approach. Alteration of components and temperature, with time and reactor length was evaluated. Results showed that the rate of formation of aromatics is becoming slower as the reactants proceed to the third reactor. The catalytic reaction rates in the reformer are well represented by the HougenWatson Langmur-Hinshelwood (HWLH) type form. The deactivation of catalyst causes the reactor behavior to continue changing over a longer period of time. This clearly seems to pay off in the scenario where coke deposition plays such a major role. It was also found that the rate of coke formation increases with the progress from first to the last bed, so keeping a decreasing inlet temperature profile from first to the last bed would lead to more uniform coke content in each bed. The production rate of reformate has a negative impact on the octane number. Temperature drop across the first reactor (~ 45C) is larger than the temperature drops across the other two reactors (10-12C). This could be related to the endothermic reaction rate which is faster in the first reactor. The results show that perfect agreement of temperatures, compositions, and fractions molar flow rate at the exit of the third reactor is obtained between predicted values and industrial values.This confirmed the reliability of the present model.


Introduction:
The Second World War played a significant role in catalyzing the technological growth of the petroleum industries and catalytic reforming was no exception. Catalytic reforming of naphtha or mixture of naphtha with a certain amount of cracking oil is a process of great interest to the petrochemical industry for the production of aromatic compounds that are raw materials for plastics, elastomers and resins manufacture. Catalytic reforming unit uses naphtha or cracking oil as feedstock to produce rich aromatic compounds and high octane value liquid products through reactions such as aromatization, cyclization, and hydrocracking. At the same time, it produces hydrogen (H 2 ) and liquefied petroleum gas (LPG) as by-products. The design or simulation of the catalytic reforming reactor is very difficult because of complicated components of catalytic reforming feedstock, higher operating temperature of the system, and the complicated reactions in the reactor [1]. A lot number of articles have appeared in the literature studying the chemistry of reforming reactions. Recently some informative papers published on the modeling of catalytic naphtha reforming would be mentioned. Miguel et al. reforming unit, the effect of benzene precursors in the feed in both laboratory and commercial reactors was also simulated. Zaidoon [3] made a one-dimensional steady-state mathematical model of a semi-regenerative naphtha catalytic reforming process including four reactors. The model incorporated a kinetic model involving 24 components, 1 to 11 carbon atoms for paraffins (n and iso) and 6 to 11 carbon atom for naphthenes and aromatics with 71 reactions.
The model presented the composition, temperature and pressure distributions along the four reforming reactors. The results showed good agreement between the reformate composition of proposed model with the experimental reformate composition. Raouf et al. [4] carried out experimental and theoretical studies to describe the reaction kinetics in heavy naphtha catalytic reforming process) on tri-metal supported on Al 2 O 3 catalysts using catalytic reforming process.
They investigated the dehydrogenation, dehydrocyclization, and hydrocracking reactions to characterize the catalysts performance toward higher activity and selectivity to desired products. The concentration, conversion, and temperature profiles have been studied and the results showed a good agreement between experimental and simulation model with a deviation ranging from 4.18% to 19.50%. Zahedi et al. [5] investigated the steady state and dynamic simulation of a fixed bed industrial naphtha reforming reactors. They used a heterogeneous model to investigate the performance of the reactor. The formulated models were validated against measured process data of an existing naphtha reforming plant. The results of simulation in terms of components yields and temperature of the outlet were in good agreement with empirical data. Cochegrue et al. [6] carried out a modeling of refining processes using metalacid bifunctional catalysts involves an exponentially increasing number of species and reactions, which may rapidly exceed several thousand for complex industrial feedstocks. They reported that due to the large number of elementary steps occurring in bifunctional catalysis, a computer generation of the reaction network according to simple rules offered an elegant solution in such a case. They stated that for several refining processes, reasonable assumptions on the equilibria between species allow to perform an a posteriori relumping of species, thus reducing the network size substantially while retaining a kinetic network between lumps that is strictly equivalent to the detailed network. They illustrated this posteriori relumping approach and successfully applied to the kinetic modeling of catalytic reforming reactions.

E 4
The present study aims to analyze the process of catalytic naphtha reforming in AL-Daura refinery and develop, via kinetic mathematical expressions, a valid reformer model that relates the input and output variables of the process and associated coefficients and apply a suitable solution technique to the mathematical statement of the problem.

2-Process Description of Catalytic Naphtha Reforming Unit in Daura Refinery:
The flow diagram of the Daura catalytic naphtha reforming unit modeled in this work is displayed in figure (1). The feedstock to the catalytic reformer are the gasoline boiling range hydrocarbons (80-180°C), collectively known as naphtha, which in the raw crude. These are separated from the other crude hydrocarbons by fractionation in the crude column. The gasoline, or naphtha, obtained directly from crude oil is always of low octane number; therefore the refiner cannot use it as a motor fuel directly, so that the purpose of the catalytic reformer process is to raise the octane number of gasoline to a higher octane number.
The naphtha used as a catalytic reformer feedstock usually contains the mixture of paraffins, naphthenes, and aromatics in the carbon number range (C 5 -C 10 ).
Tables (1,2) list the operating and geometric variables values associated with the process. As shown in figure (1), total reactor charge is heated, at first by exchange with effluent from the last reactor, and is finally brought up to the first reactor inlet temperature in the first charge heater also called as first inter heater. The reactor effluent-to-feed exchanger recovers the heat from the reactor effluent and provides it to the reactor feed. Thus it is one key to energy conservation in a catalytic reformer. The reactor effluent which may as high as (470 to 475°C) must be cooled to (250°C) for flash separation of hydrogen from reformate.
After passing through the reactor effluent-to-feed exchanger and the charge heater, the total reactor charge is (100% vapor), is up to reaction temperature, and is ready to contact the reforming catalyst. The most commonly used catalyst type is platinum on alumina support. The flow scheme figure (1) shows three reactors in series.
The reactor feed was raised to the proper temperature for the reforming reactions to occur when the charge contacts the catalyst. As shown in figure (1), total reactor charge is heated, at first by exchange with effluent from the last reactor, and is finally brought up to the first reactor inlet temperature in the first charge heater also called as first inter heater. (Effluent is total vapor flowing out of the last reactor.) The reactor effluent-to-feed exchanger recovers the heat from the reactor effluent and provides it to the reactor feed.

Modeling of a Catalytic Naphtha Reformer: -3
A kinetic model defines the rates of various reactions and the associated heat and material balance using a system of equations. Developing a kinetic model from basic reaction kinetics and fundamental engineering relationships is a major investment in engineering resources. In the present kinetic model (C 1 toC 5 ) hydrocarbons are specified as light paraffins and the (C 6 to C 8+ ) naphtha cuts are characterized as isoparaffins, normal paraffins, naphthenes and aromatics. Mole fractions of naphtha cuts with more than nine carbon number are very low and therefore are lumped together.

Assumptions of the kinetic model
The characteristics and the assumptions made in the kinetic model in this work are listed below: a-The selected reversible reaction network was proposed by (Arani et al, 2009) [7] is shown in f-Van't Hoff equation [9] would be used to evaluate equilibrium constraints of reversible reactions.
g-It is assumed that the reaction rate equations obtained by analyzing experimental conversions for (C 6 -C 8 ) hydrocarbons suggested by [8] are applicable in the present work.

Selection of deactivation model
The empirical Voorhies correlation [10] for coking in the catalytic cracking of gasoil, Where (t) is the process on-stream time. This correlation has been widely accepted and generalized beyond the scope of the original contribution [11]. Obviously coke is formed from the reaction mixture itself therefore the rate of coking must depend on the composition of the reaction mixture, the temperature, and the catalyst activity and not only on the process on stream time. Several of the overall reactions in catalytic reforming require formation of olefinic

No.15 Journal of Petroleum Research & Studies
(JPR&S) E 9 intermediates in their elementary reaction sequence. Ultimately, these olefinic intermediates lead to coke formation and subsequent catalyst deactivation. Catalyst deactivation is primarily caused by the blockage of active sites due to the coke formed from these olefinic intermediates.
For example, in the ring isomerization reaction, methycyclopentane forms a methyl cyclopentene intermediate in its reaction sequence to cyclohexane. The intermediate can be further dehydrogenated to form methylcyclopentadiene, a coke precursor. A methodology of characterizing the deactivation of a catalyst by coke deposition due to olefins formation and cyclization was developed by [12] and [8]. This approach is utilized in the present work and described as follow: Coke deposition causes a deactivation of the catalyst which can be described by introducing a deactivation function, (η) that multiplies the reaction rates at zero coke content [16].
The same approach can be taken to describe the decrease of coking rate itself where, The coke content differential equation is integrated in time to obtain the coke profiles in time.
The deactivation function was related to the coke content on the catalyst, i.e., η = f (C c ).
Amongst the several empirical expression tested for the deactivation function, an exponentially decreasing function, η = ( ), led to the best global regression of the isomerization data [12]. Various reaction paths were considered by [8]. They proposed that, the three significant contributions to coke formation which were selected to describe the rates of coke formation reactions are shown schematically in

Modeling of the fixed-bed reactors
The following assumptions would be utilized to formulate the modeling of reactor: a-The flow is assumed to be well distributed on the catalyst by good designed inlet distributor.
b-The reaction rate expressions followed the Hougen-Watson Langmur-Hinshelwood (HWLH) type form which takes into account the heterogeneous nature of the reactions, while the uncertainty associated with internal and external diffusion effects is lumped into the rate parameters [13], [14].
c-Axial dispersion of mass is not significant, since the tube length to pellet diameter ratio is greater than 50, while that of heat is also negligible if the same ratio is above 300 [15].
d-In the absence of the dispersion effects, it is possible to model the reactor as a pure "plug-flow" type reactor, in which the fluid is taken to move as a plug through the reactor.

bed reactor modeling equations -3.4 Fixed
The reactor model consists of the following differential equations for mole and energy balance, which are integrated through the reactor system. (Ergun's equation) for calculating the pressure drop is also included in the differential format, these equations are found elsewhere [16].

Model solution
The model of the catalytic naphtha reformer described in the earlier sections contains a number of parameters whose initial values were estimated from the literature data and data obtained from process log-sheet of reforming unit in AL-Doura refinery, and sometimes under varied operating conditions. The data provided contained information regarding the following operating variables. Equations 5, 6 and 7 are first order nonlinear differential equations which could be solved simultaneously by finite difference approach using Runge-Kutta method [16,22]. l properties.

Calculation of the physica
Physical properties were estimated using the empirical correlation developed by (Riazi and Daubert, 1980)

Parameters estimation 5.
The most critical step performed in the model effort was to estimate the unknown parameters on the basis of provided data. The following different types of parameters had to be determined in this work. * Parameters used for catalyst deactivation: Pre-exponential factors and activation energies of coke formation reactions, deactivation constant, the task involved was to evaluate these parameters such that for a given set of input variables the model predicted the output that matched with the output obtained from the industrial data. Since the number of reactions was (15), the number of pre-exponential factors and the activation energies themselves amounted to (30). This would give an idea about the dimensionality of the problem involved. The input variables to the system are: molar feed rate of each chemical species in feed naphtha, reactor bed inlet temperatures, reactor pressure, and recycle ratio. The adjustable parameters in the model used for fitting the data were the pre-exponential factors and activation energies, which basically decide the rate constants of reforming reactions, and the adsorption equilibrium constants accounting for adsorption of chemical species present in the reaction mixture. The kinetic scheme representing the reactions taking place consists of (15) reactions in range. Since each reaction rate is characterized by a unique set of activation energy and preexponential factor, this implies that rate parameters by themselves account for (30) constants to be evaluated. However, the number of unknown parameters can be reduced by judicious selection of certain parameters to be estimated or pre-decided based on published literature data and studies of reforming reactions carried out in the past. A standard least squares approach was adopted to solve the regression problem [18].
The minimization technique used to evaluate the model parameters is presented in equation  [8] reported that analysis of the data collected over a sufficient run length of the catalyst in a heavy naphtha reforming plant, operated at pressure=3 bar and inlet temperature = 763 K , for α=14.95, the catalyst deactivation function η vs. coke content is presented in Table (7). These values were inserted into eq.2 to account for coking rate in the present work.   This drop in temperature instantly cools the reaction mixture and decelerates the rates of all the reactions. Consequently, the temperature profile shifted down in the latter part of the bed. The total temperature drop in the first bed is about (32°C) and the reaction mixture needs to be reheated before any further reactions can be carried out. After passing through the first interheater, the reaction mixture enters the second catalyst bed. Most of the remaining naphthenes are dehydrogenated in the second bed and this leads to a net temperature drop of about (16°C).
The reaction mixture is reheated again to push the reaction rates of paraffins to naphthenes which are the most desirable reactions from reforming point of view, as far as possible and passed through the third bed. These reactions are the most difficult to carry out and slowest to proceed followed by dehydrogenation naphthenes to aromatics, creates an endothermic temperature drop in the third bed.

Fig. (3) Temperature profile across the catalyst beds.
However, the rate of temperature drop is relaxed by increasing rates of hydrocracking reactions, which are exothermic in nature. The temperature drops in third reactor are about (10 to 12°C). Our results are in agreement with the findings of [3,5]. Table (   and aromatics) as well as lighter cracked gases. As pointed out earlier, the most significant reaction in the first bed is the dehydrogenation of naphthenes. This is realized by sharp decrease in naphthene concentration through the catalyst bed and consequent increase of the aromatics formed. After the product is reheated to (486 °C) for second reactor inlet, the remaining naphthenes, particularly, five-carbon ring naphthenes are isomerized to six-membered rings, which are subsequently dehydrogenated to aromatics. As the five-carbon ring naphthenes are reacted and proceed, the +) paraffins concentration, starts to fall in the second bed.

Composition profile along the length of the catalyst bed
Isomerization of paraffins also continues to proceed in the first and the second bed. By the end of second bed paraffin isomers are at equilibrium. In third bed, the ring closure reactions of paraffins are carried out, which are also accompanied by hydrocracking reactions of paraffins.
Hydrocracking reactions proceed at very similar rates to ring closure. Consequently, concentration of paraffins keeps dropping. Although figure (4) shows a slight decline in aromatics concentration due to ring closure reactions, there is an increase in the net amount of aromatics. This may be attributed to the effect of much higher cracked products presence in the reactor. The findings of [20,23] were in agreement with our results.   Profile of octane number and reformate yield -6.3

6.5-Variation of Volumetric yield and octane number over the life of catalyst
In figure (7) variation of octane number and reformate yield over the catalyst run length is plotted. As can be seen, octane number and reformate yield always vary in opposite directions. This is quite logical because of the fact that major reforming reactions, such as dehydrogenation of naphthenes and ring closure of paraffins are accompanied by a loss in volume. As expected, the octane number keeps declining with time since the reforming reactions are adversely affected due to deficiency of catalyst activity by coke formation. The reformate yield increases at the same time, because the major source of the loss of reformate yield, i.e., hydrocracking takes place to a lower extent. In actual operation, reactor inlet temperatures are usually raised to adjust octane. So these plots, which are prepared at a constant inlet temperature, may not resemble the operating scenario. Nonetheless, they are useful from the perspective of purely studying the effect of coking on reformate quality and yield, and for model verification.

Conclusions 7.
An industrial naphtha reactor was modeled by heterogeneous model. The proposed model has been solved numerically using the 4th order Runge-Kutta approach. Alteration of components and temperature, with time and reactor length was evaluated. Some conclusions could be withdrawn from the present study: 1-The rate of formation of aromatics is becoming slower as the reactants proceed to the third reactor. This may be attributed to a continuous decrease in the concentration of reactants (i.e., paraffins and naphthenes).

2-The catalytic reaction rates in the reformer are well represented by the Hougen-Watson
Langmur-Hinshelwood (HWLH) type form.
3-The deactivation of catalyst causes the reactor behavior to continue changing over a longer period of time. This clearly seems to pay off in the scenario where coke deposition plays such a major role.
4-The rate of coke formation increases with the progress from first to the last bed, so keeping a decreasing inlet temperature profile from first to the last bed would lead to more uniform coke content in each bed 5-The production rate of reformate has a negative impact on the octane number.
6-Temperature drop across the first reactor is larger than the temperature drops across the other two reactors. This could be related to the endothermic reaction rate which is faster in the first reactor due to higher reactant concentration.
7-The reforming model contains a rich set of model features with experimental observations suggests that the model is a good platform for additional process specific refinement.