A study on naphtha catalytic reforming reactor simulation and analysis

A study on naphtha catalytic reforming reactor simulation and analysis 1 first-author †E-mail:lks-ly@163.com †E-mail:lks-ly@163.com lks-ly@163.com 1 2 School of Petrochemical Engineering, Shenyang University of Technology, Liaoyang 111003, China School of Petrochemical Engineering, Shenyang University of Technology, Liaoyang 111003, China School of Advanced Professional Technology, Shenyang University of Technology, Liaoyang 111003, China School of Advanced Professional Technology, Shenyang University of Technology, Liaoyang 111003, China A naphtha catalytic reforming unit with four reactors in series is analyzed. A physical model is proposed to describe the catalytic reforming radial flow reactor. Kinetics and thermodynamics equations are selected to describe the naphtha catalytic reforming reactions characteristics based on idealizing the complex naphtha mixture by representing the paraffin, naphthene, and aromatic groups by single compounds. The simulation results based above models agree very well with actual operation unit data. Naphtha Catalytic reforming Kinetics Simulation and analysis INTRODUCTION 1 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) and liquified petroleum gas (LPG) as its 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. Much research (Pontes et al., 1999 1999 ) on this subject had been carried out. Results have been published on the process, reactions and kinetics. A conventional naphtha catalytic reforming unit consists of 3 or 4 radial flow reactors in series operated under adiabatic conditions. The temperature and the H 2 /HC molar ratio are the most important process variables. A bifunctional catalyst with Pt and a second metal is generally used. In addition, some novel processes (Melnikov and Makarova, 1998 1998 ; Smith, 1960 1960 ; Bolz et al., 1999 1999 ) of catalytic reforming had been developed. The kinetics of catalytic reforming has been attracting the attentions of many researchers. A successful kinetic analysis proposed by Smith ( 1959 1959 ), was based on idealizing the complex naphtha mixture by representing the paraffin, naphthene, and aromatic groups by single compounds. The nature of the reforming reactions in which the number of carbon atoms was the same for the precursor and product justified this procedure. Reaction rates were derived from the assumption of a homogeneous system. Jorge and Eduardo ( 2000 2000 ) proposed a lumped kinetic model for the naphtha catalytic reforming process. The model utilizing lumped mathematical representation of the reactions taking place was in terms of isomers of the same nature. Arrhenius-type variation was added to the model in order to include the effect of pressure and temperature on the rate constants. Many other kinetic models (Jerzy, 1999 1999 ) are not described here. A catalytic reforming unit with 4 reactors in series was analyzed and simulated in this paper. Kinetics proposed by Smith was adopted to describe the reactions occurring during catalytic reforming. Simulation results were compared with the actual data. MATHEMATICAL MODEL 1 Basic assumption 2 The structure of a catalytic reforming reactor is shown in Fig. 1 1 . The following assumptions are proposed on the basis of the features of catalytic reforming reactions, reactor and its operation conditions: 1. The temperature along the axial direction (S) of the radial flow reactor is homogeneous, that is d T s =0. 2. The concentration along the axial direction of the radial flow reactor is homogenous, that is d N s =0. 3. Reactions are in homogenous phase. Based on the above hypothesis, the temperature and concentration gradients along the axial direction can be neglected and only radial direction variables are considered. In fact, the fixed bed reactor is considered as a homogeneous reactor. Dynamics 2 If some pseudo ingredients (i) are appropriately used to simplify the feedstock of naphtha catalytic reforming, four dominant idealized reactions (Smith, 1959 1959 ) can be used to describe the catalytic reforming process so that the catalytic reforming system can be simplified significantly. These four reactions are called model reactions. The chemical reaction equations are as follows Cycloalkane⇄Arene (A)+3H 2 (H) Cycloalkane (E)+H 2 ⇄Alkane (P) Hydrocracking of alkane Hydrocracking of cycloalkane According to Smith ( 1959 1959 ), hydrocracking produces nearly the same amounts of moles of C 1 ~C 6 cuts. Therefore, the model reaction Eq.( 3 3 ) and Eq.( 4 4 ) can be represented by the following formula: $${\text{C}}_n {\text{H}}_{2n + 2} + \left( {\frac{{n - 3}}{3}} \right){\text{H}}_2 \to \frac{n}{{15}}{\text{C}}_1 + \frac{n}{{15}}{\text{C}}_2 + \frac{n}{{15}}{\text{C}}_3 + \frac{n}{{15}}{\text{C}}_4 + \frac{n}{{15}}{\text{C}}_5$$ $${\text{C}}_n {\text{H}}_{2n} + \left( {\frac{n}{3}} \right){\text{H}}_2 \to \frac{n}{{15}}{\text{C}}_1 + \frac{n}{{15}}{\text{C}}_2 + \frac{n}{{15}}{\text{C}}_3 + \frac{n}{{15}}{\text{C}}_4 + \frac{n}{{15}}{\text{C}}_5$$ where, n is the number of carbon atoms. For a certain reforming feedstock, the n value can be obtained through the average mole weight M f (kg/kmol) of the feedstock and the family of component x P , x E , x A . Their relation is n = M f −2 x p +6 x A where, x p + x E + x A =1 and x P , x E , x A are mole fractions of alkane, cycloalkane, and aromatics, respectively. The corresponding empirical reaction rate equations (Rase, 1977 1977 ) are based on model reaction Eqs.( 1 1 )~( 4 4 ). Cycloalkane transformed into arene: r 1 = ηk p1 ( P E − P A P H 3 / K p1 ) k p1 = 9.87exp(23.21−34750/1.8 T ) Cycloalkane transformed into alkane: r 2 = ηk p2 ( P E P H − P P / K p2 ) k p2 =9.87exp(35.98−59600/1.8 T ) Hydrocracking of alkane: r 3 = ηk p3 P P / P k p3 =exp(42.97−62300/1.8 T ) Hydrocracking of cycloalkane: r 4 = ηk p4 P E / P k p4 =exp(42.97−62300/1.8 T ) where, η is corrective coefficient of catalyst; k p1 is reaction rate constant corresponding to r 1 , kmol/(h·kg cat·MPa); k p2 is reaction rate constant corresponding to r 2 , kmol/(h·kg cat·MPa 2 ); k p3 is reaction rate constant corresponding to r 3 , kmol/(h·kg cat); k p4 is reaction rate constant corresponding to r 4 , kmol/(h·kg cat); K p1 is equilibrium constant of cycloalkane transformed into arene, (MPa) 3 ; K p2 is equilibrium constant of cycloalkane transformed into alkane, MPa −1 ; P is pressure of system or component fractional pressure, MPa; r 1 is reaction rate of cycloalkane transformed into arene, kmol/(h·kg cat); r 2 is reaction rate of cycloalkane transformed into alkane, kmol/(h·kg cat); r 3 is reaction rate of hydrocracking of alkane, kmol/(h·kg cat); r 4 is reaction rate of hydrocracking of cycloalkane, kmol/(h·kg cat). Occurrence of hydrocracking reaction occurring in the catalytic reforming process is the main causative factor catalyst deactivation. For different periods of catalytic activity, the carbon accumulation extent are not the same. Therefore, the values η are different from each other. For new catalyst, η =1 and in other conditions, 0 < η < 1. The value of η can be calculated based on working conditions and time span of the catalytic activity. Thermodynamics 2 The most important reaction in the catalytic reforming is the transformation of cycloalkane into aromatics. It is a rapid reaction which approaches to equilibrium in very short time. For reactions of cycloalkane transformation into aromatics and alkane, their reaction equilibrium constants are as follows (Smith, 1959 1959 ): K p1 = P A P H 3 / P E =1.04×10 −3 exp(46.15−46045/1.8 T ) K p2 = P P /( P E · P H )=9.87exp(8000/1.8 T −7.12) For the above four model reactions, considering the reaction enthalpy and the number of moles of hydrogen consumed or released during the reaction, the reaction enthalpies based on the number of moles of hydrogen (Δ H 1 reaction enthalpy of model reaction (1), kJ/(kmol H 2 ); Δ H 2 reaction enthalpy of model reaction (2), kJ/(kmol H 2 ); Δ H 3 reaction enthalpy of model reaction (3), kJ/(kmol H 2 ); Δ H 4 reaction enthalpy of model reaction (4), kJ/(kmol H 2 )) are as follows: Δ H 1 =71038.06 kJ/(kmol H 2 ) (endothermic) Δ H 2 =−36953.33 kJ/(kmol H 2 ) (exothermic) Δ H 3 =−56597.54 kJ/(kmol H 2 ) (exothermic) Δ H 4 =−51939.31 kJ/(kmol H 2 ) (exothermic) The reactants and products of a catalytic reforming unit can be considered as ideal gases. The relation between its heat capacity ( C p , kJ/(kmol·°C)) and temperature is C p =4.1868( a + bT + cT 2 + dT 3 ) where, a , b , c and d can be obtained from physical properties manual and is listed in Table 1 1 (Lu, 1982 1982 ; Reid et al., 1977 1977 ). ID a b ×10 3 c ×10 6 d ×10 9 H 2 6.483 2.215 −3.298 1.826 C 1 4.598 12.45 2.680 −2.703 C 2 1.292 45.84 −16.57 2.081 C 3 −1.009 73.15 −37.89 7.678 C 4 2.266 79.13 −26.47 −0.674 C 5 −0.866 116.4 −61.63 13.67 P −1.456 184.2 −100.24 21.15 E −14.789 187.3 −106.0 22.37 A −5.817 122.4 −66.05 11.73 Model of reactor 2 The physical model for catalytic reforming radial flow reactor is shown in Fig. 2 2 . The outer radius of the catalyst bed is R o and the inner radius of the catalyst bed is R i . For a element d R at ( R , R +d R ), the inlet material stream into it is N i and the temperature is T . The outlet material stream is N i +d N i and the temperature is T +d T . With mass and enthalpy balances for the element combining kinetic and thermodynamic equations, the concentration and temperature distributions along the reactor radius can be derived as: d N A /d R =2π RLρ b r 1 d N E /d R =2π RLρ b (− r 1 − r 2 − r 4 ) d N P /d R =2π RLρ b ( r 2 − r 3 ) d N H /d R =2π RLρ b (3 r 1 − r 2 −(( n −3)/3) r 3 −( n /3) r 4 ) d T /d R =(2π RLρ b )/(Σ F i C p i )×{3 r 1 (−Δ H 1 )+ r 2 (−Δ H 2 )+[( n −3)/3] r 3 (−Δ H 3 )+( n /3) r 4 (−Δ H 4 )} Eq.( 19 19 ) to Eq.( 22 22 ) describe changing rates of arene, cycloalkane, alkane, and hydrogen along the bed layer radius. For the adiabatic fixed bed radial flow reactor, its temperature distribution is represented by Eq.( 23 23 ). F is feedstock, kmol/h; N is flow rate, kmol/h; L is height of catalyst bed in reactor, m; ρ b is bulk density of catalyst, kg/m 3 . SIMULATION OF THE REACTOR 1 Fig. 3 3 shows a flow diagram with four reactors in series. With conditions at the reactor inlet as boundary conditions, the concentration and temperature distributions can be calculated by solving grouped Eq.( 19 19 ) to Eq.( 23 23 ) with fixed step, four order Runge-Kutta method. The computing diagram is shown in Fig. 4 4 . Four catalytic reforming radial flow reactors of the Chemical Plant No.1 of the Liaoyang Petrochemical Fiber Company were analyzed and simulated with this model. Basic data 2 Liaohe straight-run distillation naphtha (Liaohe Petroleum Field, China) was selected as reforming feedstock. Its distillation temperature range is 65 °C~145 °C, d 4 20 =0.7314. Engler distillation data is listed in Table 2 2 . The feedstock of catalytic reforming reactors and recycle hydrogen content data (Yong, 1995 1995 ) is shown in Tables 3 3 and 4 4 . The inlet temperatures of all reactor I, II, III and IV were 521 °C. The operating pressure was 1.730~1.408 MPa and recycling hydrogen flow rate was 3.1372×10 4 Nm 3 /h. Distillation fraction (%) IBP 10 50 90 EBP b.p. (°C) 77 85 114 135 157 Note: B: Benzene ID Recycle hydrogen content (%, v) H 2 84.94 C 1 5.85 C 2 3.88 C 3 2.93 C 4 2.17 C 5 0.19 B 0.04 Note: T: Toluene; EB: Eshylbenzene; PX: Parexylene; MX: Meta-xylene (1,3-dimethylbenzene); OX: Orthoxylene; C 9 + A: Arene of C 9 and over ID Reforming feedstock content (%, m) C 4 0.02 C 5 0.15 C 6 4.55 C 7 31.56 C 8 18.65 C 9 4.44 C 10 + 1.57 C γ C 5 0.24 m-C γ C 5 1.03 C γ C 6 3.21 C γ C 7 20.26 C γ C 8 19.12 C γ C 9 4.54 C γ C 10 + 0.02 B 0.49 T 3.54 EB 0.35 PX 0.39 MX 2.03 OX 1.10 C 9 + A 0.74 Simulation calculation 2 Simulation calculation was carried out using data listed in Tables 2 2 – 4 4 and the results are listed in Table 5 5 . Because people usually focus on the reactor production ability, conversion rate, hydrogen generated by reforming, temperature of catalyst bed, etc., the temperature, arene, cycloalkane, alkane, and hydrogen varying with the radius of the beds are listed in Table 5 5 . Note that the data used in the simulation calculation is calibration data in the latest operation period. The catalyst correction coefficient η =0.432 at this period. Reactor ID Bed radius (mm) Bed temperature (°C) Arene (kmol/h) Cycloalkane (kmol/h) Alkane (kmol/h) Hydrogen (kmol/h) I 510 521.00 16.180 90.050 76.630 1257.870 460 496.88 32.979 71.642 76.650 1300.845 410 486.04 39.943 64.052 76.603 1320.133 360 479.48 44.096 59.601 76.565 1331.804 310 475.03 46.879 56.644 76.535 1339.686 260 471.88 48.843 54.569 76.513 1345.276 210 469.60 50.254 53.084 76.496 1349.304 170 468.26 51.084 52.213 76.486 1351.676 II 585 521.00 51.084 52.213 76.486 1351.676 535 503.79 62.982 39.260 75.398 1382.185 485 495.27 68.681 33.255 74.850 1397.267 435 489.94 72.194 29.608 74.500 1406.714 385 486.27 74.589 27.136 74.251 1413.266 335 483.60 76.334 25.363 74.066 1417.946 285 481.62 77.619 24.055 73.926 1421.453 235 480.14 78.579 23.083 73.818 1424.073 185 479.04 79.286 22.369 73.737 1426.066 170 478.78 79.457 22.196 73.717 1426.474 III 760 521.00 79.457 22.196 73.717 1426.474 710 510.04 87.774 14.685 70.250 1445.538 660 504.52 91.956 11.158 68.188 1455.138 610 501.13 94.556 9.090 66.709 1461.024 560 498.83 96.345 7.744 65.560 1464.998 510 497.17 97.652 6.811 64.670 1467.839 460 495.94 98.643 6.138 63.862 1469.948 410 495.00 99.412 5.640 63.220 1471.551 360 494.27 100.017 5.262 62.683 1472.787 310 493.71 100.495 4.979 62.237 1473.748 260 493.27 100.872 4.762 61.871 1474.491 210 492.93 101.163 4.599 61.578 1475.058 170 492.72 101.342 4.501 61.393 1475.404 IV 825 521.00 101.342 4.501 61.393 1475.404 775 519.26 103.680 4.447 56.750 1477.891 725 517.69 105.670 4.227 53.094 1480.240 675 516.33 107.329 3.977 50.175 1482.287 625 515.19 108.708 3.743 47.756 1484.025 575 514.24 109.859 3.537 45.767 1485.487 525 513.44 110.822 3.361 44.105 1486.714 475 512.77 111.630 3.213 42.708 1487.743 425 512.20 112.308 3.089 41.533 1488.603 375 511.74 112.875 2.985 40.547 1489.319 325 511.35 113.346 2.900 39.726 1489.910 275 511.03 113.732 2.830 39.050 1490.393 225 510.78 114.041 2.775 38.507 1490.777 175 510.59 114.280 2.733 38.085 1491.074 170 510.57 114.301 2.729 38.049 1491.099 RESULTS AND DISCUSSIONS 1 It is obvious, from the data listed in Table 4 4 , that in reactor I and II, the main reaction happened is cycloalkane aromatization, is an endothermic reaction and the bed layer temperature drop of these two reactors is larger. In reactor III and IV, alkane dehydrogenization, cyclization, and hydrocracking reactions are dominant there. At the same time, the cycloalkane produced there through cyclization continues to aromatize, but the amount is not large. Furthermore, a large amount of heat was released by cyclization and hydrocracking reactions. Therefore, the bed temperature drop in these two reactors, the last one especially, is smaller. It should be pointed out that under some suitable conditions the reversal reaction represented by Eq.( 10 10 ), alkane transformation into cycloalkane becomes more important. On the one hand, aromatic production yield can be increased. On the other hand, non-aromatic content in reforming arene can be decreased and aromatic compounds transformation rate can be increased. Based on data in Table 5 5 , the temperature distribution curve in the catalyst bed, that is, the temperature variation within the radius of the catalyst bed, is shown in Fig. 5 5 showing that at the inlet side (i.e. higher value of R ) of the reactors I and II, the bed temperature varies rapidly and at the outlet side (i.e. smaller R ) the bed temperature varies slowly. Because there was more catalyst in the inlet where the reactants content was higher than that at the outlet. Therefore, more aromatic compounds were generated and more heat was absorbed at the inlet side. The situation at the outlet side was totally opposite. For reactor III and IV in spite of the trend being similar to that of reactors I and II, the temperature varied more gently because the reactions were exothermic. For comparison, the reactors calibration values are also shown in Fig. 5 5 . The simulation results and actual values of the reforming product are shown in Table 6 6 . These results indicate that the model and simulation in the present paper are satisfactory. 2 Item 3 Flow rate of products (kg/h) 2 Productivity of arene (%, m) Depentane oil Arene Non-arene Actual (Yong, 1995 1995 ) 15585 11185 4399.65 57.96 Simulation 15580.96 11258.49 4322.47 58.33 Error (%) −0.026 0.657 −1.754 0.638 CONCLUSION 1 1. The three assumptions in the present model are reasonable. The results indicate that the one-dimensional homogenous model is suitable for describing catalytic reforming radial flow reactor. 2. Once pseudo ingredients were selected suitably. The present empirical kinetics model can be well applied to catalytic reforming reactor for simulation and optimization. 3. This method can be used to predict aromatic capacity, yield, transformation rate, bed temperature, and so on for new operation conditions. Therefore, the optimum operating parameters can be obtained to yield maximum profit. References CR Bolz , KR Graziani , MN Harandi , Y Ozawa , CM Sorensen . Radial-Flow Reactors Containing Multiple Catalyst Beds, Especially for Naphtha Reforming and Upgrading . 1999 . PCT Int. Appl. WO 9920384 . Radial-Flow Reactors Containing Multiple Catalyst Beds, Especially for Naphtha Reforming and Upgrading S Jerzy . On the kinetics of catalytic reforming with the use of various raw materials . Energy Fuels , 1999 . 13 ( 1 ): 29 - 39 . . On the kinetics of catalytic reforming with the use of various raw materials AJ Jorge , VM Eduardo . Kinetic modeling of naphtha catalytic reforming reactions . Energy Fuels , 2000 . 14 ( 5 ): 1032 - 1037 . . Kinetic modeling of naphtha catalytic reforming reactions H Lu . Manual of Petrochemical Industry Fundamental Data , 1982 . : Beijing Chemical Industy Press , (in Chinese) . Manual of Petrochemical Industry Fundamental Data VB Melnikov , NP Makarova . Modification of reforming: Without hydrogen circulation . Khim Tekhnol Topl Masel , 1998 . 23 ( 1 ): 9 - 13 . . Modification of reforming: Without hydrogen circulation LAM Pontes , MDC Rangel , DJM Mario . Catalytic reforming of naphtha . An Assoc Bras Quim , 1999 . 48 ( 3 ): 169 - 179 . . Catalytic reforming of naphtha HF Rase . Chemical Reactor Design for Process Plants , 1977 . Vol. 2 John Wiley & Sons, Inc , . Chemical Reactor Design for Process Plants RC Reid , JM Prausnitz , TK Sherwood . The Properties of Gases and Liquids , 1977 . 3rd Ed McGraw-Hill Co , 108 - 109 . . The Properties of Gases and Liquids RB Smith . Kinetic analysis of naphtha reforming with platinum catalyst . Chem Eng Progr , 1959 . 55 ( 6 ): 76 - 80 . . Kinetic analysis of naphtha reforming with platinum catalyst RB Smith . Reforming of Naphtha in Several Stages . US 2,965,560 . 1960 . Reforming of Naphtha in Several Stages W Yong . Calibration of E603 reforming catalyst . Liaohua Sci and Tech , 1995 . 16 ( 4 ): 16 - 19 . . Calibration of E603 reforming catalyst 10.1631/jzus.2005.B0590 2004-09-22 2004-12-17 2005-05-22 2022-01-05 Copyright © 2005, Journal of Zhejiang University Science 2005 Journal of Zhejiang University-SCIENCE B (Biomedicine and Biotechnology) 06 6B