Optimization of Semi Regenerative Catalytic Naphtha Reforming Unit to Enhance Octane Number and Reformate Yield

Transcription

1 Iranian Journal of Chemical Engineering Vol. 15, No. (Sring 018), IAChE Otimization of Semi Regenerative Catalytic Nahtha Reforming Unit to Enhance Octane Number and Reformate Yield I. Khosrozadeh 1, M. R. Talaghat *, A. A. Roosta 1 Research and Develoment Grou, Shiraz Refining Oil Comany, P.O. Box: , Shiraz, Iran Deartment of Chemical Engineering, Shiraz University of Technology, P.O. Box: , Shiraz, Iran ARTICLE INFO Article history: Received: Acceted: Keywords: Catalytic Nahtha Reforming, Octane Number Enhancement, Hydrogen Production, Otimization 1. Introduction Catalytic nahtha reforming is one of the most critical rocesses in the oil refinery industries. The rocess consists of three consecutive adiabatic reactors, and the objective is to convert virgin nahtha cuts with the boiling oint range of 0 to 00 o C into high octane gasoline, called reformate. In addition, hydrogen as a desirable roduct is roduced in this rocess [1-]. Hydrogen will be a requisite energy source in the ucoming future. Furthermore, due to environmental rotection rules, sulfur and nitrogen removal ABSTRACT Catalytic nahtha reforming is one of the most imortant rocesses in which low quality nahtha is converted into high octane motor gasoline. In this study, a mathematical model was develoed and used for investigation of the effect of temerature, ressure, hydrogen to hydrocarbon ratio on the octane number, the yield of roduct, and the undesirable henomena of coe deosition in a semi regenerative catalytic reforming unit. The result of the model was comared to the lant data to verify the model accuracy. Then, the model was used to find the otimal condition for the maximum value of octane number and yield of roduct and the minimum value of coe deosition. The otimum condition of the rocess was estimated using a genetic algorithm otimization method as an efficient otimization method. In the otimal condition, the octane number and the yield of roduct imroved by 0. % and 1. %, resectively, and the coe deosition reduced by.1 %. is one of the main uroses of refineries []. The best way to remove nitrogen and sulfur in the whole world from refined etroleum roducts and natural gas is through hydrodesulhurization and hydrodenitrogenation in the resence of a catalyst, converting them into hydrogen sulfide and ammonia. Hydrodesulhurization (HDS) is a catalytic chemical rocess widely used to remove sulfur (S) from natural gas and from refined etroleum roducts. Hydrodenitrogenation (HDN) is an industrial rocess for the removal of nitrogen from * Corresonding author: talaghat@sutech.ac.ir 5

2 Khosrozadeh, Talaghat, Roosta etroleum. Although organonitrogen comounds occur at low levels, they are undesirable because they cause oisoning of downstream catalysts. To achieve this goal, refineries use u hydrogen in large quantities for removing these contaminants [4-5]. Hence, attemts are made to enhance hydrogen roduction rate and hydrogen urity in refineries. According to the imortance of hydrotreating and hydrocracing techniques, studies should be focused on ugrading nahtha reforming rocess, which sulies a large quantity of required hydrogen for refineries. Pursuant to the nature of the reactions, the reforming rocess also roduces light ends (gases) such as roane and butane [6]. In the last decade, due to the wide economic imortance of nahtha reforming rocess, researchers have focused on finding new ways for ugrading nahtha reforming rocess. Rahimour et al. used the DE (differential evolution) otimization technique as a owerful tool to find the best oerating arameters (such as gasoline octane boosting) in the thermally couled nahtha reforming heat exchanger reactor [7]. Juarez et al. investigated the modeling and simulation of commercial reforming unit consisting of a series of four catalytic reactors for rediction of the temerature and reformate comosition rofiles [8]. Weifeng et al. roosed a multi-objective otimization strategy for an industrial nahtha continuous catalytic reforming rocess to enhance aromatics roduction rate. The rocess model is based on a 0-lumed inetics reaction networ and has been roved to be quite effective in terms of industrial alication [9]. Rahimour et al. used the differential evolution (DE) method to otimize the oerational conditions of a radial flow sherical reactor containing the nahtha reforming reactions. In this reactor configuration, the sace between the two concentric sheres is filled by catalyst. The dynamic behavior of the reactor has been taen into account in the otimization rocess [10]. In addition, many studies have been done formerly on the new catalyst rovision, showing better resistance against sintering and coe deosition on the catalyst. Benitez et al. and Boutzeloit et al. investigated the erformance of catalytic reforming rocess, through monometallic, bimetallic, and trimetallic catalysts [11-1]. Some researchers, such as Mazzieri et al. and Sugimoto et al., investigated the coe formation and regeneration of catalyst ending the catalytic reforming rocess [1-14]. The most oularly used tye of nahtha reforming unit is semiregenerative reformer, which consists tyically of to 4 consecutive fixed bed reactors. A simlified schematic for the semiregenerative nahtha reforming rocess is shown in Fig.1, and secific roerties of the feed and oerational conditions of nahtha reforming reactors are resented in Table 1. The main idea of the rocess is to convert araffins and nahthenes into aromatics. Treated nahtha is combined with a recycled gas stream containing mol % hydrogen; then, it is heated to desired temerature and asses through three consecutive adiabatic reactors and heaters between the reactors to reheat the stream into reaction temerature at design levels, before entering the next reactor [15-16]. The roduct of the last reactor is searated into liquid and gas hases in reactor roduct searator. The flashed gas hase, which is rich in hydrogen, is recycled; the liquid hases that chiefly consist of aromatics comounds and light ends are sent to a searation system to remove light gases from aromatics roducts. Iranian Journal of Chemical Engineering, Vol. 15, No. (Sring 018) 5

3 Otimization of Semi Regenerative Catalytic Nahtha Reforming Unit to Enhance Octane Number and Reformate Yield Recycle Gas Comressor L.P.G Furnace R.P.S DE C4 Nahtha Feed Reformate Figure 1. A schematic diagram of semi-regenerative catalytic reforming rocess. Table 1 Secifications of feed and roduct. TBP (true boiling oint) Nahtha feed Reformate IBP % % % % % FBP Parameter Value Unit Feed 56 m /h H /HC LHSV 1.9 h -1 Mw ave gr/mol Feed stoc (Mol %) Paraffin 49 Nahthene 6 Aromatic 15 This rocess is described by a continuous oeration over long eriods, decreasing catalyst activity due to coe deosition. The otimum reforming cycle length or the time between catalyst regenerations is determined by factors such as declining in reformate yield, secified amount of hydrogen decline, and refinery or reformer economics [15-16]. In order to maintain the conversion at a desired value, the temerature of the reactor is raised over time as the activity of catalyst decreases. The reforming oeration is shut down, and the catalyst is regenerated aroximately once each 6-4 months. The research octane number (RON) that can be achieved in semi-regenerative reforming is generally in the range of , deending on the feed stoc quality, gasoline quality, and quantity as well as the oerating conditions. Table shows the 54 Iranian Journal of Chemical Engineering, Vol. 15, No. (Sring 018)

4 Khosrozadeh, Talaghat, Roosta characterizations of nahtha reforming reactors and roerties of the catalyst used in Shiraz oil refinery. Table Reactors and catalyst characterizations. Parameter 1 st Reactor st Reactor dis (wt %) d ( m) A (m ) H (m) V (m ) W (g) L D (Kg/m ) d b (mm) Tyical roerties of catalyst Parameter Value d 1. Pt 0. Re 0.4 S a 0 ρ b 0. P v 0.6 ε 0.6 st Reactor Unit mm wt % wt % m g -1 Kgl -1 cm g Mathematical modeling A heterogeneous mathematical model is develoed by assembling the mass and energy balance on the nahtha reforming system. The mass balance rovides the variation of the concentrations, and the energy balance rovides the variation of temerature along the reactors. In addition, one of the main functions in nahtha reforming is to consider the ressure dro through the catalyst bed. The Sabri Ergun ressure dro equation considers viscous and inetic energy changes for wide flow rates to account for the ressure dro through the reforming reactors [17-18]. The following assumtions are considered during the modeling ste: 1) Ideal gas behavior is alicable. ) Plug flow attern is considered. ) All the reactors wor under adiabatic condition. 4) Axial disersion of heat is neglected. By considering comonent material balances as well as energy balance on a differential element dw of catalyst bed, the mass and energy balance are obtained as follows [19]: Mass balance: dfi dw n ε d ( ci) + γ i, j. rj = r dt j = 1 b (1) Energy balance: dt dt = nc nc nc dt ( FC ) r ( H γ i i j i i, j i = 1 dw j= 1 i = 1 nc C + ε cc cat i i rb i = 1 () Ergun equation: dp G 150(1 ε ) µ 1 = ( ) dw ρd ε d A ρ c c () Iranian Journal of Chemical Engineering, Vol. 15, No. (Sring 018) 55

5 Otimization of Semi Regenerative Catalytic Nahtha Reforming Unit to Enhance Octane Number and Reformate Yield The inetic model of Bommannan et al. based on smith's model is used, which contains four governing reactions as follows [0-1]: Dehydrogenation of nahthenes to aromatic: f1 + K (4) e1 Nahthenes A romatics H ( C H ) ( C H ) n n n n 6 Dehydrocyclization of araffins to nahthenes: Nahthenes + H Paraffins (5) f K e ( CnH n) ( CnH n + 6) Hydrocracing of nahthenes to lower hydrocarbons: ( C H ) ( C C ) Nahthenes + H f Lighter ends n n 1 5 Hydrocracing of araffins to lower hydrocarbons: ( n ) ( C H ) ( C C ) 4 Paraffins + H f Lighter ends n n+ 1 5 = ( f 1 1 )( e 1 n a h e1 r r f = ( )( e n h e f r = ( ) f r = 4 4 ( ) t t n The rate constants and heat of reactions are listed in Table. ) ) Table Rate constants and heat of reactions for nahtha reforming. Rate constant * A B E (J.mol -1 ) ΔH (J.mol -1 ) (98 K) f1 (mol.h -1.g -1 cat.mpa -1 ) f (mol.h -1.g -1 cat.mpa -1 ) f (mol.h -1.g -1 cat ) f4 (mol.h -1.g -1 cat ) e1 (MPa -1 ) e (MPa -1 ) *=Aex(B-(E/RT) (6) (7) (8) (9) (10) (11). Results and discussion.1. Model validation The oerating data of Shiraz oil refinery were used as evaluative criteria. The comarison of outlet temerature, research octane number, and yield between lant data and simulation results demonstrated the ability of the model to redict the desired oututs, as shown in Figs. and. Then, the model was alied to investigate the effect of oerative variables on 56 Iranian Journal of Chemical Engineering, Vol. 15, No. (Sring 018)

6 Khosrozadeh, Talaghat, Roosta the rocess. The main rocess variables that have the greatest effects on the nahtha reforming rocess are temerature, ressure, and hydrogen artial ressure. The effects of these arameters on the rocess are investigated in the following sections Temerature (K) Model Plant Data Inlet Plant Data Outlet Catalyst Weight ( g ) x 10 4 Figure. Temerature rofile along the reactors, a comarison between model and lant data for semiregenerative reforming at 000 Pa. R.O.N (a) Plant data Model (b) Plant data Model Figure. A comarison between model and lant data (a) research octane number versus catalyst age, (b) yield of reformate versus catalyst age. Yield (Vol %) Iranian Journal of Chemical Engineering, Vol. 15, No. (Sring 018) 57

7 Otimization of Semi Regenerative Catalytic Nahtha Reforming Unit to Enhance Octane Number and Reformate Yield.. Effect of temerature The most imortant oerating variable to control roduct quality and yields is the reactor inlet temerature. Since reformers are designed with three or more reactors in series and each reactor may contain a different quantity of catalyst, it is commonly acceted to consider the weighted average inlet temerature (WAIT). In a conventional semiregenerative unit, the loss of activity of catalyst results in a decrease in roduct octane as well as the reformate yield and recycle gas urity. As is shown in Fig. 4, an increase in the reactor WAIT results in increasing conversion of the non-aromatic comounds in the feed to aromatic, although the hydrocracing reaction is more favored than the cyclization of araffins. In addition, light ends comounds increase, and reformate yield and coe deosit increase. It is worth mentioning that the difference in the effect of temerature on the yield of the roducts is due to the difference in the heat of reactions. For instance, the nahthenic roduction (Fig. 4a) decreases with increasing the inlet temerature, because it is an exothermic reaction, while the aromatic roduction increases with temerature due to its endothermic nature. Vol % in Reformate (a) MCP N7 N8 N9 + N Temerature () Vol % in Reformate (b) Toluene A9 + A10 Xylene Benzene Temerature () 58 Iranian Journal of Chemical Engineering, Vol. 15, No. (Sring 018)

8 Khosrozadeh, Talaghat, Roosta 8 (c) Light ends vol % 6 4 C4 C C Temerature () Figure 4. Effect of WAIT on (a) volume ercent of nahthenic roduction, (b) volume ercent of aromatics roduction, and (c) volume ercent of light ends... Effect of ressure Hydrogen artial ressure is a basic variable in the nahtha reforming unit because of its inherent effect on reaction rates; however, for the sae of clarification, the total reactor ressure can be used. The lower the ressure is, the higher the yield of both reformate and hydrogen for a given octane number will be. The ressure reduction leads to an increase in the coe deosition on the catalyst and results in shorter cycle life. Higher ressures cause higher rates of hydrocracing, and more hydrocracing causes the loss of a reformate yield for a given octane number. At higher ressures, coing of the catalyst decreases, resulting in longer cycle life. The real incentive for reducing ressure in reformer reactors is more reformate yield with the added benefit of the hydrogen increase. Fig. 5 reresents the effect of ressure on R.O.N and yield in the nahtha reforming rocess. R.O.N Pa 500 Pa 000 Pa (a) (b) Figure 5. Effect of ressure on (a) research octane number, (b) yield versus catalyst age. Yield Pa 500 Pa 000 Pa Effect of hydrogen to hydrocarbon ratio The rofile of coe deosition on the catalyst in each reactor in various H /HC ratios is illustrated in Fig.6. Recycled hydrogen is necessary in the reformer oeration for uroses of catalyst stability. The hydrogen reacts with coe recursors and, thus, removes it from the catalyst before forming olycyclic aromatics, which ultimately Iranian Journal of Chemical Engineering, Vol. 15, No. (Sring 018) 59

9 Otimization of Semi Regenerative Catalytic Nahtha Reforming Unit to Enhance Octane Number and Reformate Yield deactivated the catalyst. An increase in H /HC ratio will move the nahtha through the reactors at a faster rate and suly a greater heat sin for the endothermic heat of reaction. At lower H /HC, the hydrogen artial ressure decreases and the coe formation increases. The H /HC ratio has little influence on roduct quality or yields. Hydrogen artial ressure is set by an economic balance between equiment sizing and energy saving heaters as well as recycle gas comressor and cycle duration. Coe on Catalyst wt % Coe on Catalyst wt % Coe on Catalyst wt% H/H.C = 4.1 H/H.C = 4.8 H/H.C = (a) H/H.C = 4.1 H/H.C = 4.8 H/H.C = (b) H/H.C = 4.1 H/H.C = 4.8 H/H.C = (c) Figure 6. Effect of H /HC on coe deosition (a) reactor No. 1, (b) reactor No., (c) reactor No.. 60 Iranian Journal of Chemical Engineering, Vol. 15, No. (Sring 018)

10 Khosrozadeh, Talaghat, Roosta.5. Otimization According to the revious sections, the imortant variables, which affect the nahtha reforming rocess, are the average inlet temerature of the reactors, oerating ressure, and hydrogen to hydrocarbon ratio. As mentioned reviously, the objective of the rocess is to find maximum values of octane number, yield and hydrogen roduction as well as the minimum value of coe formation. In this art of study, the otimum condition of the rocess is estimated using genetic algorithm otimization method as an efficient otimization method. After many attemts, the otimal values of decision variables were found and comared with the current status of the rocess in Table 4. The obtained otimal values of octane number, yield and coe formation for a twoyear eriod are comared with the current status in Fig. 7. By alying the otimal condition, the octane number and yield are imroved by 0. % and 1. %, resectively, and the coe formation is decreased by.1 %. Table 4 Comarison between otimal and current cases. Comarison Pressure WAIT (K) (Pa) H /HC Otimal Value Current Value Coe on Catalyst wt% (a) Current Case Otimal Condition (b) Current Case Otimal Condition R.O.N Iranian Journal of Chemical Engineering, Vol. 15, No. (Sring 018) 61

11 Otimization of Semi Regenerative Catalytic Nahtha Reforming Unit to Enhance Octane Number and Reformate Yield Yield (c ) Current Case Figure 7. A comarison between otimal and current cases: (a) coe on catalyst versus catalyst age (b) research octane number versus catalyst age, and (c) reformate yield versus catalyst age. 4. Conclusions In this study, otimization of catalytic nahtha reforming unit is successfully erformed using the genetic algorithm. The objective of the otimization is to find maximum values of octane number and yield and minimum value of coe deosition. Otimization results show that, in otimal condition, the octane number and yield are imroved significantly. Furthermore, the time eriod of the rocess can be increased due to less coe formation. By alying the otimal condition, the octane number and yield are imroved by 0. % and 1. %, resectively, and the coe formation is decreased by.1 %. Acnowledgements The author is grateful to the Shiraz University of Technology for suorting this wor. Nomenclature A area [m ]. d reactor diameter [m]. d b bed diameter [mm]. dis catalyst distribution [wt %]. d article diameter [mm]. H height [m]. H hydrogen [mol/h]. f1 forward rate constant for reaction (8) [mol h -1 g cat -1 MPa -1 ]. f forward rate constant for reaction (9) [mol h -1 g cat -1 MPa - ]. f forward rate constant for reaction (10) [mol h -1 g cat -1 MPa - ]. f4 forward rate constant for reaction (11) [mol h -1 g cat -1 MPa - ]. e1 equilibrium constant [Ma ]. e equilibrium constant [Ma ]. L D loading density [g/l]. P total ressure [Pa]. P i artial ressure of i th comonent [Pa]. P V total ore volume [cm gr -1 ]. S a surface area [m gr -1 ]. T temerature of gas hase [ o K]. V volume [m ]. W weight [g ]. Gree letters ε void fraction of catalyst bed. ρ b catalyst bul density [g m - ]. ΔH heat of reaction [J mol -1 H ]. Subscrition a aromatic. h hydrogen. n nahthene. araffin. Abbreviations 6 Iranian Journal of Chemical Engineering, Vol. 15, No. (Sring 018)

12 Khosrozadeh, Talaghat, Roosta IBP initial boiling oint [ o C]. FBP final boiling oint [ o C]. LHSV liquid hourly sace velocity (reactant liquid flow rate/reactor volume). N7 a nahthenic comound with 7 carbon atoms. N8 a nahthenic comound with 8 carbon atoms. N9 a nahthenic comound with 9 carbon atoms. N10 a nahthenic comound with 10 carbon atoms. MCP Methylcycloentame. RON research octane number. TBP true boiling oint. WAIT weighted average inlet temerature [ o C]. Pt Platinum. Re Rhenium. References [1] Aitani, A. M., Catalytic nahtha reforming: Encycloedia of chemical rocessing, Taylor & Francis,. 97 (007). [] Antos, G. J., Aitani, A. M. and Parera, J. M., Catalytic nahtha reforming: Science and technology, Marcel Decer Inc., New Yor, USA,. 409 (1995). [] Pregger, T., Graf, D., Krewitt, W., Sattler, C., Roeb, M. and Moller, S., Prosects of solar thermal hydrogen roduction rocesses, Int. J. Hydrogen Energy, 4 (10), 456 (009). [4] Alves, J. J. and Towler, G. P., Analysis of refinery hydrogen distribution systems, Eng. Chem. Res., 41 (), 5759 (00). [5] Liu, F. and Zhang, N., Strategy of urifier selection and integration in hydrogen networs, Chem. Eng. Res. Des., 8, 115 (004). [6] D'Iolito, S. A., Vera, C. R., Eron, F., Esecel, C., Marecot, P. and Piec, C. L., Nahtha reforming Pt-Re-Ge/g-Al O catalysts reared by catalytic reduction influence of the H of the Ge addition ste, Al. Catal. A Gen., 70, 4 (009). [7] Rahimour, M. R., Iranshahi, D., Pourazadi, E. and Bahmanour, A. M., Boosting the gasoline octane number in thermally couled nahtha reforming heat exchanger reactor using DE otimization technique, Journal of Fuel., 97, 109 (01). [8] Juarez, J. A., Macias, E. V., Garcia, L. D. and Arredondo, E. G., Modeling and simulation of four catalytic reactors in series for nahtha reforming, Energy & Fuels., 15, 887 (001). [9] Weifeng, H., Hongye, S., Shengjing, M. and Jian, C., Multi objective otimization of the industrial nahtha catalytic reforming rocess, Chin. J. Chem. Eng., 15 (1), 75 (007). [10] Rahimour, M. R., Iranshahi, D. and Bahmanour, A. M., Dynamic otimization of a multi stage sherical, radial flow reactor for the nahtha reforming rocess in the resence of catalyst deactivation using differential evolution (DE) method, Int. J. Hydrogen Energy., 5 (14), 7498 (010). [11] Benitez, V. M. and Piec, C. L., Influence of indium content on the roerties of Pt-Re/Al O nahtha reforming catalysts, Catal. Lett., 16, 45 (010). [1] Boutzeloit, M., Benitez, V. M., Mazzieri, V. A., Esecel, C., Eron, F., Vera, C. R., Piec, C. L. and Marecot, P., Effect of the method of addition of Ge on the catalytic roerties of Pt Re/ Al O and Iranian Journal of Chemical Engineering, Vol. 15, No. (Sring 018) 6

13 Otimization of Semi Regenerative Catalytic Nahtha Reforming Unit to Enhance Octane Number and Reformate Yield Pt-Ir/ Al O nahtha reforming catalysts, Catal. Commun., 7, 67 (006). [1] Mazzieri, V. A., Piec, C. L., Vera, C. R., Yori, J. C. and Grau, J. M., Analysis of coe deosition and study of the variables of regeneration and rejuvenation of nahtha reforming trimetallic catalysts, Catal. Today, 15, 870 (008). [14] Sugimoto, M., Muraawa, T., Hirano, T. and Ohashi, H., Novel regeneration method of Pt/KL zeolite catalyst for light nahtha reforming, Al. Catal., 95, 57 (006). [15] Anabtawi, J. A., Redwan, D. S., Al- Jaralla, A. M. and Aitani, A. M., Advances in the chemistry of catalytic reforming of nahtha, Fuel. Sci. Technol. Int., 9 (1), 1 (1991). [16] Berger, C. V., Denny, R. F. and Michalo, E., Chemistry of HC lat forming, American Chemical Society, Division of Petroleum Chemistry, Rerints, (), (1978). [17] Bird, R. B., Stewart, W. E. and Lightfoot, E. N., Transort Phenomena, nd ed., Wiley, New Yor,. 18 (194). [18] Tasar, U. and Riggs, J. B., Modeling and otimization of a semi regenerative catalytic nahtha reformer, AICHE J., 4, 740 (1997). [19] Arani, H. M., Shori, S. and Shirvani, M., Dynamic modeling and simulation of catalytic nahtha reforming, Int. J. Chem. Eng and Al., 1 (), 159 (010). [0] Bommannan, D., Srivastava, R. D. and Saraf, D. N., Modelling of catalytic nahtha reformers, Can. J. Chem. Eng., 67, 405 (1989). [1] Smith, R. B., Kinetic analysis of nahtha reforming with latinum catalyst, Chem. Eng. Prog., 55 (6), 76 (1959). 64 Iranian Journal of Chemical Engineering, Vol. 15, No. (Sring 018)