Kinetic study of catalytic cracking on the effect of reaction parameters in short-contact cyclone reactors

Kinetic study of catalytic cracking on the effect of reaction parameters in short-contact cyclone reactors

Accepted Manuscript Title: Kinetic study of catalytic cracking on the effect of reaction parameters in short-contact cyclone reactors Authors: Zhang Y...

789KB Sizes 5 Downloads 28 Views

Accepted Manuscript Title: Kinetic study of catalytic cracking on the effect of reaction parameters in short-contact cyclone reactors Authors: Zhang Yu-chun, Wang Zhen-bo, Jin You-hai, Li Zhi-he, Yi Wei-ming PII: DOI: Reference:

S0263-8762(17)30074-6 CHERD 2551

To appear in: Received date: Revised date: Accepted date:

21-6-2016 18-1-2017 24-1-2017

Please cite this article as: Zhang, Yu-chun, Wang, Zhen-bo, Jin, You-hai, Li, Zhihe, Yi, Wei-ming, Kinetic study of catalytic cracking on the effect of reaction parameters in short-contact cyclone reactors.Chemical Engineering Research and Design This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Kinetic study of catalytic cracking on the effect of reaction parameters1 in short-contact cyclone reactors Zhang Yu-chuna*, Wang Zhen-bob, JinYou-haib, Li Zhi-hea, Yi Wei-minga a College

of Agricultural Engineering and Food Science, Shandong University of Technology, Zibo 255000,

Shandong, China b

State Key Lab of Heavy Oil, China University of Petroleum, Qingdao 266580, Shandong, China

Corresponding author at: College of Agricultural Engineering and Food Science, Shandong University of


Technology, Zibo 255000, Shandong, China. Tel.: +8618615128986; E-mail: [email protected] 1

Graphical abstract

Highlights ·A 3-d model of a new type reactor for fluid catalytic cracking was built. ·The cracking reactions were simulated by using 6-lump kinetic model. ·The gradient distribution field is beneficial to increase the contact time. ·The effects of operating parameters on product yield were analyzed.

Abstract: The paper presented the research and the results obtained by the author concerning the kinetic modeling of the short-contact cyclone reactor of the catalytic cracking unit. It was an effort to model the phenomenon numerically using Euler-Euler two-phase model and 6-lump kinetic model. Geometry, boundary conditions and dimensions of industrial short-contact cyclone reactor for catalytic cracking unit were conferred for 3D simulation using commercial CFD code FLUENT 6.3. The simulated results calculated by the established model in this paper made a reasonable agreement with the experimental findings. The simulated results indicate that the gradient distribution field in the short-contact cyclone reactor is beneficial to increase the contact time between crude oil gas and catalysts, and the short-contact cyclone reactor is also capable of achieving the rapid separation between gas products and catalysts. Simulations were also carried out to determine the effects of operating parameters on product yield. The operating parameters include reaction temperature, catalyst to oil ratio and reaction time, which are helpful for optimizing the yield of desired products.


Keywords: kinetic model; catalytic cracking; reaction parameter; gradient distribution field; short-contact cyclone reactor Nomenclature CD

Drag coefficient (-)


Solid specific heat (J kg-1K-1)


Diameter of solid particles (m)


Production term of turbulent kinetic energy caused by mean velocity gradient (m2s-2)


Convection heat transfer coefficient (Wm-2K-1)


The specific enthalpy (J kg-1)


Turbulent kinetic energy (J kg-1)


Static pressure (Nm-1)


The heat flux (Wm-2)


Radial position (m)


Radius (m)


Relative Reynolds number (-)


time (s)


Temperature (K)


Velocity (ms-1)


Axial position(m)

Greek letters β

Thermal expansion coefficient(-)


Volume fraction(-) 3

Turbulent dissipation rate (m2s-3)


Gas-solid exchange coefficient (-)


Viscosity (kg m-1s-1)


Density (kg m-3)


Prandtl number corresponding with turbulent kinetic energy(-)


Stress tensor (Pa)

Subscripts i

Gas or solid phase


Gas phase


Solid phase


Introduction Combined with short-contact catalytic cracking technology and swirl separation technology, the

short-contact cyclone catalytic cracking technology was generated. As the crude oil is of poor quality, the new technology is aimed at solving the problems occurred in riser reactors. The main problems presented include core-annular flow, catalysts sliding down and back-mixing, low product yield and so on. As we all know, catalytic cracking reactions occur at high temperature, in this case the feedstock oil has become gas phase. So the flow pattern in the short-contact cyclone reactor is viewed as gas-solid two-phase flow since the catalysts are micro-scale porous particles. We have already made a detailed research on the characteristics of gas-solid gradient field and component transport properties in our previous studies (Yuchun Zhang et al., 2015; Yu-chun Zhang et al., 2013). The FCC (fluid catalytic cracking) ensures the conversion of the heavy oil into diesel and gasoline of a 4

high octane number. The short-contact reactor is the most important equipment in a FCC unit. The endothermic cracking reactions and coke deposition occur in the reactor. So it’s very complex to model a short-contact reactor because momentum, mass and heat transfer and chemical reaction process occur simultaneously in the reactor, they are interrelated and interaction each other. The convection heat transfer is the main pattern between the two phases. Besides, some other heat transfer patterns such as the collision conduction among catalyst particles, or between particles and inner wall also exist in the reactor. In this work, we assumed that the reactor wall is adiabatic and the collision conduction is ignored since the heat quantity is very little in this way. So only the convection heat transfer was taken into consideration in this research. A complete model of the short-contact cyclone reactor should include all the important physical phenomena and detailed reaction kinetics. Up to now many kinetic models for the FCC process have been proposed. Popa Cristina (2015) used a 4-lump kinetic model and a 3-lump kinetic model to represent the kinetic model of the catalytic cracking process in the riser. Muhammad Ahsan (2012) used commercial CFD code FLUENT 6.3 to simulate the fluid catalytic cracking on the 2D geometric model. The 3-lump kinetic model was used in the study. The similar study was also made by Muhammad Ahsan (2015), in the paper the 4-lump kinetic model and Euler-Euler two-phase model was used. The paper indicated that more accurate results could be predicted by implementing the model to a 3D geometry. The same conclusion was drawn by G.C. Lopes et al. (2011). In the work results showed non-uniform tendencies inside the reactor, emphasizing the importance of using three-dimensional models in FCC process predictions. They also used the 4-lump kinetic model to represent the catalytic cracking reactions. At the end of the article the author pointed out that in order to obtain more accurate simulated results, a more sophisticated kinetic scheme containing more lumps should be used. An Euler-Euler hydrodynamic model using the kinetic theory of 5

granular flow combined with a 10-lump model for the reaction kinetics was used by Waldo Rosales Trujillo et al., (2012). In the research a dynamic behavior of the particle bed was observed, with both meso- and macro-scale non-uniformities. Changning Wu et al., (2010) used a 4-lump kinetic model to simulate the complex reactions in fluid catalytic cracking processes. The simulated results captured the major features of FCC process very well either in riser or in downer, which had reasonable agreement with the experimental data in the literature. In another research about the downer, a model which took into account both cracking reaction and flow behavior through a four-lump reaction kinetics coupled with two-phase turbulent flow model was established, simulations were carried out to determine the effects of several parameters on product yield (Fei Liu et al., 2006). G.M. Bollas et al., (2007) used 5-lump kinetic model for the prediction of product selectivity in the fluid catalytic cracking process. From the above analysis we can draw a conclusion that the most widely used approach to model the catalytic cracking process in reactors is the lumping reaction kinetic model (Shuyan WANG et al., 2008;Jian Chang et al., 2012; In-Su Han et al., 2000; J.A. Souza et al., 2011). And researchers obtained satisfactory results using this model (Mohammad A. Abul-Hamayel, 2003; Jian Chang et al., 2012; Jorge Ancheyta-Juárez et al., 1999; Xingying Lan et al., 2009 ). As we all know, models which describe the products in detail have many kinetic parameters (>20) to be estimated. This is an important challenge for evaluation studies when laboratory data available for determining rate constants are usually limited. On the other hand, researches mainly focus on the kinetic mechanism of fluid catalytic cracking in the riser or downer reactors. It is a challenge for the researchers to describe the kinetic mechanism in the area of mathematical modeling in the short-contact cyclone reactor. Now commonly used models usually take the gaseous products as a lump and in some cases the gases are also lumped together with the coke yield. The prediction of coke yield separately from other lumps 6

becomes very important to perform heat integration studies. In the present work we proposed a new 6-lump kinetic model which split the light-oil product into diesel, gasoline and the light gas-lump into dry gas and LPG (liquefied petroleum gas) for catalytic cracking. This separation is very important because the key FCC products can be predicted separately. The CFD code FLUENT 6.3 was used to simulate FCC short-contact cyclone reactors. We mostly focused on the characteristics of reaction yield and the product selectivity. Special attention was paid to the product distribution feature. The objective of the current study was to investigate the reaction mechanism under the action of gradient field in the reactor. The results obtained here are theoretical basis for the development of the new technology and equipment. 2.

Mathematical modeling An Euler-Euler multiphase model was used to simulate the hydrodynamics of the multiple phases.

2.1. Conservation equations The continuity equation of phase i (i = gas, solid):    i i       i  i U i   0 t


with definition:

g  s  1


2.2. The conservation of momentum can be written as Gas phase

  g gU g      g gU gU g    g Pg   g   g   g g g   (U g  U s ) t


Solid phase

  s sU s      s sU sU s    s Pg   s   s   s s g   (U s  U g ) t


2.3. The conservation of energy for phase i yield:

P   i i Hi     ( i iUi Hi )   i i   i : Ui    q  Si t t


2.4. The convection heat transfer between catalysts and gas phase inside the reactor:

  d s3  dTs  hs d s2 Ts  Tg   Cs  6  dt

s 

(6) 7

2.5. From the Gidaspow model for the drag force formulation


CD =

 g >0.8

 s g  g U s -U g -2.65 3 K sg = CD g 4 ds

0.687 24   1+0.15  g Re s      g Res


 g  0.8

, K sg =150


 g 1- g   g g d


2 s


 g s U s -U g



2.6. k-  turbulence model Generally, it’s about turbulent flow in the FCC short-contact cyclone reactor. Therefore, it is important to use an appropriate turbulence model to describe the effect of turbulent fluctuations of velocities and scalar variables for the basic conservation equations. A k-epsilon model was used to describe the turbulent motions in both phases. In the k-epsilon model, the turbulence kinetic energy, k, and its rate of dissipation,

 , can be calculated from the following transport equations:      i i ki     ( i i kiUi )      i t ki    iGk   i i i  t  k 


      i i i     ( i i i Ui )      i t  i   i  C1 iGk  C2 i i i  t  k  k


2.7. Reaction scheme It is necessary to use kinetic models with some detailed product distribution in order to predict the behavior of commercial FCC units. However, the more lumps a model includes, intrinsically more kinetic parameters that need to be estimated and, consequently, more experimental information is required. As we mentioned above, we assumed that the major industrial products in FCC process were diesel, gasoline, LPG, dry gas and coke. So in the 6-lump kinetic model, the six lumps are heavy oil which named VGO (vacuum gas oil) (A), diesel (B), gasoline(C), LPG (D), dry Gas (E) and coke (F). This model mainly includes the following 13 reaction-paths: (1) heavy oil-diesel, (2) heavy oil-gasoline, (3) heavy oil-LPG, (4) heavy 8

oil-dry gas, (5) heavy oil-coke, (6) diesel-gasoline, (7) diesel-LPG, (8) diesel-dry gas, (9) diesel-coke, (10) gasoline-LPG, (11) gasoline-dry gas, (12) gasoline-coke, (13) LPG-dry gas. The reaction network is shown in Fig.1, and the kinetic data for cracking reactions is shown in Table 1.

2.8. Boundary conditions Fig.2 shows the structure of the short-contact cyclone reactor. At the upper part of the mixing cavity there are two axial feed inlets, two gas inlets, and two lateral inlets. The catalysts go through the two axial inlets into the reactor, the high temperature gas phase of crude oil goes through the other two axial inlets and two lateral inlets into the reactor. In the mixing cavity, oil gas contact with catalysts orthogonally first, the shear force generated by the tangential oil gas could increase the turbulence of gas-solid two-phase flow, then the catalysts spread in the mixing cavity along with the impact of oil gas. This process could increase the contact probabilities of oil gas and catalysts, enhance the heat and mass transfer efficiency during the reaction. So the reactor is divided into two parts according to the reaction intensity: one part is called the upper mixing reaction zone, in which the temperature is high, the catalyst to oil ratio is high, and the contact time is short. The other part is called the lower separation reaction zone, in which the reaction process and separation between product and catalysts are both occurred in the strong rotational flow field. We defined the coordinate system being oxyz, where x and z were the horizontal and the vertical axis respectively, y normal to xoz. The coordinate origin o was located in the section center on the top of the mixing cavity. Fig.3 shows the geometry of the reactor. GAMBIT pre-processor was used to construct the 3-D geometry. Meshing of the geometry was done by using rectangular grids. Other parameters are mentioned in Table 2. 9


Simulation set up The no-slip velocity boundary condition was applied for both gas and particle phases. The

velocity-inlet boundary condition was used at the inlet, while the pressure-outlet boundary condition was applied on the outlet. The velocity correction was realized to satisfy continuity through SIMPLE algorithm (Mohanarangam et al.,2008; Wang et al., 2011), which coupled velocity and pressure. To reduce numerical diffusion a second order discretization scheme was used for convection terms in the momentum equations while the QUICK scheme was used for the volume fraction equations. The time step and relaxation factor were determined by the convergence speed and stability. Transient CFD simulations were carried out using a time step of 5 × 10 −4s. Sufficient numbers of iterations were used to ensure numerical convergence and reduce the uncertainty caused during the calculations. On the one hand the calculated value of monitoring point doesn’t change along with the increase of computational steps while residual error of various parameters decreases and finally also doesn’t change, on the other hand when it fits with the mass and energy conservation equations, the simulation is regarded as convergent. Each simulation was performed for at least 30 s and the predictions were time-averaged for the last 10 s. 4.

Results and discussions

4.1 Model validation The calculated results were compared with the laboratory test results as shown in Table 3. The conversion rate of calculated is about 1.9% more than that of experimented. And the maximum deviation of each product yield is gasoline, the calculated value is about 1.7% more than that of experimented. Summarize the information about this table we can see that, the difference of each parameter is very small, the maximum deviation in this table is only 5%. We also made a research on the heat balance. We calculated the heat quantity the raw material took into and out of the reactor through the temperature field. 10

In theory, the heat quantity that the material took into the reactor and chemical reaction heat is closely equal to the heat quantity that the material took out of the reactor and the heat the reactor dissipated. Since the heat the reactor dissipated was ignored in the numerical simulation, so the simulated heat quantity that the material took out of the reactor is higher than that of experimented. Then the calculated temperature of the outlet should be a little higher than the experimental value. From the table below we can see that the result obtained by calculating accords with the theoretical analysis and the experimental result. In conclusion, from the comparison between the calculated and experimented results of conversion rate, product yield and so on, we can see that the model established in this paper could predict the catalytic cracking reaction behavior accurately. 4.2 Product longitudinal distribution characteristics From Fig.4 we can see that in the mixing cavity the catalyst concentration is high, other than the area around the two gas-phase lateral inlets. That’s because the catalysts were blew by the strong shear force made by the tangential feedstock gas. The high catalyst concentration increases the contact probabilities of oil gas and catalysts, and it is good for the reaction efficiency in the mixing cavity. In addition, we can see that the catalyst concentration is higher near the wall than in the central part under the action of the centrifugal force along with flowing down. Most of the catalysts go into the hopper at the end of the reactor finally. Fig.5 is the temperature field distribution. The temperature gradient is not obvious in the figure, heat exchange occurs immediately when the feedstock gas and catalysts enter into the reactor. The heat transfer efficiency is high and the temperature increases slightly while flowing down. Fig.6 shows the mass fraction distribution of heavy oil, gasoline and dry gas respectively. In the model, as we have previously written, we assumed that the raw oil has been vaporized immediately when it enters 11

into the reactor, so the catalytic cracking reactions are carried out right away. From the figure we can see that the maximum mass fraction of heavy oil gas in the mixing cavity is 20%, and the value reduces gradually in the direction of flow stream. From this figure we can also see that the heavy oil gas mainly exists near the wall in the separation cavity, and the mass fraction decreases dramatically in the middle of the separation cavity. This phenomenon indicates that the rest of heavy oil gas continues to react while flowing down. Fig.6 (b) shows the mass fraction distribution of the gasoline. Since the heavy oil cracks into different products when it enters into the reactor, we could see that the value of gasoline mass fraction is big in the mixing cavity, the variation gradient becomes distinct gradually in the separation cavity. The reason is that the heavy oil and diesel continue to react into gasoline. Meanwhile, a small percentage of gasoline occurs second cracking reactions, it could crack into shorter chain products. In this case complex changes take place constantly, the concentration distribution of each species also changes on and on. On the whole the gradient distribution of each species is the most conspicuous regularity. From the gasoline distribution of the lower area in the separation cavity we can see that the gasoline rotates into the upward flow as it belongs to the light component. So the gasoline concentration is lower in the downstream of the bottom area in the separation cavity. Fig.6(c) shows that the dry gas distribution constantly changes, the dry gas distribution rule is similar to that of gasoline, and the mass fraction of dry gas near the outlet is about 10%. 4.3 Product radial distribution characteristics It is worthwhile to make further study of the product radial distribution regularities in each part of the reactor. The radial distribution of gasoline is analyzed first of all. As seen in Fig.7, in order to observe the distribution characteristic of gasoline clearly, we made a 12

comparison on the gasoline distribution curves along the radial position among different axial sections. From Fig.7 (a) we can see that in the mixing cavity the maximum value of gasoline mass fraction is 40%, the mass fraction decreases with increasing axial distance. In addition, the gasoline mass fraction decreases with increasing radial distance on the same axial section, and the gradient becomes smaller while flowing down. From Fig.7 (b) and (c) we can see that the gasoline concentration becomes smaller gradually. The gradient product distribution field gradually forms under the influence of the gradient flow field. From the above analysis we can draw a conclusion that the gasoline distribution pattern in the reactor is basically axial symmetry, the distribution trend of gasoline is the opposite of that of catalyst particles (Yu-chun Zhang et al., 2013). That’s because the gasoline density is much smaller than the catalyst density, the heavy component moves towards the inner wall while the light moves in the opposite direction under the action of centrifugal force. So the gasoline mainly exists in the inner swirling-flow and is about parabolic distribution form, while the catalysts mainly exist in the outer-helical-flow and the concentration near the wall is higher than that in the inner area. Due to the species distribution is basically centrosymmetric, thus one side (center - wall) distribution changes were studied in this part. The distribution characteristic of heavy oil is distinct from that of products. As seen in Fig.8, the mass fraction of heavy oil decreases gradually in the mixing cavity, it continues to decrease till the separation cavity. The most striking feature is that since the heavy oil gas is denser than other gas products, the concentration near the wall is higher than that in the middle part because of the centrifugal force. It means comparing with other species, the heavy oil gas and catalysts are mostly distributed in the outer area near the wall. This extends the time that the heavy oil contacts with catalysts, and the light products could separate with catalysts at once because the light products are always distributed in the inner area away from the wall. That is to say, this distribution pattern makes a great 13

influence on improving the conversion rate of cracking reactions.

4.4 Effect of reaction parameters

The reaction temperature is an important factor that influences the catalytic cracking behavior in the cyclone reactor. It has a major effect on reaction rate, conversion rate and product selectivity especially. There are many factors that can affect the reaction temperature in the catalytic cracking unit, such as the preheated temperature of crude heavy oil, the regenerated catalysts temperature, operating capacity and catalyst circulation etc. In this work, the reaction temperature is changed by changing the inlet temperature of catalysts while the preheated temperature of crude heavy oil is constant. Fig.9 shows the mass fraction curves of six species under different reaction temperatures respectively. From the variation curve of the residual crude oil we can see that, the conversion rate increases with the increase of temperature. It means that raising temperature could enlarge the cracking depth, and the effect is more remarkable especially in lower temperature range. The influence of reaction temperature on the diesel and gasoline yield is small. From the Figure we can see that the gasoline amount decreases while the LPG and dry gas still increase at 800 K. It demonstrates that the speed of gasoline cracking into dry gas is faster than that of heavy oil cracking into gasoline at this temperature and even higher. So from the above analysis we can draw a conclusion that if we want to improve the gasoline yield, we should set the reaction temperature lower than 800 K. In addition, the coke yield decreases as the temperature increases.

As seen in Fig.10, when the reaction temperature increases, the diesel selectivity decreases gradually, the gasoline selectivity increases first and then decreases, LPG and dry gas selectivity increases while coke selectivity decreases. When the temperature is low, the heavy oil cracks into diesel, and the diesel doesn’t continue to crack at the low temperature. Then with the reaction temperature increasing, the diesel can 14

crack into other lighter products. So that’s the reason why diesel selectivity decreases in this temperature range. Comparing with diesel, the gasoline needs higher temperature to crack. Therefore the gasoline selectivity increases first and then decreases in this temperature range. In conclusion, increasing the reaction temperature is of great significance to enhance the depth of the cracking especially at low temperature. However, if the reaction temperature is too high, the cracking speed of light oil may be faster than that of heavy oil. In this case the yield of light oil will reduce, the catalytic selectivity will decrease. Fig.11 shows the final yield compared curves of each specie under different CTO (catalyst to oil ratio). The heavy oil decreases with the increase of CTO except the CTO is 30. The reason is that active centers increase with the increase of catalysts, in this case the cracking depth is enhanced. But the amount of catalysts is no longer the key factor determining cracking depth when the CTO is large enough, so the conversion rate shows no obvious change. The diesel basically decreases with CTO increasing, and the biggest gradient change appears when the CTO is 30.

The gasoline yield is slightly affected. The

reason may be more heavy oil cracks into gasoline when the CTO increases, at the same time more gasoline cracks into other products. So on the whole no significant increase is found on the amount of gasoline. Besides, the LPG and coke yield increase while the impact isn't significant for the dry gas. The increase of coke demonstrates that a lot of the useful reactions could be prompted by increasing CTO, meanwhile secondary reactions such as coking could also increase inevitably. From the product selectivity histogram we can see that the selectivity of diesel and gasoline rises first and goes down later with the increase of the CTO. The selectivity of LPG, dry gas and coke shows a general upward trend. The changes of product yield and selectivity need to be studied thoroughly combining the effect of CTO with that of reaction temperature. As previously mentioned, more 15

active centers participate in the reaction to increase the product yield when the CTO increases. The increase of CTO can lead to the increase of reaction temperature, then the dry gas and coke increase while gasoline decreases when the CTO exceeds a certain limit. At the moment increasing CTO can have detrimental effects on catalytic cracking process. In this work the reaction time refers to the mean residence time of the gas phase. As seen in Fig.13, the heavy oil deceases with the increase of reaction time. The diesel and gasoline almost have the same trend, they both rise first and then decrease. The dry gas and coke yield increase when the reaction time increases. From the change trend of the product yield we can see that, the conversion rate of catalytic cracking reaction is low when the reaction time is short and raising reaction time is favorable to improve the product yield. The secondary cracking reaction increases and the gasoline yield reduces when the reaction time is longer than 1.2 s. Fig.14 shows the related diagrams of product selectivity under different reaction time. There exists the optimal reaction time to obtain the biggest gasoline selectivity. Combined with Fig.13 we can draw a conclusion that the most suitable reaction time is 1.2s for the catalytic cracking process with the purpose of getting more light oil products. The product selectivity is worse if the reaction time is too long or too short. 5.

Conclusions In this paper the catalytic cracking was simulated by 6-lump kinetic model. The research mainly

focused on the axial and radial distributions of mass fraction of cracking products and the characteristics of product yield and selectivity. By studying the axial and radial distribution of each product we discover that, the heavy oil gas mainly distributes in the near wall region in the separation cavity, the gradient distribution is relatively low in the middle and high near the wall on the whole. Conversely, the distribution of light 16

products such as gasoline and dry gas is high in the middle and low near the wall. This gradient distribution field indicates that the short-contact cyclone reactor is beneficial to increase the contact time between crude oil gas and catalysts and is also capable of achieving the rapid separation between gas products and catalysts. All of these illustrations prove that the new reactor is a better equipment to be used in catalytic cracking in the future. The conversion rate increases with the increase of reaction temperature. However, a large number of light oil gas could crack if the reaction temperature is too high. The speed of the secondary cracking may be larger than that of catalytic cracking, in this case the light oil yield decreases and the product selectivity is bad. In this paper the best reaction temperature is 770 K. The conversion rate increases with the increase of CTO. But the conversion rate no longer increases when the CTO is 30. Besides, the temperature may increase with the increase of CTO, it causes the gas product and coke yield increase and the light oil decreases. From the simulated results in this work we can see that the most suitable CTO to get the best gasoline selectivity is 26. The conversion rate increases with the increase of reaction time. The diesel and gasoline yield rise first and then decrease. The incomplete catalytic reactions result in low conversion rate when the reaction time is short. On the contrary, the increase of the secondary reactions results in low light oil yield when the reaction time is long enough. The best reaction time is about 1.2 s in this paper. Acknowledgments This work was supported by the National Natural Science Foundation of China [grant numbers 21276281, 51406108, 51406109, 51606113]; Taishan Mountain Scholar Construction Engineering Special Fund of Shandong Province, China and the Dr. Startup Funds of Shandong University of Technology [grant number 414029]. 17


References Abul-Hamayel M A, 2004.Comparison of downer and riser based fluid catalytic cracking process at high severity condition: A pilot plant study. Petrol Sci Technol. 22(5-6), 475-490. Changning Wu, Yi Cheng, Yulong Ding, Yong Jin, 2010. CFD–DEM simulation of gas–solid reacting flows in fluid catalytic cracking (FCC) process. Chem Eng Sci. 65(1), 542–549. Fei Liu, Fei Wei, Yu Zheng, Yong Jin, 2006. CFD Simulation of fluid catalytic cracking in downer reactors. China Particuology. 4(3–4), 160–166. G.C. Lopes, L.M. Rosa, M. Mori, J.R. Nunhez, W.P. Martignoni, 2011.Three-dimensional modeling of fluid catalytic cracking industrial riser flow and reactions. Comput Chem Eng . 35(11), 2159–2168. G.M. Bollas, A.A. Lappas, D.K. Iatridis, I.A. Vasalos, 2007. Five-lump kinetic model with selective catalyst deactivation for the prediction of the product selectivity in the fluid catalytic cracking process. Catal Today.127( 1–4), 31–43. In-Su Han, Chang-Bock Chung , James B. Riggs, 2000. Modeling of a fluidized catalytic cracking process. Comput Chem Eng . 24(2–7), 1681–1687. J.A. Souza, J.V.C. Vargas, J.C. Ordonez, W.P. Martignoni, O.F. von Meien, 2011.Thermodynamic optimization of fluidized catalytic cracking (FCC) units. Int J Heat Mass Tran . 54(5–6),1187–1197. Jian Chang, Kai Zhang, Fandong Meng, Longyan Wang, Xiaoli Wei, 2012.Computational investigation of hydrodynamics and cracking reaction in a heavy oil riser reactor. Particuology.10(2),184–195. Jorge Ancheyta-Juárez, Felipe López-Isunza, Enrique Aguilar-Rodrı́guez, 1999. 5-Lump kinetic model for gas oil catalytic cracking. Appl Catal A: Gen. 177(2), 227–235. Mohammad A. Abul-Hamayel, 2003. Kinetic modeling of high-severity fluidized catalytic cracking. Fuel. 19

82(9),1113–1118. Mohanarangam, K., Tian, Z.F., Tu, J.Y., 2008. Numerical simulation of turbulent gas–particle flow in a 90° bend: Eulerian–Eulerian approach. Comput. Chem. Eng. 32 (3), 561–571. Muhammad Ahsan, 2015. Prediction of gasoline yield in a fluid catalytic cracking (FCC) riser using k-epsilon turbulence and 4-lump kinetic models: A computational fluid dynamics (CFD) approach. J of King Saud University – Eng Sci. 27(2), 130–136. Muhammad Ahsan, 2012. Computational fluid dynamics (CFD) prediction of mass fraction profiles of gas oil and gasoline in fluid catalytic cracking (FCC) riser. Ain Shams Eng J. 3(4),403–409. Popa Cristina, 2015. Four-Lump Kinetic Model vs. Three-Lump Kinetic Model for the Fluid Catalytic Cracking Riser Reactor. Procedia Eng. 100, 602–608. Shuyan WANG, Huilin LU, Jinsen GAO, Chunming XU, Dan SUN, 2008. Numerical Predication of Cracking Reaction of Particle Clusters in Fluid Catalytic Cracking Riser Reactors. Chinese J Chem Eng.16(5), 670–678. Waldo Rosales Trujillo, Juray De Wilde, 2012. Fluid catalytic cracking in a rotating fluidized bed in a static geometry: a CFD analysis accounting for the distribution of the catalyst coke content. Powder Technol. 221, 36–46. Wang, Z.B., Ma, Y., Jin, Y.H., 2011. Simulation and experiment of flow field in axial-flow hydrocyclone. Chem. Eng. Res. Des.89,603–610. Xingying Lan, Chunming Xu, Gang Wang, Li Wu, JinsenGao, 2009. CFD modeling of gas–solid flow and cracking reaction in two-stage riser FCC reactors. Chem Eng Sci. 64(17), 3847-3858. Yuchun Zhang, Zhenbo Wang, Youhai Jin, Zhihe Li, Weiming Yi, 2015. CFD simulation and experiment of residence time distribution in short-contact cyclone reactors. Adv Powder Technol. 26, 1134-1142. 20

Yu-chun Zhang, Zhen-bo Wang, You-haiJin, 2013. Simulation and experiment of gas–solid flow field in short-contact cyclone reactors. Chem. Eng. Res. Des. 91, 1768-1776.


Fig.1 6-lump reaction network model.

Fig.2 Schematic of the short-contact cyclone reactor. Left: an elevation view; right: a top view.


Fig.3 Grid arrangement diagram for simulation in the three-dimensional reactor.

Fig.4 Catalyst concentration distribution.

Fig.5 Temperature field.


(a) heavy oil

(b) gasoline

Fig.6 Distribution of different species in longitudinal section.


(c)dry gas




0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 -0.8 -0.6 -0.4 -0.2 0


z=500mm 0.31

z=50mm z=100mm z=150mm z=200mm 0.2



0.29 0.27




0.25 -0.8 -0.6 -0.4 -0.2 0



(a) mixing cavity

(b) near the guide vanes


0.3 0.28 0.26 0.24 0.22 0.2 0.18 0.16 0.14 0.12 0.1 -0.8 -0.6 -0.4 -0.2 0




z=600mm z=800mm z=900mm







(c) separation cavity Fig.7 Radial distribution curves of gasoline in different sections.







z=50mm z=100mm z=150mm z=200mm





0.5 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0


z=400mm z=500mm











(a) mixing cavity

(b) near the guide vanes



0.08 0.075






0.06 0.2











(c) separation cavity Fig.8 Radial distribution curves of heavy oil in different sections.

Fig.9 Yield curves with different reaction temperatures.




Product selectivity/%

50 45 40 35 30 25 20 15 10 5 0

730K 770K 800K 820K






Fig.10 Comparison of product selectivity with different reaction temperatures.

Fig.11 Yield curves with different CTO. 45

Product selectivity/%


CTO=16 CTO=22 CTO=26 CTO=30

35 30 25 20 15 10 5 0






Fig.12 Comparison of product selectivity with different CTO.


Product selectivity/%

Fig.13 Yield curves with different reaction time. 50 45 40 35 30 25 20 15 10 5 0

0.5s 0.8s 1.2s 1.5s






Fig.14 Comparison of product selectivity with different reaction time.


Table 1 Kinetic data for cracking reactions Activation


Pre exponential Cracking reaction

Pre exponential energy

Cracking reaction


energy factor



Heavy oil-diesel



Heavy oil-gasoline



Heavy oil-LPG



Heavy oil-dry gas



Heavy oil-coke









Diesel-dry gas









Gasoline-dry gas






LPG-dry gas



Table 2 Operate parameters list Inlet oil temperature


Inlet catalyst temperature


Catalyst-to-oil ratio


Catalyst density

1700(kg m-3)

Gas inlet velocity

5 (m s-1)

Catalyst inlet velocity

1.2 (m s-1)

Heavy oil density

10.95(kg m-3)

Particle diameter

7.0 × 10−5(m)

Solid mass flow

80(kg m-2s)

Heavy oil viscosity

7.5 × 10−6 (Pa s)


Table 1 Comparison between simulated and experimental results Parameters

Calculated value

Experimental value

Absolute deviation

Conversion rate(%)




Outlet temperature(K)
















Dry gas




Heavy oil