Modelling Of The Riser Reactor For The Technology Of Two-stage Riser Catalytic Cracking For Maximizing Propylene

Fluid Catalytic Cracking(FCC)is one of the core processes in petrochemical refineries.It plays a significant role in converting heavy oil to valuable vehicle fuel and improving light olefin production.The two-stage riser catalytic cracking for maximizing propylene(TMP)process is a novel FCC technology developed by the State Key Laboratory of Heavy Oil Processing.Although large numbers of experimental investigations on catalytic materials and catalysts,fluid dynamics,as well as optimum process conditions have been conducted,engineering studies on the novel process also need to be carried out to meet industrialization demands.If the novel technology is further completed in terms of process and equipment modelling,its competitiveness in the oil refining market must be strengthened.Therefore,this dissertation is mainly about the modelling of the riser reactor for TMP technology.According to the boiling ranges,the feedstock and products in the TMP technology were classified into ten lumps,including heavy oil,diesel oil,gasoline olefin,gasoline aromatic,gasoline saturate,(butane+propane),butylene,propylene,dry gas and coke.Reaction rules between these ten lumps were elaborated.The three layered optimization method of simulated annealing(global)-the least square(local)-simulated annealing(local)was proposed to estimate the parameters.It was found that the model could predict reasonably the product yield distributions and their compositions,with all relative errors less than 5%.Meanwhile,the model solved the problem in the 11-lump kinetic model that the yield of propylene kept increasing within the whole range of the conversion of feedstock.Moreover,the model could also be utilized to not only predict the product yields and selectivity but also determine the optimum division point for two-stage operations of TMP units.By simplifying the solving algorithm of the energy-minimization multi-scale(EMMS)model,a more convenient method to establish gas-solid drag models was put forward.The accuracy and rationality of the simplified method were verified through numerical simulations of the gas-solid flow in a conventional riser with the computational fluid dynamics(CFD)model which coupled the EMMS drag model into the two-fluid model(TFM).CFD simulations of the TMP multi-regime riser demonstrated that compared to homogeneous drag models,the non-homogeneous EMMS drag model was more appropriate.In addition,particle cluster diameter correlations affected directly the solution of EMMS drag models.Therefore,the selection of cluster diameter correlations requires much cautiousness.While the particle restitution coefficient had little effects on CFD simulations,the particle-wall specularity coefficient showed great influences.A lower specularity coefficient of 0.0001 was recommended.Besides,the laminar viscous stress model was sufficient to CFD simulations of the multi-regime riser.With the TFM-EMMS coupled model,a circulating fluidized bed(CFB)rectangular riser was simulated.Based on the CFD results,six parameters for characterizing the structure of a reactor network were proposed and subsequently determined.Hence,a novel non-ideal reactor model named equivalent reactor network(ERN)model was constructed for the CFB riser.The findings indicated that the ideal reactor network consisting of 5 CSTRs and 1 PFR could simulate reasonably the non-ideal flow behavior in the original CFB riser and obtain the residence time distributions(RTD)of the solid phase.However,reactor networks consisting of 20 or 50 CSTRs could equally predict RTD curves of the solid phase,indicating that there was not just one configuration of equivalent ideal reactor networks for a non-ideal reactor.A CFD-based ERN model,coupled with the ten-lump model,was developed for the modelling of a pilot riser reactor under reaction conditions.Through comparison studies,the developed ERN model presented advantages over the plug-flow model in terms of the model accuracy and the CFD model in the aspect of time efficiency,indicating that it is a model with relatively high-accuracy and high-efficiency.Based on the established ERN,an integrated model was constructed in the Aspen Plus simulator to numerically investigate the residue fluid catalytic cracking(RFCC)process.It was found that while the conversion of heavy oil and the yield of propylene increased with the residence time prolonged,the yields of light oil firstly increased and subsequently decreased.Besides,improving the reaction temperature gave rise to the increase in the conversion of heavy oil and the yield of propylene and the decrease in the yield of light oil.Finally,the TMP technology,which integrated the light hydrocarbons catalytic cracking process and the RFCC process together,was numerically studied using the Aspen Plus simulator.The findings indicated that the residue and light hydrocarbons had synergistic effects in the TMP riser reactor.Compared to the RFCC process,yields of both light oil and propylene in the integrated TMP technology were promoted.