The Study Of Predictive Control In A Section Of Reactor And Regenerator Of Fluid Catalytic Cracking Unit

As we all know,crude oil is a kind of mixture which is rich in high value but extremely complex,and are mainly complex hydrocarbons and non-hydrocarbons.So it must adopt various of processing methods to make it into products which meet the quality requirements.In general,after a process,using the atmospheric and vacuum distillation can obtain the light oil accounting for ten percent to forty percent,such as gasoline and diesel oil,and so on.But most of them left are hevay crude oil and residual oil and the use of them are relatively low.Due to the national economy and defense needing a lot of light oil,if the heavy crude oil is not processed again,the light oil only obtained from crude oil through the atmospheric and vacuum distillation can not meet the demand.To meet the demand,so the technology of fluid catalytic cracking have been developed.Reactor and regenerator is a core part of the fluid catalytic cracking unit(FCCU),and the optimization and control to the fluid catalytic cracking have always been the focus of researchers.As to the study of control in the section of reactor-regenerator,the paper mainly carry on the research as follows:(1)The mechanism model of the section of reactor and regenerator of FCCU have been studied.The mechanism model of the system and the solving method to the model have been studied and analyzed in the chapter.And the mathematical models needed by the optimization and control have been built,then they have been simulated and calculated.Finally the simulation and analysis to the lumped concentration and the reaction temperature of reactor and regenerator of FCCU come true.(2)The neural network model back propagation(BP) of the section of reactor and regenerator of FCCU have been built.According to the characteristics of neural network being infinitely close to the actual plant,the BP neural network model to the system have been built through programming of the software of MATLAB.And the model error is comparatively small,and the speed of convergence is very fast whentraining the network.(3)The control scheme proportion integration differentiation(PID) of the section of reactor and regenerator of FCCU have been designed and simulated.The control target,the manipulation variables and the controlled variables of the system have been confirmed.The PID control scheme of the system have been designed and simulated by using the softwore of MATLAB.The results show the effect of PID control is fine,and can satisfy the goals of industrial production of fulid catalytic cracking.(4)The predictive control of the section of reactor and regenerator of FCCU have been designed and the performance of control have been analysed.The basic principles of model predictive control(MPC) have been Explored.Combining the neural network with the predictive control,the predictive control schemes of the section of reactor and regenerator with the neural network have been designed by the toolbox of predictive control of neural network from Simulink.Then comparing the PID control,the results show the quality of the predictive control combining with the neural network is superior to the PID control,the process of oscillation is shorter and the overshoot volume is also smaller.Thus the optimization control of the system come true.