Research On Integrated Optimization And Control Method Of Pressure Swing Distillation Process

Pressure swing distillation(PSD)is a special distillation technique which uses the pressure sensitivity of azeotrope to realize its high efficiency separation.In recent years,due to the improvement of the technical standard,the research on the steady-state of PSD process has been developing towards multi-system,economic optimization,energy saving and emission reduction.However,the dynamic control is still in the traditional PID control,and it is difficult to deal with the strong coupling problem caused by complex processes such as thermal integration and the problem of on-line composition measurement.To solve the above problems,this paper starts from the steady-state process design and optimization of PSD,and carries out a series of control strategies including traditional control,active disturbance rejection advanced control and BP neural network intelligent control.The main contents include:(1)Aiming at the problem of high energy consumption and high cost of multiple azeotrope system due to the complexity of chemical process,a triple column PSD steady-state process and its thermal integration design method were proposed.Firstly,thermodynamic analysis of the azeotrope system was carried out by using the residual curve map to determine the feasibility of the high-pressure distillation separation.Secondly,the influence of process parameters was analyzed by sensitivity to determine the operating pressure,the number of stages and other variables,and the optimal separation sequence and process was determined by comparing the economic index(TAC).Finally,partial and full thermal integration schemes are designed to further reduce energy consumption and investment costs.The results show that the full thermal integration scheme has the best economic performance and environmental benefits.(2)In order to ensure the production safety and product quality of the PSD process,the PSO-PID cascade control structure was designed to realize the automatic optimization of controller parameters and precise control,aiming at the problems of feed disturbance control stability and complex parameter adjustment.Based on the composition temperature cascade control,the structure realizes the automatic optimization of composition controller parameters by using the optimization ability of PSO.Experimental simulation shows that compared with single loop temperature control and composition temperature cascade control,the PSO-PID control structure can not only realize the precise control of product concentration in the PSD process,but also improve the efficiency and reliability of controller parameter adjustment and the dynamic performance of the system.(3)By analyzing the dynamic characteristics of the PSD process and the thermal integration process,an active disturbance rejection control structure is constructed to improve the stability and robustness of the control system,aiming at the problem that PID control is difficult to deal with multi-variable and strong coupling.The extended state observer and error feedback control law in nonlinear active disturbance rejection are simplified linearly,which reduces the difficulty of controller design and parameter adjustment,and makes it more suitable for complex PSD and thermal integration processes.The test results show that compared with PID control,the control structure based on linear active disturbance rejection controller can greatly reduce the overshot,shorten the adjustment time,reduce the number of oscillations and optimize the performance index,so that the robustness and stability of the system can be significantly improved,and solve the problem of control performance degradation caused by thermal integration.4)Aiming at the problem that product concentration is difficult to be measured in real time in the process of PSD,an intelligent controller based on BP neural network is designed to achieve accurate control of composition.The input and output variables of the neural network were selected through correlation analysis.The neural network predicted and updated the temperature controller setting point online according to the variables that were easy to measure and highly correlated in the process to complete the closed-loop control of the product.It is proved that the intelligent control structure can realize the precise control of the product and guarantee the control performance without the composition measurement.