| In recent years,the extractive distillation dividing wall column(EDWC)that combines extractive distillation(ED)and dividing wall column(DWC)technology has attracted more and more attention from researchers.This integrated technology has shown significant advantages in terms of economy and energy.However,compared with ED process,EDWC is difficult to control due to its higher complexity,coupling degree and stronger variable interaction,which makes it difficult to industrialize.Therefore,based on the separation of ternary azeotropic mixtures,the steady-state models of ED and EDWC were established in this thesis.At the same time,the proportional-integral-differential(PID)scheme and the model predictive control(MPC)scheme were designed to effectively resist large flow-rate disturbance and composition disturbance for the two processes.In this thesis,tetrahydrofuran/ethanol/water,the ternary azeotropic mixture of any two components forming an azeotrope,is used as the research object.The rigorous steady-state modeling of the ED process and the EDWC process is completed using Aspen Plus software.With the minimum TAC as the objective function,the optimized parameters and the total annual cost of the two processes are obtained through the sequential iteration method.Compared with the ED process,the total annual cost of the five-year payback period of the EDWC is effectively saved by 10.7%,and the energy consumption is also reduced to a certain extent.Using Aspen Dynamics software,the PID structures of the two processes are designed respectively.And the disturbance analysis results show that the basic temperature control structure is difficult to control the stable operation of the two processes.For the ED process,the improved dual-temperature control structure can effectively resist ±20% total feed flow-rate disturbance and ±20% composition disturbance;for the EDWC,the improved control structure with component/temperature cascade and vapor reflux ratio/sensitive tray temperature cascade shows superior control performance to ±20% total feed flow-rate disturbance and ±10% composition disturbance.Since then,based on the optimal PID structure of the two processes,the corresponding MPC structure was developed by combining Matlab/Simulink and Aspen Dynamics.The dynamic response results show that the proposed MPC structure has excellent control performance under the same disturbances.Finally,the dynamic response of the optimal PID structure and MPC structure of the two processes is compared,and the results show that MPC generally exhibits better control effects.Among them,the MPC structure performs well in reducing the recovery time,reducing the overshoot and reducing the oscillation under the disturbance of the total feed flow-rate;and under the composition disturbance,the EDWC with a high degree of nonlinearity has a similar performance to the PID structure.However,the control effect of the ED process is slightly worse. |