Hybrid model predictive and active disturbance rejection control for robust and efficient control of extractive distillation process

Abstract

A robust and efficient advanced control system is crucial to ensure safe operation and improve the production efficiency of extractive distillation processes. To address the challenges posed by multi-variable, nonlinear, and coupled problems encountered in the extractive distillation process, a novel robust and efficient control strategy that fuses model predictive control (MPC) and active disturbance rejection control (ADRC) was proposed. The MPC control matrix was optimized using the multi-objective genetic algorithm, enhancing the control performance and robustness of the MPC. The ability to handle nonlinearity in the ADRC was enhanced by adopting a nonlinear error feedback control law, and the debugging process of the ADRC was simplified using a linear extended state observer. MPC and ADRC were fused based on weighted controllers. An evaluation based on the integral of squared error and integral of absolute error reveals that the MPC-ADRC fused control strategy demonstrates significantly superior dynamic control performance compared to other control structures, highlighting heightened robustness and stability. Thus, this study offers novel references and insights for the design and optimization of dynamic control strategies in extractive distillation processes.