Several New Approaches to MPC Algorithms Based on Impulse Response Model and Their Application to Chemical Processes

摘要

In this paper ,the developments of three new Model Predictive Control (MPC) algorithms based on impulse response model are presented, including their stability analyses and applications. A new MPC algorithm for system having linear constraints is developed,through the transformation of control problem into a Linear Programming (LP) problem and by using LP techniques ,and this algorithm is named Linear Programming Model Predictive Control(LP-MPC) algorithm Next, a nonlinear system MPC (NL-MPC) algorithms is developed by using nonlinear Hammerstein model. Then, an adaptive output prediction (AP) method is proposed by solving the existing problem of output prediction in the original MPC algorithm;through combination of AP and MPC algorithm ,a new Adaptive Model Predictive Control (AMPC) algorithm is developed .Theoretical analyses prove that under certain conditions , LP-MPC,NL-MPC and AMPC algorithms all have the same stability and robustness as MPC algorithm or in some cases even better Application and simulation results to the control of several chemical processes show that the algorithms of LP-MPC, NL-MPC and AMPC are simple, effective and their implementation is convenient and flexible.