Intersection analysis of input and output constraints in model predictive control and on-line adjustment of soft constraints

Abstract

The constraints of input variables and output variables commonly exit in the actual industrial production process. Due to the interference and different constraints between conflicting, the constraint conditions can not be all satisfied, appearing to look for less feasible solutions and global optimal solution and then bringing negative effects on the actual production. Based on Polyhedral pole, the constrained model predictive control feasibility and the soft constraints adjustment algorithm when infeasibility are discussed in this paper. The method in this article considers the feasibility analysis and the reasonable soft constraints adjustment before the rolling optimization in each step, which makes the whole control process meet the requirements of constraint conditions without changing the basic structure of MPC. Through the simulation results of the constrained CSTR system, the validity and feasibility of the algorithm are verified.