USING NEURAL NETWORKS AND PLS TO DESIGN MULTI-LOOP PID CONTROLLERS IN NONLINEAR MIMO PROCESSES

Abstract

An automatic tuning method of multiloop PID controllers for nonlinear multi-input multi-output (MIMO) processes is proposed. First, the dynamic PLS (DynPLS) model is derived from the decomposition structure of partial least squares (PLS) and the instantaneous linearized neural network model at each sampling time. It can decompose the MIMO process into a multiloop control system in a reduced subspace. Second, the optimum tuning PID controller with a general minimum variance self-tuning control strategy of each loop is developed. The simplicity and feasibility of this scheme provide a new approach to implementing neural network applications for a variety of on-line industrial control problems. Simulation case study is provided to demonstrate the effectiveness of the control design procedures of the nonlinear MIMO processes.