Multivariable Process Control Using Decentralized Single Neural Controllers

This paper develops a learning-type multi-loop control system for interacting multi-input/multi-output industrial process systems. The recently developed single neural controller (SNC) is adopted as the decentralized controller. With a simple parameter tuning algorithm, the SNC in each loop is able to learn to control a changing process by merely observing the process output error in the same loop. To circumvent loop interactions, static decouplers are incorporated in the presented scheme. The only a priori knowledge of the controlled plant is the process steady state gains, which can be easily obtained from open-loop test. Extensive comparisons with decentralized PI controllers were performed. Simulation results show that the presented decentralized nonlinear control strategy appears to be a simple and promising approach to interacting multivariable process control.