Simultaneous design of discrete-time inverse response predictor control systems

Abstract

Inverse response is exhibited by a wide array of industrial processes. The presence of inverse response imposes certain innate limitations on the achievable control performance of a conventional feedback control system. In order to alleviate these limitations, an additional predictor block is added to the conventional feedback system to compensate for the effects of inverse response and the modified process dynamics is controlled by a PID controller. However, all the previous model-based design methods have been developed in Laplace domain. Hence, this work proposes a data-based method for discrete-time predictive controller design for processes with inverse response using Virtual Reference Feedback Tuning (VRFT) design method. A robust stability criterion is also developed for the proposed predictive control system. Simulation results are presented to illustrate applications of the proposed design method and a comparison with the model-based predictor system and a VRFT-based conventional PID design is made.