A Self-Tuning PID Control System Based on Control Performance Assessment

Abstract

A Quasi-Newton iterative method is developed for the calculation of the best achievable PID control performance and the corresponding optimal PID setting based on the control parameters and input-output data. At the basis of the proposed method, a self-tuning PID control system is proposed for the time-variant dynamic process. When controllers performance deteriorates below the general performance, controller parameters are directly adjusted with the Quasi-Newton iterative method. When below the poor performance, it can be indirectly adjusted with the identification of the closed-loop impulse response matrix. A data-driven solution is developed for calculation of the closed-loop impulse response matrix. Based on the acquired state information, system is assessed and adjusted cyclically so that a self-tuning PID control system is finally realized. Simulation results show the practicality and utility of this method.