SPSA-based PID parameters optimization for a dual-tank liquid-level control system

摘要

PID controllers are widely used in industrial control systems. To improve the performance of a PID-type control system, the controller parameters should be optimized. However, traditional PID parameters optimization methods are usually cumbersome, time-consuming and experience-based. In this paper, a revised PID parameters sequential optimization method based on the Simultaneous Perturbation Stochastic Approximation (SPSA) algorithm was developed and implemented on a dual-tank liquid-level control system. This method searches for the performance improvement direction iteratively by perturbing all the parameters simultaneously and evaluating the corresponding control performance directly. This methodology has been tested both on the tank model and the actual equipment. Through the simulation and experimentation, its effectiveness was verified.