Tuning PID controllers for processes with inverse response using artificial neural networks

A calculation method of PID controller settings for an inverse response pro cess with time delay was presented in this study. A practical PID control a lgorithm with anti-reset windup was considered. Optimum controller tuning d ata based on a nonminimum phase model and the minimum IAE criterion were ob tained via the Powell (1964) searching technique, and these data were then empirically fitted by artificial neural networks. Thus, controller tunings based on the process model could be expediently sought from these neural ne tworks. Simulation results indicated that the proposed tuning technique can perform better than other available methods.