ONLINE PROCESS IDENTIFICATION AND AUTOTUNING FOR INTEGRATING PROCESSES

In this paper, first we analyze why the conventional PID controller is not suitable to control integrating processes mathematically and intu itively. To overcome its structural limitation, we add an internal fee dback loop, which can convert the integrating process to an open-loop stable process. Second, we propose a new and simple on-line process id entification and autotuning method for the integrating process. Here w e use-only the process input and output to obtain a process transfer f unction directly, while other previous closed-loop identification meth ods require the information of the closed-loop control structure. This identification method can be applied not only to integrating processe s but also to all those encountered in chemical engineering processes which are open-loop stable or unstable. From simulation studies, we ca n recognize that, in spite of its simplicity, the proposed on-line pro cess identification and autotuning method shows not only good model ac curacy but also superior control performances for both servo and regul atory problems. It also gives strong robustness to measurement noises and disturbances.