The Research Of PID Parameters Automatic Tuning Based On System Identification

Controlled object in the field of industrial process, such as temperature, pressure, flux in heat power plant and liquid level belong to the field of food processing or sewage treatment, which often exhibit several characteristics, such as nonlinear, time-varying, large scale time delay and strong coupling, etc. The PID control of traditional engineering experiments tuning method cannot be accustomed to the complex system. In order to strength the intelligence and automation of system, more and more controllers use auto-tuning technology. Generally, it is difficult to build up the accurate mathematical model to industrial processes. Therefore, this paper propose a new adaptive control strategy based on real-time online model identification, which combine the system identification and PID parameter tuning.This paper introduces the related contents of system identification and analyze least squares method and derivatives method. Adjustable forgetting factor method o n the basis of TC Hsia recursive least squares method is raised and good performances on MATLAB are achieved. Then, this paper analyze and emulate several PID parameters self-tuning method and discusses the differences of various kinds of algorithm, finally, decide to emulate by the didital PID tuning method, which can achieve the online self-tuning control requirements and will lead to better control effect on the pulverizing system. This paper design the experiments on water tank level platform, which contain system of lower computer and upper computer software. Small system of S7-200 PLC control PID on the under computer, upper computer develop the interface of supervise and debug by kingview, recognizing and tuning program is written in MATLAB, data transmission between the kingview interface and MATLAB program is realized b y OPC communication.The simulation and experimental results show that the accuracy of identification model parameters gained by the improved recursive algorithm is high, and good control effect with low overshoot, fast response and adapting the time-varying system is achieved.