Research Of System Identification And Its Application Based On Step Response And Open-Loop-Closed-Loop Conversion

Under the closed-loop condition, the system identification and the model based advanced process control technology have always been important aspects of research in the field of process control. It is of great theoretical significance and application value to guide the industrial production and improve teaching practice. In this thesis, focusing on the closed-loop identification and IMC-PID control algorithm study, the open-loop and closed-loop identification algorithm for the linear system are introduced, the least square method with instrumental variables is introduced to improve the closed-loop identification algorithm. System identification software is developed based on these theories, which is also used to identify actual industrial processes. At the same time, the control algorithm with fuzzy weighted settings and artificial neural network adaptive control algorithm are introduced to the IMC-PID control algorithm. In this thesis, the main research contributions are described in three aspects as follows:(1) For the closed-loop identification problems of the first-order and second-order continuous system with pure time delay, a novel identification method based on step response and open-loop and closed-loop conversion is proposed. In this method, after three steps, the dynamic parameters and pure time delay of the open-loop transfer function of the identified plant can be obtained under the closed-loop identification condition. This algorithm is also applicable to the plants with larger time delay.(2) The IMC-PID-FSW controller design method with fuzzy weighted settings is achieved. This algorithm has certain limitations for the overshoot control. On this basis, the internal model control method with adaptive adjustment of the gain and time delay is proposed. The gain and the time delay of the plants can be identified online using this algorithm with Adaline network. At the same time, the internal model and controller can be amended correspondingly.(3) C++ language is used to achieve the above various algorithms. For the liquid level tank, the flow and level are identified and controlled respectively. At the same time, in actual industrial process, for the atmospheric and vacuum distillation unit, the atmospheric tops back-flow and the product tank water drum's boundary level can be identified using this developed software.