Research On PID Controller Online Performance Assessment And Optimization Technology For Industrial Processes

In industrial processes,control loops are closely related to product quality,production efficiency and material-energy loss.Controller determines operation effect of a control loop.According to statistics,the problem of controller performance degradation is widespread.Therefore,it is indispensable to evaluate and optimize the performance of controller.Traditional model-based technology of PID controller performance assessment and optimization have the following problems:(1)The performance assessment is mainly based on offline analysis.State identification of process operation data is used for online extraction of modeling data,but many traditional state identification methods rely on parameters,are sensitive to gross errors,and perform poorly in online analysis;(2)Process modeling data is generally step response type obtained from step test,and modeling accuracy is not high;(3)In the case of changes of operating points,nonlinearity of controlled process,and uncertain time delay in the update of sensor measurements,PID controller has different degrees of performance degradation,and the robustness is not strong.Related research and optimization for above problems have been done in this article,as follows:(1)Aiming at problems of traditional state identification methods of process operation data,this paper proposes a new method for state identification.This method is with only two parameters,automatic threshold selection and consideration of gross errors.It has better performance of state identification,and can be used online.Application results of case data and three-tank process show that the state recognition accuracy of the proposed method is more than 5% higher than that of existing two typical methods.(2)In view of the problem of process modeling data is generally type of step response data obtained from step test,this paper proposes the extended integral equation approach to realize process modeling with process response data in general types,which weakens type restriction of modeling data,and modeling accuracy is improved.(3)For attenuation of PID controller performance in the case of changes of operating points,non-linear process and uncertain time delay in the update of sensor measurements,this article studies technologies of controller parameters self-tuning,parameters self-adaptation for non-linear process,and control optimization based on update of measurements to improve control performance.Corresponding simulation test results show the effectiveness of proposed methods.(4)Physical platform of spherical water tank level control system is built,and related proposed algorithms are integrated.The test results of the physical platform show effectiveness of proposed methods.