Research On Parameter Optimization Method Of Cascade Control System Based On Running Data

In the industrial process,most objects with large inertia and large delay are controlled by cascade control.When the dynamic characteristics of such controlled objects often change greatly,the conventional cascade controller is difficult to achieve the control accuracy requirements,so it is necessary to optimize the cascade controller.This paper proposes the use of data-driven control algorithm for parameter optimization of cascade controller.The specific research contents are as follows:(1)Studied how to use DFA algorithm to calculate Hurst index,analyzed the relationship between Hurst index and control system performance through simulation experiment,and verified that Hurst index can accurately evaluate the performance of control system.(2)A parameter optimization method based on data-driven cascade controller is proposed.When setting the parameters of the secondary controller,the Hurst index is calculated by collecting the output time series of the secondary loop under noise,and the parameter optimization direction is analyzed and the parameter optimization process is completed according to the relationship between the Hurst index and control performance.When setting the parameters of the main controller,the Hurst index is first used to determine the optimization direction of the parameters,and then the extended fictitious reference iterative tuning method is used to optimize the parameters of the main controller.The k-nearest neighbor algorithm is used to select the nearest neighbor information vector from the database,and the controller parameters are estimated according to the weighted controller parameters corresponding to the information vector.Finally,update database online to maintain database capacity and data vector timeliness.(3)Taking the dynamic characteristics of superheated steam temperature of a 600 MW supercritical unit under typical working conditions as the experimental object,the parameter optimization method of cascade controller based on data drive is verified.The simulation results of variable working conditions show that the cascade controller designed in this paper has good control performance when the model mismatch problem is caused by a wide range of working conditions.