Research On Generalized Minimum Variance Assessment Method For Control Systems

With the development of modern industry,there are more and more control loops in factories.Sometimes there are hundreds of control loops in a factory.However,after running for a period of operation,due to equipment wear,system failure,lack of regular maintenance,etc.,the performance of the control system will be degraded,reducing the quality of products.Therefore,in recent years,more and more attention has been paid to the research on the performance evaluation of control systems.Based on the background,significance and development of control system performance,this paper studies the generalized minimum variance evaluation method of control system.The key factor of performance evaluation of control system is how to select appropriate algorithm to calculate reasonable and accurate benchmark value.It is necessary to obtain or estimate the system model and the delay information.Therefore,this paper introduces the basis of time series analysis:data checking and preprocessing,model basic types,model parameter estimation,model adaptability test.Based on the minimum variance benchmark,a performance evaluation method for multi-disturbance system control system is proposed,and a minimum variance controller is designed.The least square fitting method is used to improve the evaluation index for multi-disturbance system with variable input.The minimum variance is further extended.Based on the generalized minimum variance benchmark,a performance evaluation method for multi-disturbance control systems is proposed.The generalized minimum variance controller is designed.In order to ensure that the performance indicators are more sensitive to the change of control performance,the generalized minimum variance performance evaluation index is improved.Combining the characteristics of minimum variance and generalized minimum variance,a comprehensive variance evaluation index is proposed.It is proved that the generalized output feedback invariants cannot be separated accurately due to the complexity of the nonlinear control system.Therefore,the generalized minimum variance benchmark cannot be obtained.Then a performance evaluation method of generalized minimum variance nonlinear control systems based on probability distribution distance is proposed.The performance evaluation index of this method is accurate,the calculation is less than variance analysis method and it is not need prior knowledge.