Research On Performance Assessment Of Control Loop Of Non-gaussian System

The performance of stochastic system control loop is directly concerned with the safety and economy of closed loop system operation.What's more,the stochastic system control loop of property evaluation shows the current statement of closed loop system operation.The whole article is derived from performance assessment of traditional index(variance).Firstly,analyzing the narrow limitation of performance assessments,which is based on variance index.Secondly,it gives importance to the defects of current performance evaluation,arising from the entropy index.In view of fact that it lays emphasis on main three research points:1.Accounting of linear Non-Gaussian system shows a new performance assessment method of computing benchmark,which is based on entropy index.Probability density function(PDF)can be approximated by using equidistant rectangles,which is has the same characters.Moreover,an approximate result of control loop performance is likely to be calculated by discrete entropy.As a such,during the simulation examples with various disturbance illustrates effectiveness and consistency of new benchmark as well as the computing method.2.Introducing the concept of unbiased frequency estimation makes the new method of estimating benchmark,which is based on entropy.The methods make full advantage of frequency and histogram concepts.In addition,both the entropy estimation method and the method of ARMA algorithm have the similarity,which only the residuals of regression are requested for related estimation.This algorithm lays a foundation for future industry application.3.Minimum entropy benchmark and its estimation algorithm expands to the given class of nonlinear system.As the method mentioned above,Simulation examples verifies the effectiveness of the algorithm in the nonlinear system.