Control Research Of Stochastic Distribution System

The research on the control of stochastic distribution system is an important part of the control theory system.Due to the non-linearity and uncertainty of most practical systems,the system output or state variables obey the non-Gaussian distribution.When the random variables existing in industrial production no longer satisfy the hypothesis of Gaussian distribution,the mean and variance of the system output cannot contain enough information to describe the controlled object.Considering that the output probability density function(PDF)can provide a more complete description of the output of the system,professor Wang Hong proposed a stochastic distribution control theory that takes the entire PDF shape of the non-Gaussian random system's output as a control object.And the theory has achieved good results in many industrial applications.In addition,with the rapid development and improvement of precision instruments,image acquisition and data processing technologies,it is now much more easier to obtain the output PDF of the system by processing the image information obtained from a graphics processing device or the like.In some actual production processes,the system control quality can be improved by studying the three-dimensional distribution information of the system.For example,for the combustion stability and the combustion efficiency in the combustion system of the boiler,the three-dimensional temperature field distribution information of the chamber can be utilized,which can be introduced into the control loop as reference information to improve the combustion conditions to a certain extent.It can prevent the flame from deflecting and further improve the combustion economy.Based on the existing research on the theory of modeling and control of stochastic distribution system,on the one hand,for a PDF model of rational square root B-spline function with the advantages of less restriction in the modeling process.Because of its design using the conventional gradient method,the controller has the disadvantage of poor tracking performance when it is expected to trace the output PDF.Then in this paper,the iterative learning controller is designed and the control algorithm is simulated.The related work enriches the output distribution theory of random system.On the other hand,aiming at the current situation of less research on the three-dimensional output distribution control theory,this paper studies the problem of establishing a three-dimensional output PDF model by using two-dimensional B-spline function.The paper gives the modeling steps and the controller design method.In order to further improve the control effect,the paper discusses the design method of the predictive controller for the model.To sum up,the paper enriches the relevant modeling and control algorithms of output distribution control theory,which provides a theoretical basis for industrial application of stochastic distributed control theory.