Font Size: a A A

Research On Multivariable Adaptive Decoupling Control Algorithm Based On Parameter Estimation

Posted on:2019-01-18Degree:MasterType:Thesis
Country:ChinaCandidate:T WuFull Text:PDF
GTID:2348330566458303Subject:Electronic and communication engineering
Abstract/Summary:PDF Full Text Request
With the rapid development of industrial production,its process has become increasingly complex and production conditions have frequently changed.The controlled object is complex and the presentation features are strongly coupled,multivariable,etc.The precise mathematical model cannot be used to describe the performance of the controlled object.These complex features make conventional PID controllers difficult to achieve a good control effect on the actual operating system,affecting production efficiency and product quality.In order to improve the control performance of multivariable control systems,the coupling between multivariables needs to be removed.In this paper,the adaptive decoupling control algorithm and parameter estimation are combined for a multivariable system with known structures and unknown parameters.A multivariable adaptive decoupling control algorithm for on-line identification of controller parameters is proposed.In this paper,the relevant theoretical knowledge about system identification is discussed at the beginning.The discrete-time stochastic linear univariate system models with known structures and unknown parameters are identified by Research experiments of Parameter identification among stochastic gradient algorithm,improved stochastic gradient algorithm and recursive least-squares algorithm.Comparing the research results of the above algorithms,the superiority of the improved stochastic gradient algorithm is confirmed;Secondly,the multivariable control system is briefly expounded,its coupling is analyzed,and three common decoupling methods are expatiated.Finally,a multivariable adaptive decoupling algorithm based on parameter estimation is proposed for the coupling problem among multivariable systems.Further,the generalized minimum variance adaptive decoupling control algorithm is applied to the coupling characteristics between multivariable control systems.By comparing the adaptive decoupling experiment before and after the comparison,adaptive decoupling and PID decoupling,it is verified that the adaptive decoupling algorithm has the best control effect and the best system stability.The adaptive decoupling control algorithm is implemented on the basis of parameter estimation.Aiming at the shortcomings of parameter estimation methods in multivariable generalized minimum-variance adaptive decoupling control,an improved stochastic gradient identification algorithm is proposed to instead the application of recursive least squares.The research of the convergence ofthe adaptive decoupling control algorithm is verified by simulation experiments,the decoupling between multiple variables was released and better control was achieved.The research indicates that the adaptive decoupling control algorithm is verified by simulation experiments results show that the improved stochastic gradient identification algorithm has higher accuracy in parameter identification.When using it in the parameter estimation of the controller,compared with the conventional PID decoupling algorithm,the improved stochastic gradient adaptive decoupling algorithm has better anti-interference performance.The experimental research results verify the superiority of the algorithm.The obtained controller parameters can quickly converge to the true value,and the decoupling performance is better.
Keywords/Search Tags:parameter estimation, multivariable system, parameter unknown, adaptive decoupling control
PDF Full Text Request
Related items