Font Size: a A A

The Application Of Predictive Control In Time-delay System

Posted on:2008-12-18Degree:MasterType:Thesis
Country:ChinaCandidate:W J JiangFull Text:PDF
GTID:2178360212983620Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
The big time-delay phenomenon within the industry process is very widespread. Its existence affects the stability of the system seriously, causes system overshot becoming big, makes adjust time become long consumedly, and even appears to flap and dissipation. The dynamic quality of the system becomes obviously bad. Many methods can overcome the bad effect of the time-delay quality in theory, but they need the precise models of the systems. In fact, it is difficult to produce the system model because of the indetermination in the industry process. So it is necessary to find a good control tactics with no depending on system model, strong anti-interference ability and good robustness. Meanwhile the predictive control has the following characters: good self-tuning ability, fast response, little overshot, short adjust-time and strong robustness, etc. It has higher control precision without the necessary of precise model, so it can control the time-delay system.In this paper, the Dynamic matrix Control studied and used in time-delay system has good control effect and compared with classic Smith control it shows good robustness. On the base of the deeply study of Generalized Predictive Control theory, it has been used in time-delay system and gives satisfactory control effect. At the same time, the impact to the system response of the parameters of the GPC is studied through simulation. It is hard to produce the precise model of the industry process because of the nonlinear, so it is difficult to be used in real system. Neural network has nonlinear convergence ability, so the predictive control based on neural network has been proposed. In the common two networks, BP neural network has low convergence speed, local convergence, training slowly and inefficient, while RBF can overcome these shortcomings. Based on this, DMC based on RBF is proposed by connecting the RBF and DMC algorithm. The simulation results demonstrate that the presented control method can be used to control time-delay system and the response of the system shows good control performance.
Keywords/Search Tags:Time-delay system, Predictive Control, Generalized Predictive Control, Dynamic Matrix Control, Neural Network
PDF Full Text Request
Related items