Font Size: a A A

Study On Identification By Dynamic Recurrent Fuzzy Neural Networks And Robust Control For Electro-hydraulic Servo System

Posted on:2007-02-24Degree:DoctorType:Dissertation
Country:ChinaCandidate:Y W ZhangFull Text:PDF
GTID:1118360185986676Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
Electro-hydraulic servo system(EHSS) is a kind of system with the characteristics of time-variant, serious nonlinearity, parameter and structure uncertainty, and uncertain load disturbance in most cases. These characteristics make it very difficult to realize high accuracy control for it by conventional methods. In recent years, intelligent control theory represented by fuzzy control, neural network control, knowledge based expert control and genetic algorithms and their integration with modern robust control theory represented by adaptive control, variable structure control and H_∞, control provide effective ways to solve the modeling and control of EHSS with the characteristics listed above. EHSS is taken as the studied object in this paper, and the capability of fuzzy system's integrating the knowledge of expert, the learning capability of neural network, and the capability of the adaptive control, variable structure control and H_∞, control to stabilize the system are used to study the cross-corporation control problem of the above control theory and methods. The main work is summarized as follows:1) The traditional fuzzy neural network(TFNN) is a kind of static neural network and the controlled plant in practice, such as EHSS, expresses the dynamic behaviors of the system, so the static neural network isn't suited for the identification of it. Dynamic recurrent fuzzy neural network(DRFNN) and its related learning algorithm are proposed, and its stability, and through simulation its identification effectiveness for dynamic nonlinear system are analyzed. Simulation results show that DRFNN proposed in this paper have better identification accuracy than TFNN because of the usage of both current and historical data fully.2) Based on the identification for the nonlinearity and uncertainty in EHSS by DRFNN, three kinds of variable structure control algorithms in according with the knowledge degree of the system are studied to avoid realization difficulty because of too large control effort by conservative...
Keywords/Search Tags:Electro-hydraulic servo system, Nonlinear dynamic system, Dynamic fuzzy neural network, Gain adaptive variable structure control, Friction partition compensation, H_∞control, L2 gain design, Robust control, Die-forging hydraulic press
PDF Full Text Request
Related items