Research On Theories And Methods Of Adaptive Sliding Mode Control For AC Servo Systems

With the benefits of performance, power ratings and reliability, AC servo systems play an important role in high-tech regions. Whereas, AC machines are of a class of complex multivariable systems with higher order, nonlinearity and strong coupling, thus, analysis and design are usually conducted based on a simplified model, and the position tracking performance in practical applications may be degradated due to the presence of parametric uncertainties, unmodelled dynamics and extern disturbance. Consequently, new robust control design of servo systems has received more attention than ever.Based on the review of domestic and overseas literatures, theories and design methods for improved robustness and transient/steady-state performan are considered in detail for an induction motor servo system.A comprehensive survey on the advances and trends of sliding mode control is given, research productions and existent deficiencies in servo control and correlative regions are also summarized.The mathematical model and general design requirements of AC servo systems are introduced in brief. A DASMC-â… controller based on the combination of direct adaptive method and traditional sliding mode control is proposed, robustness against parametric uncertainties and extern disturbance are both achieved via online parameters estimation and switching control, respectively, the chattering is also suppressed effectively. Then, a DASMC-â…¡, which is the extention of DASMC-â… , is proposed for the situation where the upper bound of high frequency component of the lumped disturbance is unknown in advance, futhermore, an improved estimation method for this bound is proposed to solve the common problem in existing methods, i.e. the estimated bound grows infinity with time. The stability of the proposed strategies is proved in Lyapunov sense. the parameterized form of steady-state and transient performance is also derived.Initial parameters estimation error, sudden change in the parameter values and non-persistent-excitation may all lead to transient processes. Therefor, better transient response, which means smaller peak error and faster recovery, is important for high performance servo systems. Based on a detailed analysis on the characteristics of typical adaptive methods, an noval composite adaptive sliding mode control(CASMC) is developed, Lyapunov analysis shows that exponential convergence can be conserved in large range under persistent excitation, which implies the more clear transient character. A multiple model composite adaptive control(MMCASMC), which combines multiple model with CASMC, is also proposed to confront sudden changes in the parameter values, in this design, an noval identification and switching method based on dual-channel filters is proposed to provide further improvment in transient response.To deal with the impact of parameter dynamics and structural disturbances, the idea of Adaptive Approximation based Control is incorporated to construct an Neural Network based enhanced adaptive sliding mode control(EASMC). In this design, a universal 3-layer feedforward Neural Network topology is proposed to achieve better approximation in realtime applications, then, the control scheme using structural compensation is established. The boundness of estimation error of all weights is guaranteed by using the modified adaption laws based on discontinuous projection, moreover, an improved adaptive switching control is designed to confront the lumped equivalent disturbance, which is composed of reconstruction error, higher-order-tails of Taylor expansion and nonstructural disturbance. The stability of the proposed scheme is proved in Lyapunov sense, and the advantages of EASMC is futher illustrated in the sense of time-variation.A simulation system is developed using Matlab/Simulink, lots of simulation experiments are conducted to demonstrate the effectiveness of the proposed schemes. An Ethernet based embedded AC servo system experimental platform is also established, some of the proposed schemes are realized in the platform to test the feasibility.