Design Methods Of Repetitive Controller For Discrete-time Systems

There are lots of dynamic behaviors of repeated operations in the actual industrial production process.At the same time,the non-repeated uncertain disturbance such as change of operating parameters and measurement noise presents difficulties for precision control system.The conventional control methods which either consider the repeated operations or uncertain disturbance are more and more difficult to meet the increasing demand of manufacturing industry for product quality.As the repetitive feature and uncertainties are inseparable,the theories and methods of automatic control for repetitive system have important application value for promoting industrial automation.This thesis focouses on the tracking performance of the repetitive control method for periodic signals,and designs the discrete sliding mode repetitive controller with the ability to restrain uncertainty.The error-dynamics approach for discrete-time sliding mode repetitive controller is presented with controlling tracking error to origin directly.The major achievements are summarized as follows:1.A discrete sliding mode repetitive control method is presented for tracking control of repetitive system with periodic disturbance.The convergence characteristics of exponential reaching law and modified reaching law include monotonically decreasing region,absolute attractive layer,qusi-sliding band width and convergence time are analysed with different disturbance upper bounds.Further more,the reaching proess of discrete power reaching law is analysed and the convergence performance of system status with bounded disturbance is discussed.For convenience of application,the monotone convergence domain,absolute attractive layer and qusi-sliding band are deduced when the exponential parameter of reaching law takes typical values.The sliding mode repetitive control method helps the repetitive system overcome the problems caused by periodic disturbance.2.An integral sliding mode repetitive control method based state feedback is presented using no-switching reaching law.According to the optimal control theory,the feedback gain matrix is optimized to avoid the unreasonable control energy by high gain feedback.In order to improved the control accuracy,the influence of sampling period on the steady-state performance of switching function and tracking error is analysed.The simulation experiment of servo motor tracking control illustrates that the integral sliding mode repetitive control method is valid for tracking control of discrete repetitive system.3.The control force and its changing velocity of actual actuator are limited by electric power and inertia of moving parts.In order to adapt to the actual condition,a rare-variant attracting law is presented.A disturbance rejection term is embedded in it.Then the error dynamic equation is deduced for designing discrete-time repetitive controllers.The relationship between the parameters of rare-variant attracting law and the control force of the actuator is analysed.The convergence time,convergence domain and steady-state band of tracking error is discussed.The simulation and actual experiment of servo motor illustrates that the rare-variant attracting law based repetitive controller can be used for limiting the control force.It is benefit to protect actual actuator and improve system stability.4.Based on no-switching error attracking law,a repetitive controller is designed for the tracking control of uncertain discrete systems,when the reference signals is periodic.The monotone convergent behavior,convergence time and dynamic error compensation methods are discussed.The steady-state band of error of uncertain system with bounded disturbances is deduced.The simulation and actual experiment of permanent magnet synchronous motor illustrates the theoretical analysis results.5.The fractional periodic repetitive controller is designed for a kind of repetitive system with fractional periodic symmetrical characteristic.Using the symmetry of the desired trajectory in time domain,the system response time and the occupied memory is reduced.In order to evaluate the tracking performance of repetitive system,the monotonic convergence of sine switch attracking law is analysed.Then the influence of switching boundary and switching step length of attracking law on steady-state error of tracking system is also analysed.6.The RBF neural network is introduced to solve the tracking control problem of a kind of repetitive system with model uncertainty.In the thesis the RBF neural network controller and RBFR controller are designed.The steady-state convergence of the system with designed controller is analysed and the speed tracking control simulation experiment of permanent magnet synchronous motor verifies the results.