Research On Model Predictive Control Algorithm For AC Servo System With Backlash And Disturbance

The permanent magnet synchronous motor(PMSM)has been widely used in AC servo systems due to its advantages such as high power density,torque-to-inertia ratio and efficiency.To meet the increasingly high requirements for the control performance of the servo systems,modern AC servo systems usually utilize the advanced control algoithms to replace the conventional PI control.Model predictive control(MPC),as one of the advanced control algorithms,has been extensively studied,because of the ability to effectively deal with the constraints in the system and obtain the optimized control performance.This paper studies the speed regulation problem of the PMSM servo system.Based on the analysis of the PMSM mathematical model,a baseline model predictive control and an integral model predictive control for the PMSM servo system are studied.Simulation and experiment results show that these two MPC algorithms can obtain optimal tracking performances.The PMSM servo systems are confronted with various forms of disturbances,e.g.,parameter uncertainties,load torque and unmodeled dynamics.It is more appropriate to treat these disturbances as the polynomial form,thus a generalized proportional integral observer(GPIO)is designed to estimate the lumped disturbances.By embedding the estimation of disturbances into the prediction model,an MPC algorithm is proposed based on the disturbance estimation.The integrated design of MPC and disturbance observer can consider the effects of disturbances in the receding optimization procedure,so the optimal control performance will be obtained.Simulation and experiment results show that the proposed composite MPC algorithm can effectively attenuate the effects of disturbances on the servo system.The servo systems rarely run with the single motor,but drive the load through a transmission mechanism such as the gearbox and the screw to convert the energy.Backlash existing in the transmission mechanism brings the problems such as limit cycles,impact and rigid collision to the system.As a nondifferentiable strong nonlinearity,the backlash is difficult to be estimated and compensated by the disturbance observer.To this end,the dead zone model of backlash is obtained by the identification method firstly,and then the transmission torque of backlash is estimated from the identified model.The estimated torque value is embedded into the prediction model to suppress the influence of backlash on the servo system.At the same time,considering the disturbances such as parameter uncertainties and backlash modeling residual error in the system,a GPIO is designed to estimate the lumped disturbances.By integrating the estimation of backlash transmission torque and disturbances into the prediction model,an MPC algorithm is proposed based on the backlash cancellation and disturbance estimation.Simulation results show that the proposed composite MPC algorithm can significantly eliminate the influence of backlash and disturbances on the servo system.