Research On Model Predictive Position Control In Repetitive Motion

In the modern industrial field,many equipment with Permanent Magnet Synchronous Motor(PMSM)control system as the core,such as precision machine tools,gearboxes,etc.,need to maintain a specific position trajectory for repetitive periodic movement.However,the nonlinear factors such as parameter perturbation,periodic disturbance and friction encountered in the operation process will seriously interfere with the position tracking accuracy of the repetitive motion of the motor system.Therefore,how to maintain fast and accurate repetitive motion under a variety of disturbances is of great significance.Model predictive control,as a control strategy that has attracted wide attention in recent years,has excellent tracking ability and accuracy on the basis of knowing the accurate model.Based on the above factors,the repetitive motion and model predictive control of PMSM servo system are studied.The main research work of thesis is as follows:(1)Aiming at the problem of fast response of servo system,a direct speed predictive control method is designed.On this basis,considering the weight factor design problem of finite set model predictive control,an improved sequential predictive control method is adopted.Because finite set model prediction requires high model accuracy,a sliding mode observer is designed to estimate load changes and feed back to the controller.Finally,aiming at the uncertainty of parameters in the model,the proportional integral method is used to compensate at the output end.(2)A data-driven model prediction direct position control method is proposed to solve the problem of parameter perturbation and uncertain disturbance in the position tracking process of repetitive motion servo system.The traditional three-loop control is merged into single-loop control to improve the response ability of the system.Firstly,a PMSM model with lumped disturbance was established,and a disturbance observer was designed to estimate the change of the disturbance.Secondly,by collecting the current and voltage data,the corresponding table of voltage vector and current increment was constructed,and a novel update method of the data table was designed.On this basis,a position single-loop controller is designed,and a cost function containing current,speed and position information is designed to solve the problem of uncontrollable speed under the direct position model prediction method.(3)A data-driven model predictive position control method based on active disturbance rejection control is proposed in this chapter,which is aiming at the problems of parameter perturbation and friction in the repetitive motion servo system.Firstly,the IPMSM model containing friction is established.Secondly,friction and mismatching disturbances are designed as lumped disturbances,and an active disturbance rejection controller is designed for second-order systems.Real-time current data and voltage data are collected to build a hyperlocal model about motor current and voltage,and based on the hyperlocal model,a current model prediction controller is designed to eliminate the influence of parameter perturbation.In summary,the thesis studies the servo system control problem based on the model prediction method for the friction and parameter perturbation problems of PMSM in servo control repetitive motion,which improves the rapid response ability and anti-interference performance of the servo system.In addition,the effectiveness of the designed algorithm is verified by using d SPACE and RT-Lab co-simulation experiment platform.