| Permanent magnet synchronous motor（PMSM）has attracted more and more attention in recent years.One reason is that the rare earth materials used to make permanent magnet synchronous motors have long been discovered,and rare earth resources are not scarce,at least in China,where they are abundant.The second reason is that permanent magnet synchronous motor has many advantages.For example,the volume and mass is small,and it is convenient to install in the drive system of some small space so that it can be controlled by people flexibly;The PMSM requires no reactive power,so it has a high power factor.Since permanent magnet synchronous motor plays an important role in people’s life at present,scholars from all walks of life have been studying permanent magnet synchronous motor with high enthusiasm,and studying what kind of control algorithm can improve the control performance of the motor.The design of the speed and current controller of the traditional vector control system is based on the PI controller.This control method is simple and can achieve the effect of no static difference in the steady state.However,the d-q axis is not fully decoupled,which causes certain influence on the dynamic and static performance of the current.The time of overshoot and steady-state recovery cannot be optimized at the same time.If the sliding mode control algorithm is used to design the speed controller,the performance of the speed loop control is not affected by the change of motor parameters.So this paper proposes the speed sliding mode controller applied in the speed ring.With the development of control theory,predictive control theory has been applied in PMSM vector control system by most scholars to solve the shortcomings of traditional PI control.In this paper,the minimum beat current predictive control theory is used to design the current loop controller,because the minimum beat current predictive controller can reduce the error of current to zero in the minimum sampling period,so as to realize the high bandwidth operation of the system.The minimum beat current prediction controller can improve the dynamic response of PMSM,reduce current harmonics,and improve the accuracy of current prediction and the performance of the control system.However,due to the sampling of signal variables,the traditional minimum beat current predictive control has a large error in the prediction results.Therefore,this paper proposes an improved minimum beat current predictive control algorithm which takes the average system variable as the predictive value in the sampling interval to improve the control accuracy of the system.But current predictive control algorithm design of permanent magnet synchronous motor vector control system of the current loop controller,the internal electromagnetic parameter value depends on the motor,but in the process of the motor running,affected by external environment and its own heating,electromagnetic parameters may change,but the design of the controller when parameters of motor used for the parameters of the motor when the motor nameplate,this makes the design of the parameters of the controller and motor actual runtime parameters of the controller and motor parameter mismatch.In order to solve the influence of parameter mismatch on the low control performance of the control system,this paper identifies the motor parameters to solve the problem.In this paper,two methods for identifying motor parameters are proposed.A differential evolution algorithm is combined with motor parameter identification,based on the idea of optimization to get the value of the parameter to be identified.And aiming at the shortcomings of the standard differential evolution algorithm,an improved adaptive differential evolution algorithm is proposed.Simulation results show that the improved differential evolution algorithm identification precision is higher.The other method is to use Adaline neural network to identify motor parameters.As long as the d-q axis current,voltage and motor speed signals when4)（9）=0 and4(（9）≠0 are collected,the parameter values to be identified are taken as the weights of the neural network.The weight adjustment algorithm is used to adjust the parameter values to be identified to minimize the difference between the motor reference model and the output of the neural network.The final parameter identification value is obtained.In addition,an improved strategy is proposed for the shortcomings of the traditional Adaline neural network algorithm.The experiment verifies that the improved neural network can identify PMSM parameter values with higher accuracy and smaller steady-state error. |
PDF Full Text Request