Research On Simplified Three-Vector Model Predictive Current Control Of PMSM Based On Predictive Error Compensation

Permanent magnet synchronous motor(PMSM)has been widely used in the industrial field due to its advantages of high power density,wide speed regulation range and large torque-to-current ratio.As an advanced motor control strategy,PMSM model predictive control(MPC)strategy has the advantages of simple principle,easy handling of nonlinear constraints,and fast dynamic response.However,the steady-state control performance of single-vector MPC strategy is poor,while the multi-vector MPC strategy can improve the steady-state control performance while increasing the calculation amount.In addition,the MPC strategy has a strong dependence on the prediction model.When the motor parameters in the prediction model do not match the actual motor parameters,a prediction error will occur between the predicted value and the actual value,which will affect the control performance of the strategy.Aiming at the above problems,this thesis takes the model predictive current control(MPCC)strategy as an example to study reducing the computation amount of multi-vector MPC and suppressing the adverse effects of parameter mismatch on MPC.Firstly,the mathematical model of permanent magnet synchronous motor is established,and the principles of traditional single-vector MPCC,double-vector MPCC and three-vector MPCC are analyzed.A simplified three-vector MPCC method is proposed to solve the problem of large computation in traditional three-vector MPCC method.In this method,only one sector needs to obtain the corresponding voltage vector action time by solving the current deadbeat equation,the action time of the voltage vector corresponding to other sectors can be directly obtained by simple addition and subtraction operations according to the principle of vector synthesis,the amount of calculation is reduced.In addition,a cost function with voltage error as a constraint is introduced,which eliminates the need to perform current prediction on the voltage vector to be selected,further reducing the amount of calculation.The simulation results show that the simplified three-vector MPCC method has the same dynamic and steady-state control performance as the traditional three-vector MPCC method while reducing the amount of calculation.Secondly,aiming at the problem that the prediction error caused by motor parameter mismatch leads to the decline of strategy control performance,the prediction error compensation method is studied.In the prediction error compensation method,the stability of the compensation coefficient is easily affected by the current sampling noise,so a processing method of the compensation coefficient is designed in this thesis.In this method,the compensation coefficient is filtered by the first-order low-pass filter function with variable filter coefficient,and the filter coefficient is taken according to the reliability of the compensation coefficient in the current control period.A simulation model is built by combining the prediction error compensation method with the simplified three-vector MPCC strategy.The simulation results show that the compensation coefficient processing method designed in this thesis has a good suppression effect on the compensation coefficient pulsation caused by the current sampling noise,and the processed compensation coefficient can effectively compensate the prediction error of the simplified three-vector MPCC strategy when the parameters mismatch.Finally,a PMSM platform is built to verify the proposed method.The analysis of the experimental results shows that the simplified three-vector MPCC strategy has good dynamic and steady-state control performance,after combined with the prediction error compensation method,the d-q axis current can accurately track the given value when the parameters are mismatched,and the pulsation amplitude does not increase obviously compared with the parameter matching.