Research Of Prediction Error Analysis And Compensation Of Predictive Control For PMSM

Predictive current control(PCC)is a model-based control method,which is quite sensitive to motor parameters due to the large number of parameters in motor prediction model.In practical applications,motor parameters cannot be accurately measured and may change under different working conditions,which will cause a deviation between parameter in prediction model and actual motor.Parameter deviations will cause an inaccurate prediction model,resulting in prediction errors and poor control performance of predictive control system.Therefore,the research of predictive error analysis and compensation of PCC is particularly important.This paper mainly analyzes the current prediction error of PCC for permanent magnet synchronous motor(PMSM),and proposes an online parameter correction method based on the relationship between inductance mismatch and current prediction error to reduce the current prediction error and to achieve the indirect compensation of prediction error.Firstly,the prediction error of PCC for PMSM is analyzed.The main source of current prediction error is obtained by motor prediction model.For the effect of motor parameter deviation on current prediction error,the mapping relationship is derived according to prediction model of PMSM.Then,the effects of stator inductance deviation,stator resistance deviation and magnetic flux deviation on current prediction error are analyzed.The influence law and influence degree of three motor parameter deviations are analyzed and summarized.Then,in order to realize the compensation of prediction error,an online parameter correction method of PCC for PMSM is proposed in this paper:(1)the relationship between inductance deviation and prediction error is derived,and the absolute value of inductance correction value is obtained by proportional regulator;(2)the relationships between fluctuation amplitudes of predicted current and actual current and polarity of inductance deviation are analyzed,then,the judgment principle of polarity for inductance correction value is obtained;(3)based on the mechanical rotation period of PMSM,a correction period that can be adaptively adjusted according to motor speed is set to ensure the accuracy of inductance correction value.Finally,simulation and experimental results under different load,speed and inductance deviation conditions are carried out to verify the effectiveness of proposed online parameter correction method.Results show that the proposed method can significantly reduce current prediction error,achieve good error compensation effect and improve control performance.At the same time,the proposed method only needs a proportional regulator to achieve the correction of inductance value,which has a simple structure and low computational complexity.And it is easy to implement in practical applications.