Research On The Online Parameter Estimation Of Permanent Magnet Synchronous Machine

The vector control system of permanent magnet synchronous motor depends on the accurate knowledge of the motor parameters.However,the motor parameters may change because of the motor temperature and magnetic saturation,thus affecting the control performance.Accurate motor parameters have many benefits for improving the dynamic and steady-state performance of the current controller,reducing the estimation error of the sensorless position and reducing the torque harmonics,and real-time monitoring of the motor operating state.This article aims at online identification of the parameters of permanent magnet synchronous motors,and following research is conducted:This paper analyzes the reasons for the changes of the motor parameters and its influence on the vector control system of the motor.Aiming at the problem of under-rank of the steady-state voltage equation itself,the method using dynamic voltage equation as the identification model for parameter identification is proposed.In this respect,the basic principles of forgetting factor least squares,extended Kalman filtering algorithm and Kalman filtering algorithm are studied.And the implementation of three algorithms in solving parameter identification problems are introduced.In view of the dynamic models used for these three algorithms,the conditions that need to be met to realize multi-parameter identification are discussed.The non-ideal factors existing in the motor vector control system are analyzed,with the causes and effects of control delay,rotor position initial angle deviation and inverter dead zone deeply studied.Based on the basic principle of least squares,a quantitative calculation method of identification error is proposed to analyze the influence of noise in the identification model on the identification result.Corresponding compensation measures are proposed for these three non-ideal factors to improve the accuracy of the identification results.The multi-parameter identification method described in this paper is verified on the permanent magnet synchronous motor experiment platform.The experimental results indicate that the multi-parameter real-time online identification of the motor can be achieved,with the least square method and Kalman filter algorithm taking the shortest time,and the extended Kalman filter algorithm having the best anti-noise performance.The experimental results before non-ideal factor compensation verifies the quantitative calculation method of identification error proposed in this paper,and the results after compensation proves that the compensation method can significantly improve the accuracy of the results.