Abstract:In the sensorless control system for interior permanent magnet synchronous motor based on the extended Kalman filter (EKF) approach, accurate estimation of motor speed and rotor position is important. To address the issue of decreased estimation accuracy for speed and rotor position due to phase current gain errors, a method for compensating position estimation errors is proposed. Firstly, through theoretical derivation, it is shown that phase current gain errors can cause estimated speed and rotor position to oscillate at twice the supply frequency during the EKF estimation process. Secondly, a novel structure for the EKF observer is designed by adding a gain error coefficient solving section to the traditional EKF observer, which reduces the pulsations in estimated speed and rotor position caused by phase current gain errors, thereby improving estimation accuracy. Finally, simulations and experiments validate the feasibility of the proposed compensation method.