Virtual Active Flux Injection-Based Online Full-Parameter Estimation of IPMSMs Decoupling From Sensorless Control

Both online full-parameter estimation and flux linkage based sensorless control depend on the machine model of interior permanent magnet (PM) synchronous machines (IPMSMs). A coupling influence between these two processes exists, leading to degradation in estimation accuracy of dq-axis inductances, stator resistance, and PM flux linkage, as position errors increase. In this paper, a new signal injection method, i.e., square wave virtual active flux injection, is proposed to decouple parameter estimation from sensorless control. When a square wave virtual active flux is injected into the estimated active flux in a position observer, it is found that the position error caused by parameter mismatches keeps constant and is only related to the ratio of q-axis to d-axis voltage fluctuation. Therefore, the ratio of q-axis to d-axis voltage fluctuation is independent of unknown parameters and can be minimized to compensate the position error. Consequently, the developed full-rank parameter estimation model is decoupled from sensorless control. Moreover, the injected virtual active flux is independent of rotor speed, which can enhance the performance of full-parameter estimation at low speed, down to 2% of the rated speed under flux-linkage based sensorless control. The experiments have validated the proposed method.