Sensorless Controlled PM Flux Linkage Estimation of Direct-Drive DTP-SPMSGs Under Variable Speed Control

Permanent magnet (PM) flux linkage is critical for monitoring the state of permanent magnet in dual three-phase surface-mounted permanent magnet synchronous generators (DTP-SPMSGs). However, for direct-drive DTP-SPMSGs under sensorless control where the blade speed is variable and completely depends on the wind speed, this leads to the variation in position error. This makes conventional PM flux linkage estimation methods ineffective, as the position error is no longer constant but dynamically coupled with operating speed. This paper proposes a virtual flux linkage (VFL) injection-based method for online PM flux linkage estimation of DTP-SPMSGs under sensorless variable-speed operation. By injecting opposite virtual flux linkage pulses into two position observers corresponding to two winding sets simultaneously, the response of difference in observed positions between two sets can be utilized to construct a PM flux linkage estimation model, where the unknown terms including time-varying position error, stator resistance, stator inductance, and inverter nonlinearity are totally cancelled. The proposed method has been validated experimentally.