Virtual Signal Injection-Based Online Full-Parameter Estimation of Surface-Mounted PMSMs Without Influence of Position Error and Inverter Nonlinearity

Abstract

In this article, novel virtual signal injection methods are proposed for full-parameter estimation of surface-mounted permanent magnet (PM) synchronous machines (SPMSMs) under sensorless control, including virtual resistance, virtual inductance, virtual voltage/current, virtual back electromotive force (EMF), and virtual flux linkage. When a positive and a negative virtual signal is injected into a flux observer, respectively, it is found that the position error is only related to the ratio of ${q}^{e}$ - and ${d}^{e}$ -axis voltage fluctuation. Therefore, the developed winding inductance, resistance, and PM flux linkage estimation model can be completely independent of position error and inverter nonlinearity at different speeds and loads. Moreover, compared with other virtual signals, the injected virtual flux linkage is independent of rotor speed, which can enhance the estimation performance at low speed to a minimum of 2% of the rated speeds under flux observer-based sensorless control. The experiments are given to validate the proposed method.