Position-Sensorless Offline Parameter Identification Method for Permanent Magnet Synchronous Motors Considering Inverter Nonlinearity

To address the issue of electrical parameter identification methods for Interior Permanent Magnet Synchronous Motors (IPMSMs) being affected by inverter nonlinearity, this paper proposes a position-sensorless offline parameter identification method for IPMSMs that considers the influence of inverter nonlinearity. For resistance identification, a DC voltage-current method with nonlinearity compensation is employed. By injecting a DC signal into the d-axis and compensating for the error voltage, the motor resistance parameters are accurately identified. For inductance identification, a high-frequency injection method is utilized, where high-frequency voltage components are injected into the dq-axes, and the fundamental high-frequency current components are extracted to identify the dq-axis inductances. For flux linkage identification, a novel method is proposed by decomposing the flux linkage onto a virtual dq-axis, enabling flux linkage identification without position information. Experimental results based on a 3 kW IPMSM demonstrate that the proposed method achieves sensorless parameter identification with advantages such as fast convergence, effective compensation for inverter nonlinearity, and ease of engineering implementation.