Linear Regression-Based Dead-Time Identification Strategy for PMSM Using Static Voltage Injection

This paper analyzes the dead-time effect and error voltage model of the inverter in permanent magnet synchronous motors (PMSMs). Based on the current response to d-axis voltage injection when the motor is stationary, a linear regression-based identification model for equivalent dead-time using d-axis voltage injection is proposed. The identification method does not require knowledge of the motor's electrical parameters and offers rapid identification speed. It can accurately identify the equivalent dead time across different motors. Experimental results demonstrate that the proposed method can accurately identify the equivalent dead time under both nominal and non-nominal motor parameters, consistent with the actual inverter dead-time error.