Sensorless Position Estimation, Parameter Identification and Control Integration for Permanent Magnet Synchronous Machines using Current Derivative Measurements

This paper presents the position and speed estimation techniques for IPMSM in which the current derivative measurements during each PWM cycle are used. This technique requires no additional current injection. Instead, the rates of rise and fall of machine currents due to certain voltage-vector excitations during each PWM switching period are used for estimating the speed and position. Furthermore, it has been shown that these derivatives can also be used to track the inductance parameter variations L d and L q , of which the latter is known to vary significantly with load. The use of current derivative based position, speed and inductance parameters have fast response time compared to existing methods such as signal injection, and have potential for improving model based sensorless control, direct torque control, model predictive control and for more accurate trajectory (such as MTPA and MTPV) following under FOC and sensorless control. This paper presents the theory and experimental evaluation of current derivative based estimators for sensorless IPMSM control over wide speed range from zero to rated speed.