Sensorless PMSM Drive Inductance Estimation Based on a Data-Driven Approach

In the permanent magnet synchronous motor (PMSM) sensorless drive method, inductance is a decisive parameter for rotor position estimation. Due to core magnetic saturation, current easily invokes variation and degrades estimation accuracy. For constant load torque, saturated error in method are constant. Inductance results minor degradations, such as less optimal torque current, but no speed error. periodic periodically fluctuates and, result, also fluctuate. Periodic makes regulation compensation especially difficult. This paper presents an estimator based on polynomial neural network (PNN) machine learning PMSM with period compensator. By applying estimator, we proposed saturation minimize fluctuation. Simulation experiments were conducted validate by confirming improved accuracy reduced system vibration against torque.