Application of Deep Learning in Parameter Estimation of Permanent Magnet Synchronous Machines

Abstract

This paper presents a novel method for real-time identification of four parameters of the permanent magnet synchronous machines (PMSM) namely stator resistance, d-axis inductance, q-axis inductance and the rotor flux linkage. The proposed method is based on the utilization of the deep neural network to solve the problems of the existing model-based parameter estimation methods, which are caused by the non-linearity of the inverter and the inaccuracy of the measured rotor position. Extensive numerical simulations and experimental studies have been conducted to evaluate the robustness and the accuracy of the proposed online parameters identification solution, compared with the conventional methods such as recursive least square, extended Kalman filter and Adaline neural network.

Keywords

• Online parameter identification
• PMSM
• deep learning
• neural network
• recursive least square
• extend Kalman filter