Enhanced Startup and Steady-State Performance in Sensorless PMSM Control Based on Flux Observer and Gradient Descent-Optimized MTPA

作者：Wentao Geng, Xiaoyan Huang · 发表于：Vehicle Power and Propulsion Conference · 年份：2025 · DOI：10.1109/VPPC66000.2025.11392982

In this paper, we address key challenges in sensorless control of permanent magnet synchronous motors (PMSMs) at startup and steady-state, and propose a novel method to improve performance in these regimes. The proposed approach integrates rotor initial state detection with a Maximum Torque Per Ampere (MTPA) control strategy to optimize motor operation. By employing a gradient descent algorithm for real-time optimization, the method reduces computational complexity and accelerates convergence, thereby improving motor efficiency and minimizing power losses. Simulation results demonstrate that the proposed method significantly enhances performance, confirming its effectiveness for real-world PMSM applications.