Performance Evaluation of High-Frequency Injection Techniques for 48-V PMSM Sensorless Control

Abstract

There is a strong adoption of high-frequency signal injection (HFI) methods to low-speed and zero-speed sensorless control of permanent magnet synchronous motors (PMSMs), in crossover to applications, where back-EMF-based estimation is unusable. The paper gives detailed analysis of comparative performance analysis of numerous HFI strategies used on a 48 V PMSM drive platform that would be utilised in low-voltage industrial and mobility applications. It analytically evaluates three kinds of injection methods, constant-amplitude injection, heuristic adaptive amplitude control, and learning-assisted amplitude optimization in different operating conditions, such as ultra-low speed, startup, and dynamic load disturbance. An integrated synchronous demodulation and phase-locked loop (PLL)-based position extraction system is used to achieve consistent performance in comparison with the other approaches. Important Key performance indicators are: rotor position estimation error (RMSE and peak error), torque ripple factor, current total harmonic distortion (THD), convergence time and power losses due to incremental injection. The experimental and analytical findings show that adaptive amplitude control is key to strongly improve the robustness of estimations and the limitations of torque oscillations in relation to the alternative of fixed-amplitude injection, and the learning-assisted scheme is additionally able to optimise stability to variations in the parameters. Also, the adaptive strategies minimise unnecessary high-frequency excitation, and fewer harmonic distortion and more efficient results are achieved. The research gives both quantitative information and design principles toward application of robust and low-loss HFI-based sensorless operation in 48-V PMO in electric mobility and miniaturisation industrial drive applications.