Novel Random Square-Wave Voltage Injection Method Based on Markov Chain for IPMSM Sensorless Control

Abstract

The square-wave voltage signal injection into the estimated <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">d</i> -axis is an effective and novel sensorless control scheme for interior permanent magnet synchronous motor at zero speed and low speed. However, the notable acoustic and electromagnetic noises induced by the conventional fixed-frequency signal injection are very harsh and shrill to hear, which limits the practical application. Aiming at reducing unnecessary acoustic noise, a novel random voltage injection method based on the Markov chain (MC) is proposed in this article. Two high-frequency square-wave voltages with different types of frequencies and amplitudes are randomly injected into the estimated <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">d</i> -axis by MC rules, and this way has spread spectrum characteristics, which can effectively reduce the unnecessary noise. In addition, a new signal demodulation compensation method considering the digital delay effect in high-frequency signals is proposed, which effectively reduces the position and velocity estimation errors. Then, the power spectra density of the induced high-frequency currents under the fixed frequency injection method and the proposed MC method are compared and analyzed. The high performance of the proposed MC method has been validated by the experiment drive platform and compared with the traditional fixed frequency under different control conditions.