High-sensitive state perception method for inverter-fed machine turn insulation based on FrFT-Mel

Abstract

Amidst the swift advancement of new power systems and electric vehicles, inverter-fed machines have progressively materialized as a pivotal apparatus for efficient energy conversion. Stator winding turn insulation failure is the root cause of inverter-fed machine breakdown. The online monitoring of turn insulation health can detect potential safety risks promptly, but faces the challenge of weak characteristics of turn insulation degradation. This study proposes an innovative method to evaluate the turn insulation state of inverter-fed machines by utilizing the fractional Fourier transform with a Mel filter (FrFT-Mel). First, the sensitivity of the high-frequency (HF) switching oscillation current to variations in turn insulation was analyzed within the fractional domain. Subsequently, an improved Mel filter is introduced, and its structure and parameters are specifically designed based on the features intrinsic to the common-mode impedance resonance point of the electrical machine. Finally, an evaluation index was proposed for the turn insulation state of inverter-fed machines. Experimental results on a 3kW permanent magnet synchronous machine (PMSM) demonstrate that the proposed FrFT-Mel method significantly enhances the sensitivity of turn insulation state perception by approximately five times, compared to the traditional Fourier transform method.