Novel IGBT Module Switching Time Monitoring Method for Motor Inverters via Adaptive Wavelet‐Based Compressed Sensing Strategy

ABSTRACT

Accurate monitoring of insulated gate bipolar transistor (IGBT) switching transients is critical for ensuring the reliability, efficiency, and fault resilience of permanent magnet synchronous motor (PMSM) inverters. However, conventional high‐speed sampling methods for capturing nanosecond‐scale switching events impose stringent requirements on data acquisition systems, leading to excessive storage, bandwidth consumption, and computational overhead. To address these challenges, this paper proposes a novel adaptive wavelet‐based compressed sensing (AWCS) strategy for real‐time precise monitoring of IGBT switching times in motor inverters. By leveraging the sparsity of switching voltage/current waveforms in the wavelet domain, the proposed approach enables faithful reconstruction of transient waveforms from significantly undersampled measurements, thereby reducing sampling rates by up to 98% without compromising timing accuracy. A tailored compressed sensing framework and an oscillation‐aware iterative reconstruction algorithm, incorporating adaptive wavelet basis selection and oscillation‐sensing weighted orthogonal matching pursuit (OSW‐OMP), are developed to enhance recovery fidelity and robustness against noise in switching transient signals. Experimental validation on an IGBT double pulse test rig and the PMSM three‐phase IGBT‐based inverter demonstrates that the proposed method achieves switching time monitoring errors below 5 ns at a compression ratio (CR) = 50, outperforming standard CS and wavelet‐thresholding baselines method.