Abstract:

Motors are typically operating at varying speed conditions, especially when they are driven by inverters for high efficiency. However, it is challenging to detect motor faults under varying operating conditions due to the frequency variation of fault signatures, spectrum distortion, and other interference. Even the fault signature is extracted, the fault severity is often underestimated and not robust. To address these issues, we propose a sparsity-driven method to iteratively extract speed- dependent fault frequency components from the stator current, considering frequency variation due to varying speed. Experiments show that our method can extract frequency signatures of different faults with significantly better results compared to other state-of-the-art methods. The proposed method can be applied to inverter-fed motor drive system for robust fault detection.