Meticulous prediction of vacuum contactor fault with denoising and decomposition based temporal model

Vacuum contactor is a crucial electrical component for power conversion and distribution, and thus its safe performance is paramount in practical industry. Its fault prediction helps to avoid serious accidents. The complete operational cycle of a vacuum contactor generally consists of two stages, namely closing and opening. Most published researches detect faults of vacuum contactor after the whole opening and closing stages, making it challenging to forecast faults before the contactor opens. To address this issue, the present paper proposes a meticulous fault prediction method of vacuum contactor with denoising and decomposition based temporal model. The original current signal is denoised and decomposed to obtain the filtered signal components, which are input to corresponding serial models separately. Therefore, the current components for opening stage can be predicted at the closing stage. The predicted components are further reconstructed for fault prediction. The experimental results indicate that the proposed method achieves a higher fault recognition rate over related approaches, demonstrating its considerable practical value.