8 / 2024-03-21 13:57:14

Degradation trend characterization and stage division of AC contactor contact system based on ICEEMDAN-PE-WDR-Wasserstein-Pelt

AC contactor; Data denoising; Deterioration characterization; Change point detection

The stage division of AC contactor degradation state is an important prerequisite for realizing electrical self-perception. However, the degradation state of AC contactor has the characteristics of non-stationary and multi-attribute, and there is a lot of noise interference in the contactor characteristic parameter data, which will aggravate this feature, resulting in that the traditional method cannot effectively divide similar and overlapping characteristic states when dividing the degradation stage. As a result, the division area is too broad, and precise judgment cannot be made in the identification of the AC contactor health state. Aiming at the above problems this paper proposes a characterization of degradation trend and stage division method of AC contactor contact system based on ICEEMDAN-PE-WDR-Wasserstein-Pelt. Firstly, the characteristic parameters were obtained through the full life test of AC contactor, and the ICEEMDAN method was used to decompose the original characteristic parameters once to obtain the exact Intrinsic Mode Function (IMF), and the complexity analysis was carried out by permutation entropy (PE). Secondly, the denoised signal was obtained by Wavelet Denoising Reconstruction (WDR). Then, the Spearman Rank Correlation Coefficient (SRCC) was introduced to obtain the optimal feature subset. Thirdly, the Wasserstein distance was used to quantify the degradation trend of the AC contactor. Finally, the PELT algorithm was used to divide the degradation trend into stages. At the same time, other AC contactor samples of the same type are taken as examples to verify that the method proposed in this paper has good versatility and high accuracy.