An Intelligent Recognition Method of a Short-Gap Arc in Aviation Cables Based on Feature Weight Enhancement

The cases of loose pin of electrical connectors and broken wire harnesses easily occur in aviation cables due to vibration and thus cause the short-gap arc fault. When the vibration frequency exceeds 30 Hz, the short-gap arc fast burning and extinguishing properties make the changes in time domain and frequency domain of fault signals weak, thus increasing the difficulty in the identification of such a fault. In this article, an arc recognition method based on weight enhancement (WE) and principal component analysis (PCA) is proposed to solve the identification problem of the short-gap arc fault caused by its unapparent frequency-domain characteristics. In the proposed method, the currents under different conditions were measured and analyzed based on the considerations of different vibration conditions and loads. The measured current signals were first decomposed with the empirical modal decomposition (EMD) method, and then, optimal intrinsic mode functions (IMFs) were selected based on the correlation coefficient and kurtosis index to reconstruct the signals with WE algorithm. The time-domain, frequency-domain, and sample entropy characteristics extracted from reconstructed signals were used to establish the 3-D eigenvectors with direct fusion algorithm. Finally, the short-gap arc fault was identified with a principal element analysis and dimension reduction method. The experimental results proved that WE could improve the selection of the optimal direction of PCA dimensional reduction projection and that the short-gap arc fault of aviation cables could be quickly and effectively identified with the proposed method.