A Fault Types Recognition System of Power Cable Joint by Hilbert–Huang Transform with Fractal Feature Enhancement Through Mountain Clustering

A defect type recognition system based on the Hilbert–Huang Transform (HHT) with fractal feature enhancement and the mountain clustering approach through partial discharge (PD) signal analysis for 25 kV XLPE underground power cable joint is proposed. First, the eligible product and three types of defects are constructed according to common fault types of power cable joints. The acoustic emission (AE) signal collection system of PD is set for on line testing. The collected AE signal is applied by HHT to transform them into a time-frequency spectrum. Fractal dimension and lacunarity based on fractal theory, are used to extract features from the HHT time-frequency spectrum. Then, the mountain clustering methods are applied to recognize defect types from the test model. The recognition results has been analyzed for appliances. As the field detection of fault cable under voltage is often disturbed by external noise, an empirical model decomposition is utilized to filter the noise in the PD signal. To demonstrate the effectiveness of the proposed approach, the identification ability is investigated on 120 sets of field-tested PD patterns of XLPE power cable joints. The study results indicate that, under a 15% noise level, the recognition rate of the proposed system can achieve approximately 77%, signifying a great potential in applying the proposed recognition system to field measurements in the future.