Insulator leakage current analysis based on time-frequency and HHT methods

Leakage current (LC) passing through the contaminated insulator surface under wet conditions was frequently analyzed by FFT, STFT, spectrogram and wavelet transform, in order to reflect the surface condition of insulators, while more or less, these methods have some limitations on nonstationary and nonlinear signals. A recently reported technique, termed HHT, was applied to provide time-frequency information about typical LC waveforms. Comparative study found HHT is more adaptive and has higher frequency resolution than other time-frequency methods, hence more suitable to study the discharge property of LC waveforms. It was shown that the decrease in the ratio between RMS and absolute maximum value of LC indicates the presence of partial arcs. According to this, this paper first separated the nonlinear LC into transient impulses and noisy signal, then employed HHT and wavelet transform to analyze each parts separately. Present study showed that wavelet transform is sensitive to high frequency impulses, and the discontinuity of fundamental harmonic together with the dramatically enlarged high frequency harmonics indicate LC gets into danger region.