联合时频分析应用于局部放电信号处理的研究

Study on Pulse Current of Partial Discharge by Using Joint Time-Frequency Analysis Methods

将短时傅立叶变换谱、Wigner分布及 Gabor谱等 3种联合时频分析方法应用于离散时间域 ,分析并处理了油中沿面放电波形。结果表明 ,时频分析方法可以较好地描述局部放电信号频谱的时变特性。其中 ,Gabor谱不仅在时频域具有和 Wigner分布相同的高分辨率 ,优于短时傅里叶变换谱 ,而且能有效地消除 Wigner分布中交叉干扰的影响。对时频谱的分析还发现局部放电信号的频率分量可达到 GHz,为超高频和超宽频带局部放电检测的研究提供了理论依据。

This paper applied three joint time frequency analysis (JTFA) methods to a surface discharge signal in oil, i.e., Short time Fourier transformation (STFT) based spectrogram, Wigner distribution (DWD) and Gabor spectrogram (DGS). Experimental results show that JTFA methods can describe clearly how the spectral content varies with time. It is also found that the DGS has high resolution in both frequency domain and time domain without the cross term defects, so it is more suitable for PD signal analysis than the STFT based spectrogram and the DWD. From the time frequency spectrum, we can know that the frequency spectrum of a PD signal contains harmonics up to 1GHz, in some cases>1GHz, which makes a reference for UHF and UWB detection in theory. It is useful for laying foundations on setting PD pattern recognition and increasing the measuring accuracy in insulation detection.