变压器局部放电在线监测信号处理技术的研究

Signal processing technique for on-line partial discharge monitoring of power transformers

针对运行中变压器局部放电在线监测信号中的强烈窄带干扰,分别研究了基于快速Fourier变换(FFT)和时域分析算法的滤波器,对放电高频信号进行信号处理,以抑制干扰。处理结果表明,高频放电信号样本的信噪比可以提高15～20dB。对处理后的高频信号和同时采集到的超高频信号进行了综合分析,发现两种放电样本中的放电脉冲在时延和幅值上均存在着十分明确的相关关系,而干扰样本中则不存在这种对应关系。这一结果可以为局部放电信号的识别提供分析判断的依据。

The strong narrow band interference in data obtained through partial discharge online monitoring systems was reduced using filters based on an FFT and time domain analysis algorithm to process highfrequency partial discharge signals. Processing results showed that the signalnoise ratios (SNR) of highfrequency partial discharge of signals could be improved 15～20 dB. Analysis of the highfrequency signals and ultra highfrequency signals sampled at the same time showed a strong correlation between the two discharge time delays and amplitudes. Interference samples did not have the same kind of correlation. The signal correlation provides a criterion for discriminating PD signals from interference.