中文摘要：

为准确识别高压电气设备中被噪声淹没的局部放电信号,提出了一种基于相关概率小波变换的局部放电信号检测方法。该方法首先对采集到的原始信号进行小波变换,利用局部放电信号与噪声信号相关性的不同对各层的小波系数进行预处理,然后基于分位数的概念在处理后的各分解层上设置若干个多尺度阈值,根据这些阈值计算原始信号各处为局部放电信号的概率,最后根据概率值的大小来判断该处是否发生局部放电。利用该方法对仿真及实测信号进行分析,并与传统小波变换方法的处理结果进行比较。结果表明,该方法能够更为有效地抑制局部放电在线监测中的噪声干扰,全面、可靠地检测到强噪声背景下的微弱局部放电信号,具有一定的工程应用价值。

英文摘要：

In order to accurately identify the partial discharge signal overwhelmed by noise in high voltage electrical equipment, a method based on correlated probabilistic wavelet transform is proposed for the accurate detection of partial discharge signal in this paper. First, the original signal measured is decomposed by wavelet transform, and the wavelet coefficients of each level is preprocessed according to the difference of correlation between the partial discharge signal and noise signal, and then the multi-scale thresholds at each processed decomposition level are set based on quantile, and the probability which indicates the possibility of each point in the original signal to be partial discharge signal is computed based on these thresholds, finally determine whether there is partial discharge or not according to its probability. The method presented in this paper was applied on the simulation and measurement signal, and compared with the result of traditional wavelet transform. The results show that the method can suppress the noise more effectively in partial discharge online monitoring, which are more comprehensive and reliable to detect weak partial discharge signal in the strong noise background, and has certain engineering value.