基于声发射技术的输电线路绝缘子放电模式识别

Discharge Mode Recognition of Transmission Line Insulator Discharge based on Acoustic Emission

为正确判断输电线路绝缘子放电的类型,首先在时域状态下求出反应放电产生的超声波信号(放电声发射)波形特征的18个特征指标,利用主成分分析法,在这18个特征指标中提取出了四个互不相关的主成分的表达式,并解释每个主成分所代表的含义,以降低指标维数和信息重叠,提高放电模式识别的速度和准确性。其次,根据得到的特征指标,应用BP神经网络对放电模式进行识别。识别结果表明,利用放电声发射信号进行绝缘子放电模式的识别,可以有效地判断绝缘子放电的类型,该方法为电力设备的在线监测与故障诊断提供了一种新思路。

In order to recognize the discharge type of transmission line insulator , first of all, 18 characteristic indexes which can reflect the features of discharge acoustic emission wave were calculated in time domain state, then, the expression of the 4 principal components which are independent with each other are extracted from the 18 characteristics through principal component analysis, and the meaning of each principal component is explained, so that the target dimension and duplication of information were reduce, and the rate and the accuracy of identification for fault or exception would be improved. Secondly, according to the characteristic indexes, the BP neural network was used for the recognition of discharge pattern. The recognition results indicated that by using the signals of discharge acoustic emission for recognizing the discharge mode of insulators, the insulator discharge type mode could be judged effectively, a new thread was provided by this method for the on-line monitoring and fault diagnosing of power equipment.