基于交叉小波变换的变压器振动信号分析

摘要

Analysis of correlation of vibration signal spectrum on each measuring point of the transformer and re⁃moval of spectrum interference of other measuring points have important significance for accurate extraction of characterization of transformer working parameters and realization of on⁃line monitoring. In the paper, cross wavelet transform is used to analyze the correlation of the signals at different measuring points of the transformer in time⁃space domain, so as to obtain the rule of the variation of versus time and frequency, and also the spectrum compo⁃nents that are strongly correlated with the signals. The obtained results are compared with the spectrum components of the original signals, and the spectrum components that are independent of other signals of the measuring points can be obtained. The results thus obtained can provide effective reference for monitored transformer’ s operating conditions.%分析变压器各测点上振动信号频谱之间的相关性，去除其他测点信号的频谱干扰，对于准确提取表征变压器工作状态的特征参数、实现变压器的在线监测具有重要意义。本文应用交叉小波变换在时⁃频域空间中分析变压器上不同测点振动信号间的相关性，得出振动信号间的相关性随时间和频率变化的规律及信号间相关性强的频谱成分，再与原信号的频谱成分相比较后可提取出该信号独立于其他测点信号的频谱成分，为振动法在线监测变压器的运行状态提供参考。