基于振动信号统计特性的变压器绕组状态监测

Condition monitoring of transformer windings based on statistics feature of vibration signal

运行中的变压器振动信号与其绕组状态密切相关,为深入挖掘变压器振动信号的变化规律,实现绕组状态的准确监测,本文从在运变压器的振动监测信号特性出发,通过对预处理后的变压器振动信号进行系统聚类,得到了变压器绕组振动特征曲线,进而根据变压器振动信号的统计特性即T2控制图对绕组状态进行监测分析。对某500k V变压器振动在线监测信号的分析结果表明,所提出的系统聚类方法能够有效地提取变压器振动信号的特征,依据振动信号的T^2控制图可以准确地判断变压器的绕组状态。研究结果可为基于振动信号的变压器绕组状态监测提供重要依据。

Vibration signals of the operating transformer are closely related to its winding condition.To further explore the variation tendency of vibration signals of power transformer and realize the condition monitoring of transformer winding,hierarchical cluster is applied to obtain the vibration feature curves of transformer winding based on the processed vibration signals of the operating transformer winding.The T 2 control chart of vibration signals are proposed to judge the winding condition of transformer.The calculated results of the vibration signals of a 500kV transformer have shown that the features of vibration signals are effectively described by the proposed method of hierarchical cluster.The T 2 control chart is capable of accurately assessing the winding condition of transformer.The results could provide the important basis for the on-line condition monitoring of transformer windings.