基于DGA技术与BP神经网络融合的变压器故障诊断

Transformer Fault Diagnosis Based on DGA Technology and BP Neural Network

运用油中气体数据分析(DGA)技术是目前诊断变压器故障方法中最普遍也是运用最广泛的,它在变压器的预防诊断试验中具有很重要的作用。但由于故障的多样性,DGA技术不能全面的准确地检测出多重故障。因而将加入BP神经网络理论使之与DGA技术融合来判断变压器的故障,能够更全面更精确地检测出故障。

Using data analysis technology of gas oil（DGA） is currently the most common method for transformer fault diagnosis and is also the most widely used. It has a very important role in the prevention of transformer＇s test. However, due to the diversity of faults, DGA technique can not accurately detect all of multiple faults. So we use the amalgamation of DGA method and BP neural network to determine the transformer＇s failure, and make it more comprehensive and accurate to be detected.