基于神经网络综合分析的变压器油色谱在线监测系统

Online Monitoring System of Transformer Oil Chromatography Based on Neural Network Integrated Analysis

油色谱在线监测是电力变压器在线监测领域常用的方法之一。变压器故障诊断的结果将直接作为变压器是否需要检修的决策依据,鉴于变压器故障原因的复杂性,仅靠单一的故障诊断方法很难满足故障诊断的要求,故将传统故障诊断方法与BP神经网络方法通过Borda模型相结合,以提高变压器故障诊断的准确率,最后用C#语言设计开发了故障诊断系统。该系统改变了以往的定期试验模式,实现了变压器状态在线监测和分析。

Oil chromatography monitoring is one of the commonly used methods in the online monitoring field of power transformers. The results of fault diagnosis for a transformer will provide a direct basis for determining whether the transformer needs an overhaul. Due to the complex reasons of transformer faults, it is difficult for a single diagnosis method to meet the requirements of fault diagnosis. In order to improve the ac- curacy of fault diagnosis, this paper combined the traditional fault diagnosis method with the back propagation （BP） neural network method by the Borda model. The fault diagnosis system developed by C# language improved the traditional mode of routine test and achieved online monitoring and analysis for the state of power transformers.