一种基于模糊神经网络的变压器故障检测

阐述了基于模糊神经网络变压器故障检测的方法及数学模型．从传统的ＢＡＭ 网络入手，结合模糊理论，根据变压器油的气相色谱分析，运用基于模糊Ｈｅｂｂｉａｎ 学习律的模糊联想记忆（ＦＡＭ），进行变压器故障类型和严重程度检测的方法步骤．模型算法分为两个步骤：激励阶段和冲突解决阶段．试验表明，该方法精度较高，应用方便．

Transformer Fault Detection Based on Fuzzy Neural Network

The fault detection and diagnosis for transformer has become more and more crucial because a serious fault means a significant loss in a blink and may also lead catastrophic events. To prevent these events in advance, the routine maintenance and surveillance is necessary. Recently, a new transformer oil gas chromatography is adopted by many engineers. To upgrade this method, in this paper, the principle of transformer fault detection based on fuzzy neural network and its numerical model was presented, and a fuzzy Hebbian based learning algorithm (FAM) was applied. The study of the paper differs from the previous architecture for multi fault detection. The experimental results show satisfactory agreement and this method can be applied easily with fair high precision.