摘要
介绍了反向传播人工神经元网络在变压器故障诊断中的应用 ,建立了一种新型的基于综合特征输入的溶解气体分析DGA模型 ,说明了它的工作原理、样本分类方法及计算步骤 结果表明 ,该网络模型在变压器故障诊断中 ,经过不断地自适应训练 。
Application of ANN(BP) used for transformer fault diagnosis is introduced. A new dissolved gas analysis model based on comprehensive characteristic input is set up. Its work fundamental principle, the sample classification method and the main calculation steps are presented. The result shows that network model can be used for transformer fault diagnosis and can improves obviously the diagnosis accuracy with self adapted training well.
关键词
故障诊断
气体分析
电力变压器
神经网络
Fault diagnosis
Gas analysis
Power transformers
Neural networks
