大型变压器故障的机理难以动用准确的数学模型加以描述。故障现象与故障原因之间存在着很多不确定性。在利用模糊理论处理问题的同时，本文以特征气体法为例，引入了人工神经网络（ＡＮＮ）理论，建立一个基于ＡＮＮ变压器故障诊断模型；并探讨了以此模型为基础进行故障诊断的可能性与准确性 The mechanism of large transformer failure is difficult to describe with accurate mathematical models. There is a lot of uncertainty between the cause of the malfunction and the cause of the malfunction. While using the fuzzy theory to deal with the problem, this paper takes the characteristic gas method as an example, introduces artificial neural network (ANN) theory to establish a fault diagnosis model based on ANN transformer, and discusses the possibility of fault diagnosis based on this model With accuracy