针对电力变压器故障的特点及传统诊断方法在变压器故障诊断中的局限性,提出了基于灰色神经网络的变压器故障诊断方法.首先将典型油中气体浓度样本集作为参考序列,挖掘出样本集中的故障信息,然后利用灰色神经网络进行变压器的故障诊断.通过大量的实例,并将诊断结果与IEC三比值法和改良三比值法的诊断结果相比较,表明基于matlab灰色神经网络的诊断方法具有更高的精确度. Aiming at the limitation of power transformer fault and the limitation of traditional diagnosis method in transformer fault diagnosis, a fault diagnosis method of transformer based on gray neural network is proposed.Firstly, a sample set of gas concentration in typical oil is taken as a reference sequence, Information, and then use gray neural network for fault diagnosis of transformers.Through a large number of examples, and compare the diagnostic results with the diagnostic results of IEC three ratio method and improved three ratio method, it shows that the diagnosis method based on matlab gray neural network has a higher Accuracy