In order to improve efficiency and accuracy of the faultdiagnosis of neural networks, and in view that wavelet has favorable time-frequency characteristic, the paper combines wavelet with neural networks and uses it in power transformer fault diagnosis. Good effects are obtained.%为提高神经网络系统故障诊断的效率和准确率,考虑到小波变换的良好的时-频特性,本文结合小波变换和神经网络并应用于电力变压器的故障诊断中,取得了较好的效果。
展开▼