摘要
提出采用多层感知器模型应用于变压器故障诊断系统,将遗传算法全局搜索能力强的特点和梯度下降法局部搜索能力强的特点有效结合,增强了神经网络模型的识别效果。
This paper proposed the applying of a multilayer perceptron model in transformer fault diagnosis system,combining the strong global searching ability characteristics of genetic algorithm and the strong local search ability characteristics of gradient method can enhance the recognition effect of neural network model.
出处
《能源与节能》
2012年第11期102-103,107,共3页
Energy and Energy Conservation
关键词
变压器
故障诊断
多层感知器
梯度法
power transformer
fault diagnosis
multilayer perceptron
gradient method
