基于神经网络的变压器故障诊断方法研究

针对概率神经网络(PNN)模型强大的非线性分类能力,PNN能够很好地对变压器故障进行分类;文章通过对PNN神经网络的结构和原理的分析,应用PNN概率神经网络方法对变压器故障进行诊断;通过实例仿真表明,PNN网络的训练时间比BP网络少,比之预测准确度也要高,而且还具有高度的泛化能力,这使得PNN网络可以有效地运用到变压器故障诊断中,具有一定的可操作性。

Investigation of Transformer Fault Diagnosis Method Based on PNN

For the probabilistic neural network(PNN) model of strong non-linear classification capability,PNN can be good for transformer fault classification.Based on the PNN neural network structure and principles of analysis,application PNN probabilistic neural network approach to transformer fault diagnosis.simulation and by example,we can draw,PNN network compared with the traditional BP neural networks,not only better in training speed and accuracy advantages,but also has a high degree of generalization ability,which makes the PNN network can be effectively applied to the transformer fault diagnosis,has a certain operational and practical value.