概率神经网络在母线故障识别中的应用研究纪

摘要:

提出一种利用概率神经网络来进行母线故障识别的新方法。首先用PSCAD/EMTDC 软件对母线故障状态进 行仿真，提取仿真数据，进行数据预处理，然后建立概率神经网络模型，训练网络后判断母线的区内、区外故障及正确 识别各种故障类型，从而验证基于概率神经网络的母线故障识别的准确性。

Abstract:

A new method based on probabilistic neutral network ( PNN) is proposed to identify the fault of bus． Firstly，the software of PSCAD/EMTDC is used to simulate the state of the fault bus，and the data is abstracted and preprocessed． Then amodel of PNN is established． Sequentially，the accuracy of bus fault identification based on PNN can be verified by discriminatingall kinds of bus faults correctly after training．