摘要
为解决牵引供电系统故障识别预测难题,建立一种BP神经网络模型,从系统运行状态信号中提取故障特征参数,并利用MATLAB生成数据集开展模型训练,从而有效提高故障识别预测精度。结合模型预测精度,合理确定测试集占比、神经网络隐藏层节点数和模型算法,并根据模型检验结果修正模型,最终确认模型可以准确识别F-R、F-T和T-R故障。结果表明,模型可有效识别预测故障信号,保证牵引供电系统的运行安全性。
In order to solve the problem of fault identification and prediction of traction power supply system,a BP neural network model was established to extract fault characteristic parameters from the system running state signal,and MATLAB was used to generate data set to carry out model training,which effectively improved the accuracy of fault identification and prediction.Combined with the prediction accuracy of the model,the proportion of test set,the number of hidden layer nodes of the neural network and the model algorithm were reasonably determined,and the model was modified according to the model test results.Finally,it was confirmed that the model could accurately identify F-R,F-T and T-R faults.The results show that the model can effectively identify and predict fault signals,ensure the operational safety of traction power supply system.
作者
张镇鸿
ZHANG Zhenhong(Basic Department of Guangxi Coastal Railway Co.,Ltd.,Nanning,Guangxi 530000,China)
出处
《自动化应用》
2025年第2期122-124,127,共4页
Automation Application
关键词
BP神经网络
牵引供电系统
故障识别预测
BP neural network
traction power supply system
fault identification and prediction
