摘要
分析了基于零序直流原理的单处故障的电缆测距模型,由于模型中的参数复杂,提出了基于RBF网络的故障电缆距离的预测模型,从而实现对基于零序直流原理的故障电缆测距模型中的复杂参数进行辨识。通过仿真采集大量的原始数据,建立RBF神经网络预测模型,与BP(Back Propagation)预测方法进行仿真比较,结果表明RBF神经网络具有更好的预测效果,实现了故障电缆的准确定位,为进一步对该模型的研究与开发奠定了基础。
This paper introduces a fault cable location model for single fault based on the principle of Zero-sequence DC.Since some of the parameters in the model is complex,RBF neural networks is applied to create the prediction model of fault cable in order to identify unknown quantity in the model.By training RBF network prediction model with a large number of raw data collected through experiments and comparing with forecasting method of BP（Back Propagation）,the results show RBF neural network has better predictive performance,achieving accurate fault location of cable and making foundation for researching and developing the model further.
出处
《电测与仪表》
北大核心
2010年第10期41-45,共5页
Electrical Measurement & Instrumentation
