基于模糊神经网络的25Hz相敏轨道电路故障诊断研究

Fault Diagnosis Research for 25 Hz Phase Sensitive Track Circuit Based OnFuzzy Neural Network

针对传统模型对25Hz相敏轨道电路故障诊断网络求解过程中普遍收敛速度慢,诊断精度不高,文章提出BF神经网络模型来求解,达到对25Hz相敏轨道电路进行故障诊断的目的。考虑到轨道电路特征量的复杂性、模糊性,对特征参量进行预处理,并模糊下。然后,利用BF神经网络相关参数,基于模糊神经网络模型进行网络诊断研究。由仿真结果可知,用文章所构建的模糊神经网络模型诊断精度高,且扩展性强。

In view of the traditional model of 25 Hz phase sensitive track circuit fault diagnosis network is generally slow convergence speed in the process of solving, diagnostic accuracy is not high, this article put forward the BF can be solved by neural network model, up to 25 Hz phase sensitive track circuit fault diagnosis. Considering the characteristics of track circuit complexity, fuzziness, the preprocessing, feature parameter and fuzzy. Then, using the BF neural network parameters, based on fuzzy neural network model for network diagnosis research. The simulation results shows that using fuzzy neural network model is constructed in this paper high diagnostic accuracy, and strong expansibility.