摘要
针对传统铁路列车车-地无线通信设备网络故障诊断模型结构复杂,诊断精度不高等问题,运用粗糙集理论(RS)、模糊系统(FS)和神经网络(NN)相融合的方法进行铁路列车车-地无线通信设备故障诊断研究。首先对原始样本数据进行模糊化处理,建立故障诊断样本数据表,基于粗糙集理论对故障样本数据进行约简,去除冗余属性,减少样本输入,然后利用约简后的数据训练神经网络,建立基于粗糙集与模糊神经网络车-地无线通信设备故障诊断系统模型结构;最后,将该模型运用于故障诊断中。试验结果对比表明,此方法简化了网络的结构,缩短了训练所需要的时间,提高了故障诊断的精度,从而验证了该方法的可行性。
In view of the complexity and low diagnostic accuracy of the traditional railway train vehicle-ground communication equipment of network fault diagnosis model structure, this paper merges rough set theory （ RS）, fuzzy systems （FS） and neural networks （NN） to conduct fault diagnosis research on railway train vehicle-ground wireless communicating equipment. First, the origin data are fuzzified and then fault diagnosis sample data table is established. Fault sample data are simplified based on rough set theory,redundant attribute is removed,sample input is reduced, and then data are reduced to train the neural network and establish fault diagnosis system model structure based on rough sets and fuzzy neural network vehicle-ground communication equipment. Finally, the model is applied to fault diagnosis. The comparison of test results shows that this method simplifies the network structure, shortens the training time and improves the precision of fault diagnosis, which verifies the feasibility of this method.
出处
《铁道标准设计》
北大核心
2017年第6期180-184,共5页
Railway Standard Design
