摘要
针对ZPW-2000型轨道电路故障诊断难、诊断效率低等问题,提出一种基于模糊c均值聚类算法(FCM)和广义神经网络(GRNN)结合的ZPW-2000型轨道电路故障诊断方法。首先采用模糊c均值聚类对故障样本数据分为9类,并得到每类的聚类中心和个体模糊隶属度矩阵,再采用广义神经网络对样本数据作近一步判断,最后采用现场故障数据进行验证,得到较好的诊断精度,因此该法能够为现场维护人员提供诊断辅助,提高了诊断效率。
In order to solve the problems of ZPW-2000 track circuit fault diagnosis,such as the difficulty of fault diagnosis and the low efficiency of fault diagnosis,a method of ZPW-2000 track circuit fault diagnosis based on the combination of fuzzy c-means clustering algorithm(FCM)and generalized neural network(GRNN)is proposed.Firstly,the fuzzy c-means clustering is used to divide the fault sample data into 9 categories,and the clustering center and the fuzzy membership matrix of each category are obtained.Then,the generalized neural network is used to make a further judgment on the sample data.Finally,the field fault data is used to verify,and the better diagnosis accuracy is obtained.Therefore,the method can provide diagnosis assistance for the field maintenance personnel and improve the diagnosis efficiency.
作者
孙彤
褚俊英
刘玉杰
Sun Tong;Chu Junying;Liu Yujie(Seaside Transportation College,Cangzhou Hebei,061100)
出处
《电子测试》
2020年第5期62-64,共3页
Electronic Test
基金
沧州市重点研发计划指导项目(2014AA110501)
北京交通大学海滨学院大学生创新创业训练计划项目资助(201914202056)
