基于EDBN-SVM的高速列车故障分析
发布时间：2019-05-20 16:49
【摘要】：深度学习作为机器学习领域的新热点,为故障诊断技术领域的研究开拓了新的思路。针对高速列车进行故障分析的重要性,将深度学习和集成学习相结合,提出一种基于EDBN-SVM(EnsembleDeep Belief Network-Support Vector Machine)的故障诊断模型。首先对高速列车振动信号进行快速傅立叶变换,其次分析确定了EDBN-SVM模型的参数,然后将信号的FFT系数作为EDBN-SVM模型的可视层输入,并逐层学习高层特征,最后利用多个SVM分类器进行识别并对识别结果进行集成。为评估该方法的有效性,采用实验室数据和仿真数据进行实验测试,并与传统的几种故障分析方法进行对比。结果表明,该方法的故障识别效果优于传统的故障分析方法,同时稳定性更好。
[Abstract]:As a new hot spot in the field of machine learning, deep learning opens up new ideas for the research of fault diagnosis technology. In view of the importance of fault analysis of high-speed trains, a fault diagnosis model based on EDBN-SVM (EnsembleDeep Belief Network-Support Vector Machine) is proposed by combining deep learning with integrated learning. Firstly, the fast Fourier transform is carried out on the vibration signal of high-speed train. Secondly, the parameters of EDBN-SVM model are analyzed and determined, and then the FFT coefficient of the signal is used as the visual layer input of EDBN-SVM model, and the high-level features are learned layer by layer. Finally, multiple SVM classifiers are used to identify and integrate the recognition results. In order to evaluate the effectiveness of this method, laboratory data and simulation data are used for experimental testing, and compared with several traditional fault analysis methods. The results show that the fault identification effect of this method is better than that of the traditional fault analysis method, and the stability of the method is better than that of the traditional fault analysis method.
【作者单位】： 西南交通大学信息科学与技术学院;西南交通大学电气工程学院;
【基金】：国家自然科学基金项目(61134002,61572407)资助
【分类号】：U279.33
本文编号：2481793
[Abstract]:As a new hot spot in the field of machine learning, deep learning opens up new ideas for the research of fault diagnosis technology. In view of the importance of fault analysis of high-speed trains, a fault diagnosis model based on EDBN-SVM (EnsembleDeep Belief Network-Support Vector Machine) is proposed by combining deep learning with integrated learning. Firstly, the fast Fourier transform is carried out on the vibration signal of high-speed train. Secondly, the parameters of EDBN-SVM model are analyzed and determined, and then the FFT coefficient of the signal is used as the visual layer input of EDBN-SVM model, and the high-level features are learned layer by layer. Finally, multiple SVM classifiers are used to identify and integrate the recognition results. In order to evaluate the effectiveness of this method, laboratory data and simulation data are used for experimental testing, and compared with several traditional fault analysis methods. The results show that the fault identification effect of this method is better than that of the traditional fault analysis method, and the stability of the method is better than that of the traditional fault analysis method.
【作者单位】： 西南交通大学信息科学与技术学院;西南交通大学电气工程学院;
【基金】：国家自然科学基金项目(61134002,61572407)资助
【分类号】：U279.33
【相似文献】
相关期刊论文 前10条
1 张建华;;浅谈铁路给水设备维护及常见故障分析[J];科技创业家;2013年10期
2 高明;;东风_4型机车一起水温高的故障分析[J];内燃机车;1992年12期
3 严万千,孟蕊;运用科学方法进行故障分析[J];铁道通信信号;2002年02期
4 陈启成;东风_4型机车新式轴箱轴承内圈崩裂故障分析及防止措施[J];内燃机车;1995年09期
5 原毅;提速客车运行二年来故障分析及对策[J];上海铁道科技;1999年03期
6 冯捷;120型紧急阀故障分析[J];铁道机车车辆;2000年04期
7 翟玉江;燃油箱进水造成柴油机爆燃振动的故障分析[J];内燃机车;2003年10期
8 高伟东;东风_4型机车一起冷却风扇转速低的故障分析[J];内燃机车;1994年04期
9 张勤,智强;轨道车故障分析与控制综述[J];铁道技术监督;2005年03期
10 郝春霖;;运行客车检修故障分析[J];内蒙古科技与经济;2006年18期
相关会议论文 前2条
1 孙新根;;302D-Z型调速器故障分析与改进[A];河南省铁道学会2007年学术活动月优秀论文选集[C];2007年
2 孙新根;;302D—Z型调速器故障分析与改进[A];河南省铁道学会2007年学术活动月优秀论文集[C];2007年
相关硕士学位论文 前1条
1 黄冬;轨道车辆门可靠性数据管理及故障分析系统的研究与实现[D];南京理工大学;2007年
，本文编号：2481793
本文链接：https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/2481793.html
