首页|基于MDS和改进SSA-SVM的高速铁路道岔故障诊断方法研究

基于MDS和改进SSA-SVM的高速铁路道岔故障诊断方法研究

扫码查看
原文链接
  • 万方数据
  • 维普
针对高速铁路道岔设备故障频繁,现场维修工作量大等问题,提出基于多维尺度缩放法(MDS)和改进麻雀搜索算法(SSA)优化支持向量机(SVM)的高速铁路道岔故障诊断模型.首先以ZDJ9道岔转换功率曲线为研究对象,总结现场典型道岔故障类型及故障原因,分别提取道岔功率曲线的时域、频域特征指标以及小波包能量熵,组成特征指标向量;其次采用MDS方法进行多维特征指标的降维优化,建立道岔故障特征指标样本数据库;最后利用改进Circle混沌映射初始化种群,并通过自适应t分布增强麻雀种群的多样性,再以改进SSA算法优化SVM模型中的惩罚因子和核函数方差2个关键参数,构建改进SSA-SVM的道岔故障诊断模型.故障诊断结果表明,本模型的故障诊断正确率高达96.25％,诊断效果优于其他方法,可以为道岔设备的故障维修提供理论依据.
Research on Fault Diagnosis Method for High-speed Railway Turnouts Based on MDS and Improved SSA-SVM
Aiming at the frequent faults of high-speed railway turnout equipment and the heavy workload of on-site main-tenance,a fault diagnosis model of high-speed railway turnouts was proposed based on multiple dimensional scaling(MDS)and support vector machine(SVM)optimized by improved sparrow search algorithm(SSA).Firstly,based on the switching power curve of ZDJ9 turnout and the summary of the fault types and causes of typical turnouts in the field,the time-domain and frequency-domain characteristic indexes of the turnout power curve and the wavelet packet energy entropy were extracted respectively to form the characteristic index vector.Secondly,by using the MDS method to opti-mize the dimension reduction of multidimensional feature indicators,the sample database of fault diagnosis feature indi-cators was established.Finally,the improved Circle chaotic map was used to initialize the population,before the diversi-ty of sparrow population was enhanced through the adaptive t distribution.By optimizing the two key parameters of penal-ty factor and kernel function variance in the SVM model using the improved SSA algorithm,the improved SSA-SVM turnout fault diagnosis model was constructed to realize the turnout fault diagnosis.The fault diagnosis results show that this model,with the fault diagnosis accuracy as high as 96.25％,demonstrates better diagnostic effect than other meth-ods,which can provide a theoretical basis for the fault maintenance of turnout equipment.

high-speed railway turnoutfault diagnosisimproved SSA-SVMCircle chaos mappingself-adaptive t-distri-butionwavelet energy entropymultiple dimensional scaling

王彦快、米根锁、孔得盛、杨建刚、张玉

展开 >

兰州交通大学铁道技术学院,甘肃兰州 730000

兰州交通大学 自动化与电气工程学院,甘肃兰州 730070

中国铁路兰州局集团有限公司兰州电务部,甘肃兰州 730000

北京全路通信信号研究设计院集团有限公司,北京 100070

国网甘肃省电力公司电力科学研究院,甘肃兰州 730070

展开 >

高速铁路道岔 故障诊断 改进麻雀搜索算法-支持向量机 Circle混沌映射 自适应t分布 小波包能量熵 多维尺度缩放法

甘肃省科技计划兰州交通大学青年科学基金

21JR7RA3052021027

2024

10.3969/j.issn.1001-8360.2024.01.009

铁道学报
中国铁道学会

铁道学报

CSTPCD北大核心
影响因子：0.9
ISSN：1001-8360
年,卷(期)：2024.46(1)
  • 22