基于超球体高斯分布的悬浮系统异常检测  

Anomaly Detection for Suspension Systems Based on the Gaussian Distribution of Hyperspheres

作　　者：王平 梅子[2] 龙志强 WANG Ping;MEI Zi;LONG Zhiqiang(Facility Design and Instrumentation Institute,China Aerodynamics Research and Development Center,Mianyang,Sichuan 621000,China;College of Intelligence Science and Technology,National University of Defense Technology,Changsha,Hunan 410073,China)

机构地区：[1]中国空气动力研究与发展中心设备设计与测试技术研究所,四川绵阳621000 [2]国防科技大学智能科学学院,湖南长沙410073

出　　处：《机车电传动》2021年第6期9-17,共9页Electric Drive for Locomotives

基　　金：国家“十三五”重点项目课题(2016YFB1200600);国家自然科学基金项目(61801479,62003049)。

摘　　要：悬浮系统的异常指在中低速磁浮列车的实际运行中,悬浮系统工作状态与期望状态不匹配,但是系统没有发生故障。准确预测悬浮系统的异常有助于合理地分配有限的监控资源、提前安排预防性维护计划、减少计划外维修成本、降低事故率。虽然根据《中低速磁浮交通车辆悬浮控制系统技术条件》(CJ/T 458—2014)可获得基于悬浮间隙的经验阈值,但在工程应用中,该方法受到额定悬浮间隙不唯一和外界扰动的影响,常会产生漏报问题,同时悬浮间隙数据的平衡问题增加了异常检测的难度。因此,本文提出了一种基于超球体高斯分布的悬浮系统异常检测方法。该方法首先利用快速沃尔什变换(Fast Walsh-Hadamard Transform,FWHT)技术提取样本特征,并使特征分布空间近似超球体;然后根据样本特征与球心的欧氏距离确定异常阈值。现场试验数据结果表明,与基于经验阈值、基于主成分分析(Principal Component Analysis,PCA)和基于支持向量数据描述(Support Vector Data Description,SVDD)的方法相比,提出的方法在系统异常检测能力方面更具有优越性。The abnormality of the suspension system means that the operating state of the suspension system does not match the expected state,and the system has no faults in the actual operation of the middle-low speed maglev train.Accurately detecting the abnormality of the suspension system helps managers allocate limited monitoring resources reasonably,arrange preventive maintenance plans in advance,reduce unplanned maintenance costs,and reduce accident rates.Although an empirical threshold based on the suspension gap can be obtained according to the"Technical Conditions for the Suspension Control System of Middle-low Speed Maglev Trains CJ/T458—2014",it is affected by the non-unique rated suspension gap and external disturbances,which will cause false negatives in engineering applications.Meanwhile,the problem of the balance of suspension gap data increases the difficulty of anomaly detection.Therefore,an anomaly detection method for suspension systems based on the Gaussian distribution of hyperspheres was proposed.Firstly,FWHT(Fast Walsh-Hadamard Transform)technology was used to extract sample features,and the feature distribution space was made approxim ate to a hypersphere.Then,the abnormal threshold was determined according to the Euclidean distance between the sample feature and the center of the sphere.The experimental results of the field data show that,compared with the methods based on empirical threshold,the principal component analysis(PCA)and support vector data description(SVDD),the proposed method is more superior in anomaly detection ability.

关 键 词：磁浮列车 悬浮系统 悬浮间隙 异常检测 FWHT PCA SVDD 轨道不平顺 

分 类 号：U237[交通运输工程—道路与铁道工程]