刘磊^1,2,4,,
蒋一然^3,4
1. 中国科学院电子学研究所, 北京 100190
2. 中国科学院空间信息处理与应用系统技术重点实验室, 北京 100190
3. 北京大学地球与空间科学学院, 北京 100871
4. 高铁地震学联合研究组, 北京 100029

基金项目: 国家自然科学基金项目（61701478）资助


详细信息

作者简介: 刘磊, 男, 1989年生, 2011年本科毕业于清华大学自动化系, 2016年在清华大学自动化系获得博士学位.目前主要从事地球科学大数据机器学习方面的研究工作.E-mail:lliu1@mail.ie.ac.cn

中图分类号: P631

收稿日期:2018-12-10
修回日期:2019-05-22
上线日期:2019-06-05

Attribution extraction and feature analysis for large amount of high-speed-train seismic events

LIU Lei^1,2,4,,
JIANG YiRan^3,4
1. Institute of Electronics, Chinese Academy of Sciences, Beijing 100190, China
2. Key Laboratory of Technology in Geo-spatial Information Processing and Application System, Chinese Academy of Sciences, Beijing 100190, China
3. School of Earth and Space Sciences, Peking University, Beijing 100871, China
4. The Joint Research Group of High-Speed Rail Seismology, Beijing 100029, China


MSC: P631

--> Received Date: 10 December 2018
Revised Date: 22 May 2019
Available Online: 05 June 2019


摘要
摘要:列车在高铁上运行能够产生大量的高铁地震事件，这些地震事件既包含了作为震源的列车本身的结构信息，也蕴含了高铁线路的地质环境信息.本文提出了一种频率-空间-时间（FXT）属性描述大量高铁地震事件之间的关联.在FXT属性上，高铁地震信号之间的关联表征为在频率坐标轴F上相互对齐、在空间位置坐标轴X和事件到达时间坐标轴T上连续变化的分立谱线.本文对课题组在深圳为期两天的观测采集得到的4500余条高铁地震信号进行了FXT属性的提取，并对FXT属性进行了统计特性的分析：方差分析表明，高铁线路的地质环境的变化对高铁地震信号特征的影响大于不同列车的结构差异对高铁地震信号特征的影响；聚类分析表明，高铁地震信号的特征按接收点位置能够无监督地聚类为和高铁线路途经地质环境相关联的不同类别.通过上述分析表明，高铁地震信号中蕴含着丰富的可解释的高铁运行地质环境信息，具备对高铁列车运行安全进行监测的潜力.
关键词: 高铁地震事件/
属性/
特性分析/
方差分析/
聚类分析

Abstract:Trains on high-speed-railway generate plenty of seismic events, which not only include information related to the body of trains themselves, but also involve geological environmental information along the railway. This article proposes a frequency-space-time (FXT) attribution to describe the relationship between multiple high-speed-train seismic events. The relationship between events are described by discrete spectral lines on the proposed attribution, which register along frequency axis, and change continuously along space axis and time axis. This article extracts FXT attribution from 4500 high-speed-train seismic signals acquired in Shenzhen area, and then takes statistical analysis on the extracted attribution. The variance analysis proves that the geological environmental changes around the railway is the main fact to effect the features of seismic signals, compared with high-speed-train body structural changes. The clustering analysis proves that the features of high-speed-train signals at different receiving positions cluster unsupervised as classes related to geological environments along railway. The analysis explains the potentiality of security monitoring for trains with the application of high-speed-train seismic signals, which contain abundant and interpretable geological environmental information.
Key words:High-speed-train seismic event/
Attribution/
Feature analysis/
Variance analysis/
Clustering analysis



PDF全文下载地址:

http://www.geophy.cn/data/article/export-pdf?id=dqwlxb_15028