摘要
针对高速铁路沿线同一监测点两台风速风向计风速监测数据差异较大的情况,基于数据挖掘技术研究了异常数据的判识方法。首先采用相关性分析方法对两台风速风向计风速监测数据进行判断,若数据异常,再采用差分阈值法分别对两台风速风向计风速监测异常数据进行判识。根据风速对列车运行和基础设施的影响将高速铁路风速监测数据分为6个等级,根据历史数据确定各等级下的差分阈值。根据当前风速自适应动态选择差分阈值识别风速异常数据,并结合风向进一步研究异常数据产生的原因。对京张高速铁路风速监测数据的分析验证表明该判识方法快捷有效。
In view of the large difference of wind speed monitoring data between two anemometers at the same monitoring point along the high speed railway,the identification method of abnormal data was studied based on data mining technology.Firstly,the correlation analysis method is used to judge the wind speed monitoring data of the two anemometers.If the data are abnormal,the difference threshold method is used to judge the abnormal data in wind speed monitoring of the two anemometers respectively.According to the impact of wind speed on train operation and infrastructure,the wind speed monitoring data of high speed railway is divided into six levels,and the difference threshold under each level is determined according to the historical data.According to the current wind speed,the adaptive dynamic selection difference threshold is used to identify the abnormal wind speed data,and the causes of the abnormal data are further studied combined with the wind direction.The analysis and verification of wind speed monitoring data of Beijing-Zhangjiakou high speed railway shows that the identification method is fast and effective.
作者
包云
李亚群
马祯
陈中雷
白根亮
BAO Yun;LI Yaqun;MA Zhen;CHEN Zhonglei;BAI Genliang(Institute of Computing Technology,China Academy of Railway Sciences Corporation Limited,Beijing 100081,China;Beijing Jingwei Information Technologies Co.Ltd.,Beijing 100081,China)
出处
《铁道建筑》
北大核心
2021年第10期154-157,共4页
Railway Engineering
基金
中国国家铁路集团有限公司科技研究开发计划(K2020T003)。
关键词
高速铁路
风速监测
异常数据判识
差分阈值
风速
风向
支持向量机算法
孤立森林算法
high speed railway
wind speed monitoring
abnormal data identification
differential threshold
wind speed
wind direction
support vector machine algorithm
isolated forest algorithm
