摘要
为了对运营高速铁路沉降监测数据合理利用进而为科学预警提供准确的依据,采用Kalman滤波算法对其数据进行分析。并基于初始状态参数中随机干扰因子设置的不同,提出了静态、常速、加速3种Kalman滤波模型,利用上述3种模型对高铁实际沉降监测数据进行滤波处理,得到静态、常速、加速三种模型的滤波残差最大值分别为8.28 mm、0.749 mm、0.0181 mm;单位权方差最大值分别为94.0330 mm^(2)、0.7256 mm^(2)、0.3213 mm^(2);均方根误差最大值分别为2.5911 mm、0.0536 mm、0.0076 mm;平均绝对误差最大值分别为1.7600 mm、0.0364 mm、0.0066 mm。实验数据表明,3种滤波模型中,Kalman加速滤波模型能更好地拟合了变形体的变化趋势,滤波结果最为可靠。
In order to make reasonable use of the settlement monitoring data of operating high-speed railways and provide an accurate basis for scientific early warning,Kalman filter algorithm was used to study and analyze the data.The working principle and formula of Kalman filtering were discussed in details.Based on the different random interference factor settings in the initial state parameters,three Kalman filtering models were proposed:static,normal speed,and acceleration.Three models were used to monitor the actual settlement data of high-speed railways.The maximum filter residuals of the static,normal speed and acceleration models were:8.28 mm,0.749 mm,0.0181 mm;the maximum error of unit weights were 94.0330 mm^(2),0.7256 mm^2,0.321 mm^(2);the maximum rms errors were 2.5911 mm,0.0536 mm,0.0076 mm;the average absolute errors were 1.7600 mm,0.0364 mm,0.0066 mm.The results show that the Kalman accelerated filtering model can better fit the change trend of the deformed body,the filtering result is the more reliable.
作者
焦雄风
谭社会
张献州
陈铮
郑旭东
Jiao Xiongfeng;Tan Shehui;Zhang Xianzhou;Chen Zheng;Zheng Xudong(School of geoscience and environmental engineering,southwest jiaotong university,Chengdu 611756,China;China railway Shanghai bureau group Co.,Ltd.,Shanghai 200071,China;National and local joint engineering laboratory of space information technology for high-speed railway operation safety,southwest jiaotong university,chengdu,611756,China)
出处
《铁道勘察》
2021年第3期38-42,共5页
Railway Investigation and Surveying
基金
中央高校基本科研业务费专项资金(SWJTUO10ZT02)。
关键词
运营高速铁路
沉降监测
KALMAN滤波
质量评价
high-speed railway during operation
settlement monitoring
Kalman filtering
quality evaluation
