期刊文献+

基于XGBoost的应变传感器温变在线计量模型研究

In-situ Calibration Model for Strain Sensor Temperature Variation Based on XGBoost
原文传递
导出
摘要 传统计量方法由于条件要求不适用于桥梁结构健康监测系统用于应变传感器的计量工作,在线计量方法可以解决难题,现场环境因素影响、服役传感器种类繁多的问题为在线计量有效应用带来了极大挑战。为有效保障桥梁结构健康监测系统长期可靠运行,论文以桥梁用电阻式应变传感器和振弦式应变传感器为例,搭建室内应变试验平台。设计多个温度条件和荷载等级,开展室内温度-荷载应变试验。利用Pearson相关分析研究传感器与温度、荷载的影响规律,依靠有限元模型进行理论分析验证,最后通过XGBoost模型实现不同种类传感器间的量值传递。结果表明,温度对电阻式应变传感器呈正相关影响,对振弦式应变传感器呈弱于电阻式的负相关影响,均不呈线性;有限元模型分析仿验证了实测数据的有效性;利用XGBoost搭建的量值溯源模型预测值与实际值误差小,可以较好地实现电阻式、振弦式传感器间量值传递。 The traditional metrology methods,due to their environmental requirements,are not suitable for calibrating strain sensors in bridge structural health monitoring systems.The in-situ calibration approaches offer solutions to this problem;however the varied environmental factors at the site and multitude of in-service sensor types pose significant challenges to their effective application.To ensure the long-term reliable operation of bridge structural health monitoring systems,the study focuses on the resistive and vibrating wire strain sensors,constructing an indoor strain experimental platform.A series of temperature conditions and load levels were designed to carry out indoor temperature-load strain experiments.Pearson correlation analysis was employed to investigate the influence patterns of sensors with temperature and load,while theoretical analysis validation was performed by using finite element models.Subsequently,the transference of measurement values among different types of sensors was realized through XGBoost model.The result indicates that the temperature has a positive correlation with resistive strain sensors and a weaker negative correlation with vibrating wire strain sensors with both correlations being non-linear.The finite element model analysis corroborates the validity of empirical data.The metrological traceability model built by using XGBoost yielded predictions with small errors compared to actual values,facilitating the transfer of measurement values between resistive and vibrating wire sensors.The error between predicted value and actual value of the traceability model built with XGBoost is small,which can realize the value transfer between the resistive and vibrating wire sensors.
作者 冯笑凡 唐煜 李斌 修钊 彭璐 FENG Xiao-fan;TANG Yu;LI Bin;XIU Zhao;PENG Lu(Research Institute of Highway,Ministry of Transport,Beijing 100088,China;Guizhou Provincal Quality Safety Traffic Engineering Monitoring and Inspection Center,Guiyang,Guizhou 550081,China;National Center for Materials Service Safety,University of Science and Technology,Beijing 100083,China)
出处 《公路交通科技》 CAS CSCD 北大核心 2023年第S02期143-150,共8页 Journal of Highway and Transportation Research and Development
关键词 桥梁工程 应变测量传感系统 在线计量 Pearson相关性分析 XGBoost bridge engineering strain measurement sensing system in-situ calibration Pearson correlation analysis XGBoost
  • 相关文献

参考文献6

二级参考文献27

共引文献39

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部