Ground-based radar monitoring building thermal expansion and contraction data applied to machine learning multi-feature-deformation regression prediction task

Under the background of continuous increase of surface temperature caused by global warming, high temperature events occur frequently, posing a potential threat to natural ecosystems and economic and social development. China 's coastal areas have dense infrastructure and are high-risk areas for this weather. Due to its high flexibility and low fundamental frequency characteristics, super high-rise buildings are highly sensitive to temperature action, and structural damage or even failure is prone to occur under high temperature loads. However, the traditional thermal expansion and contraction deformation measurement technology relies on sensors. This method has problems such as high cost, limited coverage, and easy damage to structures, and it is difficult to meet the needs of large-span dynamic monitoring. This experiment is based on the portable ground-based radar interferometer ( GPRI ) system developed by Swiss GAMMA company. This technology has the advantages of non-contact measurement, all-weather observation, high spatial and temporal resolution, sub-millimeter high precision and wide monitoring range. It provides an innovative solution for the monitoring of thermal expansion and contraction deformation of super high-rise buildings in high temperature weather. At the same time, the meteorological instrument is used to monitor the corresponding meteorological data in real time. Finally, the deformation data and monitoring data are used as the data set of multi-feature deformation prediction of machine learning model. The final conclusion is that the correlation coefficient between the predicted value and the true value is very high, which is greater than 0.96, and the mean square error is very low, which is less than 0.007, which verifies the reliability of the GPRI system.在全球变暖导致地表温度持续升高的背景下，高温事件频发，对自然生态系统以及经济社会发展构成潜在威胁。中国沿海地区基础设施密集，是此类高温天气的高风险区域。超高层建筑因具备高柔性与低基频特性，对温度作用高度敏感，在高温荷载作用下易发生结构损伤甚至失效。然而，传统的热胀冷缩变形测量技术依赖传感器，该方法存在成本高昂、覆盖范围有限、易对结构造成损伤等问题，难以满足大范围动态监测的需求。本实验采用瑞士GAMMA公司研发的便携式地面雷达干涉仪（ground-based radar interferometer，GPRI）系统，该技术具备非接触式测量、全天候观测、高时空分辨率、亚毫米级高精度以及监测范围广等优势，为高温天气下超高层建筑的热胀冷缩变形监测提供了创新解决方案。同时，本实验借助气象仪器实时采集对应的气象数据，最终将变形数据与监测数据作为机器学习模型多特征变形预测的数据集。最终实验结论显示：预测值与真实值的相关系数极高（大于0.96），均方误差极低（小于0.007），验证了GPRI系统的可靠性。