大跨度桥梁的车辆荷载时空分布识别与建模研究

It is well-recognized that data mining and analysi for the massive volume of measurement data collected from structural health monitoring (SHM) system of long-span bridge structures is increasingly becoming a world-wide research focus. For the long-span bridges, the vehicle load is one of the critical external loads, and the heavy trucks, especially the overloaded trucks, may result in severe structural damages such as the fatigue damage or even bridge collapse. Aiming to the key issues of data mining of SHM and solving the moving vehicle loads identification problem, this proposal is to develop the approaches of identification and modeling of time-spatial distribution of vehicle loads on long-span bridges based on the compressive sensing theory, data fusion, video image technology, and the random field theory. An identification method for time-spatial distribution of the vehicle loads on long-span bridges using compressing sensing and cable forces monitoring data is first studied. Furthermore, identification method based on compressive sensing and multi-type sensor data fusion technique is proposed for time-spatial distribution of vehicle loads on long-span bridges to improve the identification accuracy. Third, identification approach for time-spatial distribution of vehicle loads on long-span bridges using the combined video and weigh-in-motion technologies is proposed. Finally, the modeling of spatial distribution of vehicle loads based on random field theory is investigated. This proposal is anticipated to obtain comprehensive and detailed time-spatial distribution of vehicle loads on long-span bridges and aid to completely understand the real relationship between vehicle loads and responses of structures. This research will develop the identification and modeling theories of moving vehicles loads, which have important scientific significance and practical sense for SHM of bridges.

结构健康监测数据分析与挖掘是当前国内外土木工程领域的热点研究方向，车辆是大跨度桥梁的主要荷载之一，重车尤其是超重车是造成桥梁破坏和疲劳累积损伤的主要因素之一。面向健康监测的数据挖掘科学问题，突破移动车辆荷载识别的瓶颈，采用压缩感知理论、数据融合方法、视频图像技术和随机场理论，研究大跨度桥梁车辆荷载时空分布的识别方法和建模方法。首先，研究基于压缩感知理论和索力监测信息的车辆荷载时空分布识别方法；其次，研究基于压缩感知理论和多种类传感器信息融合的车辆荷载时空分布识别方法；然后，研究基于视频监控信息与动态地秤监测信息集成的车辆荷载时空分布识别方法；最后研究基于随机场理论的车辆荷载空间分布建模方法和模型。本项目的研究将实现大跨度桥梁车辆荷载时空分布的全面透彻感知与识别，为深刻认识车辆荷载作用与结构响应真实的映射关系奠定基础，丰富和发展数据挖掘理论和结构健康监测理论，具有重要的科学意义和实用价值。

重车尤其是超重车是造成桥梁疲劳损伤甚至倒塌的主要因素之一。尽管桥梁上一般安装了车辆称重系统，但该系统只能测量某一个固定截面处的车辆轴重，无法获得车辆空间分布。针对大跨桥梁车辆荷载空间分布识别难题，本项目首先对实际桥梁一年的车辆荷载监测数据进行了统计分析，获得了真实车辆荷载数据并分析了车辆超载对桥梁整体性能的影响；然后，提出了基于斜拉索索力监测信息的大跨度桥梁车辆荷载时空分布识别方法；提出了基于视频与动态地秤监测信息融合的大跨度桥梁车辆荷载时空分布识别方法；最后，建立了基于Markov随机场理论的重车荷载空间分布模型。以上研究成果实现了从单一极值至车辆荷载空间分布模型的跨越，为结构体系整体安全评定奠定基础。项目研究成果发表论文13篇，其中SCI论文8篇，邀请报告1次，出版专著1部；参编交通部技术规程1部，获黑龙江省自然科学一等奖1项（排名第3）；授权国家发明专利1项，申请2项。