Excessive Bridge Deflections: Inverse Analysis to Identify Multi-Decade Creep Properties of Concrete

本课题的研究目的是利用实际的过大徐变挠度数据，找到一个符合实际的几十年混凝土徐变规律。帕劳KB桥是世界跨径纪录节段箱梁桥，其垮塌引起了人们对其前蠕变挠度的关注，其前蠕变挠度达到1.61m，接近设计挠度的5倍。西北大学的研究人员被迫公开了这座桥的技术数据，得出了标准设计建议已经过时的结论。这激发了广泛的搜索，到目前为止，已经找到了64座大跨度桥梁，在20到40年内过度偏转，而本应使用100年。这种过时的主要原因是大多数实验室蠕变数据的持续时间为6年。将桥梁挠度数据合并到一个综合数据库中，并将箱梁挠度的三维有限元反徐变分析与实验室数据库相结合，以校准改进的混凝土徐变材料规律。利用复杂的统计概念和遗传算法，从混凝土的组成特性和强度来预测材料蠕变规律的参数。研究结果将作为工程学会的标准设计建议。尽管改造或关闭大量过度偏转的桥梁的社会成本是巨大的，但他们的数据代表了进步的金矿。该研究将阐明如何避免重复，以免造成使用性能损失和腐蚀开裂。一种改进的蠕变模型对现代超高层建筑中柱的差异缩短设计也很重要。

The resaerch objective of this project is to use actual data on excessive creep deflections to find a realistic multi-decade concrete creep law. The fatal collapse of KB Bridge in Palau, a segmental box-girder bridge of world record span, called attention to the preceding creep deflections, which attained 1.61m, nearly 5-times the design deflection. Forced unsealing of the technical data on this bridge allowed Northwestern University researchers to conclude that the standard design recommendations were obsolete. This motivated a broad search which has so far led to a collection of 64 large-span bridges deflecting excessively within 20 to 40 years, while supposed to serve 100 years. The main cause of this obsoleteness is that most laboratory creep data have durations 6 years. The bridge deflection data are combined in one aggregate database, and an inverse three-dimensional finite-element creep analysis of box-girder deflections is combined with the lab database to calibrate an improved material law for creep of concrete. Sophisticated statistical concepts and a genetic algorithm are used to predict the parameters of material creep law from the concrete composition characteristics and strength. The results will be proposed for standard design recommendations of engineering societies.Although the societal costs of retrofits or closings of numerous excessive deflecting bridges are enormous, their data represent a gold mine for progress. The research will clarify how to avoid repetitions, which causes serviceability losses and cracking with corrosion. An improved creep model will also be important for design against differential shortening of columns in modern super-tall buildings.