基于多源数据融合的钢桥腐蚀疲劳系统可靠度分析方法及试验研究

Targeting to the issue of fatigue cracking in orthotropic steel bridge deck, this project explores the fatigue disaster mechanism of steel bridges under the coupled action of traffic loading and corrosive environment. Based on the 15-year long-term structural health monitoring data of the Tsing Ma Bridge, a corrosion fatigue probabilistic analytical model of structural components in steel bridges is established considering multi-source data fusion. The fatigue performance degradation analysis method of steel bridge structures is developed based on temporal and spatial corrosion distribution model. A three-phase fatigue life analysis model of steel bridge components is established considering the coupled action of corrosion and fatigue, and the corresponding fatigue life calculation methods are established by combining the corrosion-fatigue interaction relationships at different stages. Through conducting the corrosion fatigue tests of cruciform welded specimens and the segment model of typical orthotropic steel bridge deck, the corrosion-fatigue interaction mechanism of steel bridge components is revealed. A method for real-time identification and dynamic tracking of multi-point fatigue cracks is proposed based on digital image pattern recognition. The fatigue crack propagation rule is investigated under the coupled action of corrosion and fatigue. A framework for analysis of corrosion-fatigue system reliability of steel bridge structures is formulated based on Bayesian network, and applied to the corrosion-fatigue system reliability assessment of the orthotropic steel deck system of the Tsing Ma Bridge. The results of this project will provide a reference for the probability-based fatigue design theory and assessment method of steel bridges, which have a significant theoretical value and practical significance.

本项目针对正交异性钢桥面板疲劳开裂问题，探究交通荷载和腐蚀环境耦合作用下钢桥疲劳灾变机理。基于青马大桥15年长期结构健康监测数据，建立考虑多源数据融合的钢桥结构构件腐蚀疲劳概率解析模型。发展基于时空分布腐蚀模型的钢桥结构疲劳性能劣化分析方法，建立考虑腐蚀和疲劳耦合作用的钢桥构件疲劳寿命三阶段分析模型，结合各阶段腐蚀-疲劳相互作用关系建立相应的疲劳寿命计算方法。通过十字形焊接试件腐蚀疲劳试验和典型正交异性钢桥面板节段模型腐蚀疲劳试验，揭示钢桥结构构件腐蚀-疲劳相互作用机理，提出基于数字图像模式识别的多点疲劳裂纹实时识别和动态跟踪方法，研究腐蚀-疲劳耦合作用下疲劳裂纹扩展规律。构建基于贝叶斯网络的钢桥结构腐蚀疲劳系统可靠度分析框架，并应用于青马大桥正交异性钢桥面板系统的腐蚀疲劳系统可靠度评估。本项目的研究成果为基于概率的钢桥疲劳设计理论和评估方法提供依据，具有重要的理论价值和现实意义。

本项目采用理论分析、试验研究和现场监测等研究方法，以长期现场监测数据为基础，针对交通荷载和腐蚀环境耦合作用下钢桥疲劳灾变机理这一关键科学问题开展了系统性的研究。主要成果及科学意义包括：. (1)基于长期现场监测数据，研究了交通荷载和环境因素共同作用下钢桥构件应力谱特征，发现了车辆荷载与风场环境对疲劳应力谱的影响规律，建立了疲劳应力谱随机特性概率建模方法，提出了基于不同分布函数和参数估计算法的疲劳应力谱最优概率模型和判断依据。. (2)建立了钢桥焊接构件两阶段腐蚀-疲劳耦合计算模型，研究了荷载频率、腐蚀环境、腐蚀坑形貌对不同腐蚀区域钢桥构件疲劳性能的劣化规律，建立了不同腐蚀区域的疲劳强度退化分析模型，研发了正交异性钢桥面板腐蚀-疲劳耦合试验装置，提出了基于计算机视觉和深度学习的疲劳裂纹智能识别和动态跟踪方法。. (3)提出了正交异性钢桥面板焊接节点腐蚀疲劳可靠度分析模型，考虑焊接节点之间疲劳失效的相关性，推导了正交异性钢桥面板系统疲劳可靠度分析模型，建立了系统疲劳可靠度动态贝叶斯网络分析框架，提出了求解贝叶斯网络系统和构件失效概率的计算方法。. 本项目的研究成果可以为基于概率的钢桥疲劳设计理论和评估方法提供依据，具有一定的理论价值和工程意义。通过本项目资助，发表学术论文25篇，其中SCI收录24篇，EI收录17篇，授权发明专利10项，培养博士研究生2名、硕士研究生2名。