非平稳车载及抗力劣化进程下混凝土桥梁时变可靠性评估

为实现同时考虑车载过程及抗力劣化进程非平稳性的在役混凝土桥梁构件时变可靠性评估，首先，联合时域内的动态广义极值分布模型及蒙特卡洛模拟实现对连续非平稳车载过程的极值建模，介绍基于Gamma过程的在役混凝土桥梁构件抗力非平稳劣化模型的建立及更新；其次，综合考虑边际救生成本准则、个体风险准则及社会风险准则对运营阶段目标可靠度指标取值进行讨论，为时变可靠性评估提供基准安全边界；最后，在基于风险函数的时变可靠性分析方法框架之下建立同时考虑车载及抗力非平稳性的时变可靠性分析方法，其中借助高斯数值积分及泰勒级数展开解决时变可靠性的求解问题，并采用一个实桥分析案例对上述分析流程的应用进行说明。研究结果表明：当荷载参数截口分布呈现多峰形态时，可采用广义极值分布函数族中的极值Ⅰ型分布对其年最大分布进行描述；交通量的持续增长将导致变量年最大分布位置参数的不断提升及尺度参数的不断下降；综合考虑3种可靠度指标分析准则，建议在役混凝土桥梁构件运营阶段年目标可靠度指标取为3.98，具体评估工作中不能忽略基准期对目标可靠度指标的影响；通过时变可靠性评估工作的开展，可获取构件在未来较长服役期内可靠度指标的变化情况、服役状态达到临界安全水平所对应的时间节点以及构件可靠性冗余度的时变情况；该类结果的获取可为在役桥梁全寿命维养策略制定等工作提供直接参考。

Time-dependent Reliability Assessment of Concrete Bridges Considering Non-stationary Vehicle Load and Resistance Deterioration Processes

This study attempts to realize a time-dependent reliability assessment of components of existing concrete bridges by considering the processes related to non-stationary vehicle load and non-stationary resistance simultaneously. First, a dynamic generalized extreme value distribution within a time domain and a Monte Carlo simulation were combined to conduct extreme value modeling of the non-stationary vehicle load process. A Gamma process based non-stationary resistance deterioration modeling and corresponding updating procedure were then introduced. Second, the principles of marginal life-saving cost, individual risk, and social risk were considered to determine the target reliability index in the operational stage as these can provide basic safety criteria for a time-dependent reliability assessment. Finally, a risk-function-based analytical method was established to realize the time-dependent reliability analysis by including non-stationary factors. For this method, the Gaussian numerical integration and Taylor series expansion were introduced to overcome the solving problems. A case study was conducted to illustrate the application of the proposed analytical procedure. The results indicate that the extreme type I distribution can be used to describe the annual maximum distribution when the sectional distribution of load parameters exhibits a multimodal distribution. Traffic growth can lead to an increase in the number of location parameters and a decrease in the number of scale parameters of the annual maximum distribution. It is suggested that the annual target reliability index of existing concrete bridge components can be 3.98 when considering the three types of principles and that the effects of the reference period must not be neglected. When the time-dependent reliability assessment is conducted, the following information can be derived:the variation in the reliability index over a long period, the critical moment that the safety level of a component reaches the safety criterion, and the variation in the reliability redundancy. The obtained information can provide some direct references to plan lifetime maintenance strategies for existing concrete bridges.