大跨度混凝土斜拉桥多层次安全评定方法研究

The evaluation methods of bridge safety in the current codes and standards are mainly based on the structural member level and subjective experience, considering the structural system level less for the complex bridge structures. Many concrete cable-stayed bridges have been built in China, and their multi-level safety during the subsequent service life are urgent to be evaluated due to rapidly degraded performances under the influences of multiple factors (e.g. unfavorable environment and overloaded vehicles). This proposal tries to carry out researches on evaluation methods based on multi-level safety for long-span concrete cable-stayed bridges. The main research work is as follows: a method of performance sensing based on smart aggregate is studied for key parts of main beam in concrete cable-stayed bridges, to obtain sensing data; experiments on performance degradation are carried out for concrete cable-stayed bridges, to reveal the laws; a multi-level resistance model is proposed based on Bayesian updating theory and relevant data; a strategy for iterative computation of design points is studied based on multiple response surfaces and zero residual fitting method to efficiently obtain the representative sampling points corresponding to the main failure mode, and a method for searching failure mode is proposed based on cluster analysis; a method is studied to determine the key structural members and parameters for concrete cable-stayed bridges, and the evaluation indexes and criterions are also proposed based on multi-level safety. Then, evaluation methods based on multi-level safety are obtained for long-span concrete cable-stayed bridges, and scientific evidences are also provided for life predictions and maintenance decisions of these bridges.

现行规范和标准中桥梁安全评定方法主要考虑构件层次，且以主观经验为主，对复杂桥梁结构体系层次的安全考虑较少。我国已建造了一大批混凝土斜拉桥结构，受不利环境和超载等多因素影响，这些桥梁的性能快速退化，其后继服役期内的多层次安全性水平亟需评定。本项目拟开展大跨度混凝土斜拉桥多层次安全评定方法研究。主要内容有：研究基于智能骨料的混凝土斜拉桥主梁关键部位性能感知方法，以获取数据；开展混凝土斜拉桥性能退化试验，揭示其性能退化规律；利用贝叶斯更新理论和相关数据，建立结构多层次抗力模型；采用多响应面零残差拟合方法，研究验算点迭代求解策略，高效获得与主要失效模式对应的代表性样本点集，基于聚类分析提出失效模式搜寻方法；研究混凝土斜拉桥关键构件、参数确定方法，提出结构多层次安全评定指标与准则。在此基础上，形成大跨度混凝土斜拉桥多层次安全评定方法，为服役大跨度混凝土斜拉桥寿命预测和维修决策提供科学依据。

本项目开展了基于压电传感的斜拉桥预应力管道密实性检测和优化布置研究，提出了一种基于应力波的压电陶瓷传感器主动传感预应力导管注浆质量控制方法。设计的注浆密实度分别0%、50%、75%、90%和100%。试验通过电压信号、功率谱密度(PSD)能量和小波包能量法，建立了注浆密实度和检测信号间的映射关系，发展了压电传感检测在斜拉桥服役过程中的感知方法。.为研究典型损伤下预应力混凝土斜拉桥性能退化规律，以合江长江二桥为原型，基于相似比理论和刚度相似原理，设计了缩尺比为1:40的斜拉桥节段模型，研究了拉索断索和主梁损伤对斜拉桥性能退化的影响；建立了典型损伤工况下模型桥和原型桥的有限元模型，在原型桥上进行拉索损伤或断索的模拟，获取了不同程度损伤或断索后其它拉索索力、主梁应力、主梁变形和主塔变形情况，揭示了典型损伤工况下斜拉桥性能退化的一般规律；以张力指标、模型桥跨中位移、模型桥跨中弯矩、辅助墩及塔墩梁固结处的节点支反力等5个指标为输入向量，发展了3种三层BP神经网络，分别对斜拉桥主梁损伤位置、斜拉桥拉索损伤位置和斜拉桥拉索损伤程度进行识别与预测；以双塔三跨式斜拉桥为例进行计算模型，推导建立了拉索损伤和主梁损伤后斜拉桥性能退化模型。.基于多响应面模型，提出了一种可稳步收敛的结构失效方程零残差拟合方法，高效获得含验算点的较佳代表性样本点集，归纳合并失效模式，得到大跨结构主要失效模式；提出了模糊失效准则下拉索锈蚀可靠度分析模型，并采用将服役期离散化的方法简化了时变可靠度的计算问题。.发展了一种基于支持向量机与遗传算法相结合的可靠度寻优求解算法，研究了运营期结构性能退化后的可靠性变化规律，并应用于一座大跨度预应力混凝土斜拉桥；建立了大跨桥梁多层次安全评定指标体系，综合考虑安全、适用、耐久因素，构建了特大跨桥梁安全性评定准则，开发了基于时变可靠度理论的结构安全性评定系统。