为有效利用桥梁监测系统积累的大量数据进行桥梁性能分析，首先，以西堠门大 桥为研究对象，建立多尺度非线性有限元模型，并探讨了利用监测系统所获得的风速、温 度、车辆及位移等多类监测数据对所建模型的可靠性进行验证的方法;然后，基于 WIM 系 统采集的海量车辆信息数据，通过统计分析确定了 6 类疲劳标准车，将其施加到多尺度有 限元模型，获得了钢箱梁焊接细节的应力响应，结合 Palmgren-Miner 线性累积损伤理论， 计算钢箱梁构造细节疲劳寿命的理论值，并采用滤波、去噪等大数据分析方法对应力监测 数据进行分析，基于实测数据计算钢箱梁焊接细节疲劳寿命，以确定理论疲劳寿命的修正 系数;最后，根据修正后的理论疲劳寿命，将钢箱梁横向不同位置的构造细节划分为 4 个 疲劳等级． 文中研究为获取未布置传感器的构造细节疲劳寿命提供了新的思路，可指导大 跨桥梁钢箱梁的日常养护管理工作．

In order to effectively use the mass data accumulated by bridge monitoring system to analyze the perform- ance of a bridge,first,by taking Xihoumen Bridge as an object,a multi-scale nonlinear finite element model is established,and a method to verify the reliability of the proposed model according to such monitoring data as wind speed,temperature,vehicle load and displacement is presented.Next,on the basis of mass data recorded by the WIM system,six types of standard fatigue vehicles are determined and applied to the multi-scale finite element model to calculate the stress response of the details of steel box girder.Then,by combining Palmgren-Miner linear cumulative damage theory with the stress amplitude,the theoretical fatigue life of structural details in the steel box girder is obtained.Moreover,by means of filtering,denoising and other big data processing methods,the moni- tored stress data are analyzed to calculate the actual fatigue life of the weld details and to determine the correction coefficient of the theoretical fatigue life.Finally,according to the corrected fatigue life,the structural details of the steel box girder in different positions are divided into four fatigue grades.This research provides a new ap- proach to obtaining the fatigue life of the details without laying sensors.It also serves as a guidance for the daily maintenance and management of steel box girders in long-span bridges.