现有大跨径桥梁有限元模型修正(finite element model updating,FEMU)方法一般未考虑运营荷载对结构动力特性的影响,导致修正后模型的参数变异性大.鉴于此,提出了一种考虑运营荷载的层次贝叶斯有限元模型修正方法,该方法包含考虑温度和交通荷载的概率参数修正、概率响应预测和结构状态评估.首先,根据监测数据的相关性分析结果确定了计算理论频率时需要考虑的荷载.随后,建立了温度-弹性模量线性关系,并基于动态称重(weigh-in-motion,WIM)数据,提出一种车辆荷载估计方法,以在有限元模型中定量考虑运营荷载对结构频率的影响.同时,引入两阶段马尔科夫链蒙特卡洛(Markov chain Monte Carlo,MCMC)采样方法和响应面代理模型,以提高概率模型修正的计算速率.该方法在一座采集了两年监测数据的大跨径拱桥上得到了验证.结果表明,在考虑运营荷载、参数不确定性和建模误差后,实测频率基本处于预测频率的95％置信区间内.最后,基于实测响应和预测响应置信区间提出了一个结构状态指标,并利用该指标检测出该桥的路面铺装更换过程.
Hierarchical Bayesian model updating method for long-span arch bridge considering operational loads
Existing finite element model updating(FEMU)method for large-span bridge generally do not consider effects of operational loads on dynamic characteristics of structure to cause large parametric variability in the updated model.Here,hierarchical Bayesian FEMU method considering operational loads was proposed.This method could include probability parametric updating considering temperature and traffic loads,probability response prediction and structural state evaluation.Firstly,loads needed to consider when calculating theoretical structural frequency were determined according to correlation analysis results of the monitored data.Subsequently,a linear relation between temperature and elastic modulus was established,and based on weighing-in-motion(WIM)data,a vehicle load estimation method was proposed to quantitatively consider effects of operational loads on structural frequency in a finite element model.At the same time,a two-stage Markov chain Monte Carlo(MCMC)sampling method and a response surface surrogate model were introduced to improve computational speed of probability model updating.This method was verified on a large-span arch bridge where monitoring data were collected for two years.The results showed that after considering operational loads,parametric uncertainty and modeling errors,the measured frequency is basically within the 95％confidence interval of the predicted frequency.Finally,a structural state index was proposed based on confidence intervals of measured and predicted responses,and this index was used to detect the pavement replacement process of this bridge.
finite element model updating(FEMU)hierarchical Bayesianoperational loadslong-span arch bridgestructural state evaluation
万方数据
维普
