基于贝叶斯后验估计的桥梁动态称重算法理论与试验研究  

Bridge weighing-in-motion algorithm theory based on Bayesian posterior estimation and tests

作　　者：张龙威 原璐琪 陈宁[1] 袁帅华[1] 张龙 ZHANG Longwei;YUAN Luqi;CHEN Ning;YUAN Shuaihua;ZHANG Long(College of Civil Engineering,Hunan University of Science and Technology,Xiangtan 411201,China;Hunan Provincial Key Lab of Structural Engineering for Wind Resistant and Vibration Control,Hunan University of Science and Technology,Xiangtan 411201,China)

机构地区：[1]湖南科技大学土木工程学院,湖南湘潭411201 [2]湖南科技大学结构抗风与振动湖南省重点实验室,湖南湘潭411201

出　　处：《振动与冲击》2024年第1期20-27,共8页Journal of Vibration and Shock

基　　金：国家自然科学基金项目(52378509);湖南省自然科学基金项目(2023JJ40290)。

摘　　要：桥梁动态称重(bridge weigh-in-motion,BWIM)利用过桥车辆对桥梁产生的动力响应快速识别车辆轴质量。由于实测的动力响应包含测量误差,在一定程度上降低了传统BWIM算法的轴质量识别精度。为了解决这一问题,提出基于贝叶斯后验估计的桥梁动态称重算法。该算法考虑了测量误差对轴质量识别精度的影响,假设测量误差和轴质量服从高斯分布,利用测量误差的标准差和轴质量标准差得到能抑制测量误差的约束因子,推导出新的轴质量求解方程。基于数值仿真和实桥试验,分别得到传统BWIM算法和贝叶斯算法的轴质量识别精度,并进行对比分析。试验结果表明:相比于传统BWIM算法,贝叶斯算法能够有效抑制测量误差的影响,明显改善轴质量识别精度。Bridge weigh-in-motion(BWIM)algorithm can find vehicle axle weights with measured responses of a bridge acted by passing vehicles.However,since measured responses contain measurement errors,the accuracy of the traditional BWIM algorithm is reduced to a certain extent.Here,to solve this problem,a novel BWIM algorithm based on Bayesian posterior estimation was proposed.The proposed algorithm could consider negative effects of measurement errors on axle weight recognition accuracy.Firstly,it was assumed that measurement errors and axle weights both obey Gaussian distribution.Then,the standard deviation of measurement errors and axle weights’standard deviation were used to obtain constraint factor which could suppress measurement errors.Finally,the new solving equation of axle weight using BWIM algorithm based on Bayesian posterior estimation was derived.Based on numerical simulation and actual bridge tests,recognition accuracies of axle weights using the traditional BWIM algorithm and the proposed BWIM algorithm based on Bayesian posterior estimation were obtained,respectively.Both of them were analyzed contrastively.The results showed that compared to the traditional BWIM algorithm,the proposed BWIM algorithm based on Bayesian posterior estimation can effectively suppress effects of measurement errors,and obviously improve the recognition accuracy of vehicle axle weights.

关 键 词：桥梁动态称重(BWIM) 贝叶斯后验估计 最小二乘 测量误差 实桥试验 

分 类 号：U446.1[建筑科学—桥梁与隧道工程]