基于区间分析的非接触式影响线识别方法及复合反演机制研究  

Research on contactless influence line identification method and composite inversion mechanism based on interval analysis

作　　者：周云[1,2,3] 胡健鑫 周赛 裴熠麟 程依婷 ZHOU Yun;HU Jianxin;ZHOU Sai;PEI Yilin;CHENG Yiting(Hunan Provincial Key Laboratory of Damage Detection(Hunan University),Changsha 410082,China;College of Civil Engineering,Hunan University,Changsha 410082,China;Research Center of New Structural System,Academician Zhou Xuhong,Hunan University,Changsha 410082,China)

机构地区：[1]工程结构损伤诊断湖南省重点实验室(湖南大学),湖南长沙410082 [2]湖南大学土木工程学院,湖南长沙410082 [3]周绪红院士湖南大学新型结构体系研究中心,湖南长沙410082

出　　处：《地震工程与工程振动》2021年第2期24-34,共11页Earthquake Engineering and Engineering Dynamics

基　　金：国家自然科学基金项目(51878264);湖南省重点研发计划项目(2017SK2220)。

摘　　要：近年来,一种不需要在桥梁结构布置传感器的非接触桥梁动态称重(cBWIM)方法被提出。为了对cBWIM系统的实现提供支持,文中验证了基于区间分析的支持向量机(SVM)识别影响线方法的效果。考虑到实际应用中,会出现汽车总重量无法布满总重估计区间而导致对轴重区间过宽估计的情况,文中通过车桥数值模拟发现轴重区间过宽估计会降低SVM影响线识别的精确性。基于影响线区间与轴重区间的相互映射关系,提出了大数据影响线识别的复合反演校验机制。所提出方法可以有效减小区间过宽估计造成的SVM影响线识别误差,并通过数值模拟对所提出方法进行了验证,其中影响线的识别误差从初始的26.09%降低到1.36%,达到了提高影响线识别准确性的目的。In recent years,a contactless bridge dynamic weighing(cBWIM)method is proposed,which does not need to arrange sensors in bridge structures.In order to support the realization of cBWIM system,this paper verifies the effect of the support vector machine(SVM)method of identifying influence lines based on interval analysis.Considering that in practical application,the total weight of automobile can't be filled with the total weight estimation interval,which leads to the over-wide estimation of axle weight interval,this paper finds that the over-wide estimation of axle weight interval will reduce the accuracy of SVM influence line recognition.Based on the mutual mapping relationship between the influence line interval and the axial load interval,a compound inversion verification mechanism for identifying the influence line of big data is proposed.The proposed method can effectively reduce the recognition error of SVM influence line caused by over-wide interval estimation,and the proposed method is verified by numerical simulation,in which the recognition error of influence line is reduced from the initial 26.09%to 1.36%,thus achieving the purpose of improving the accuracy of influence line recognition and non-contact dynamic axle load recognition.

关 键 词：结构健康监测 桥梁影响线 桥梁动态称重 非接触式识别 复合反演 

分 类 号：P315.9[天文地球—地震学]