极值分析方法在车辆荷载评估中的应用与比较  

Application and Comparison of Extreme Value Analysis Methods in Vehicle Load Assessment

作　　者：韩大建[1] 潘玲[1] 

机构地区：[1]华南理工大学土木与交通学院,广东广州510640

出　　处：《建筑科学与工程学报》2011年第2期8-13,共6页Journal of Architecture and Civil Engineering

基　　金："十一五"国家科技支撑计划项目(2006BAJ01B07-05)

摘　　要：针对现有车辆荷载安全评估中采用拟合得到的近似分布作为极值估计的底分布而导致的极值估计结果误差较大的问题,引入了极值统计的新方法,即阈值模型、次序统计量模型、点过程模型,并分别采用这些模型对广州北环高速公路的5型车车质量样本进行统计分析,得到重车荷载的概率分布尾部及最大值分布,将新方法的结果与现行方法得到的结果进行了比较。结果表明:阈值模型和点过程模型与样本的一致性较好,现行方法得到的结果与样本点相差较大,且现行方法在一定程度上低估了车辆荷载的极值。In accordance with the existing problems,that current analysis methods for vehicle loads using an approximate distribution obtained by fitting as a base distribution of extreme value estimation,which might cause a significant error for extreme values,new methods for extreme value statistics were introduced,including threshold model,order statistic measurement model and point process model.In order to obtain the probability distribution tail and the maximum value distribution of heavy truck loads,these new models were respectively adopted to analyze the samples of type 5 vehicle weight on Guangzhou North-ring Expressway.The comparisons between results from the new extreme value statistic models and those from the current method were carried out.The results show that the estimations from the threshold model and the point process model are more consistent with the measured samples,whereas the results from the current method show errors.Moreover,to a certain extent,the value of estimation load from the current method undervalues the extreme value of vehicle loads.

关 键 词：车辆荷载 极值估计 阈值模型 次序统计量模型 点过程模型 

分 类 号：U441.2[建筑科学—桥梁与隧道工程]