机场复合道面性能预估的非线性混合效应模型  

Nonlinear mixed effect model for airport composite pavement performance prediction

作　　者：张宏[1] 袁捷[2] 杜浩[3] 

机构地区：[1]内蒙古大学交通学院,内蒙古呼和浩特010070 [2]同济大学道路与交通工程教育部重点实验室,上海200092 [3]同济大学地下建筑与工程系,上海200092

出　　处：《长安大学学报（自然科学版）》2015年第6期56-62,84,共8页Journal of Chang’an University(Natural Science Edition)

基　　金：国家自然科学基金项目(51468047;51308412);内蒙古教育厅科学研究项目(NJZY13018;NJZY13016)

摘　　要：针对中国机场复合道面使用性能数据少、刚性道面数据较多的实际情况,采用非线性混合效应模型,并且联合2个数据源进行参数估计,建立了机场复合道面使用性能预测模型;通过对中国60多个机场的数据计算分析,估计了该模型参数,并分析了拟合效果。研究结果表明:非线性混合效应模型能够有效分解道面使用性能面板观测数据间存在的异方差,使残差满足正态假设并获得满意的拟合效果;非线性混合效应模型和联合估计方法一方面能解决部分个体观测数据不足的问题,另一方面提高了参数估计的稳定性和精度;非线性混合效应模型既可获得数据样本的总体趋势,也可得到个体的性能曲线,而且比传统方法的预估精度更高,为分析道面使用性能面板数据提供了新的技术手段。With regard to the fact that there are few data about airport composite pavement use performance while more data about rigid pavement in China,the nonlinear mixed effect model was adopted and two data sources were combined for parameter estimation,which led to the building of airport composite pavement use performance prediction model. The model＇s parameters were estimated and its imitative effect was analyzed based on the data set of the over60 airports in China.The results show that the heterogeneity originated from panned observation data of pavement use performance can be decomposed effectively,making residual error normal distributed as is assumed achieving the satisfactory imitative result.Nonlinear mixed effect model and jointed estimation method can be reasonable solutions to the problem of partial observation data deficiency,and can also improve the stability and precision of the parameter estimation.The total trend of data sample is got by nonlinear mixed effect model;the performance curve is alsoachieved.Moreover,compared with the traditional nonlinear regression,nonlinear mixed effects model provides a more realistic prediction.4tabs,15 figs,14refs.

关 键 词：道路工程 混合效应 模型 机场道面 性能预测 

分 类 号：U412.6[交通运输工程—道路与铁道工程]