摘要 半刚性基层沥青路面是我国高等级公路的一种主要结构形式。近年来，由于材料和设计体系的原因，沥青路面早期损坏现象非常严重表现为横向裂缝、纵向裂缝、网裂等。这些结构性破损使得路面结构逐渐丧失整体性，削弱承载力，影响行车安全性和舒适性。目前，针对这种结构性破损路面的检测、评价和养护已成为路面管理的重要工作。FWD(落锤弯沉仪)作为固定采样、动态弯沉类的高自动化机械弯沉测试仪，是目前应用较为广泛的弯沉检测设备，代表了弯沉检测和结构性能评价的发展方向。本文正是基于FWD的测试数据，构建了适于半刚性基层沥青路面的结构性能评价体系。 本文首先重点研究沥青路面的动力特性，针对路面的各种开裂缝模式建立了有限元分析模型，其中对路面结构和裂缝的形态都作了较大的简化，计算分析了在FWD冲击荷载下沥青路面在不同裂缝问距、宽度和开裂深度情况下和层间出现滑动情况下的动力响应，通过与完好状态下的结果比较，分析了这些破损状态对路面动力计算的影响。计算过程采用Fortran90编程，从前处理到动力方程求解，均编制了相应的子程序，为后文理论弯沉盆参数分析提供数据来源。 然后本文基于理论弯沉数据进行有效弯沉盆参数分析。将实际路面结构简化为三层结构开裂模型j在土基状态分析时，通过人工动态弯沉盆数据得到了土基模量预测参数；用同样的方法得到了两层路面上层和三层路面的面层、基层结构破损状态的弯沉贫预测参数。然后利用这些参数训练新建的人工神经网络， 最后利用实测弯沉数据评价路面结构性能，包括路面结构层的动态模量以及路段的剩余寿命等。相应地，本文基于Matlab2006平台编制了神经网络反演程序ToEsg．m 和剩余寿命预估程序NNSEm，并提供了结构性能评价分析的算例。最后是本文理论在试验路上的应用情况。首先对107国道保定一石家庄段的高等级半刚性基层沥青路面进行了全面的路况调查，着重分析了反映结构性能的两项指标：路面状况指数和路面强度指数，然后基于实测FWD数据，利用人工神经网络进行了模量反算和剩余寿命预估。根据评价结果提出了基于三层体系的补强设计方法。另外，利用本文理论对宝山实验路实测FWD数据进行了剩余寿命计算，用于指导补强设计。 关键词：半刚性基层沥青路面，有限元分析，动力响应，神经网络，补强设计，动态模量，剩余寿命。 ABSTRACT Semi-rind base asphalt pavement is the main structural type of highway．In recent years，because of the defects of semi—rigid base’s character and design system，the premature failures are very serious in this pavement，showing as transverse crack， longitudinal crack，and alligator crack，etc．All types of crack will gradually deteriorate the integrity,reduce the beating capacity,and greatly affect driving safety and amenity．Detection，evaluation and maintenance to this type of pavement have become most important of the pave management system．As an advanced detector, FWD(falling weightght deflectometer)has been widely used in the world at the present time．Therefore，on the basis of the data of FWD，constructional evaluation system was built in this dissertation，which is fit for Semi．rigid base asphalt pavement． Firstly,the dynamic character of the material and structure in the asphalt pavement Was analyzed．Three double—layered models of damaged asphalt pavement were built，in which lOaded crack，reflected crack and inter-layer slide were considered respectively．Then transient finite element analysis(FEA)was applied to acquire the dynamic response of deflection under FWD load．In the process．the shapes of structure and the crack were simplified，and various spacing，width and depth are considered，as well as the sliding condition between layers．The impact of distress Was analyzed compared with intact pavement. All subroutines were compiled in Fortran90，which supplied theoretical data for research of DBP． Secondly, research for deflectometer basin parameter Was conducted，in which the actual pavement was simplified as three．1ayered model．To evaluate the subgrade．predictor parameters for dynamic modulus were generalized through theoretical deflecttion database．These parameters，together with the thickness of structure were input into BP ANN(Artificial Neural Network)to train it．with the trained ANN，the accrual dynamic modulus and sectional remaining 1ife can be directly obtained．Correspondingly,ToEsg．m and NNSP．munder Matlab2006 were compiled to evaluate the modulus and remaining 1ife respectively．Furthermore，one trial pavement section Was evaluated the program． Finally, the theory Was applied to trial section．In this phase，overall survey was carried out to Bao．shi section in 107 national highway．Two indices．PCI and SSl were emphasized．Then based on the FWD data,modulus and remaining life were respectively computed．According to the results，a three-layered designed method for reinforcement Was developed relied on current design specifications．In the mean time, the remaining 1ife of Baoshan was computated using field data．. Key Words: emi—rigid base pavement，FEA，dynamic response，ANN， reinforcement design，dynamic modulus，remaining life |
博士论文