节点文献

沥青路面车辙破坏模式识别、多维度指标评价与预测研究

Failure Pattern Recognition,Multi-dimensional Indicators Evaluation and Prediction of Rutting in Asphalt Pavement

【作者】 惠冰

【导师】 蔡宜长;

【作者基本信息】 长安大学 , 道路与铁道工程, 2013, 博士

【摘要】 随着我国高速公路里程的不断增长,交通量特别是重载交通量的持续快速增加,使得公路养护管理任务繁重而紧迫。车辙作为沥青路面最主要的损坏类型,不仅对道路结构产生影响,同时会对交通安全造成危害。全面综合、可靠快速的掌握车辙状况不仅能为养护管理决策提供依据,而且能及时、有效的预防由车辙引起的道路安全事故;对于保障道路通行能力和服务水平,提升公路养护决策的科学化水平和养护资金使用效率有重要意义。目前国内外广泛采用单一的深度指标衡量车辙的严重程度,但由于车辙成因的复杂性造成横断面形态的多样性,同一深度数值可能对应若干种不同的横断面形态,因而无法准确反映不同类型车辙对路面结构的影响程度,同时也无法反映路面与车辆作用的真实情况;使得分析现有的车辙深度检测数据无法为管理者提供全面、真实、准确的车辙信息,制约了公路养护管理的科学化、精细化发展。论文通过大量收集相关资料和文献,以陕西省某高速公路为依托,在车辙现场表观调查的基础上,结合钻芯取样分析,确定了四类车辙路段的主要破坏层位;结合国内广泛采用的多点激光位移传感器车辙检测设备所量测的路面横断面高程数据绘制了横断面图形,通过分析不同破坏层位车辙的横断面特征,在前人研究的基础上,实现了部分车辙一、二维评价指标的自动、准确计算;并且基于灰关联技术提出了车辙起终点确定和变异横断面剔除方法以及车辙凹凸体积变形量计算方法,实现了车辙三维体积指标的建立。在车辙多维度评价指标建立的基础上,采用主成分分析与相关性分析对指标进行筛选,确定了最大车辙深度、负面积和正负面积比三个包含信息大且独立性较高的评价指标,结合径向基神经网络建立了车辙层位识别模型;针对现行规范在车辙评价中存在的不足,从车辙对道路功能性与结构性危害两个角度出发,根据车辙破坏所涉及到的道路结构层位以及车辙形态对车辆行驶造成的影响,分别构建可量化的多维度评价指标,并提出了相应的评价等级和范围,建立了基于灰色-层次分析法的车辙评价模型;根据车辙多维度评价指标数据的等时距、小样本、高维度和多量纲的特点,考虑了多指标间相互影响、相互制约的关系,初步尝试采用MGM(1,n)模型建立车辙多指标的预测模型的可行性。通过建立车辙多维度评价指标并对其应用进行研究,是对我国现有车辙分析技术的完善和补充。研究有助于我国高速公路沥青路面病害检测分析水平的提高,对推进公路养护管理的科学化、精细化建设有重要的意义。

【Abstract】 With the rapid development of highway mileage in our country and the increasing of thetraffic volume especially heavily-loaded traffic, the task of highway maintenancemanagement is becoming heavily and urgently. Rutting as the main disease type of asphaltpavement, does harm to the road structure, affects the service quality and even causesconcern for traffic safety. Therefore, knowing the rutting situation comprehensively, reliablyand rapidly can not only provide a basis for maintenance management decision-making, butalso prevent traffic accidents caused by rutting timely and effectively. It is important for theprotection of road capacity and service levels, improving the scientific level of roadmaintenance decision-making and the fund usage efficiency.At present, depth as the single index to evaluate the serious degree of rutting is widelyused at home and abroad, but the same rut depth may correspond to different transversesection, so it cannot reflect the influence of the rutting to pavement structure accurately andthe real interaction between the pavement and the vehicle. Therefore, analyzing the text dataof rut depth cannot provide overall, real and accurate information for managers, which limitsthe development of the highway maintenance management to be scientific and meticulous.With a large collection of relevant information, on-the-spot survey and drilled core analysis,the destruction layers of three rutting-serious sections are determined, relying on a highwayin Shannxi province. With the cross-section elevation data measured by the multi-point laserdisplacement sensor rut detection equipment, the paper draws a cross-sectional graphics andproposes the multi-dimensional rutting evaluation indexes by analyzing the cross-sectionalcharacteristics with different destruction layers. The method of determining rutting start andend points and excluding the variation of cross-sectional and rutting bump volumedeformation calculation method are proposed based on Grey Theory and then thethree-dimensional rutting characteristic index are established.Based on the Multidimensional Evaluation rutting index, the paper selects the maximumrut depth, negative area and positive and negative area ratio, which contain more informationand independence characteristic, using the principal component and correlation analysismethods, and establishes rutting horizon identification model combining radial basis function neural network (RBF); Aiming at the shortcomings of the existing norms in the rut evaluation,the paper proposes an evaluation index system which can reflect the impact degree of therutting morphology and destruction layers, from the perspectives of the road functional andstructural hazards, and the rutting evaluation model based on gray-Analytic HierarchyProcess is established; Under the small sample size, high dimension and multi-dimensionalfeatures of multi-dimensional evaluation indicators data and considering the interactionbetween the indexes, the paper tries to build multi-index prediction model applying toshort-term rutting type using MGM (1, n) model.Studying the establishment of multi-dimensional evaluation index of the rutting and itsapplications can improve and complement our existing rutting analysis techniques. The studycontributes to the improvement of the analytic level of asphalt pavement disease detection,and has an great significance in promoting the scientific and meticulous development of roadmaintenance management.

  • 【网络出版投稿人】 长安大学
  • 【网络出版年期】2014年 05期
节点文献中: 

本文链接的文献网络图示:

本文的引文网络