【作者】 苑苗苗；

【导师】 张肖宁；

【作者基本信息】 华南理工大学 ， 道路与铁道工程， 2013， 博士

【摘要】 疲劳破坏是沥青路面的主要破坏形式，也是沥青路面结构基础理论与设计的本源性问题。研究表明，路面行车荷载的随机性、环境因素的复杂性，以及沥青混合料细观结构的变异性导致了沥青路面疲劳响应的分散性，从而使得室内疲劳试验结果与实际路面疲劳损伤累积过程存在较大的差异。因此有必要从细观层次上研究沥青混合料的疲劳破坏机理。本文基于光学测量方法——数字散斑相关方法（DSCM）及Matlab平台自行编程，实时观察并测量了疲劳加载过程中沥青混合料的位移场和应变场，通过DSCM方法将图像信息转化成了能够表征材料力学性能的变形信息，并对变形场的局部化特征进行了分析。然后引入了沥青混合料疲劳损伤阈值的概念，结合四点弯曲疲劳试验，验证了沥青混合料微损伤的愈合效应，计算了沥青混合料的疲劳损伤阈值。基于沥青混合料等应力区损伤累积的随机性，提出沥青混合料存在局部失效强度，并给出了局部失效强度的理论推导过程及破坏概率函数。最后对沥青混合料的疲劳失效判据进行了探讨及验证。提出了基于可变步长的耗散能相对变化率的突变点作为疲劳失效判据，并给出了理论解释。本文综合考虑现有试验条件及设备，主要完成了以下工作：1）为了获得沥青混合料在循环荷载作用下的变形分布及演化特征，详细探讨了光学测量方法——数字散斑相关方法（DSCM）的基本原理及数值计算方法。2）基于Matlab平台结合DSCM数值计算方法自行编程，实现了亚像素位移及全场应变的计算，并对其进行了数值模拟验证，然后对实际试验环境中的位移和应变测量精度进行了分析。3）分别对两种粒径的基质沥青混合料梁试件进行了四点弯曲疲劳试验，通过DSCM测量系统，实时监测了沥青混合料疲劳加载过程中的位移场及应变场，并对其断裂过程区进行了试验观察，得出了沥青混合料在四点弯曲疲劳加载下的变形分布特征。4）根据失稳断裂前应变场的分布特征及结合数字图像，对沥青混合料断裂过程区的尺寸进行了定义，并对沥青混合料疲劳裂纹扩展路径与集料的相对位置进行分析。在弯拉强度试验中，通过DSCM测量系统，捕捉到了沥青混合料的起裂时刻，发现沥青混合料在峰值荷载之前起裂，并计算了起裂应变。5）根据断裂统计理论，在分析材料的破坏行为时，必须预先给定一个达到破坏的损伤阈值。因此基于沥青混合料疲劳损伤阈值的基本思想，结合四点弯曲疲劳试验及弯拉强度试验，计算了沥青混合料的疲劳损伤阈值，并对沥青混合料微损伤的愈合效应进行了试验验证。6）在对未预制切口的试件进行四点弯曲疲劳加载过程中，通过DSCM测量系统发现试件中段等应力区，不同位置存在不同程度的损伤累积，因此推断沥青混合料存在局部失效强度，并给出了局部失效强度的理论推导过程，此理论推导同时考虑了沥青混合料的黏弹性、裂纹尖端应力重分布及荷载的间歇性等因素。这与沥青路面在真实交通荷载作用下的情况相符。7）采用统计断裂力学中的基本模型——最弱链环模型，对沥青混合料断裂破坏过程进行了的概率分析，推导出了沥青混合料在某一应力作用下的断裂概率函数。8）鉴于目前国内外对于沥青混合料的疲劳失效判据未达成共识，指出疲劳失效判据的不统一是导致沥青混合料疲劳试验数据分散性大的原因之一，提出了疲劳失效判据应具备的条件。然后通过对不同配合比的混合料进行不同荷载控制模式的四点弯曲疲劳试验，提出了基于可变步长的耗散能相对变化率突变点作为疲劳失效判据，并对其物理意义进行了理论解释。更多还原

【Abstract】 Fatigue failure due to repeated loading is one of the major distress in asphalt pavements.That is also the major issue in basic theory of strcture and design in asphalt pavement. Studieshave shown that traffic load randomness, environment complexity and the mesostructurevariability of asphalt mixture lead to the randomness of asphalt pavement fatigue response. Sothere are big differences between indoor fatigue test and fatigue damage accumulation processof the actual pavements. Therefore, it is necessary to acquire more insight into the fatiguebehavior of asphalt mixture from the mesoscopic level,to obtain a better understanding of thefatigue failure mechanism of asphalt mixtures.The displacement fields and strain fields of asphalt mixtures in four-point bending fatigueloading process were obtained in this paper using digital speckle correlation method andMatlab platform. Furthermore, the local characteristics of the strain fields were analyzed; localfailure strength of asphalt concrete was proposed; the asphalt mixture fatigue failure criterionwas discussed. And finally this paper proposed the ratio of dissipated energy change toindicate the damage accumulation in the mixture as asphalt mixtures fatigue failure criterion,giving a theoretical proof. Considering the existing test conditions and equipment, the maininnovative work completed in this paper is included as follows:1) In order to obtain the deformation distribution and evolution characteristics of asphaltconcrete under cyclic loading, the basic principle of the digital speckle correlation method(DSCM) and numerical methods were introduced in detail.2) The subpixel displacement and strain fields were calculated based on Matlab platformcombining numerical calculation method. And then Numerical simulations were carried out onthe calculation results; the accuracy of displacement and strain measurement in actual testenvironment was discussed.3) Four-point bending fatigue test was carried out on asphalt concrete beam specimens oftwo particle size matrix. By the DSCM measuring system, the displacement field and strainfield of asphalt concrete were monitored in real time under4PB test. And then the deformationdistribution characteristics of asphalt concrete were obtained under4PB test.4) The size of asphalt concrete in the fracture process zone was defined, according to thedistribution characteristics of the strain field before the unstable fracture and digital images.The relative position between fatigue crack growth path and aggregates was analized. In4PBstrength test, the crack initiation time of asphalt concrete was captured by DSCM measuringsystem, awaring of the asphalt concrete starting to crack before peak load and then failure strain was calculated.5) Based on the fracture statistical theory, damage threshold must be pre given beforeanalyzing material fracture behavior. Therefore, based on the basic idea of the asphalt mixturefatigue damage threshold, combining4PB fatigue test and flexural strength test, microdamagehealing effect of the asphalt mixture was verified and fatigue damage threshold was calculated.6) During4PB fatigue test, there are different degrees of damage accumulated at differentlocations in the middle of specimen which is equal stress zone. Therefore,the presence ofpartial failure strength in asphalt concrete was inferred, giving the approach of calculatingpartial failure strength.This approach appeared to be rational and to provide the followingadvantages: strain rate was used in the determination of the dissipated creep strain energy limit,which accounts for the effects of viscoelasticity during repeated loading test;the actual stressdistribution near the crack tip was considered;dissipated creep strain energy per cycle wasdetermined by integrating the stress and strain functions associated with the actual loadingcondition(i.e.,the harvesine load).7)The failure probability of asphalt concrete fatigue failure was analized using statisticalfracture mechanics model named weakest link model. And then the fracture probabilityfunction was derived under some stress level.8) Presently, it is not reached consensus for asphalt mixtures fatigue failure criterion.Themost universally accepted fatigue criterion is a decrease in initial stiffness by50percent inflexural fatigue test. However, the50percent reduction in initial stiffness is an arbitrarynumber with no relation to damage accumulation in the material. This paper presents ratio ofdissipated energy change to indicate the damage accumulation in the mixture as asphaltmixtures fatigue failure criterion, giving a theoretical proof. This procedure shows similaritybetween constant stress and constant strain modes of testing. Finally, this paper verifies thefatigue performance prediction model, based on the ratio of dissipated energy change, isindependent of the load control mode.更多还原