Failure Pattern Recognition,Multi-dimensional Indicators Evaluation And Prediction Of Rutting In Asphalt Pavement

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.