基于人工神经网络的沥青路面龟裂预测研究

Prediction of alligator cracking of asphalt pavement based on artificial neural network

沥青路面在车辆重复荷载作用下,面层或稳定基层破坏严重,易发生龟裂,是力学-经验路面设计方法中应考虑的重要因素。通过数据挖掘技术,提取了美国长期路面性能4 490组数据,分析了不同类型因素对路面龟裂的影响规律,尤其是非线性因素。为准确预测沥青路面在使用过程中龟裂破坏程度,利用人工神经网络模型法对龟裂破坏进行了分析,模型综合考虑了交通量、路面结构与施工、气候条件等影响因素,预测结果与实际值较为接近。研究成果可在沥青路面设计时对龟裂破坏预测建模提供理论支持,并对优化路面结构设计有一定指导作用。

Under the action of repeated vehicle loads,the surface layer or stable base layer of asphalt pavement is seriously damaged and prone to alligator cracking,which is an essential factor to be considered in the mechanical-empirical pavement design guide.Through data mining technology,4 490 sets of long-term pavement performance data onto the United States were extracted,and the influence of different types of factors of pavement alligator cracking was analyzed especially nonlinear factors.In order to accurately predict the degree of cracking and damage to asphalt pavement during use,the paper uses an artificial neural network model to analyze the alligator cracking damage.The model comprehensively considers factors such as traffic volume,pavement structure and construction,and climatic conditions.The predicted value is close to the actual value.The research results can provide theoretical support for the prediction and modeling on crack failure of the design of asphalt pavement and have a specific guiding role in optimizing the design of pavement structure.