نوع مقاله : علمی - پژوهشی
نویسندگان
1 استادیار، دانشکده مهندسی عمران، دانشگاه صنعتی شاهرود ، شاهرود ، ایران
2 دانشجوی کارشناسی ارشد راه و ترابری، دانشگاه صنعتی شاهرود ، شاهرود، ایران
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
The most important part of a pavement management system is the pavement performance prediction model. The efficiency of maintenance and rehabilitation is dependent on the accuracy and validation of pavement performance prediction model. In family models, several pavement sections with similar properties and deterioration fall into a family. So, performance model is made for all pavements of a family. Modeling based on pavement family can provide high accuracy results with fastest and costless way and minimum data is needed. In this paper, two pavement families are defined in City of Sari streets including family 1 with high traffic loading and more asphalt thickness and family 2 with low traffic loading and less asphalt thickness. All arterial and other important streets are evaluated, and then pavement condition index (PCI) and pavement age is determined. For each pavement family a regression model is made and finally a third degree model is developed. Family 1 of pavements presents R2= 0.90 and family 2 has a regression coefficient R2= 0.84. In addition, for each family, the pavement performance is predicted by using a multi-layer perceptron neural network. In both families a regression coefficient can be seen around of R2 = 0.93. Models are made based on just one pavement evaluation, so we can define their accuracy as very good. High accuracy model is resulted by pavement family idea. The results show that combining between family model and artificial neural network (ANN) can provide more accurate prediction than regression method. So, this method is recommended for movement management system in earlier stages.
کلیدواژهها [English]