Predicting Rutting Development of Pavement with Flexible Overlay Using Artificial Neural Network

Pavement maintenance and repair is a crucial part of pavement management systems. Accurate reliable performance prediction the prerequisite for making reasonable decisions selecting suitable schemes. Rutting deformation, as one most common forms asphalt failures, key index evaluating performance. To ensure accuracy commonly used models, input parameters models need to be understood, coefficients should locally calibrated. This paper investigates rutting development pavements with flexible overlays based on data Canadian Long-Term Performance (C-LTPP) program. that may related were extracted from survey Dipstick analysis. Then, an artificial neural network (ANN) was adopted analyze factors affecting rut depth, establish model overlays. The results sensitivity analysis indicate not only affected by traffic climatic conditions, but it also greatly thickness surface layer voids in mixture. Finally, evaluation provided describe severity, threshold time proposed results. These provide basis predicting maintenance.