Application of Artificial Neural Network (ANN) to PA Lifespan: Forecasting Models

More than 60% of the Dutch motorways are covered with porous asphalt (PA) which has many advantages; however, its lifespan is mostly short causing high maintenance costs. The PA lifespan is defined by the combination of mixture properties, historical damage, construction condition, and environmental factors. Due to the complexity of this combination, Dutch road engineers needs a more powerful modeling technique than the conventional ones to analyze PA lifespan. In this study, we developed artificial neural network (ANN) models for PA lifespan. The study used 102 PA road sections obtained from SHRP-NL database containing ten years data (1991-2000). The ANN forecasting resulted in many valuable results such as the proper percentage range for each item in asphalt mixture to increase PA lifespan. The study showed that ANN could be the powerful modeling technique that Dutch road engineers seek. However, ANN models are as good as allowed by the data.