Modeling Pavement Temperature Prediction Using Artificial Neural Networks

Abstract

One of the key elements in the mechanistic-empirical (M-E) design system of flexible pavements is a pavement temperature prediction model. Many regression models have been developed to predict the temperature of asphalt concrete (AC) layer using air temperature with some other input parameters. Some of these models are old and cannot be applied to various site locations with accuracy. Others are quite accurate but they require many input data parameters that may not be available to the ordinary practitioner. Therefore, this paper discusses the feasibility of applying artificial neural network (ANN) technology in predicting the AC layer temperature. The neural network has been trained and tested using NeuroSolutions 5.0 software through actual field data obtained from Long-Term Pavement Performance (LTPP), Seasonal Monitoring Program (SMP) - DataPave3.0 software. Two ANN models have been created: The first is based on air temperature together with some other parameters, and the second is based only on air temperature for simplicity. Results indicated that the developed ANN-based pavement temperature prediction models can be used in predicting AC layer temperature with high accuracy as compared with measured values. This outcome is considered crucial to the pavement design especially the second ANN model where some input parameters may not be available.