Analysis of Regression-Based Models for Prediction of Depth Temperature of Asphalt Layers – A Review

Due to the viscoelastic behavior of asphalt mixtures, depth temperature of asphalt layers is very important in structural evaluation of flexible pavements. Depth temperature could be measured directly in the field, or may be predicted using prediction models. This paper presents a comprehensive analysis of different twelve regression-based models for prediction of depth temperature of asphalt layers. With reference to the literature, required input parameters, sensitivity analysis, evaluation of prediction performance, as well as a comparison of goodness of these models were discussed. Furthermore, calibrated models for different local conditions were presented. This is due to the fact that the original models were usually developed in specific geographical regions and climatic conditions. Results show that the regression-based models have a good performance and high accuracy in predicting the depth temperature of asphalt layers. Among the investigated models, according to the variety of data (or parameters) used in the model development, performance, considering the effect of various parameters, BELLS model introduced as one of the best regression-based models for prediction of depth temperature of asphalt layers. The model developed by Solatifar et al. as a new version of BELLS model, showed very good accuracy for newly constructed pavements. In addition, with applying some modifications, it could be possible to use these models in different pavements and local conditions.