Regression Predictive Models for Depth Temperature of Asphalt Layers in Iran

Abstract

Abstract Depth temperature of asphalt layers is one of the most predominant required factors for asphalt pavements analysis, design, maintenance, and rehabilitation purposes. In this study, using the results from field experiments in six asphalt pavement sites located in different climatic conditions in Iran, the depth temperature of asphalt layers was investigated. By employing the four well-known regression-based predictive models including, Gedafa et al., Albayati and Alani, BELLS, and Park et al., the depth temperatures of asphalt layers were predicted. Two statistical criteria, accuracy and bias, have been adopted for evaluating the performance and capability of these models in predicting the depth temperature of asphalt layers. Results exhibited a fair correlation between the predicted depth temperature values and those measured during the Falling Weight Deflectometer (FWD) testing. However, it is necessary to enhance the prediction accuracy and diminish its bias by calibrating the mentioned models for determining the depth temperature of asphalt layers in pavements in Iran. For this reason, these existing prediction models were calibrated, and new predictive models were developed. Performance evaluation and validation of the newly developed models showed an excellent correlation between predicted and measured values. In addition, results exhibit the capability of the developed models in predicting the depth temperature of asphalt layers with very good prediction precision (R2 = 0.94) and low predictive bias.