Calibration of Regression Models Based on Viscoelastic Principles for Prediction of Dynamic Modulus of In-Service Asphalt Layers

Dynamic modulus is the viscoelastic behavior property of asphalt materials. Mechanistic–Empirical Pavement Design Guide (MEPDG) uses this modulus as an important input parameter in the design and rehabilitation of asphalt pavements. Dynamic modulus predictive models as an alternative method for laboratory determination of this parameter, are a major part of developing dynamic modulus of in-service asphalt layers. It should be noted that these models were developed based on laboratory data and there is a need for developing new models for prediction of dynamic modulus of in-service asphalt layers in different traffic and climatic conditions. In this research, ten asphalt pavement sites were selected in Khuzestan and Kerman provinces in Iran. At each site, Falling Weight Deflectometer (FWD) test was done and core samples were taken for extraction their binder and aggregate gradation. Using the test results, regression models for predicting dynamic modulus of asphalt mixes including Global Model and Simplified-Global Model, that developed based on viscoelastic principles, were used for analysis. These two models were calibrated and new models entitled as “In-situ Global Model” and “In-situ Simplified-Global Model” were constructed for predicting in-situ dynamic modulus of asphalt layers. Performance evaluation and validation of models showed very good correlation between predicted and measured values (R2=0.96). In addition, new in-situ models have a very good prediction accuracy and very low bias.