Quantifying Uncertainty in Dynamic Modulus for Asphalt Pavement Design: A Colorado Case Study

Abstract: This study focuses on the viscoelastic modeling and uncertainty analysis of asphalt mixtures in Colorado, aimed at improving the prediction of pavement performance. Dynamic modulus (|E*|), a critical property for long-term pavement durability, was analyzed using data from 160 asphalt mixtures with varying aggregate gradations and binder types. The research utilized a combination of sixteen master curve and shift factor models to assess the viscoelastic behavior of these mixtures under different temperature and frequency conditions. An in-depth error and uncertainty analysis was conducted to evaluate the accuracy of the viscoelastic models and identify key parameters influencing the reliability of dynamic modulus predictions. Specifically, the impact of error minimization techniques on master curve fitting was evaluated. Further, the impact of dynamic modulus uncertainty on pavement performance predictions using the AASHTOware Pavement-ME software was studied. Initial results suggest that certain combinations of viscoelastic models and fitting methods can significantly reduce uncertainty, leading to more precise predictions of pavement behavior. This enhanced accuracy in dynamic modulus prediction provides a stronger foundation for the design of more resilient pavements. Future research will refine these error minimization approaches and further evaluate the role of model parameters in error propagation. The study’s findings are expected to guide infrastructure design and resource allocation, offering improved reliability in predicting pavement performance across diverse conditions in Colorado.