Framework for Evaluation of Asphalt Mixture Cracking Resistance within the Balanced Mix Design Process during Production

Abstract State agencies require rigorous testing to ensure the quality and durability of asphalt mixtures used in pavement construction. Standard protocols mandate frequent evaluation of key mixture properties, including gradation, binder content, volumetrics, and performance metrics to ensure pavement durability. However, time-consuming laboratory tests often delay corrective actions, creating a need for efficient predictive tools to streamline quality control. This study developed a data-driven framework to estimate the cracking resistance of asphalt mixtures, as quantified by the cracking tolerance (CT) index from the indirect tensile cracking test (IDT-CT). Variable selection and model robustness were guided by rigorous statistical methods, including correlation analysis, variance inflation factors, and principal component analysis. Three diverse data sets spanning laboratory- and plant-produced mixtures were compiled for model development. Key predictors included asphalt content, aggregate specific gravity, the Glover–Rowe parameter, and gradation (as measured by the No. 30 sieve). A comprehensive suite of statistical diagnostics, combined with forward feature selection (FFS), was employed to establish a practical, statistically robust modeling framework. Multiple linear regression was selected for its interpretability, target-oriented optimization, and compatibility with regulatory design practices. The proposed model demonstrated reliable predictive performance across training, validation, testing, and independent verification stages. A case study further demonstrated the practical application of the proposed framework in proactively identifying, addressing, and minimizing risks associated with current and potential upgrades to acceptance specifications. To further support the use of the model, future work should incorporate additional data to enhance its ability to estimate the cracking potential of asphalt mixtures and improve operational efficiency.