Correlating field performance to laboratory dynamic modulus from indirect tension and torsion bar

Dynamic modulus has several useful functions in flexible pavements, including stress/strain characterization, rutting and cracking characterization, an input into several analytical and numerical models, and a primary input into PavementME Design. While the traditional dynamic modulus test is run in the uniaxial configuration, this is not possible for field cores. Therefore, the Indirect Tension dynamic modulus (IDT |E∗|) and torsion bar shear modulus (torsion bar |G∗|) have been developed. However, there has been limited research looking at analyzing the data from field cores for these two geometries, comparing modulus data from the two geometries, examining in-service aging of dynamic modulus, and quantifying pavement conditions using dynamic modulus. This research examines ten field sections in Arkansas, comprised of four "good" performing sections, two "medium" performing sections, and four "poor" performing sections in an attempt to address these four questions. First, this research found that using AASHTO T 342 and AASHTO R 62 can lead to irrational coefficients but provide rational results. Second, while the IDT |E∗| and torsion bar |G∗| values were similar at high modulus values, the IDT |E∗| values began to increase as the modulus decreased compared to the torsion bar |G∗| values, increasing to over a decade of difference. Third, a noticeable difference was observed between the modulus values of the bottom surface layer and top surface layer, with the bottom surface layer showing higher modulus values in all cases. While the upper surface layer showed higher oxidation, other weathering effects such as moisture and traffic appear to have overwhelmed the oxidation effect and pavement deterioration has reduced the integrity of the mix. Finally, both the IDT |E∗| and torsion bar |G∗| were not able to quantify a noticeable difference between poor and medium performing sections, and medium and good performing sections, but were able to quantify a difference between the poor and good behaving sections. Overall, the IDT |E∗| and torsion bar |G∗| tests were able to produce consistent master curves, correlate to each other, identify differences between surface course lifts, and quantify differences in field performance.