Development of Laboratory to Field Shift Factors for Hot-Mix Asphalt Resilient Modulus

Resilient moduli of different surface mixes placed at the Virginia Smart Road were determined. Testing was performed on Field cores (F/F) and laboratory-compacted plant mixed (F/L), laboratory mixed and compacted per field design (L/L), and laboratory designed, mixed, and compacted (D/L) specimens. The applied load was chosen to induce a strain ranging between 150 and 500 microstrains. Two sizes of laboratory compacted specimens (100-mm in diameter and 62.5-mm-thick and 150-mm in diameter and 76.5-mm-thick) were tested to investigate the effect of specimen size on the resilient modulus. At 5°C, the measured resilient moduli for both specimen sizes were similar. However, the specimen size has an effect on the measured resilient modulus at 25 and 40°C, with larger specimens having lower resilient modulus. At 5°C, HMA behaves as an elastic material; correcting for the specimen size using Roque and Buttlar's correction factors is applicable. However, at higher temperatures, HMA behavior becomes relatively more viscous. Hence, erroneous resilient modulus values could result when elastic analysis is used. In addition, due to difference in relative thickness between the 100- and 150-mm diameter specimens, the viscous flow at high temperature may be different. In general, both specimen sizes showed the same variation in measurements. Resilient modulus results obtained from F/L specimens were consistently higher than those obtained from F/F specimens. This could be due to the difference in the volumetric properties of both mixes; where F/F specimens had greater air voids content than F/L specimens. A compaction shift factor of 1.45 to 1.50 between the F/F and F/L specimens was introduced. The load was found to have no effect on resilient modulus under the conditions investigated. However, the resilient modulus was affected by the load pulse duration. The testing was performed at a 0.1s and 0.03s load pulses. The resilient modulus increased with the decrease of the load pulse duration at temperatures of 25°C and 40°C, while it increased at 5°C. This could be due to the difference in specimen conditioning performed at the two different load pulses. Finally, a model to predict HMA resilient modulus from HMA volumetric properties was developed. The model was tested for its fitting as well as predicting capabilities. The average variability between the measured and predicted resilient moduli was comparable to the average variability within the measured resilient moduli.