Subgrade resilient modulus testing at Tennessee instrumented pavement sites : standard and alternative methods

The resilient modulus is a basic material property that is used in the current 1993 AASHTO Pavement Design Guide to characterize subgrade soils under vehicular loading. The resilient modulus is used in the mechanistic empirical design method to determine strains at the top of the soil subgrade and to predict strains developed in the overlying asphalt layer. The new 2002 AASHTO Pavement Design Guide, which is moving toward a more rational mechanistic design method, will continue to use the resilient modulus to characterize the subgrade. The resilient modulus is often estimated since the standard repeated load triaxial test is time consuming, not always economically feasible, and a limited number of laboratories have the capability to perform the test. Therefore many researchers have made attempts to develop other methods. A laboratory investigation was conducted on four fine-grained subgrade soils from various instrumented pavement sites across Tennessee. For each of the four site soils, index properties were determined. Several specimens were remolded for each site at optimum moisture and density. Repeated load triaxial tests were performed in accordance with the AASHTO T 307-99 procedure. A log-log equation proposed by Schwartz (2001) and Andrei (1999) was used to model the results. The results from the triaxial tests were found to be consistent and the model was found to have a strong correlation with the data obtained. The resilient modulus was also determined using the ATM (alternative test method) developed by Li (1992). Improvements were made to the ATM prototype device and to the data processing. Calculations based on the single degree of freedom mass spring system produced inconsistent and counterintuitive results. The data was reanalyzed by integrating the complete time-acceleration history to find the recoverable strain. The results from the double integration technique produced values of resilient modulus greater than the repeated load triaxial test. However the results were more consistent than those obtained with the previous analysis approach. Loading rate effects and differing states of stress between the repeated load triaxial test and the ATM, can explain the difference in values. A hammer weight and reduction factor was recommended for using the ATM to estimate resilient modulus. It was suggested that additional testing be performed on ATM specimens without confinement and the effects of sample length be investigated. While it is easy to find resilient modulus values for different stress states with the ATM by using different combinations of drop heights and weights, the model developed in the standardized triaxial test is still preferable and more compatible with pavement design methods. With the automated data collection and analysis system, the need for an alternative method may not be as great.