Multivariate Volumetric Specifications and Dynamic Modulus as a Quality Measure for Asphalt Concrete Materials
Files
TR Number
Date
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The Virginia Department of Transportation (VDOT) has worked toward end-result specifications (ERSs) in asphalt concrete since the mid-1960s. As stated by Hughes et al. (2007), true ERSs can lead to a reduction in VDOT's overall inspection force resulting in considerable savings and allow for the reallocation of inspection resources to key construction and placement processes that cannot be measured upon delivery (e.g., joint tacking and construction platform preparation). The latest efforts toward this end were conducted by Hughes et al. (2007) who suggested expanding the quality measures for asphalt concrete acceptance to include the asphalt concrete volumetric properties of voids in total mix (VTM) and voids in mineral aggregates (VMA), along with the already used asphalt content (AC) and gradation. This report builds on that and further investigates, through the use of the asphalt concrete dynamic modulus, how performance-related ERSs can be introduced into a quality assurance (QA) plan. Specifically, the report 1) documents the current variability of VTM, VMA, and AC; 2) explores different QA specification plans; and 3) develops and applies a method to predict asphalt concrete rutting performance from asphalt concrete dynamic modulus test results using the mechanistic-empirical pavement design guide (MEPDG). Contractor volumetric test results (for the years 2006 through 2008) for VTM, VMA, and AC were obtained from VDOT's central database for production asphalt concrete. Statistical measures of mean, variance and covariance were calculated. The experimental distribution of test results for each of the three volumetric measures was obtained and compared to the normal (Gaussian) distribution. This research used these data and exploratory analysis to present alternative QA plans, which ranged from a simple univariate plan to a multivariate percent within limits (PWL) plan. The choice of a specific plan to implement depends, among other criteria, on the variable-more specifically on the correlation between these variables-that are included as part of this plan. The PWL method for uncorrelated variables (in this case VTM and AC) is recommended as it presents a sound statistical approach that avoids the complexities that result from incorporating correlated variables. With advances in mechanistic-empirical pavement design methods (specifically the new MEPDG), a framework for performance-related ERSs is now available. The dynamic modulus as a function of temperature and frequency is the main asphalt concrete material input property in the MEPDG. It has significant influence on distress prediction, which makes it a quality candidate test for performance-related ERSs. A principal technical barrier to using the dynamic modulus test is the time required to perform the test temperature sweep. To address this obstacle, this report presents a method to reduce the required number of tests to characterize asphalt concrete rutting characteristics. It demonstrates that a single dynamic modulus test is sufficient to estimate asphalt concrete rutting potential as calculated by the MEPDG. This is an initial step towards using the dynamic modulus in performance-related ERSs. However, actual implementation still depends on broader acceptance and use of the dynamic modulus testing equipment and procedures, as well as the proper calibration of the MEPDG distress models to reflect observed field performance. If and when this is accomplished, the method can be extended to fatigue cracking.