Repeatability and Reproducibility Analysis for Functional Test Results from Asphalt Mixture Performance Tester Cyclic Fatigue Test

Abstract

The standards for asphalt mixture performance tester (AMPT) cyclic fatigue testing (AASHTO T 400-22 and AASHTO TP 133-21) enable pavement engineers to predict asphalt mixture performance over a wide range of loading and climate conditions. To enable widespread implementation of AMPT cyclic fatigue testing, precision statements that define the repeatability and reproducibility of the tests are needed. ASTM E691-20 and ASTM C670-15 provide guidance for conducting an interlaboratory study to determine the precision of a test method. However, the standards are written for test methods that yield a single numerical figure as the test result, whereas the cyclic fatigue test yields the damage characteristic curve—a functional relationship—as a primary test result. In this study, a vnorm index was established to quantify variation in the entire damage characteristic curve. Repeatability and reproducibility analysis was conducted using vnorm, six single-point measures of damage characteristic curve variation, and one single-point measure for the fatigue test failure criterion ( D R). The major findings are: (1) the outcomes of repeatability and reproducibility analysis using single-point measures of the damage characteristic curve vary with the chosen point of reference; (2) vnorm demonstrates clear increases in within- and between-laboratory variations with increasing mixture nominal maximum aggregate size (NMAS) values; (3) within-laboratory variation of D R results increases with increasing NMAS for a given specimen geometry while the reproducibility of D R does not depend on the NMAS. Correspondingly, precision limits for both vnorm and D R were recommended and evaluated using a broad set of test results.