Assessment of Asphalt Specimen Homogeneity

Abstract

The inhomogeneity of aggregates in an asphalt concrete pavement can be a primary factor in its deterioration. Statistical tests are useful for judging the extent of inhomogeneity. However, their use requires knowledge of both the distribution function of the statistic and the power of the test. The distribution is used in deciding whether or not a pavement is homogeneous, while the power of a test is used to select the test that is most appropriate for a particular gradation. Simulation is used to derive the distributions and powers of four tests, two based on aggregate frequencies and two on aggregate separation distances. Simulation must be used for evaluating the distribution of the homogeneity indices because they cannot be derived analytically, which is generally the case for asphalt aggregate specimens. A nearest-neighbor statistic is shown to be the most powerful for one homogeneous-inhomogeneous condition. Simulation is also used to show that independent estimates of a statistic can be obtained from a single core, which will increase the number of values of a statistic from a given number of specimens.