Abstract
Reclaimed Asphalt Pavements (RAP) and Recycled Asphalt Shingles (RAS) are two of the most commonly used reusable materials in asphalt industry because of their environmental and economic benefits. One of the biggest disadvantages in incorporating RAP and RAS into hot mix asphalt is their asphalt binder which is oxidized to high extents as they were in service for several years. The aged asphalt binder when mixes with virgin materials, e.g. new aggregates and binder, creates an unknown state in which only some part of it blends with the virgin binder. As the binder oxidizes, it becomes stiffer and more brittle which is beneficial at high temperature but unfavorable at intermediate and low temperature regimes. In this study, asphalt mixtures were designed with varying proportions of RAP and/or RAS and went through the mixing and compaction procedures. By measuring the dynamic modules of asphalt mixtures at varying temperatures and frequencies, master curves are constructed at a unique reference temperature. Additionally, asphalt binder from RAP and RAS was extracted and recovered and then blended with different percentages of virgin binder. Performance Grade (PG) of all the blended binders were measured and their associated master curves were constructed at the same reference temperature considered for mixtures. Hirsch Model is one of the most popular formulations that makes use of volumetric properties of asphalt mixtures to correlate moduli of mixtures and binders. Constants of this model have been calibrated based on a control mixture to match with the aggregate skeleton designed in this study. Substitution of a binder master curve into this locally calibrated model represents a master curve for its associated mixture. Each master curve for the artificially blended binders would then result in an estimation of mixture master curve, using the Hirsch Model. Scatter of measured versus estimated mixture master curves can predict the closest degree of blending occurred during mixing, conditioning and compaction. Least Sum of Squared Errors and regression through origin are two statistical methods to discover the closest prediction to estimation. The output is a percentage of RAP or RAS binder that diffuses into the virgin binder and forms a binder with new rheological characteristics. The PG of the resultant degree of blending is of high importance since it is one of the key parameters to predict asphalt pavement distresses at ME Design software. A number of mechanical tests on some of the mixtures has also been carried out to confirm the methodology implemented in this study. Investigation on low temperature cracking potential, fracture energy, and fatigue behavior of asphalt mixtures matched well with PG of the resultant degrees of blending. It is worth mentioning that the method presented in this study is best applicable to mixtures made with RAS and low percentages of RAP since uncertainties on aggregate properties may play as a source of error when high amounts of RAP is introduced into mixtures.