Asphalt Binder Linear Amplitude Sweep Test: Contribution Related to the α-Value Estimation

Abstract

This paper calls attention to important issues that have not been considered regarding linear amplitude sweep (LAS) test in previous publications: Is the relationship between the storage modulus logarithm [logG′(ω)] and frequency logarithm [log(ω)] linear in the whole range of experimental frequencies? If not, what is the influence on the fatigue life of asphalt binders calculated using the AASHTO protocol? In which range of frequencies is the relation between the storage modulus logarithm [logG′(ω)] and frequency logarithm [log(ω)] linear in the whole range of experimental frequencies? Does fatigue life follow a normal distribution? Several tests were conducted using three different types of asphalt binder [neat, crumb rubber–modified (CRM), and styrene–butadiene–styrene (SBS)–modified asphalt]. Residual analysis showed a lack of fit, indicating the inadequacy of the linear model in the whole range of experimental frequencies. This inadequacy may suggest that the protocol produces an overestimation of the binder fatigue life. In addition, a sequential search reveals that the relationship between logG′(ω) and log(ω) is linear in a limited subset of frequencies. The empirical distribution of the fatigue life is built by an extensive Monte Carlo simulation.