Stochastic Identification of Linear-Viscoelastic Models of Aged and Unaged Asphalt Mixtures

Abstract

The behavior of asphalt mixtures is typically described using linear viscoelastic models at low-strain applications. The time and temperature dependency of asphalt mixture properties is described by forming a master curve that includes three components: time-temperature shift factors, and storage and loss moduli (or compliances). Mathematical models are needed to describe the master curve, which are used in calculating asphalt pavement responses to load and also to compare the overall properties of various mixtures at a wide range of temperatures. This paper proposes a rigorous approach to mathematically describe the master curves using stochastic identification techniques. These techniques have the advantage over current deterministic methods in their ability to account for the uncertainty associated with the constructed models, which could be contributed to variation in the material properties of the asphalt mixture phases as well as their spatial distribution, measurement errors, modeling errors, and inadequate available information. Consequently, uncertainty can be accounted for in the analysis and design of asphalt pavements. The stochastic approach is used successfully in this paper to identify the linear viscoelastic master curve for asphalt mixtures that have been aged to different time durations.