Probabilistic approach to characterise laboratory rutting behaviour of asphalt concrete mixtures

• Published in: The international journal of pavement engineering Vol. 21; no. 3; pp. 384 - 396
• Main Authors: Singh, Priyansh, Swamy, Aravind Krishna
• Format: Journal Article
• Language: English
• Published: Abingdon Taylor & Francis 23.02.2020; Taylor & Francis LLC
• Subjects: Asphalt; Asphalt concrete; eegression; hamburg wheel testing; Laboratories; probabilistic approach; rut depth; rutting; Scattering; Statistical analysis; Test procedures; Tracking devices; Weibull distribution; wheel tracking test
• ISSN: 1029-8436; 1477-268X
• DOI: 10.1080/10298436.2018.1480780

Even under extremely controlled laboratory conditions, wheel tracking test results shows significant scatter. This scatter can be attributed to variations in constituent materials, specimen fabrication, aggregate skeleton and testing practices. This paper presents a probabilistic approach to characterise the scatter found in rutting test results. For this purpose, two distinct AC mixtures were designed and evaluated. Several slabs were tested for rutting susceptibility using wheel tracking device. Initially, wheel rut test data from each slab was fitted with Francken model to smoothen the data. The smoothened data was further used to predict (i) rut depth or (ii) number of passes at prespecified locations. The data thus generated (at a specific rut depth or number of passes) was fitted with normal, log normal and Weibull distribution. Using predicted values (rut depth or number of passes) made for a particular probability, probabilistic rutting curve (PRC) were constructed. Based on location of PRCs, it can be concluded that PRCs obtained assuming Weibull distribution results in conservative rut estimates while those based on lognormal distribution results in overestimation. Further, significant differences were found between conventional ordinary least squares approach and PRC's developed in this research. The proposed methodology combines advantages of traditional testing protocol and probabilistic approaches.