Abstract
The propagation of cracks in in-service asphalt pavements is closely related to the complicated traffic loading patterns over time. However, typical traffic-related variables capture only the overall traffic level without being able to account for the load-time history. Therefore, this study aims to investigate the effects of traffic load sequence on the cracking performance of asphalt pavement from both field and laboratory perspectives. A load amplitude sequence (LAS) index was developed to characterize the traffic loading sequence in the field. Two machine learning (ML) algorithms, namely artificial neural network (ANN) and random forest regression (RFR), were applied to correlate the LAS index with field pavement cracking performance. The two-block semi-circular bending (SCB) test was developed to characterize the non-linear fatigue damage accumulation of asphalt mixtures. It was found that heavier traffic loads in early stages are detrimental to the long-term pavement cracking performance. The LAS index plays a crucial role in the prediction and development of pavement cracks. The laboratory test results reveal that a loading sequence starting with a higher stress may shorten the fatigue life and vice versa. The outcomes of this study may contribute to a better understanding of the traffic loading characterization of in-service asphalt pavements.
Original language | English |
---|---|
Article number | 2152027 |
Journal | International Journal of Pavement Engineering |
Volume | 24 |
Issue number | 1 |
DOIs | |
Publication status | Published - Jan 2023 |
Keywords
- Load amplitude sequence
- machine learning
- pavement cracking
- traffic loading characterisation
- variable amplitude fatigue test
ASJC Scopus subject areas
- Civil and Structural Engineering
- Mechanics of Materials