Coupling asphalt construction process quality into product quality using data-driven methods

Coupling asphalt construction process quality into product quality using data-driven methods Qinshuo Shen, Faridaddin Vahdatikhaki, Seirgei Miller, Andre Doree Pages 349-356 (2023 Proceedings of the 40th ISARC, Chennai, India, ISBN 978-0-6458322-0-4, ISSN 2413-5844) Abstract: The long-term quality of the asphalt layer is crucial for maintaining the functionality of roads. Despite extensive research on predicting pavement failure modes and the effect of design and road use on the quality of the asphalt layer, there is limited understanding of how the quality of road construction impacts the long-term quality of asphalt pavement. This paper presents a data-driven approach to studying the impact of construction process quality on the International Roughness Index (IRI) of roads. Two machine learning models (Random Forest and Gated Recurrent Unit) were compared in a case study, with the GRU model (R2 of 0.8284) outperforming the RF model (R2 of 0.5498). Results showed that construction process quality was the third most significant factor affecting IRI. Keywords: Asphalt construction, construction process quality, international roughness index (IRI), data-driven methods, regression, machine learning DOI: https://doi.org/10.22260/ISARC2023/0048 Download fulltext Download BibTex Download Endnote (RIS) TeX Import to Mendeley Presentation Video: https://youtu.be/vdBVP5kklLU