DATA-DRIVEN ASSESSMENT OF PAVEMENT PERFORMANCE AND RESILIENCE UNDER EXTREME WEATHER USING THE LONG-TERM PAVEMENT PERFORMANCE PROGRAM
AI and Data Science Journal · 2024
DOI: 10.61784/adsj03037 ↗Kurz gefasst
Fachartikel in der Zeitschrift AI and Data Science Journal, 2024. Verfasst von ShuXin Zhang und 1 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref). Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Extreme weather driven by a changing climate is placing unprecedented stress on road infrastructure, yet the quantitative link between weather extremes and observed pavement deterioration—and the resilience of pavements to that stress—remains poorly characterized at scale. This paper presents an end-to-end, reproducible framework that couples the U.S. Federal Highway Administration Long-Term Pavement Performance (LTPP) database with interpretable machine learning and a novel pavement resilience formulation. From 376 asphalt-concrete test sections monitored between 1988 and 2016, we construct a panel of 2,656 section-year records enriched with engineered extreme-weather indicators (freeze–thaw cycles, extreme-heat and freezing days, intense-precipitation days, and thermal range). A variance decomposition shows that 48% of the variance in surface cracking is attributable to unobserved site-specific factors, explaining why absolute distress in unseen sections is intrinsically hard to predict from climate alone (transfer R² ≈ 0). A linear mixed-effects model that controls for site and pavement age isolates intense-precipitation days (standardized β = 1.10, p < 0.001) and extreme-heat days (β = 0.70, p = 0.043) as the statistically significant weather accelerators of cracking, whereas a condition-informed forecasting model reaches R² = 0.53 with prior condition dominating the SHAP attribution. We introduce a Climate Stress Index (CSI) and a two-dimensional stress–impact positioning that yields a Demonstrated Resilience Score, separating climate-driven vulnerability from intrinsic fragility. The framework offers actionable screening for climate-adaptive pavement asset management.
Zitieren
APA
Zhang, S., & Qiu, L. (2024). DATA-DRIVEN ASSESSMENT OF PAVEMENT PERFORMANCE AND RESILIENCE UNDER EXTREME WEATHER USING THE LONG-TERM PAVEMENT PERFORMANCE PROGRAM. AI and Data Science Journal. https://doi.org/10.61784/adsj03037
Vancouver
Zhang S, Qiu L. DATA-DRIVEN ASSESSMENT OF PAVEMENT PERFORMANCE AND RESILIENCE UNDER EXTREME WEATHER USING THE LONG-TERM PAVEMENT PERFORMANCE PROGRAM. AI and Data Science Journal. 2024. doi:10.61784/adsj03037
BibTeX
@article{Zhang2024,
author = {Zhang, ShuXin and Qiu, Lei},
title = {DATA-DRIVEN ASSESSMENT OF PAVEMENT PERFORMANCE AND RESILIENCE UNDER EXTREME WEATHER USING THE LONG-TERM PAVEMENT PERFORMANCE PROGRAM},
journal = {AI and Data Science Journal},
year = {2024},
doi = {10.61784/adsj03037}
}
EUVIMED