Prediction of Mechanical Properties of Spent Coffee Ground Ash Mixed Asphalt Using Experimental Investigation and Machine Learning Approaches
Advances in Civil Engineering · 2026-01 · 2026 (1)
DOI: 10.1155/adce/9281929 ↗Kurz gefasst
Fachartikel in der Zeitschrift Advances in Civil Engineering, online erschienen am 10.07.2026. Verfasst von Shiva Kumar G. und 5 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 73 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
This study investigates the feasibility of utilizing spent coffee ground ash (SCGA) as a sustainable bio‐modifier for asphalt binders and mixtures through an integrated experimental and computational framework. The novelty of this research lies in the adoption of a dual‐strategy approach that combines comprehensive experimental validation with explainable artificial intelligence (XAI) for predictive performance assessment. Unlike previous investigations that focused separately on binder‐level or mixture‐level behavior, the present study integrates laboratory characterization, mechanical performance evaluation, machine learning prediction, and SHAP‐based interpretability within a single unified framework. SCGA was incorporated at dosages of 2%, 4%, 6%, and 8% by weight of binder to evaluate its effects on physical, rheological, durability, and mechanical properties of asphalt mixtures. Experimental investigations included penetration, softening point, viscosity, storage stability, Marshall stability, indirect tensile strength (ITS), tensile strength ratio (TSR), rutting resistance, fatigue performance, and ultrasonic pulse velocity (UPV) assessments. Results demonstrated that SCGA significantly enhanced binder stiffness and high‐temperature performance, with penetration decreasing by 21% and viscosity increasing by nearly 49% relative to the control binder (CB). The optimum performance was achieved at 6% SCGA, where Marshall stability increased by 19%, ITS improved by 10%, rutting resistance increased by ~12%, and fatigue life (FL) improved by nearly 12% compared with the conventional mixture. However, excessive SCGA content (8%) reduced moisture resistance and fatigue performance because of increased brittleness and excessive stiffening. To strengthen predictive capability, decision tree and XGBoost regression models were developed, where the decision tree model achieved superior prediction accuracy with an R 2 value of 0.98. Furthermore, SHAP analysis identified ITS and modulus‐related parameters as the most influential factors governing asphalt mixture performance, demonstrating the effectiveness of XAI in supporting sustainable pavement material design and engineering decision‐making. This study highlights SCGA as a promising eco‐friendly additive for enhancing the properties of asphalt pavement, contributing to Sustainable Development Goals (SDGs) 9 (Industry, Innovation, and Infrastructure), 11 (Sustainable Cities and Communities), 12 (Responsible Consumption and Production), and 13 (Climate Action).
Zitieren
APA
G., S. K., Kumar, D. H., Ujwal, M. S., Mahesh, R., Ramaraju, H. K., & Pandit, P. (2026). Prediction of Mechanical Properties of Spent Coffee Ground Ash Mixed Asphalt Using Experimental Investigation and Machine Learning Approaches. Advances in Civil Engineering, 2026(1). https://doi.org/10.1155/adce/9281929
Vancouver
G. SK, Kumar DH, Ujwal MS, Mahesh R, Ramaraju HK, Pandit P. Prediction of Mechanical Properties of Spent Coffee Ground Ash Mixed Asphalt Using Experimental Investigation and Machine Learning Approaches. Advances in Civil Engineering. 2026;2026(1). doi:10.1155/adce/9281929
BibTeX
@article{G2026,
author = {G., Shiva Kumar and Kumar, Doma Hemanth and Ujwal, M. S. and Mahesh, Raghavendra and Ramaraju, H. K. and Pandit, Poornachandra},
title = {Prediction of Mechanical Properties of Spent Coffee Ground Ash Mixed Asphalt Using Experimental Investigation and Machine Learning Approaches},
journal = {Advances in Civil Engineering},
year = {2026},
volume = {2026},
number = {1},
doi = {10.1155/adce/9281929}
}
EUVIMED