Loading...
Development of data-driven framework for the geotechnical behavior of xanthan gum-treated clay reinforced with polypropylene fibers
Author(s)
Onyelowe, Kennedy C.
Baldovino, Jair De Jesús Arrieta
Kumar, S. Anandha
Ebid, Ahmed M.
Hanandeh, Shadi
Arunachalam, Krishna Prakash
Date Issued
January 1, 2026
Type
Article
Abstract
This study develops a robust data-driven modeling framework to predict the unconfined compressive strength (UCS) and stiffness (Go) of low-plasticity clay soils stabilized with Xanthan Gum (XG) and Polypropylene Fiber (PPF), aiming to advance sustainable geotechnical design. A total of 108 soil specimens were prepared with varying XG dosages and cured over different periods, and predictive models were constructed using a Decision Table algorithm optimized with six bio-inspired optimization techniques. Among these, the Firefly-optimized model consistently provided the highest accuracy, demonstrating reliable agreement between predicted and measured values. Sensitivity analysis identified XG dosage, curing time, and dry density as the most influential factors governing UCS and Go. These findings highlight the strong potential of the proposed machine learning framework to guide field engineers in optimizing mix design parameters for improved mechanical behavior of bio-treated soils, reducing reliance on time-consuming and costly laboratory tests while promoting environmentally sustainable foundation practices. The need for this study arises from the growing demand for green soil stabilization techniques that minimize the use of cement and lime while still ensuring reliable performance in construction. Its applicability extends to real-world geotechnical projects such as embankments, road subgrades, and shallow foundations, where predictive modeling can significantly streamline design decisions and improve long-term sustainability.
Citation
Composites and Advanced Materials, 35, 1-31, 2026