Effect of polypropylene fibre, moisture content, relative density, and vertical stresses on mechanical properties of poorly graded sandy soil: A geotechnical and hybrid XGBoost-based investigation
Construction and Building Materials, cilt.493, 2025 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 493
- Basım Tarihi: 2025
- Doi Numarası: 10.1016/j.conbuildmat.2025.143286
- Dergi Adı: Construction and Building Materials
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, Aerospace Database, CAB Abstracts, Communication Abstracts, Compendex, INSPEC, Metadex, Veterinary Science Database, Civil Engineering Abstracts
- Anahtar Kelimeler: Extreme gradient boosting, High relative density, Polypropylene fibre, Role of vertical stress, Shear strength parameters, Tiki-Taka Optimization
- Süleyman Demirel Üniversitesi Adresli: Evet
Özet
This research presents the effect of polypropylene fibre, moisture content, relative density, and vertical stress on the mechanical properties, i.e., shear stress (SS), cohesion (C), and internal friction angle (ɸ), of reinforced soil specimens. For that purpose, the soil specimens were collected from the Chlef River in Algeria and classified as poorly graded sandy soil. The effect of moisture content (MC) and relative density (Dr) was analyzed, and it was observed that the SS of the soil specimen was decreased with increasing moisture content (2, 4, and 6 %) and relative density (Dr 40–80 %). Still, the soil specimen achieved higher SS at 300kPa vertical stress (VS). Conversely, the effect of fibre length and content was analyzed for soil specimens prepared with 55 % Dr and 3 % MC. The results demonstrated that the SS of poorly graded sandy soil increased with fibre parameters (i.e., length and content) and vertical stress. Moreover, a significant improvement in C and ɸ was observed at vertical stress of 100 and 200kPa. To analyze the effect of fibre parameters, MC, Dr, and VS on the shear strength, 78 soil specimens were tested in a laboratory, and a database was created. The database comprises fibre length & content, VS, Dr, MC, and SS of 78 soil specimens. For the first time, Arithmetic Optimization (AOA), Brainstorm Optimization (BOA), Quantum-inspired Evolutionary (QEA), Tiki-Taka Optimization (TTA), and Whale Optimization (WOA) algorithms-optimized extreme gradient boosting (XGBoost) models were employed, analyzed, and compared in predicting the SS of reinforced soil. The performance comparison, score analysis, accuracy metric, and reliability analysis introduced the XGBoost_TTA model as a robust and reliable hybrid XGBoost model to predict the SS. In addition, the multiple linear regression model demonstrated the role of VS in determining the SS of reinforced soil, with an agreement of 0.9589 between actual and predicted SS.