Expansive soils normally exhibit high shrinkage–swelling potential that threatens the stability of transportation and infrastructure systems. This study aimed to develop a multivariate predictive model for Plasticity Index (PI) reduction and to validate the stabilization mechanism through physicochemical and microstructural analyses of expansive soils treated with nano-lime (NK), nano-silica (NS), and chitosan biopolymer (B). Three expansive soil types classified as CH, CV, and CE were treated using NK (0% – 5%), NS (0% – 5%), and B (0% – 0.20%) and analyzed using Atterberg Limits, Methylene Blue Value (MBV), Cation Exchange Capacity (CEC), Scanning Electron Microscopy (SEM), ImageJ pore structure analysis, and Energy Dispersive Spectroscopy (EDS). The results showed substantial PI reductions, ranging from approximately 56–67% for CH soil, 43–53% for CV soil, and 39–51% for CE soil, with optimum compositions of 2% NK + 4% NS + 0.10% B for CH soil, 3% NK + 1% NS + 0.05% B for CV soil, and 4% NK + 3% NS + 0.05% B for CE soil. The multivariate regression model demonstrated high predictive performance with R² = 0.986, indicating strong correlations between liquid limit, additive dosage, and PI reduction. Validation using soils from five additional locations produced prediction errors below 20% for soils with PI values of 40 – 80. SEM, ImageJ, and EDS analyses confirmed pore refinement, denser particle bonding, and variations in Ca/Si ratios associated with the stabilization process. This study proposes an unprecedented integration of multivariate modeling with quantitative microstructural and physicochemical validation to develop a reliable predictive framework for nano-biopolymer stabilization of expansive soils.
Copyrights © 2026