Multi-objective optimisation and synergistic mechanisms of expansive soil improvement using organic fertiliser, slow-release fertiliser, and rice straw

This study systematically investigated the synergistic improvement ofexpansive soil using organic fertiliser (OF), slow-release fertiliser (SRF), and rice straw (RS) through Box-Behnken design (BBD) and response surface methodology (RSM). Key findings include: the quadratic models demonstrated high statistical significance (root density: R² = 0.765, F = 25.84; shear strength: R² = 0.885, F = 18.65; swelling rate: R² = 0.20, F = 15.23; all P< 0.001) with low prediction errors (root content: ± 0.08mg/cm3; shear strength: ± 0.58 kPa; swelling rate: ± 0.38%); The combination of12.30% OF + 0.7kg/m3 SRF + 0.4% RS achieved 58% improvement inshear strength, 32% improvement inroot content, 42.7% reduction inswelling rate; OF exhibited negative linear effects onroot density (β = – 0.18, P= 0.002) with >10% dosage reducing root growth by9.0%; SRF showed positive linear impacts onshear strength (β = +0.25, P= 0.001) and root density (β = + 0.12, P= 0.023); RS enhanced shear strength below 0.5% (β = + 0.08, P= 0.042) but impaired root density due topore clogging (β = – 0.15, P = 0.008). The optimised formulation, validated bytriplicate centre-point tests (coefficient ofvariation ≤ 2.1%), isrecommended for slope stabilisation while limiting OF to≤ 10% toprevent performance degradation. This data-driven approach provides actionable insights for balancing agricultural waste utilisation and geotechnical performance inexpansive soil improvement.