This paper gives an overview of several studies that the authors have been involved in the field of physical modelings and associated finite element analyses for geosynthetic-reinforced soil retaining walls under static and earthquake loadings. Full-scale and centrifuge model tests were considered as benchmarks for verifying the numerical procedures. The use of simple nonlinear elastic models and development of advanced plasticity models of different complexities in simulating the monotonic and subsequently cyclic loading behavior of soils and geosynthetic materials was illustrated. The calibrations of the models and details of finite element analysis were presented. The responses, such as accelerations, wall deformations and geogrid tensile forces obtained from the analyses were compared with the experimental results. The analyses were able to reproduce many aspects of the behavior of geosynthetic-reinforced soil retaining walls under cyclic loading, as validated through full-scale shaking table tests. The study showed that existing design specifications (L/H = 0.75) are conservative when reinforced soil retaining walls with block facings were subjected to very strong earthquakes.
This study developed an integrated machine learning framework for predicting the factor of safety (FOS) of soil slopes using Extreme Gradient Boosting (XGB) optimized with three metaheuristic algorithms: Particle Swarm Optimization (PSO), Tiki Taka Optimization Algorithm (TTOA), and Whale Optimization Algorithm (WOA). A dataset comprising six influential parameters, namely cohesion, unit weight, friction angle, slope angle, slope height, and pore pressure ratio, was employed for model development. The dataset was normalized and divided into training and independent testing subsets, while five-fold cross validation was adopted during model optimization to improve robustness and reduce overfitting. Model performance was assessed using multiple statistical indicators, scatter plots, Regression Error Characteristic curves, and Taylor diagrams. The results demonstrate that all optimized XGB models effectively captured the nonlinear relationship between the input parameters and FOS, while the proposed framework provides an accurate, interpretable, and computationally efficient tool. To improve model transparency, SHAP, LIME, and permutation feature importance analyses were incorporated. The explainability results identified slope angle, cohesion, and slope height as the dominant predictors, consistent with their governing roles in slope geometry and shear strength. The proposed framework provides an accurate, interpretable, and computationally efficient tool for data driven slope stability assessment.
This study investigated the characteristics of construction and demolition (CDW) waste and excavated soils to evaluate their recovery potential. CDW waste was sorted into granular fractions of 0/D, 0/16 mm, 16/40 mm and 40/80 mm. The material was classified into recycled aggregates, recycled sand, and soils. Testing of materials showed that the fines content of excavated soils ranges from 4
The present study presents a comparative machine learning (ML) framework for assessing shear wave velocity (Vs) utilizing five geotechnical parameters obtained from field investigations in Jamshedpur, India. A dataset comprising 99 observations was developed using Standard Penetration Test N-value, depth of investigation starta, bulk density, groundwater-table depth, and engineering soil description as predictors. Multiple Linear Regression (MLR), Support Vector Regression (SVR), Artificial Neural Network (ANN), Gradient Boosting Regression (GBR), and Random Forest Regression (RFR) models were developed and evaluated using an 80:20 train-test split, 10-fold cross-validation, and 30 repeated random train-test splits. Based on the performance, RFR showed the highest average predictive performance (test R² = 0.65 ± 0.12; RMSE = 24.0 ± 4.2 m/s; MAE = 18.9 ± 3.1 m/s across repeated splits), while MLR and SVR (R² ≈ 0.63 to 0.64) showed comparable performance. Feature-importance analysis showed that depth and SPT-N together accounting for 68 to 91
An engineering-zoning framework was developed for the continuous permafrost zone of Russia to express slope-insulation demand as the minimum required thermal resistance (Rins) of embankments built under the frozen construction principle. Complete meteorological records from 137 stations for 1995–2024 were combined with a Stefan-based analytical thaw-depth model, and Rins was computed iteratively for a reference embankment geometry under the condition that thaw depth did not exceed embankment height. The territory was classified into six engineering zones, and maps were produced for six consecutive five-year sub-periods, for 30-year mean conditions, and for worst-case conditions. A pronounced north–south gradient in insulation demand was identified, together with clear temporal non-stationarity: the combined area of the two highest insulation classes increased from 17.6
