Rajasthan Technical University

This study presents a novel, integrated physics-informed machine learning framework that simultaneously addresses UCS prediction, uncertainty quantification, and mix design optimization for chemically stabilized soils, combining a rigorously curated 990-mixture dataset, a Physics-Informed Neural Network (PINN) embedding geochemical monotonicity constraints, Monte Carlo-based reliability quantification, and a multiobjective Pareto optimization pipeline within a single unified framework. Various machine learning models including linear regression, random forests, support vector regression, gradient boosting, and artificial neural networks (ANN) were tested. Out of the models tested, the ANN produced the best results with an R2 value of 0.919 and RMSE of 1.124 MPa. A Physics-Informed Neural Network (PINN) was further developed to enforce physically consistent trends, providing improved interpretability despite marginally reduced accuracy. Reliability-based UCS estimation was possible as predictive uncertainty was quantified using Monte Carlo simulation. Along with defined geo-chemistry principles, SHAP analysis identified the three most important controlling factors to be curing time, GGBS content, and alkaline activators. A multiobjective optimization framework revealed the possible design of economically viable and environmentally friendly stabilization mixes. The developed framework represents a real-world decision-making support system for chemically stabilized soil systems.

Granular piles are widely used for ground improvement in soft soils but often suffer from bulging failure in their upper sections under load. This study presents a comprehensive numerical investigation into the effectiveness of combined mitigation strategies, top stiffening of the pile and geosynthetic reinforcement within the granular pad to control bulging and enhance load-settlement performance. Ordinary Granular Piles with diameters of 80 mm, 100 mm, and 120 mm, and length-to-diameter ratios of 4, 5, and 6, were analyzed under a prescribed settlement of 50 mm. Results show that bulging is most severe within the top 4D zone and increases with pile diameter and L/D ratio. Introduction of stiffening in the upper pile segment (1.5D, 2D, and 3D) significantly reduced bulging, with up to 22.96

In this paper, an Adaptive Hybrid Compressive Sensing framework with Hierarchical Optimized Multi-Domain Security (AHC–HOMS) is proposed for secure and efficient image transmission. The proposed approach jointly integrates signal sparsification, adaptive compression, and multi-layer encryption within a unified pipeline. A hybrid DWT–DCT sparsifying transform is employed to enhance signal sparsity across spatial and frequency domains, while an adaptive block-wise compressive sensing strategy dynamically allocates measurements based on local texture complexity to balance compression efficiency and reconstruction fidelity. To ensure security, a hierarchical dual-layer encryption mechanism is introduced, where the most significant transform coefficients are protected using AES and the compressed measurements are selectively scrambled using chaos-seeded random mapping. Extensive experiments conducted on multiple benchmark images with varying structural and textural characteristics demonstrate the robustness and effectiveness of the proposed framework. For legitimate users, AHC–HOMS achieves PSNR values in the range of 24.08–29.43 dB, SSIM values up to 0.84, and consistently low reconstruction error, indicating reliable perceptual and structural recovery. In contrast, unauthorized reconstructions exhibit severe degradation, with PSNR values close to 10 dB and SSIM well below 0.2, rendering the recovered content visually meaningless. Security analysis further confirms strong resistance against differential attacks, with NPCR values exceeding 98

The present work introduces an accurate computational model by comparing Artificial Bee Colony (ABC), Genetic (GA), Grey Wolf (GWO), Harris Hawks (HHO), Particle Swarm (PSO), and Salp Swarm (SSO) optimized least squares support vector machine (LSSVM) models to estimate the bearing capacity of reinforced concrete (RC) piles (PU). Based on a substantial dataset, key geotechnical and structural parameters were identified as both highly influential and multicollinear. The Particle Swarm Optimization-based model demonstrated superior performance (variance accounted for = 95.49, root mean square error = 70.801, correlation = 0.9858, and mean absolute error = 51.521) compared to other hybrid variants, demonstrating enhanced robustness and generalization. Analyses revealed that most alternative models exhibited either underfitting or overfitting, primarily due to problematic multicollinearity among critical input features. This research presents a novel analysis of how feature multicollinearity affects model-fitting behavior in geotechnical predictive modeling.

Rising international power requirements has enabled a sudden need of renewable energy-waste heat recovery solutions demand efficient thermodynamic models competent of binding solar and thermal energy successfully. This study proposes a novel integrated heliostat-based solar thermal power generation system coupled with an absorption refrigeration cycle, employing high initial heat source temperature to enhance overall performance. A comprehensive thermodynamic assessment is implemented to assess system behavior under varying irradiation levels, evaporator temperatures, pump work, and absorption characteristics. A priority-weighted machine learning method is employed to identify the dominant contributors to system losses and to highlight critical operating parameters. Results show that the central receiver and heliostat account for 31.37 and 27.40