Sambhram Institute of Technology

Fused Deposition Modelling (FDM) has emerged as a transformative additive manufacturing technology enabling rapid fabrication of complex thermoplastic components, yet systematic understanding of how infill density simultaneously influences mechanical durability and tribological performance remains incomplete. This study presents a comprehensive comparative investigation of FDM-fabricated PLA and ABS components through integrated mechanical, tribological, and microstructural characterization across the complete volumetric infill density spectrum 0%, 25%, 50%, 75%, and 100%. The research establishes quantitative material-process-property relationships to support rational material selection and infill optimization. Experimental evaluation encompassed tensile testing, compressive testing, hardness measurement, pin-on-disc tribological analysis under dry sliding conditions, and scanning electron microscopy characterization of microstructural evolution. Results demonstrated that PLA exhibited superior tensile performance with strength increasing 150% from 20.3 +/- 1.1 MPa at 0% infill to 49.7 +/- 1.8 MPa at 100% infill, alongside exceptional tribological properties characterized by specific wear rate reduction from 2.24 & times; 10-4 to 1.60 & times; 10-5 mm3/Nm and friction coefficient decreasing from 0.65 to 0.28. Conversely, ABS demonstrated modest tensile gains 18.2 to 27.3 MPa, 50% improvement but excelled in compressive loading at intermediate densities, achieving 62.3-67.8 MPa at 50-75% infill 15-22% superior to PLA attributed to its amorphous structure enabling extensive plastic deformation and energy absorption through progressive cellular densification mechanisms. Statistical analysis ANOVA with Tukey HSD post-hoc testing confirmed significant material-infill interactions (p < 0.001) for all measured properties, establishing that optimal performance requires material-specific infill selection. The 50-75% infill range emerged as a cost-effective solution delivering 70-85% of maximum mechanical properties while achieving 25-35% material savings compared to full-density configurations.

Aluminum alloys are widely used in aerospace structures, necessitating continuous improvement in their mechanical properties. Micro alloying with nickel can enhance these properties and improve the coefficient of thermal expansion. This study investigates the influence of nickel on the microstructural and thermal expansion characteristics of AA2024 and AA7175 aluminum alloys, both commonly used in aerospace. AA2024 primarily contains copper, while AA7175 has zinc; both are heat-treatable and possess excellent strength. In this research, alloys were stir-cast with varying percentages of nickel and 0.2% strontium, which improves grain structure. The alloys were homogenized at 480°C for 15 hours, quenched in water, and subjected to tensile testing, EDS, XRD, and microstructural and thermal expansion analyses. Results showed that nickel addition increased strength to 215 MPa for AA2024 and 284 MPa for AA7175 with 5% nickel. XRD and EDS revealed the formation of intermetallic compounds like Al2Ni3, Al3Ni, and Al3NiCu. However, increasing nickel beyond 5% led to undesirable needle-like structures. Thermal expansion studies indicated a reduction in the coefficient of linear thermal expansion by 14.1% for AA2024 and 16.5% for AA7175 with 2% nickel, reducing thermal stress under loading up to 350°C.

Our research work extends the application of ensemble learning models for crop yield prediction by incorporating interpretability techniques to explain stacked regression models. Building on top of the prior work focused on predicting yields in the Davangere district using ensemble methods like Ridge Regression, Random Forest, XGBoost, CatBoost, LightGBM, and Gradient Boosting, our work emphasizes model transparency and explainability. While ensemble methods excel in predictive accuracy, their complexity often obscures the reasoning behind predictions, limiting their practical adoption by stakeholders in agriculture. To address this challenge, Local Interpretable Model-Agnostic Explanations (LIME) and SHapley Additive explanations (SHAP) were applied to a stacked regression framework combining the strengths of multiple ensemble models. These interpretability techniques were used to identify key features influencing crop yields, such as soil health, weather conditions, and crop type. The stacked regression model achieved an impressive accuracy of 99.64

Tussar and bivoltine silk fibres are from different silkworm species and possess different physicochemical properties because of the differences in their chemical composition and molecular organization. In this study, a comparative analysis of tussar and bivoltine silk fibres was conducted to study the effect of chemical composition on crystalline structure and surface morphology. The crystallinity and β-sheet structure identification were evaluated by X-ray diffraction (XRD), the chemical bonding and functional groups were determined by Fourier Transform Infrared (FTIR) spectroscopy, and the surface morphology was observed by Scanning Electron Microscopy (SEM). XRD patterns of the fibres exhibited characteristic peaks corresponding to silk fibroin. Tussar silk showed relatively broader peaks indicating lower crystallinity while bivoltine silk showed sharper and more intense peaks corresponding to higher β-sheet.

Smallholder agricultural systems face growing threats to their productivity because of climate changes, soil erosion, unpredictable rainfall patterns, and restricted availability of local advisory services. The existing agricultural artificial intelligence solutions need substantial computing power and permanent internet access which makes them unsuitable for use in remote agricultural areas. The research introduces AgriSense which functions as a lightweight artificial intelligence system that provides crop yield prediction and farmer advisory services through its explainable design. The system uses a machine learning framework that combines meteorological data with soil information and satellite NDVI satellite data. The research team developed the model using historical agricultural data NASA POWER weather data SoilGrids soil data and satellite vegetation information. The research team analyzed four machine learning algorithms which included Linear Regression Decision Tree Random Forest and XGBoost by using three performance indicators RMSE MAE and R2. The research results show that using both soil and NDVI data for forecasting weather conditions enables better prediction results when compared to using only weather data. XGBoost produced its highest results through an RMSE measurement of 7.8 an MAE calculation of 5.9 and an R2 value of 0.89. The research team confirmed that their model maintained reliable performance across diverse farming ecosystems by testing it in more than thirty agricultural regions. The system includes a mobile advisory application which functions offline and provides support to users in rural areas with limited internet access.