Nowadays, Unmanned Aerial Vehicles (UAVs) increasingly support applications in surveillance, logistics, and environmental monitoring, yet their operational capability is fundamentally restricted by the limitations of onboard energy sources. The commonly used lithium-polymer (Li-Po) battery provides low energy density, voltage degradation under various loading conditions, and short endurance, typically limiting the flight duration to approximately 12 minutes. Hydrogen PEMFCs exhibit higher energy density, stable voltage outcomes, and sustainable operation, making them a suitable alternative for long-endurance operations. To examine this inherent property, a 300 W air-cooled PEMFC is integrated into a 450 mm quadcopter test platform and experimentally evaluated under controlled static (non-flight) bench-test conditions at different electrical loading levels, and its performance is compared to a conventional power source. During the bench-scale UAV evaluation, the regulated UAV power bus maintained a stable output voltage of approximately 12.5 V with a current below 4.6 A, demonstrating stable power delivery to the propulsion system. The fuel cell polarization curve shows that voltage reduces from 35.6 V to 32.2 V as current magnitude increases, while power output reaches approximately 14 W. In contrast, the Li-Po battery provides a 0.8 V voltage drop during operation and significantly shorter endurance. During 30-minute operation, the stack temperature increased from 25 °C to 40 °C, indicating stable thermal operation. Comprehensive analysis shows that fuel cell stacks provide better voltage stability, constant power delivery, comparatively stable electrical performance relative to the Li-Po battery, ensuring that it is suitable for potential for long-duration operation UAV operations and improved operational reliability.
Soil organic contamination, which causes sickness, arises from unintentional discharges of crude oil, pesticides, industrial effluents, and sewage, presenting a significant environmental issue due to the infiltration of contaminants into groundwater. One possible technique for cleaning up and removing pollutants from soil is electrokinetic remediation. This research involved an electrokinetic analysis of clay loam soil to examine the influence of electrolyte type, electrode configuration, voltage gradient, and pH on energy usage and the efficacy of phenol removal from contaminated soil. Nine experiments occurred following the optimal (custom) design, each spanning seven consecutive days. Each experiment entailed the treatment of one kilogram of industrially polluted soil with a phenol amount of 50 ppm. The findings indicate that pH, voltage gradient, and removal duration substantially influence phenol removal efficiency and energy usage. The phenol removal efficiency attained roughly 66.53%, with an energy expenditure of 0.6350 kWh /1kg of soil.
Accurate day-ahead load forecasting in commercial buildings is essential for demand-side management and energy flexibility. This paper proposes a knowledge-based feature engineering (KBFE) framework for day-ahead forecasting of office building electricity demand, in which solar geometry and irradiance decomposition are used to derive physically meaningful predictors from numerical weather prediction data, while the learning model itself remains purely datadriven. The framework is evaluated with extreme gradient boosting (XGBoost) on real metering data from an office building in northern Portugal and compared with artificial, recurrent, long short-term memory and gated recurrent neural networks, support vector regression, randomforests, and XGBoost models trained on historical load data. The KBFE model reduces the mean absolute percentage error by 36.3% relative to the mean of these baselines and by 33.3% relative to XGBoost, the strongest one, and is markedly more stable across seasons.
Early detection of defects in rotary machinery is critical to preventing system downtime and cost losses. This study proposes a hybrid deep learning model combining 1D CNN, Bidirectional LSTM, and Multi-Head Attention mechanism to classify healthy and fractured tooth conditions using vibration signals obtained from the SpectraQuest simulator. The model, which works directly on raw signals, eliminates manual feature extraction while achieving noise resilience through rich data enhancement techniques such as amplitude scaling and frequency masking. Training results show that training, validation and testing accuracy reached 99.92%, 97.00% and 97.13%, respectively. This approximately 2.8% generalization difference demonstrates that the model avoids overlearning and adapts to varying operating conditions in industrial environments. These achievements have been supported by 5-fold cross-validation and are presented in detail and comparatively through durability tests performed at different noise levels (5-20 dB SNR) and ablation studies revealing the contribution of architectural components.
Continuous and impulsive occupational noise constitutes a relevant physical risk in the Armed Forces, with potential implications beyond clinical auditory damage. This systematic review with narrative synthesis, guided by PRISMA 2020, evaluated the impact of military noise exposure on hearing health and its implications for functional availability. Searches in Scopus, Web of Science, and PubMed (2019–2026) identified 29 eligible articles, with methodological quality assessed through an adapted CASP matrix. The evidence suggests that noise-induced hearing loss and tinnitus may affect speech perception, sleep quality, cognition, communication, and situational awareness. Although functional availability was rarely assessed as a primary outcome, the findings support a cautious functional interpretation of military noise exposure
