
Garment industries in Bangladesh are one of the key sectors that account for 83.9 percent of all exports, employing more than 4 million people. This vital sector often faces power shortages that reduce productivity and hinder economic activity. This crisis is largely driven by dependence on imported fossil fuels, highlighting the need for reliable and sustainable power sources. Therefore, the prime purpose of this research is to evaluate the energy performance and techno-economic feasibility of an industrial rooftop photovoltaic (PV) system using performance indicators obtained from the PVsyst simulation tool. The suggested system has 1007 monocrystalline PV modules and four grid-tied inverters. It can make 713 MWh/year with a performance ratio of 77.85%. This PVsyst-based performance analysis demonstrates an approximately 503.5 kWp rooftop PV system that can be implemented on the roof area, with an estimated levelized cost of energy (LCOE) of US$ 0.05 per kilowatt-hour (kWh). This means the investment will pay off in 9.9 years and yield an estimated lifetime ROI of 123% over the 25-year project period. The environmental assessment shows that the system will cut CO2 emissions by a total of 8348.35 tons over its 25-year lifetime. The proposed rooftop PV system can help mitigate the ongoing energy crisis and improve energy reliability, thereby supporting productivity in Bangladesh's garment sector and in other developing economies.
The increasing penetration of inverter-based renewable energy resources introduces complex voltage, frequency, and power-dynamic interactions that challenge conventional smart-grid stability assessment methods. This study proposes a Physics-Informed Graph-Attention Feature Fusion Network (PI-GAFFN) for topology-aware, interpretable, and multi-task smart-grid stability assessment. The investigation is organized into two complementary tracks. First, a detailed renewable-integrated electrical system comprising a 100 MWp photovoltaic plant, a 60 MWh battery energy storage system, a virtual-synchronous-machine-based grid-forming converter, medium-voltage collection networks, and a 230 kV grid interface is developed in MATLAB/Simulink–Simscape Electrical. Its dynamic behavior is examined under renewable-power variation, load switching, grid outage, temporary three-phase fault, and reactive-power-support conditions. Second, the UCI Electrical Grid Stability Simulated Data benchmark is transformed into a four-node physics-consistent graph and used independently for PI-GAFFN training, validation, and testing. The proposed network integrates physics-aware node encoding, physics-guided graph attention, feature fusion, and a multi-task objective for simultaneous continuous stability-index regression and stable/unstable classification. PI-GAFFN achieves 99.50% accuracy, 99.48% precision, 99.74% recall, 99.61% F1-score, and 99.98% ROC–AUC, outperforming LR, SVM, RF, XGBoost, CNN, BiLSTM, and CNN–BiLSTM under identical test conditions. For continuous stability estimation, the clean-data evaluation yields an RMSE of 0.012 and an (R2) of 0.9987. Robustness tests retain 93.00% accuracy under 10% Gaussian noise and 92.80% under 10% false-data injection, while the model requires approximately 0.07 ms per sample for inference. Attention and ablation analyses further demonstrate the contribution of topology-aware interactions and physics-informed feature learning. The results indicate that PI-GAFFN provides a promising physics-guided framework for accurate, robust, and interpretable stability assessment of renewable-integrated smart grids.
The need for secure, sustainable, long-lasting, and high-performing electrochemical energy systems has grown rapidly due to advances in distributed storage, renewable energy integration, and electric mobility. To meet these objectives, sophisticated operando monitoring instruments have been rapidly developed in recent years. These instruments identify localized mechanical, chemical, and thermal processes that govern battery performance, longevity, and health. Conventional electrical and electrochemical sensors degrade quickly and struggle to operate in harsh environments, and embedding them in small, compact next-generation batteries is challenging. Optical fiber sensors (OFSs) have recently emerged as a transformative platform for real-time, non-invasive battery diagnostics. By harnessing light–matter interactions, OFSs transduce local variations in temperature, strain, refractive index, or chemical composition into optical signals with exceptional sensitivity and resolution. Recent advances in fiber Bragg gratings (FBGs), interferometric architectures, and distributed spectroscopic techniques, such as optical frequency/time-domain reflectometry (OFDR/OTDR) and φ-OTDR-based sensors, have demonstrated multi-parameter sensing in battery monitoring systems. These approaches enable spatially resolved measurements of temperature, pressure, strain, and electrochemical parameters, which are critical for early detection of degradation and thermal runaway.With intrinsic advantages such as electromagnetic immunity, chemical robustness, miniaturization, and multiplexing capability, OFSs represent a compelling and increasingly validated platform for battery diagnostics. This review consolidates recent advances in operando optical sensing, emphasizing transduction mechanisms, structural designs, and integration strategies. It further discusses key challenges in sensor durability, calibration, sensor intrusiveness and scalability, and outlines future directions toward photonic battery management systems with predictive diagnostic capability, integrating optical diagnostics with predictive control for next-generation energy storage technologies.
Lithium-ion batteries (LiBs) are widely recognized for their high energy density, long cycle life, and favourable electrochemical performance, making them a fundamental technology for applications such as electric transportation, renewable energy integration, and grid-scale energy storage. In renewable residential systems, they store excess energy from solar panels or wind turbines, ensuring stable power supply. However, their safe operation is fundamental, as they are susceptible to thermal runaway and mechanical deformation, which can lead to severe failures. Monitoring temperature and strain is essential to ensure battery safety. This study presents a novel approach for continuous and prolonged safety monitoring of LiBs using Fiber Bragg Grating (FBG) sensors, which offer high metrological qualities and the ability to measure temperature and strain simultaneously. Residual strains and changes in the baseline temperature may indeed be the cause of unexpected issues and therefore be considered as warning parameters. An automated system was developed to operate in remote control to charge/discharge the battery according to desired protocols. Eight FBG sensors were mounted on a polymer lithium-ion battery shell. Four of them were pre-tensioned and glued to measure cell’s temperature and deformation, while the other four were applied only with heat-conducting paste to capture temperature variations. Repeated charge–discharge cycles at different rates (0.5ℂ and 1ℂ) were performed on a lithium-ion system. FBG sensors revealed higher mechanical deformations along the x-direction (∼152 μɛ) than along the y-direction (∼42 μɛ). A 24-hour domestic grid scenario was then simulated for 4 consecutive repetitions, replicating typical household power demand with integrated renewable energy. Results show a progressive reduction of the maximum strain values the cycles, suggesting a stable mechanical behaviour of the POLI cells under the investigated operating conditions. FBG sensors are effective in detecting both strain and temperature variations, indicating that also mechanical deformations could potentially be used to improve state-of-charge (SoC) estimation.
Power transformers play a key role in the generation, transmission, and distribution of electrical energy. The stability and continuity of electrical service depend on their efficient operation. During operation, transformers are subjected to several types of faults, which if not corrected, can lead to premature failure. This study presents a structured pipeline for power transformer fault diagnosis based on the Voltage-Current Technique and machine learning, including a bayesian model selector that adaptively assigns input data to the optimal classifier according to the environmental noise levels. Three types of faults are simulated: turn-to-turn short circuit, buckling stress, and axial displacement. The diagnosis pipeline integrates transformer modelling, data collection, feature engineering, data transformation, classification using multiple machine learning algorithms, and a model selector. The proposed model provides accurate results in both clean and highly noisy environments, with accuracies ranging from 70% in very noisy conditions (7 dB) to 89% in clean conditions.