This paper analyzes CNNs and spatiotemporal transformers to predict future oceanic circulation patterns and examine meridional circulation in the North Atlantic Gyre, aiming to improve existing models and provide a new tool for analysis.
This study introduces a novel Genetic Algorithm (GA) designed to optimize artificial lighting in vertical farms to enhance Light Use Efficiency (LUE). The proposed GA seeks to identify the optimal spectral composition of Red, Green, and Blue (RGB) LEDs, aiming to maximize crop productivity by evaluating characteristics such as height, width, fresh weight, and leaf count. The algorithm operates through ten stages, including initialization of a population, actuation of RGB values, fitness evaluation, and iterative processes of selection, crossover, mutation, and validation. By comparing RGB treatments with a reference cold white light treatment, the algorithm refines lighting conditions to improve crop performance at different growth stages. Detailed methodologies for fitness evaluation, crossover, mutation, and validation are provided, highlighting the practical steps for implementing this approach in vertical farming environments. This research aims to contribute to more energy-efficient and productive vertical farming practices, supporting the broader goal of sustainable agricultural development.
The integration of renewable energy sources (RES) is essential for sustainable energy management systems (EMSs) in residential areas. However, the adoption of traditional EMSs remains constrained due to inefficiencies and limited adaptability to varying energy demands. This study presents a Solar PV/battery/EV-based EMS for home load that enhances energy efficiency, adaptability, and cost-effectiveness. The system converts solar energy into DC power using photovoltaic panels, which are then stored in a battery bank. An intelligent controller optimizes energy distribution by prioritizing essential loads and reducing reliance on grid power. The proposed advanced EMS model, developed using MATLAB Simulink optimization, demonstrates an energy efficiency ranging from 85% to 90%, resulting in expected energy savings of around 80% over traditional Home Energy Management System (HEMS) and improved user convenience by automating energy distribution. The developed model assumes that a standard residential load of around 10 kWh can be sustainably managed, ensuring uninterrupted power supply during peak hours and minimizing grid dependency.
Prediction of cryptocurrency prices relatively
accurate remains a formidable challenge due to inherent
volatility associated with it and the absence of traditional
valuation metrics. This research explores the performance of
Long Short-Term Memory (LSTM), Support Vector Machine
(SVM), and a hybrid model of LSTM+SVM for this complex
task. LSTM has demonstrated potential in capturing short-term
price fluctuations, while the hybrid model aims to combine the
strengths of temporal dependencies of LSTM and pattern
recognition of SVM. To evaluate the models’ performance,
comprehensive evaluation framework has been employed,
considering generalization ability of the models, robustness,
computational efficiency, and interpretability. Historical daily
price data for five leading cryptocurrencies, Ethereum, Solana,
BNB, Tether, and Bitcoin was collected from 2020 to 2024.
This data was used to evaluate the model performance of
LSTM, SVM, and a hybrid model of them, using metrics such
as R-Square, Root Mean Square Error (RMSE), and Mean
Absolute Error (MAE). The findings from the study indicate
that LSTM generally outperformed both SVM and the hybrid
model in terms of these evaluation metrics. Moreover, the
hybrid model demonstrated competitive performance,
particularly when considering its statistical significance and
ability to generalize across different volatile conditions. While
SVM model has potential, it requires meticulous
hyperparameter tuning and feature engineering to reach optimal
performance. This research offers a comparative analysis of
machine learning models for cryptocurrency price forecasting,
detailing the strengths and limitations of LSTM, SVM, and
hybrid approaches. The insights provided are valuable for both
practitioners and researchers. Future studies could explore more
advanced hybrid architectures considering different algorithms,
incorporate additional data sources, and assess how varying
market conditions may affect model performance.
data characterization in federated learning
The evolution of smart grid technology necessitates sophisticated methods for fault detection to ensure system reliability and efficiency. Monitoring a complex power grid with phasor measurement units (PMUs) continuously transmitting data at high velocities. Rapidly and accurately analyzing this extensive data to detect faults presents a major challenge for grid operators. This study introduces a novel approach for fault classification in smart grids by utilizing Convolutional Neural Networks (CNNs) based on the architectures of VGG16 and VGG19. VGG is capable of rapidly classifying various grid events, such as faults, generation losses, and synchronous motor switching, with high efficiency. The system detects faults swiftly, allowing operators to minimize downtime and prevent significant damage by enabling prompt responses. The study meticulously examines the performance metrics of each model, including accuracy, precision, recall, and F1 score. Evaluations reveal that the VGG16 model outperforms VGG19, achieving an impressive accuracy of 98.75% and consistent precision, recall, and F1 scores of 0.99. In contrast, the VGG19 model attained a lower accuracy of 95.00%, with slightly diminished performance metrics. These findings highlight the efficacy of advanced deep learning techniques in improving fault detection accuracy within smart grid systems, suggesting that VGG16 offers a more reliable and accurate solution compared to VGG19.
1. The authors proposed the design of an MV microgrid with traditional DG and a considerable PV plant control WECC model.
2. The proposed system has a better controller time constant, which can guarantee the effectiveness and robustness of the system.
3. The voltage profile of critical buses is improved, which results in a 1% steady-state error.
This paper aims to contribute to the body of knowledge towards finding alternative solutions for phishing detection by developing a novel approach to convert phishing URLs to images, using text-to-image generation methods, and demonstrate the applicability of CNNs on images generated from URLs, as an effective tool to classify phishing websites.
This paper investigates the critical challenge of detecting prohibited items in student dormitories, and we proposed a new deep-learning model to detect prohibited items automatically. To address the lack of an existing dataset for this task, we developed a new dataset, PISD (Prohibited Items in Student Dormitories). Our model achieved competitive performance, with the lowest GFLOPS and inference time, the highest FPS, and strong results in terms of precision, and recall highlighting its efficiency and effectiveness. This demonstrates the model’s capability to reliably detect and classify prohibited items in student dormitory environments.
The compact voltage doubler rectifier incorporates a rectangular square wave-shaped DC filter to smooth the rectified signal. With dimensions of 33 × 13 mm, it is well-suited for wearable biomedical devices. The rectifier demonstrates high sensitivity at low input power levels, effectively balancing size and power conversion efficiency (PCE), making it ideal for WBDs where both factors are crucial. This work contributes to the advancement of efficient RF energy harvesting (RFEH) solutions for wearable biomedical applications.