Arabian Journal for Science and Engineering最新文献

In this study, the synthesis of a novel magnetic Fe₃O₄/ZnO composite embedded in porous carbon matrix, namely Fe₃O₄/ZnO/C was conducted. Fe₃O₄/ZnO/C (FZC) was synthesized by a simple method by the calcination of mixed FeCl₂ and ZIF-8 at 600 °C under N₂ flow. It was found that ZFC had a porous structure owning a reasonable surface area of 184 m² g⁻¹. The adsorption performance of the material was demonstrated through the elimination of methyl red (MR) in water. The MR adsorption process was investigated under various conditions, reaching the highest removal efficiency to 95% within 180 min at pH 7, an initial MR solution concentration of 10 mg L⁻¹, and an adsorbent dosage of 0.2 g L⁻¹. Furthermore, the Elovich kinetic model and the Bangham isotherm models were found suitable for describing the MR adsorption on this material. The Q_max value was 88.8 mg CR per gram of FZC adsorbent. The proposed adsorption mechanisms were π–π interaction, n–π interaction, electrostatic interaction and hydrogen bonding. With three cycles of reuse and maintaining good stability, and adsorption capacity, FZC may be highly appropriate for a promising application in practical dye treatment systems.

Additive manufacturing methods are increasingly popular for producing parts with complex geometries due to advantages such as cost reduction and shorter production times. Among these methods, material extrusion is widely used in automotive, aerospace, and biomedical industries for its fast, precise, and low-cost production capabilities. However, the strength of 3D-printed parts remains a critical challenge, primarily due to weak interlayer adhesion, which significantly affects mechanical performance. Previous studies have extensively analyzed factors such as design parameters, adhesive properties, adhesive thickness, and part geometry on joint strength. This study investigates the effects of cohesive zone model (CZM) parameters on the interlayer adhesion strength of single-lap joints produced with a 3D printer. Simulations were conducted using the CZM in Abaqus finite element software and were experimentally validated. The maximum force in the simulation results was obtained at only a 1.35% error rate (1200.1 N in the simulation, 1184.16 N in the experiments). Taguchi analysis was employed to determine the behavior of design factors with a minimal number of simulations. Cohesive stiffness (K), damage initiation (σ_f), and fracture energy (G_IIc) were selected as design factors, while maximum force and displacement served as output parameters. Analysis of variance (ANOVA) was used to determine the effect ratio of these design factors on the output parameters. Additionally, the influence of different damage element types, damage stabilization, and fracture energy on the force–displacement behavior of the material was investigated. The results showed that maximum force and displacement increased with higher damage initiation and fracture energy, while cohesive stiffness had a variable effect. Moreover, damage initiation, cohesive stiffness, and fracture energy were ranked in order of their impact on maximum force and displacement.

Reducing NOx emissions from fuels is critical to reducing air pollution and promoting environmental sustainability, particularly in the maritime industry. This study investigates the effectiveness of NO absorption in a hydrogen peroxide solution as a potential method for reducing NO emissions in marine exhaust systems. Numerical analyses were performed using the computational fluid dynamics (CFD) method to simulate the chemical interaction between exhaust gases and H₂O₂ under varying conditions. In this context, the study analyzed the effect of varying exhaust gas velocity, temperature and NO mass fraction in the ranges of 0.01–10 m/s, 288–573 °C and 0.001–0.5, respectively, on NO absorption efficiency and nitric acid formation. The results indicate that higher temperatures enhance NO conversion efficiency, reaching its peak at 573 °C due to accelerated reaction kinetics. However, increasing exhaust gas velocity negatively affects NO absorption, reducing its overall effectiveness. These findings provide valuable insights for optimizing NOx removal processes in marine applications and contribute to the development of strategies compliant with IMO Tier III emission regulations.

Semantic segmentation of the hippocampus aims to enhance the efficiency and accuracy of medical image analysis by precisely segmenting the hippocampus, thereby assisting doctors in making more accurate diagnoses and developing effective treatment plans. However, in 3D medical image segmentation tasks realized through downsampling using transformers within conventional U-shaped networks, the sequence-to-sequence prediction problem often neglects the connection between channel information and spatial information, leading to partial loss of channel information. To address this issue, we propose a network for hippocampal 3D image segmentation based on a hybrid attention mechanism with cross-dimensional interactions, termed CDI-Unet. The cross-dimensional interaction block (CDIB) is first introduced to address the separation of channel attention and spatial attention by capturing both spatial and channel dimensions of the input tensor, thereby enabling more comprehensive extraction of key segmentation semantic information from 3D medical images. The multilayer fusion block (MLFB) is then employed to replace the jump connections in the Unet in order to tackle the problems of category imbalance and feature loss. Additionally, a new space and channel reconstruction convolution block (SCRC) is designed to eliminate the substantial redundancy that arises when CNNs extract features, thus reducing computational load and redundant features. Experimental results on current mainstream datasets demonstrate that our CDI-Unet outperforms existing methods in all metrics, achieving better scores and performance.

In the space of Internet filtering, Saudi Arabia has been opening its digital borders in a deliberate new era toward openness. Internet filtering is routinely used to restrict access to websites and services that promote content that is deemed inappropriate with respect to governing laws, values, or policies. Here, we present a comprehensive longitudinal study of digital filtering in Saudi Arabia over a period of seven years. This study aims to comprehensively assess the evolution and impact of digital filtering focusing on both mobile apps and website access. We monitor access to: (a) 18 social media and communications mobile apps such as WhatsApp, Facetime, and Skype; and (b) Alexa’s top 500 websites in 18 different categories. In addition, we investigate and characterize the technical mechanisms and the network topology used in the implementation of the filtering. Furthermore, we conduct measurements from multiple vantage points covering the three largest telecommunications companies and five major cities in Saudi. Our results show that Saudi has indeed made significant progress toward opening its digital borders. For example, Internet filtering decreased by 3.4% and 2.2% in Adult and Shopping, respectively, which are the most two blocked categories. Also, we find that many of the blocked mobile apps in 2017 are accessible today. Moreover, we undertook further analyses to examine the effect of the COVID-19 pandemic on digital filtering and find that Saudi authorities have stepped up their actions to eliminate misinformation, aiming to maintain the clarity of public health communications. Finally, we find that changes in the filtering policy reflect the wider geopolitical dynamics of the region.

The earthquake warning system (EWS) is a critical technological advancement that prevents significant loss of life and infrastructure, particularly considering the current inability to predict earthquakes. Few developed nations, like the USA, Japan, etc. have implemented EWS. India has also made efforts to implement EWS for metro rails and Uttarakhand. Expanding EWS in developing nations necessitates the development of cost-effective and reliable solutions. Seismic sensors based on geophones and MEMS accelerometers have emerged as promising, cost-effective alternatives to expensive conventional sensors due to their performance. However, challenges persist related to the suboptimal acceleration response of these sensors. Therefore, experiments have been conducted on a tri-axial shake table to analyse the acceleration response of MEMS-based Raspberry Shake 4D (RS4D), geophone-based Raspberry Shake 3D (RS3D), with the conventional strong-motion sensor Guralp CMG-5TC by capturing simulated seismic waveforms. Based on the observations, machine learning (ML) techniques have been leveraged to establish the correlations between sensor responses in predicting acceleration using RS4D and RS3D sensors. Out of 41 regressors of Lazy Predict API, stochastic gradient descent (SGD) yields the best performance with a root-mean-square error (RMSE) of 0.2 and an R² score of 97.46. This predictive capability is significant and vital for multilevel earthquake warnings, significantly when acceleration profoundly impacts the operational management of critical installations. An Internet of Things (IoT)-based framework utilising ML-based acceleration predictions for EWS realisation has also been discussed. Integrating heterogeneous sensors with IoT technology provides an innovative framework to develop a cost-effective, reliable EWS.

Left ventricular ejection fraction (LVEF) is cardiovascular function's most important clinical parameter. The accuracy in estimating this parameter depends on precisely segmenting the left ventricle (LV) structure at the end-diastole and -systole phases. Therefore, developing robust algorithms for precisely segmenting the heart structure during different phases is crucial. In this work, an improved 3D UNet model is proposed to segment the LV and myocardium while excluding papillary muscles, as per the recommendation of the Society for Cardiovascular Magnetic Resonance. For the practical testing of the proposed framework, 8,400 cardiac MRI images were collected and analyzed from the military hospital in Tunis (HMPIT) and the popular ACDC public dataset. The Dice coefficient and the F1 score were used as performance metrics to validate the LV and myocardium segmentations. The data was split into 70%, 10%, and 20% for training, validation, and testing, respectively. It is worth noting that the proposed segmentation model was tested across three axis views: basal, medio basal and apical, at two different cardiac phases: end-diastole and -systole instances. The experimental results showed a Dice coefficient of 0.965 and 0.945 and an F1 score of 0.801 and 0.799 at the end-diastolic and -systolic phases, respectively. The clinical evaluation outcomes revealed a significant difference in the LVEF and other clinical parameters when the papillary muscles were included or excluded. The proposed framework outperforms state-of-the-art methods by around 0.1 in terms of Dice coefficient, demonstrating its accuracy in assessing the left ventricular function precisely.

In low-light conditions, image quality deteriorates, posing significant challenges for advanced computer vision tasks. The three main goals of Low-Light Image Enhancement (LLIE) are to enhance visual quality: restoring natural color and brightness, maintaining textures and edges, and minimizing noise and artifacts. Convolutional Neural Networks (CNNs) and Self-Attention (SA) processes, which have demonstrated excellent performance in low-level vision tasks, have been the focus of recent research on deep learning-based methodologies. CNNs excel at capturing local patterns with translational equivariance, while SA mechanisms are effective at modeling distant dependencies. However, both approaches have limitations—CNNs suffer from a restricted receptive field and limited feature diversity, while SA mechanisms struggle with local feature associations. To overcome these challenges, we propose a Multi-Level Integration and Disintegration Network (MIDNet) for LLIE. It leverages the strengths of both CNNs and SA by introducing an uneven dual-path architecture that facilitates mutual feature representation and progressive enhancement. This design enables effective decomposition and association of normal light and low-light features. Comprehensive assessments show that MIDNet outperforms state-of-the-art methods in LLIE across several benchmarks. The repository is available on GitHub at https://github.com/SATHISHMOTHE/MIDNet

This study investigates the effects of renewable resource management in scenarios involving autonomous battery energy storage systems (BESS) controlled by an energy management strategy (EMS) to manage the timing of BESS charging and discharging as an independent entity. A cost-minimization problem is proposed, along with multiple objective, to improve the system’s reliability and performance by mitigating interruption costs for both the distribution system and consumers, to reduce the power loss costs occurring in line and distribution transformers (DT), and to enhance DT operation by minimizing aging costs. The Pareto-search (PS) algorithm is employed to optimize the allocation of solar photovoltaic (SPV) and wind turbine (WT) generators, along with BESS units, within a mixed-integer nonlinear programming framework. The research also focuses on reducing peak power demand and stabilizing power consumption during nonpeak periods. Three different case studies are validated using Bus 4 of the Roy Billinton Test System (RBTS). The modeling approach considers the fluctuating power outputs of renewable energy sources and the variation in load profile over time. The results demonstrate that the proposed approach can effectively mitigate system congestion during real-time operations, decreasing strain on conductors and equipment caused by peak power demand. Additionally, it leads to a decrease in system outages and power loss.

The critical task of detecting anomalies in multivariate time-series data faces challenges due to the lack of anomaly labels and the unpredictable nature of the data. Despite deep learning’s advancements in improving anomaly detection, few models adeptly handle these complexities. This study introduces a novel deep learning model called DASTAD, which is designed to effectively detect anomalies in multivariate time-series data. Unlike current models that focus on time-based differences but overlook the connections between different univariate time-series, DASTAD utilizes transformer-based attention mechanism to process the time-series data both across time and between features. This dual approach ensures a deeper understanding of the data’s dynamic patterns. The model further overcomes the limitations of conventional encoder–decoder approaches by incorporating self-conditioning and adversarial training, ensuring reliable feature extraction and training stability. The model has been extensively tested on six publicly accessible datasets. The results show that it outperforms the current leading approaches in terms of detection and diagnostic accuracy. Additionally, the model achieves these results while using less data for training. Our approach specifically enhances F1 scores by a maximum of 7.97%.