Bulletin of Electrical Engineering and Informatics
Not a member yet
2885 research outputs found
Sort by
Machine learning based annual solar energy forecasting for enhanced grid integration of photovoltaic systems
Get PDFThe increase in electricity demand is witnessed by many nations due to the rise in population and ongoing developments. To cope with energy requirements, countries are looking towards cleaner alternatives to reduce overreliance on energy generation from conventional resources. The introduction of artificial intelligence (AI) in real-world applications is acknowledged positively by experts as it enhances the performance and efficiency of the system. This paper reports the advancement of AI in harnessing renewable energy sources (RESs) to their true potential by leveraging their response when the grid is not able to fulfill the power requirement from conventional resources. Moreover, the prediction also remains a challenge with renewables due to their volatile behavior, especially with solar-based energy generation. This issue is also addressed by interfacing AI-enabled applications and the difference between true and predicted values for one year is observed. The result reveals that the true response aligns with the predicted response, which ensures the ability of AI to harness solar energy by consuming minimal time. The proposed approach is also promising from the utility operators’ and end users’ perspectives in designing any large-scale renewable projects for sustainable development and also encourages the utilization of renewables to a larger extent
Adaptive voltage controller based on extreme learning machine for DC-DC boost converter
Get PDFThis study presents an adaptive voltage controller for a DC-DC boost converter using the extreme learning machine (ELM) algorithm to address the limitations of conventional control techniques under varying load and reference voltage conditions. The ELM is implemented to predict the optimal parameters of a PI controller (Kp and Ki), enabling real-time adaptability of the system. Simulation results in MATLAB/Simulink demonstrate that the proposed ELM-based proportional-integral controller (PI-ELM) outperforms both traditional PI controllers and those optimized using metaheuristic algorithms. Specifically, the controller achieved a maximum absolute error of only 0.0185 for Kp and 0.0294 for Ki across a range of operating conditions, with corresponding mean squared errors (MSE) of 0.01861 and 0.02798, respectively. These findings confirm the effectiveness of the ELM in enhancing the dynamic response and robustness of boost converter voltage regulation systems
Artificial intelligence in smart home security: balancing innovation with ethics
Get PDFBecause of the evolution of artificial intelligence (AI), home security has progressed from a basic security system to an active architecture that is responsive and adaptive to real world situations. Due to the rapid adoption of AI in smart systems, there is increasing suspicion surrounding privacy issues and ethical ambiguity, as well as gaps when it comes to regulating these technologies. We provide an overview of AI in smart home security applications and examine the area of security, access control, intrusion detection, human action recognition, and research on intelligent automation. We summarize the last decade of evolution, with some summaries of previous on computer vision, authentication systems, and finding unusual patterns recently. Our key findings include the development of approaches to improve real time security monitoring, dramatic reductions in false alarms, and customization of home access using AI. Improvements in security have also increased risk with respect to ethical ambiguity as well as technical issues in certain cases. In this paper, we offers pathways for improved AI system design, proposed formal data protection regulations, and examples of simplifying complex system for user comprehension, which also establishes the groundwork for future efforts. Home security should balance new opportunities with ethical considerations
An efficient clustering approach in electrical energy consumption patterns
Get PDFA comprehensive understanding of electrical energy consumption patterns is essential for strategizing and monitoring the use of energy resources. Industry and business customers of electrical have energy consumption patterns that vary widely depending on the type of industry, business size, and operating hours. This research uses clustering analysis to obtain electrical energy consumption patterns in industrial and business electricity customer groups by grouping data into similar groups. The variables used in this research are daytime, active power (kW), apparent (kVa), and power factor (PF). The objective of this research is to determine the efficacy and benefits of each clustering technique employed in load profile analysis. The clustering algorithm approach used in this research is k-means and fuzzy subtractive clustering (FSC). The trials carried out on these two approaches provide valuable knowledge regarding the effectiveness and superiority of each algorithm in producing significant clusters from the data used in this research. The evaluation conducted using the Davies-Bouldin index (DBI) indicates that the quality value for FSC is 0.25 for business customers and 0.31 for industrial customers. On the other hand, the quality value for k-means is 0.55 for business customers and 0.56 for industrial customers
Deflection enhancement of ferrite magnetic core-based microactuator
Get PDFMicroactuators play a vital role in several microelectromechanical systems (MEMS) that generate forces or deflections necessary to accomplish functions such as scanning, tuning, manipulation, or delivery. Utilizing a ferrite magnetic core has shown the potential to enhance the deflection of the microactuator. However, the previous study presented a complex fabrication method with high power consumption unsuitable for micropump application. Herewith, we report the impact of ferrite core length on the deflection generated by a microactuator with a simple fabrication method. The deflection behavior shows that the corresponding magnetic core length is inverse to the deflection improvement. The force reduction generated led by a longer magnetic core because of the farther distance to the coil. Our study can be used as a reference to support the development of micropump or active micromixer devices, which require compact devices with simple fabrication and high deflection, achieving ultra-high flow rate and high mixing index
Efficient diabetic retinopathy detection using deep learning approaches and Raspberry Pi 4
Get PDFDiabetic retinopathy (DR) is a leading cause of vision loss, predominantly affecting individuals aged 25-74 with diabetes mellitus. Timely medical intervention can protect against irreversible blindness in over 90% of cases, emphasizing effectively identifying and treating DR. In the scope of deep learning (DL), the possibility of using them in DR screening has garnered a lot of interest. Specifically, we adopted the densely connected convolutional networks (DenseNet) model because to its capacity to acquire complex features and learn from diverse datasets. Developing the computational model on retinal images labelled with varying phases of DR are obtained from databases such as Messidor and Kaggle. To enhance accessibility and user-friendliness, we integrated the DenseNet model into a Raspberry Pi 4, a compact, affordable and widely accessible computing platform. The proposed approach resulted in an impressive classification accuracy of 88%, demonstrating its proficiency in distinguishing between different phases of DR progression. The study aims to assist in the early detection and diagnosis of the disease, providing a potential resource that could help medical practitioners and ophthalmologists to evaluate the extent of DR in a timely manner
Development of distance formulation for high-dimensional data visualization in multidimensional scaling
Get PDFThis research aims to produce a new method called pasca-multidimensional scaling (pasca-MDS) by modifying the multidimensional scaling (MDS) method, the developed model comes as a solution to overcome the problem of data complexity by reducing its description dimension without losing important information. This model, offers an innovative approach in dealing with these problems. Pasca-MDS not only focuses on reducing the dimensionality of data, but also retains the essence of relevant information from each data point. As such, it allows for easier and more efficient analysis without compromising the accuracy of the information conveyed. The main advantage of pasca-MDS lies in its ability to produce simpler visual representations while maintaining the original structure of complex data. This provides clarity and ease in understanding the patterns or relationships hidden within. By using adjustment techniques after the MDS process, this model can provide more optimized results. This process allows the adjustment of data points to achieve a better representation in a lower dimensional space, resulting in a more intuitive and easy-to-understand interpretation. The developed distance formula has the ability to minimize stress compared to other distance formulas in MDS space, with the aim of improving the accuracy of high-dimensional data visualization
Graphene based nano-antenna for wireless communication systems at terahertz band
Get PDFThe need for nano-antennas with decreased size and the capacity to operate at mid-infrared frequencies to enable adequate coverage of signal is being investigated. In this paper, we present graphene-based nano-antenna and analysed at the resonating frequency 33 THz using gallium arsenide material as a substrate having dielectric constant 11.35 and a loss tangent of 5.6×10-4 for terahertz (THz) frequency. The height of substrate is optimized to 108 nm and in-plane dimension being 1,700×1,400 nm. Graphene was used as a rectangular patch with dimension 850×450×5 nm and ground having chemical potential=1.4 eV, and relaxation time=1 ps, to achieve high gain and bandwidth. Impact of slot width variation on the antenna parameters have been reported in terms of reflection coefficient (S11), voltage standing wave ratio (VSWR), radiation pattern and gain. Reported beam width being 90.4° for both electric and magnetic planes. Proposed antenna achieved a return loss of -18.38 dB, VSWR less than 2, indicating good match with load, highest gain of 8.8 dBi and bandwidth of 500 GHz at the target resonance frequency making it suitable for 5G/6G mm wave wireless communication
Advanced drug recommendation using long short-term memory and type-2 fuzzy logic integration
Get PDFThis research on hybrid models for drug recommendation systems proposes long short-term memory (LSTM) and type-2 fuzzy logic (T2FL) to make its recommendations more accurate and reliable. The model leverages LSTM's ability to capture temporal patterns in medical data while addressing the inherent uncertainty through T2FL. Evaluation metrics such as mean absolute error (MAE), root mean squared error (RMSE), coefficient of determination (R²), accuracy, precision, recall, F1-Score, and area under the curve-receiver operating characteristic (AUC-ROC) demonstrate that the proposed model significantly outperforms traditional models like LSTM without fuzzy, linear regression, and random forest. Integrating these two methods results in more accurate and consistent predictions, making the model highly effective in handling complex and uncertain data. Practical implications include the potential for improving personalized treatment plans and patient outcomes in clinical settings. Future research directions involve applying this hybrid approach to larger, more diverse datasets and exploring additional hybrid methods that enhance prediction accuracy and model robustness. The findings suggest that the LSTM+T2FL model is a promising tool for advancing drug recommendation systems in the medical field
Optimizing cloud infrastructure efficiency through advanced multimedia data deduplication techniques
Get PDFOrganizations worldwide commonly utilize cloud infrastructure to manage large volumes of data, making the optimization of storage crucial for enhancing cloud performance. One effective optimization technique is data deduplication, which identifies duplicate objects and ensures that only one copy of unique data is stored in the cloud. While several deduplication schemes currently exist, there is a pressing need to improve efficiency in cloud storage through innovative approaches. In this paper, we propose a new system model designed to facilitate an efficient deduplication process. Our algorithm, called deduplication in cloud infrastructure (DCI), offers a systematic and effective method for handling deduplication challenges related to redundant data storage. DCI focuses on hash generation, metadata comparison, and pointer-based deduplication, providing a comprehensive strategy for optimizing cloud storage resources and minimizing duplication. This ultimately enhances both the efficiency and cost-effectiveness of cloud-based data management. A simulation study using CloudSim and the Hadoop distributed file system (HDFS) simulator demonstrates that the proposed deduplication method is effective. Experimental results show that our algorithm outperforms many existing solutions, achieving the highest deduplication ratio of 6.7 and saving 85.09% of storage space due to its efficient deduplication approach. The proposed system can be used in cloud infrastructures for efficiency