Search results

Conventional isolated dc–dc converters offer an efficient solution for performing voltage conversion with a large improved voltage gain. However, the small-signal analysis of these converters shows that a right-half-plane (RHP) zero appears in their control-to-output transfer function, exhibiting a nonminimum-phase stability. This RHP zero can limit the frequency response and dynamic specifications of the converters; therefore, the output voltage response is sluggish. To overcome these problems, the purpose of this study is to analyze, model and design a new isolated forward single-ended primary-inductor converter (IFSEPIC) through RHP zero alleviation.

At first, the normal operation of the suggested IFSEPIC is studied. Then, its average model and control-to-output transfer function are derived. Based on the obtained model and Routh–Hurwitz criterion, the components are suitably designed for the proposed IFSEPIC, such that the derived dynamic model can eliminate the RHP zero.

The advantages of the proposed IFSEPIC can be summarized as: This converter can provide conditions to achieve fast dynamic behavior and minimum-phase stability, owing to the RHP zero cancellation; with respect to conventional isolated converters, a larger gain can be realized using the proposed topology; thus, it is possible to attain a smaller operating duty cycle; for conventional isolated converters, transformer core saturation is a major concern, owing to a large magnetizing current. However, the average value of the magnetizing current becomes zero for the proposed IFSEPIC, thereby avoiding core saturation, particularly at high frequencies; and the input current of the proposed converter is continuous, reducing input current ripple.

The key benefits of the proposed IFSEPIC are shown via comparisons. To validate the design method and theoretical findings, a practical implementation is presented.

This paper aims to obtain a symmetrical step-down topology with lower equivalent capacity and wider step-down range under the condition of the same output. Two new symmetrical step-down topologies of star-connected autotransformers are proposed in this paper. Taking the equivalent capacity as the main parameter, the obtained topologies are modeled and analyzed in detail.

This paper adopts the research methods of design, modeling, analysis and simulation verification. First, the star-connected autotransformer is redesigned according to the design objective of symmetrical step-down topology. In addition, the mathematical model of two topologies is established and a detailed theoretical analysis is carried out. Finally, the theoretical results are verified by simulation.

Two symmetrical star-connected autotransformer step-down topologies are designed, the winding configurations of the corresponding topology are presented, the step-down ranges of these three topologies are calculated and the influence of step-down ratio on the equivalent capacity of autotransformer are analyzed. Through analysis, the target step-down topologies are obtained when the step-down ratio is [1.1, 5.4] and [1.1, 1.9] respectively.

Because the selected research object is only a star-connected autotransformer, the research results may lack generality. Therefore, researchers are encouraged to further study the topologies of other autotransformers.

This paper includes the implications of the step-down ratio on the equivalent capacity of autotransformers and the configuration of transformer windings.

The topologies designed in this paper enable star-connected autotransformer in the 12-pulse rectifier to be applied in step-down circumstances rather than situations of harmonic reduction only. At the same time, this paper provides a way that can be used to redesign the autotransformer in other multi-pulse rectifier systems, so that those transformers can be used in voltage regulation.

Automatically extracting stance information from natural language texts is a significant research problem with various applications, particularly after the recent explosion of data on the internet via platforms like social media sites. Stance detection system helps determine whether the author agree, against or has a neutral opinion with the given target. Most of the research in stance detection focuses on the English language, while few research was conducted on the Arabic language.

This paper aimed to address stance detection on Arabic tweets by building and comparing different stance detection models using four transformers, namely: Araelectra, MARBERT, AraBERT and Qarib. Using different weights for these transformers, the authors performed extensive experiments fine-tuning the task of stance detection Arabic tweets with the four different transformers.

The results showed that the AraBERT model learned better than the other three models with a 70% F1 score followed by the Qarib model with a 68% F1 score.

A limitation of this study is the imbalanced dataset and the limited availability of annotated datasets of SD in Arabic.

Provide comprehensive overview of the current resources for stance detection in the literature, including datasets and machine learning methods used. Therefore, the authors examined the models to analyze and comprehend the obtained findings in order to make recommendations for the best performance models for the stance detection task.

Dielectric barrier discharge (DBD) is widely used in the treatment of skin disease, surface modification of material and other fields of electronics. The purpose of this paper is to design a high-performance power supply with a compact structure for excimer lamps in electronics application.

To design a high-performance power supply with a compact structure remains a challenge for excimer lamps in electronics application, a current-source type power supply in a single stage with power factor correction (PFC) is proposed. It consists of an excitation voltage generation unit and a PFC unit. By planning the modes of the excitation voltage generation unit, a bipolar pulse excitation voltage with a high rising and falling rate is generated. And a high power factor (PF) on the AC side is achieved by the interaction of a non-controlled rectifier and two inductors.

The experimental results show that not only a high-frequency and high-voltage bipolar pulse excitation voltage with a high average rising and falling rate (7.51GV/s) is generated, but also a high PF (0.992) and a low total harmonic distortion (5.54%) is obtained. Besides, the soft-switching of all power switches is realized. Compared with the sinusoidal excitation power supply and the current-source power supply, the proposed power supply in this paper can take advantage of the potential of excimer lamps.

A new high-performance power supply with a compact structure for DBD type excimer lamps is proposed. The proposed power supply can work stably in a wide range of frequencies, and the smooth regulation of the discharge power of the excimer lamp can be achieved by changing the switching frequency. The ideal excitation can be generated, and the soft switching can be realized. These features make this power supply a key player in the outstanding performance of the DBD excimer lamps application.

Electrical characteristics of transformer oil (TO) have been studied during normal and thermal aging conditions. In this paper, breakdown voltage (BDV), partial discharge (PD), heat transfer results and the physical mechanisms considering the impact of varying the diameter of Al₂O₃ nanoparticles (NPs) have been investigated. Different quantities of the two sizes of Al₂O₃ were added to the oil using a two-step method to determine the positive effect of NPs on the electrical and thermal properties of TO. Finally, the physical mechanisms related to the obtained experimental results have been performed.

The implementation of nanoparticles in this paper was provided by US Research Nanomaterials, Inc., USA. The provided Al₂O₃ NPs have an average particle size of 20–80 nm and a specific surface area of 138 and 58 m²/g, respectively, which have a purity of over 99%. Thermal aging has been done. The IEC 60156 standard has been implemented to calculate the BDV, and a 500-mL volume test cell (Apar TO 1020) has been used. PD test is performed according to Standard IEC 60343, and a JDEVS-PDMA 300 device was used for this test.

BDV tests indicate that 20 nm Al₂O₃ is more effective at improving BDV than 80 nm Al₂O₃, with an improvement of 113% compared to 99% for the latter. The analysis of Weibull probability at BDV indicates that 20 nm Al₂O₃ performs better, with improvements of 141%, 125% and 112% at probabilities of 1, 10 and 50%, respectively. The results of the PD tests using the PDPR pattern also show that 20 nm Al₂O₃ is superior. For the heat transfer test, 0.05 g/L of both diameters were used to ensure fair conditions, and again, the advantage was with 20 nm Al₂O₃ (23% vs 18%).

The effect of Al₂O₃ NP diameter (20 and 80 nm) on various properties of virgin and aged TO has been investigated experimentally in this paper to examine the effect of proposed NP on electrical improvement of TO.

This study aims to examine the impact of socioeconomic factors on electricity usage and assess the feasibility of implementing a mini-grid system in Kasangazi, Malawi. The primary aim is to understand the community’s current and potential utilisation of electrical equipment.

A mixed-methods approach was used to collect quantitative and qualitative data. Information was gathered through structured questionnaires, and energy audits were conducted among 87 randomly selected households from 28 Kasangazi communities. Data analysis relied on descriptive statistics using IBM SPSS version 28.

The study indicates that every household in Kasangazi uses non-renewable energy sources: 60 households use disposable batteries for lighting, 20 for radios and all use firewood, freely sourced from local forests, for cooking and heating water. The study shows that firewood is the community’s preferred energy source, illustrating the challenges faced in the fight against deforestation. Most household income comes from farming, with smaller contributions from businesses, employment and family remittances. Access to higher education is scarce, with only one out of 349 family members receiving tertiary education. Despite the constraints of low education levels and income, there is a demand for larger electrical appliances such as stoves and refrigerators. This underscores the need for mini-grid solutions, even in less technologically advanced, agriculture-dependent communities.

This study underscores that in Sub-Saharan Africa, factors like household size, income and education levels do not significantly influence the electricity demand but should be taken as part of the fundamental human rights. Rural populations express a desire for electricity due to the convenience it offers, particularly for appliances like refrigerators and stoves. Mini-grids emerge as a viable alternative in regions where grid electricity provision is challenging. It is concluded from this paper that the issue of using renewable energy should not only be taken for environmental preservation but also to promote energy access, augmenting efforts in supplying electricity to the remotest parts of the country.

This study aims to regarding the application of traditional pulse frequency modulation control full-bridge LLC resonant converters in wide output voltage fields such as on-board chargers, there are issues with wide frequency adjustment ranges and low conversion efficiency.

To address these issues, this paper proposes a fixed-frequency pulse width modulation (PWM) control strategy for a full-bridge LLC resonant converter, which adjusts the gain by adjusting the duty cycle of the switches. In the full-bridge LLC converter, the two switches of the lower bridge arm are controlled by a fixed-frequency and fixed duty cycle, with their switching frequency equal to the resonant frequency, whereas the two switches of the upper bridge arm are controlled by a fixed-frequency PWM to adjust the output voltage. The operation modes of the converter are analyzed in detail, and a mathematical model of the converter is established. The gain characteristics of the converter under the fixed-frequency PWM control strategy are deeply analyzed, and the conditions for implementing zero-voltage switching (ZVS) soft switching in the converter are also analyzed in detail. The use of fixed-frequency PWM control simplifies the design of resonant parameters, and the fixed-frequency control is conducive to the design of magnetic components.

According to the fixed-frequency PWM control strategy proposed in this paper, the correctness of the control strategy is verified through simulation and the development and testing of a 500-W experimental prototype. Test results show that the primary side switches of the converter achieve ZVS and the secondary side rectifier diodes achieve zero-current switching, effectively reducing the switching losses of the converter. In addition, the control strategy reduces the reactive circulating current of the converter, and the peak efficiency of the experimental prototype can reach 95.2%.

The feasibility of the fixed-frequency PWM control strategy was verified through experiments, which has significant implications for improving the efficiency of the converter and simplifying the design of resonant parameters and magnetic components in wide output voltage fields such as on-board chargers.

COVID-19 continues to spread, and cause increasing deaths. Physicians diagnose COVID-19 using not only real-time polymerase chain reaction but also the computed tomography (CT) and chest x-ray (CXR) modalities, depending on the stage of infection. However, with so many patients and so few doctors, it has become difficult to keep abreast of the disease. Deep learning models have been developed in order to assist in this respect, and vision transformers are currently state-of-the-art methods, but most techniques currently focus only on one modality (CXR).

This work aims to leverage the benefits of both CT and CXR to improve COVID-19 diagnosis. This paper studies the differences between using convolutional MobileNetV2, ViT DeiT and Swin Transformer models when training from scratch and pretraining on the MedNIST medical dataset rather than the ImageNet dataset of natural images. The comparison is made by reporting six performance metrics, the Scott–Knott Effect Size Difference, Wilcoxon statistical test and the Borda Count method. We also use the Grad-CAM algorithm to study the model's interpretability. Finally, the model's robustness is tested by evaluating it on Gaussian noised images.

Although pretrained MobileNetV2 was the best model in terms of performance, the best model in terms of performance, interpretability, and robustness to noise is the trained from scratch Swin Transformer using the CXR (accuracy = 93.21 per cent) and CT (accuracy = 94.14 per cent) modalities.

Models compared are pretrained on MedNIST and leverage both the CT and CXR modalities.

Genre classification of songs based on lyrics is a challenging task even for humans, however, state-of-the-art natural language processing has recently offered advanced solutions to this task. The purpose of this study is to advance the understanding and application of natural language processing and deep learning in the domain of music genre classification, while also contributing to the broader themes of global knowledge and communication, and sustainable preservation of cultural heritage.

The main contribution of this study is the development and evaluation of various machine and deep learning models for song genre classification. Additionally, we investigated the effect of different word embeddings, including Global Vectors for Word Representation (GloVe) and Word2Vec, on the classification performance. The tested models range from benchmarks such as logistic regression, support vector machine and random forest, to more complex neural network architectures and transformer-based models, such as recurrent neural network, long short-term memory, bidirectional long short-term memory and bidirectional encoder representations from transformers (BERT).

The authors conducted experiments on both English and multilingual data sets for genre classification. The results show that the BERT model achieved the best accuracy on the English data set, whereas cross-lingual language model pretraining based on RoBERTa (XLM-RoBERTa) performed the best on the multilingual data set. This study found that songs in the metal genre were the most accurately labeled, as their text style and topics were the most distinct from other genres. On the contrary, songs from the pop and rock genres were more challenging to differentiate. This study also compared the impact of different word embeddings on the classification task and found that models with GloVe word embeddings outperformed Word2Vec and the learning embedding layer.

This study presents the implementation, testing and comparison of various machine and deep learning models for genre classification. The results demonstrate that transformer models, including BERT, robustly optimized BERT pretraining approach, distilled bidirectional encoder representations from transformers, bidirectional and auto-regressive transformers and XLM-RoBERTa, outperformed other models.

Crack detection of pavement is a critical task in the periodic survey. Efficient, effective and consistent tracking of the road conditions by identifying and locating crack contributes to establishing an appropriate road maintenance and repair strategy from the promptly informed managers but still remaining a significant challenge. This research seeks to propose practical solutions for targeting the automatic crack detection from images with efficient productivity and cost-effectiveness, thereby improving the pavement performance.

This research applies a novel deep learning method named TransUnet for crack detection, which is structured based on Transformer, combined with convolutional neural networks as encoder by leveraging a global self-attention mechanism to better extract features for enhancing automatic identification. Afterward, the detected cracks are used to quantify morphological features from five indicators, such as length, mean width, maximum width, area and ratio. Those analyses can provide valuable information for engineers to assess the pavement condition with efficient productivity.

In the training process, the TransUnet is fed by a crack dataset generated by the data augmentation with a resolution of 224 × 224 pixels. Subsequently, a test set containing 80 new images is used for crack detection task based on the best selected TransUnet with a learning rate of 0.01 and a batch size of 1, achieving an accuracy of 0.8927, a precision of 0.8813, a recall of 0.8904, an F1-measure and dice of 0.8813, and a Mean Intersection over Union of 0.8082, respectively. Comparisons with several state-of-the-art methods indicate that the developed approach in this research outperforms with greater efficiency and higher reliability.

The developed approach combines TransUnet with an integrated quantification algorithm for crack detection and quantification, performing excellently in terms of comparisons and evaluation metrics, which can provide solutions with potentially serving as the basis for an automated, cost-effective pavement condition assessment scheme.