IEEJ Transactions on Electrical and Electronic Engineering最新文献

This paper investigates two studies on the magnetic circuit model incorporating the play model. First, the calculation speed of several general-purpose circuit simulators is compared when analyzing the magnetic circuit model incorporating the play model. Second, the calculation accuracy and speed of the two modeling methods of the play model on a general-purpose circuit simulator are compared. © 2024 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.

In order to solve the contradiction between the rapid isolation of low voltage DC faults and the low bounce of small stroke vacuum switches, taking the fast vacuum switch for 400 V DC power distribution system as the research object, the electromagnetic buffer simulation and experimental research in the opening process are carried out. The influencing factors of the dynamic and reaction mechanism and buffer output characteristics are analyzed, and the electromagnetic buffering effect under typical working conditions is simulated, which proves that electromagnetic buffer is a necessary part of 400 V fast vacuum switch and the motion characteristics of fast switching are optimal when applying 220 V electromagnetic buffer. According to the simulation calculation results, the engineering prototype was made, the experimental research on the electromagnetic buffer under typical working conditions was carried out, and the design criteria of the small stroke electromagnetic buffer were proposed, which solved the problem of fast switching of small stroke and reliable control and low bounce. © 2024 Institute of Electrical Engineers of Japan and Wiley Periodicals LLC.

Current sensors based on the tunneling magnetoresistive effect (TMR) are widely used for current measurement due to their high sensitivity, small size, and low power consumption. This paper proposes an effective error correction model to rectify the eccentricity of the transmission line, which can cause a significant measurement error in the ring-array single-axis TMR sensor. The model employs a convolutional neural network (CNN) to identify the relationship between the conductor eccentricity and the output of three sensors. The resulting correction factor is then fed back to eliminate the error associated with wire eccentricity. Concurrently, the Sparrow search algorithm (SSA) is employed to optimize the hyperparameters of the convolutional neural network (CNN) in order to enhance the model's performance. The experimental results demonstrate that the maximum error of the ring-array single-axis TMR current sensor, corrected by SSA-CNN, is less than 0.42%, which markedly enhances the precision of the measurement. © 2024 Institute of Electrical Engineers of Japan and Wiley Periodicals LLC.

An improved EA-YOLO object detection algorithm based on YOLOv5 is proposed to address the issues of drastic changes in target scale, low detection accuracy, and high miss rate in unmanned aerial vehicle aerial photography scenarios. Firstly, a DFE module was proposed to improve the effectiveness of feature extraction and enhance the whole model's ability to learn residual features. Secondly, a CWFF architecture was introduced to enable deeper feature fusion and improve the effectiveness of feature fusion. Finally, in order to solve the traditional algorithm's shortcomings it is difficult to detect small targets. We have designed a novel SDS structure and adopted a strategy of reusing low-level feature maps to enhance the network's ability to detect small targets, making it more suitable for detecting some small objects in drone images. Experiments in the VisDrone2019 dataset demonstrated that the proposed EA-YOLOs achieved an average accuracy [email protected] of 39.9%, which is an 8% improvement over YOLOv5s, and [email protected]:0.95 of 22.2%, which is 5.2% improvement over the original algorithm. Compared with YOLOv3, YOLOv5l, and YOLOv8s, the [email protected] of EA-YOLOs improved by 0.9%, 1.8%, and 0.6%, while the GFLOPs decreased by 86.4%, 80.6%, and 26.7%. © 2024 Institute of Electrical Engineers of Japan and Wiley Periodicals LLC.

In this study, we propose self-retrieval learning, a self-supervised learning method that does not require an annotated dataset. In self-retrieval learning, keywords extracted from documents are used as queries to construct training data that imitate the relationship between query and corpus, such that the documents themselves are retrieved. In the usual supervised learning for information retrieval, a pair of query and corpus document is required as training data, but self-retrieval learning does not require such data. In addition, it does not use information such as reference lists or other documents connected to the query, but only the text of the documents in the target domain. In our experiments, self-retrieval learning was performed on the EU and UK legal document retrieval task using a retrieval model called DRMM. We found that self-retrieval learning not only does not require supervised datasets, but also outperforms supervised learning with the same model in terms of retrieval accuracy. © 2024 Institute of Electrical Engineers of Japan and Wiley Periodicals LLC.