Bolt fasteners play a critical role in securing, connecting, and sealing petrochemical equipment. Long-term vibration, high temperatures, corrosion, and other harsh environmental factors frequently lead to loosening or missing fasteners, degrading equipment performance and creating safety hazards that may ultimately trigger serious accidents. At present, defect detection and hidden-danger screening for bolt fasteners still rely mainly on manual periodic inspections, which suffer from low efficiency, heavy labor loads, and high rates of missed or false detection. In this study, a proprietary dataset was built from historical inspection images and on-site photographs of defective fasteners on petrochemical equipment. To mitigate the problems of insufficient defect samples and class imbalance, an orthogonal experiment was designed to investigate the sensitivity of dataset quality to common image-augmentation techniques. A tailored pre-processing and augmentation strategy for fastener-defect datasets was proposed, significantly enhancing sample diversity and overall dataset quality, thereby establishing a specialized dataset for petrochemical fastener defects. A mobile-oriented detection method for loosened or missing bolt fasteners in petrochemical equipment was developed based on YOLOv5s. Specifically, the BiFormer attention mechanism was incorporated into the YOLOv5s backbone to strengthen the extraction of small-target features. The original PANet in the neck was replaced with a Bi-FPN structure to improve multi-scale feature fusion and inference speed. A Focal-EIoU loss function was adopted to accelerate convergence and enhance the localization accuracy of small objects. Finally, the improved YOLOv5s model was converted into an NCNN lightweight model for mobile deployment. Experiments on the self-built dataset show that the improved YOLOv5s outperforms the original YOLOv5s by 7.9 % in precision (P), 8.8 % in mean average precision at IoU = 0.5 (mAP@0.5), and 8.2 % in recall (R). Although the detection speed decreases by 8.8 FPS, the frame rate still meets real-time requirements, enabling real-time, high-accuracy identification of missing or loosened bolt fasteners on explosion-proof equipment.

抄録全体を表示

2008年 Google 社を中心とする OHA（Open Handset Alliance）より世に出された Android は、スマー トフォンを中 心に世界に普及し日本でも本格的な普及期を迎えている。 Android はオープンソースであり、大学等教育機関でも 自由に開発を行うことができる。 本稿では学生の Java 言語習得のために 2010年度、 2011年度に実施した Android アプリケーションコンテストの実施結果 （2010 年度） 及び実施状況 (2011年度） について報告する。

抄録全体を表示

Blood pressure measurement by auscultatory method is a compulsory skill that is required by all healthcare practitioners. During the measurement, they must concentrate on recognizing the Korotkoff sounds, looking at the sphygmomanometer scale, and constantly deflating the cuff pressure simultaneously. This complex operation is difficult for the new learners and they need a lot of practice with the supervisor in order to guide them on their measurements. However, the supervisor is not always available and consequently, they always face the problem of lack of enough training. In order to help them mastering the skill of measuring blood pressure by auscultatory method more efficiently and effectively, we propose using a sensor device to capture the signals of Korotkoff sounds and cuff pressure during the measurement, and display the signal changes on a visualization tool through wireless connection. At the end of the measurement, the learners can verify their skill on deflation speed and recognition of Korotkoff sounds using the graphical view, and compare their measurements with the machine instantly. By using this device, the new learners do not need to wait for their supervisor for training but can practice with their colleagues more frequently. As a result, they will be able to acquire the skill in a shorter time and be more confident with their measurements.

抄録全体を表示

Maintenance engineers (ME) are responsible for delivery, installation and maintenance of computer-related equipment in the field, where efficiency and accuracy are required. In this paper we propose an Integration Method of Phone and PC systems to realize improving efficiency and accuracy of maintenance engineer's works by using smart devices, such as smartphones and tablets. These systems are accomplished by fusing the function of the Web browser function and SIP phones. We developed an easy method to switch tab format information acquired from the device in the field, and information obtained by accessing to maintenance center server. And also we implemented features of synchronization and coordination of information between ME and the administrator in the maintenance center.

抄録全体を表示