http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
딥러닝을 이용한 고압 수소에 의하여 손상된 Acrylonitrile Butadiene Rubber (NBR) 단면 균열 탐지
이상민(Sangmin Lee),최병률(Byeonglyul Choi),최병호(Byoung-Ho Choi) 대한기계학회 2021 대한기계학회 춘추학술대회 Vol.2021 No.11
Demand for understanding material behaviors in hydrogen rich environment rises due to the extended use of hydrogen gas. Organic material such as rubber used in O-rings for hydrogen containers displays pore-like cracks when it is exposed extensively to high pressure hydrogen. To find the optimal organic material in a certain hydrogen rich environment, a method for localizing and analyzing pore-like cracks must be devised. This paper deals with a new method to detect high pressure hydrogen induced cracksusing deep learning algorithms. In this study, acrylonitrile Butadiene Rubber (NBR) was exposed to hydrogen of 96.6MPa for 24 hours. Images were taken of the appearing pore like cracks magnified by 100. For bounding box labeling a semi-automated labeling was used using maximally. stable extremal region (MSER) features on graphical contours. The crack detection model was trained adopting the structure of faster R-CNN using ResNet50 as its backbone network. The resulting artificial intelligence model showed robust and accurate detecting ability.