Research

DL for extrusion process

본 연구에서는 마이크로카테터 튜브 제조 공정을 위한 온라인 압출 각도 추정 시스템이 개발되었다. 작업자의 변화에 따른 품질 변화를 최소화하기 위해 압출 각도를 정량화 하는 컨볼루션 신경망이 개발되었다. 제안된 시스템은 전면과 측면을 촬영하는 두 대의 RGB 카메라와 이미지를 처리하는 딥러닝 시스템으로 구성된다. 카메라에서 캡처된 이미지는 컨볼루션 신경망으로 분할되고 분할된 이미지를 기반으로 주성분 분석을 통해 압출 각도를 추정한다.

In this study, an online extrusion angle estimation system was developed for a microcatheter tube manufacturing process. To minimize the variation of quality from the operator actions, a system was developed using a convolutional neural network for quantifying the extrusion angle. The proposed system consists of two RGB cameras that capture the front and side surfaces and a server that processes the images. The captured images from the cameras are segmented by a convolutional neural network. Finally, based on the segmented image, an extrusion angle is estimated by using principal component analysis.

Related papers

S. H. Jeong, J. W. Lee, G. H. Yoon*, and D. H. Choi, “Topology optimization considering the fatigue constraint of variable amplitude load based on the equivalent static load approach”, Applied Mathematical Modelling. [LINK]

S. H. Jeong, D. H. Choi, and G. H. Yoon*, “Fatigue and static failure considerations using a topology optimization method”, Applied Mathematical Modelling. [LINK]

A. Takezawa*, G. H. Yoon, S. H. Jeong, M. Kobashi, and M. Kitamura, “Structural topology optimization with strength and heat conduction constraints”, Computer Methods in Applied Mechanics and Engineering. [LINK]

S. H. Jeong, D. H. Choi, and G. H. Yoon*, “Separable stress interpolation scheme for stress-based topology optimization with multiple homogeneous materials”, Finite Elements in Analysis and Design. [LINK]

S. H. Jeong, G. H. Yoon*, A. Takezawa, D. H. Choi, “Development of a novel phase-field method for local stress-based shape and topology optimization”, Computers & Structures. [LINK]

S. H. Jeong, S. H. Park, D. H. Choi, and G. H. Yoon*, “Toward a stress-based topology optimization procedure with indirect calculation of internal finite element information”, Computers & Mathematics with Applications. [LINK]

S. H. Jeong, S. H. Park, D. H. Choi, and G. H. Yoon*, “Topology optimization considering static failure theories for ductile and brittle materials”, Computers & Structures. [LINK]

Development of CPT boom weld process optimization program

본 연구에서는 자동화된 용접 해석을 기반으로 CPT 붐 용접 순서를 최적화하는 프로그램을 개발했다. 개발된 프로그램은 사용자가 생성한 기본 용접 모델을 이용하여 자동으로 용접 순서를 변경하고, Sysweld 해석 입력 파일을 생성하여 용접 해석을 수행한다. 또한, 유전자 알고리즘을 이용하여 사용자가 지정한 관심영역의 변형량을 최소화하도록 개발되었다.

In this study, a CPT boom weld sequence optimization program was developed based on automated weld distortion analysis procedure. The developed program uses the basic welding model created by the user to automatically change the welding sequence, creates Sysweld analysis input files, and performs welding analysis. In addition, it was developed to minimize the amount of deformation of the region of interest designated by the user using a genetic algorithm.