repository
Special Graduate Schools
Graduate School of Science and Technology
Department of Industrial Engineering
4. Theses(Ph.D)
선박부재의 네스팅에서 페어링을 위한 2차원 폴리곤 분류 및 쌍별 클러스터링
- Alternative Title
- Two-Dimensional Polygon Classification and Pairwise Clustering for Pairing in Ship Parts Nesting
- Author(s)
- 나건열
- Advisor
- 양정삼
- Department
- 일반대학원 산업공학과
- Publisher
- The Graduate School, Ajou University
- Publication Year
- 2022-08
- Language
- kor
- Abstract
- 조선산업에서 네스팅은 강판의 활용률을 높이고 스크랩 비율을 줄이기 위해 선박부재의 절단 패턴을 배치하는 과정이다. 선박부재는 많은 수, 다양한 크기 그리고 복잡한 형상으로 인해 높은 네스팅 복잡도가 유발된다. 강판의 낭비를 최소화하도록 배치위치를 탐색하는 알고리즘들은 긴 계산시간으로 인해 두 부재를 짝지어 네스팅 복잡도를 감소시키는 페어링 없이 네스팅에 적용되기 어렵다. 본 논문에서는 네스팅의 복잡도를 줄이기 위해 딥러닝 기반 페어링 방법을 제안한다. 페어링을 위해 선박부재는 분류되고 쌍별 클러스터링 된다. 선박부재의 분류를 위해 딥러닝 모델이 폴리곤을 레스터화 없이 학습하는 방법과 쌍별 클러스터링을 위해 서로 다른 형상의 선박부재를 짝짓는 방법이 제안된다. 265개의 선박부재를 이용하여 제안된 방법과 조선사에서 사용하는 페어링 방법을 비교한 결과 82개의 부재를 더 짝지었으며, 분류 시간, 배치 시간 및 스크랩 비율은 평균 44.1%, 47.5% 및 11.0% 감소시켰다. 제안한 방법은 선박부재를 빠르고 정확하게 페어링하여 효율적인 네스팅에 기여한다.
- Alternative Abstract
- In the shipbuilding industry, nesting is the process of devising the patterns of ship parts to be cut to increase the utilization rate of steel sheets and reduce the scrap rate. Its complexity is high because of the large number of ship parts with complex shapes at various sizes. Arrangement algorithms for minimizing steel-sheet wastage cannot be readily applied to nesting without pairing two parts for reducing the nesting complexity because of the considerable computation time required. In this study, a pairing method to reduce the nesting complexity is proposed. Ship parts were classified, and pairwise clustering was applied for pairing. A method for classifying ship parts, in which a deep neural network architecture learns polygons without rasterization, and a method for pairing ship parts of different shapes for pairwise clustering were proposed. Using 265 actual ship parts, the performance of the proposed method was compared with that of existing pairing methods involving shape-based matching algorithms that are used by shipbuilding companies. Subsequently, 82 more parts were paired, and the classification time, arrangement time, and scrap rate decreased by 44.1, 47.5, and 11.0%, respectively, on average. Pairing based on deep learning classifiers and pairwise clustering can rapidly and accurately pair ship parts, thereby contributing to efficient nesting.
- Fulltext
- Appears in Collections:
- Special Graduate Schools > Graduate School of Science and Technology > Department of Industrial Engineering > 4. Theses(Ph.D)
- Files in This Item:
- There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.
-
- License
- STATISTICS
- Total Visit :3,772,618
- Total Download :1,819
- Today View :9,119
