主催: 一般社団法人 日本機械学会
会議名: 生産システム部門研究発表講演会2021
開催日: 2021/03/08 - 2021/03/09
A method of presenting similar piping design failures in shipbuilding by applying sub-graph isomorphism to CAD data at design stage is discussed. The CAD data are converted as graph data structures. In recent years, the importance of automation in design work has been increasing as the design lead time has tended to be shortened. Particularly in shipbuilding, the design period of outfitting components such as piping is short and a pressing issue. On the other hand, CAD checking functions, such as collision checking, have been enhanced, and research on automation, such as automatic pipe design, has been conducted. The purpose of this study is to support efficient feedback of past failure information in the current design. In this paper, at first, a method of expressing and converting a pipe in a graph data structure from CAD data is considered. Using sub-graph isomorphism and calculating similarity based on rule, a suggesting the similar failures in a ranking format in design stage. Verification based on the actual ship CAD data is conducted and it is confirmed that the faults can be detected with an accuracy close to that of human detection. However, these are only some of the defects covered in this study, and their versatility is a subject for future study.