[3Xin2] Poster session 1

[3Xin2-23] An Investigation of a Method Using Fault Knowledge Created by Functional Decomposition Trees and Ontologies for Improving Efficiency in Automobile Design Work

〇Kunika Iida¹, Risa Kubo¹, Munehiko Sasajima¹ (1.University of Hyogo)

Keywords:ontology, functional decomposition tree, fault reasoning

Automobile design work is usually conducted by modifying the function of previously designed automobile components. Specifically, this process is time-consuming because it requires exhaustive reasoning and enumeration of information about past faults of the component and their impact on other components and the automobile as a whole. We supposed that skilled designers perform design change work and consider what functions the components they are focusing on are working and what causes the functions to stop working (faults). Therefore, we will create a mechanism to create a knowledge model of an automobile from the viewpoint of function and fault, and then reason about faults, so that even non-skilled designers can perform the exhaustive reasoning necessary for information-collecting work, which is usually performed by skilled designers.
This study uses the drive unit of an automobile as an example, we examine methods that can improve the efficiency of information collecting work in automobile design work by a knowledge model created using functional decomposition trees and ontologies. In this paper, we have developed a prototype mechanism for "fault reasoning" that uses a functional decomposition tree and ontology to model the knowledge of skilled designers, allowing them to derive fault knowledge of possible faults from the function of a component.