Disassembling Process Inference Using Positional Relations Matrix for Complicated Machines

Abstract

Disassembling of a part is required for maintenance of machinery in case. However, the disassembling process is often not explained in the operation manual, or the explanation of the disassembling does not cover all the situations of all the individual parts, even though, such disassembling could be dealt with by operators that are not familiar with the mechanism of machine. Operators themselves have to determine the disassembly process in such a case. Therefore, it is crucial to develop a system that helps inexperienced operators to find out a proper disassembling process. We focus on the disassembling of a specific part referred to as a target part. The approach is based on the positional relation information among the parts. The positional relations matrix that obtained from the contact states of any two parts in all possible directions and can be generated from the ordinary CAD data. This study proposed a method to infer a disassembly process of a specific part based on the positional relation matrix. The method deduces the disassembly process of the target part with the shortest steps, in the condition of one-part-at-a-time manner. We also introduced an integration of disassembling parts based on the obtained process. A case study was conducted and the result confirmed the feasibility of the proposed method; the effectiveness of the integration approach was also demonstrated.