Automatic Assembly Precedence Detection in Axisymmetric Products

Assembly sequence planning starting from CAD models turned out to be a relevant task in the industrial manufacturing field. To have a successful assembly sequence, the relations between the assembly's parts and the possible interferences during the assembly operations deserve to be investigated. In particular, the collision analysis is the phase in which the movement of each part along some directions is evaluated to assess if it is obstructed by any of the other parts, and according to which the precedence matrix is computed. A lot of work has been done to address the problem, however, existing algorithms need to be improved yet. Among the open issues, the following three result to be the most challenging: the combinatorial explosion of the problem complexity, the limited choice of the assembly directions, and the engineering meaning of parts that is not taken into account, or it is manually given in input by experts. In this paper, an automatic assembly sequence planning approach is introduced. The focus is on the collision detection and precedence analysis for engineering meaningful subassemblies, namely the axisymmetric. Information automatically extracted relying on geometric processing and engineering knowledge, such as parts features and semantic interpretation of mechanical components, is first exploited to identify the subassemblies and, then, to choose the feasible assembly direction, as well as to treat fasteners and deformable parts in a more realistic way. An industrial CAD model of a gearbox is selected as case study to illustrate the approach, also emphasizing the importance of axisymmetric subassemblies.