Integrating CAD files and automatic assembly sequence planning

Abstract

In this research study, a fully automated assembly sequence planner was developed, which automatically extracts geometrical information directly from STEP CAD files and then generates feasible assembly sequences, with a minimum number of assembly direction reorientations. The effectiveness of using the planner to reduce assembly time was also verified. The research study was completed in three parts: (1) methods were developed for automatically extracting and translating geometric information contained in CAD files into matrices, (2) the capabilities developed in part 1 of the research study were integrated into a genetic assembly planner, which was used to select optimal or near-optimal assembly sequences automatically, with respect to a user-defined assembly sequence evaluation criterion, number of assembly direction reorientations, and (3) the effect of generated assembly sequences, with different numbers of reorientations, on overall assembly time was verified, for both robot assembly and human operator assembly. In the first part of the research study, algorithms and software were developed for extracting geometrical information contained in STEP CAD files and for detecting potential collisions between parts during assembly motions along principal-axis assembly directions, based upon the extracted geometrical information. The developed software directly takes a STEP CAD file of a designed product assembly, as input, and generates, as output, six interference-free matrices, which represent potential collision relationships between parts of an assembly, for assembly directions along the six principal axes. In the second part of the research study, the software developed in part 1 of the research study was integrated into a genetic algorithm-based assembly sequence planner. The