Welding Sequence and Robot Path Optimization for Spot Welding Robots to Prevent Deformation of Sheet Metal Parts

摘要

There is an increasing demand for production automation, while there is a growing shortage of workers to teach robot motions. Spot welding in particular requires not only teaching of the robot motion, but also a determination of the welding sequence, which means that the teacher must have extensive experience and has led to a severe personnel shortage. Hence, we propose a robot teaching automation method for spot welding robots. The optimization problem of the welding sequence is treated as a traveling salesman problem (TSP), with additional requirements that the robot motion must satisfy a time requirement while avoiding collisions and the welding sequence must satisfy sequential constraints to prevent deformations in specific shapes within the target work. Our approach focuses on characteristics unique to spot welding, where in that the robot motion time of the large hand-sized spot welding robot is shorter when the normal directions of the planes to which the welding points belong are similar and that specific shapes can be expressed as specific combinations of planes. Specifically, by analyzing the shape of the target work and extracting planes from it, and by grouping together the welding points that belong to the same plane, the large TSP is divided into small TSPs, one for each group. Then, welding sequences are generated in which groups of planes with similar normal directions are welded consecutively. Furthermore, specific combinations of planes are extracted from the target work, and robot paths are planned only for welding sequences that satisfy the sequential constraints. We validated the proposed method in four case studies where the method reduced the robot motion time by up to 31% compared with a conventional method and satisfied the two requirements.