Surface and Process Modeling and Off-Line Programming for Robotic Spray Painting of Curved Surfaces

摘要

Abstract An algorithm and a computer program is developed for off-line programming of industrial robots for spray painting of curved surfaces. The study is initiated to find a more productive and repeatable, and thus higher quality method for painting of car body components in car body repair shops, where only the deformed region or part is straightened or replaced, and then painted. Body components of a specific car are modeled, and painting strategies, parameters and paths are determined. CAD models of relatively simple surfaces are formed by using the surface generation tools of the software. For parts with more complex surfaces, point data related to the part is collected by using a laser scanner, and this data is used to form the CAD model of the part surface. For paint (coating) thickness distribution analysis, the surface is divided into small elements, and their centroid coordinates and unit normals are determined. These properties are then used together with the spray distance, painting velocity and the paint flow rate flux for paint thickness analysis. Paint flow rate flux through the spray gun is determined by using experimental paint thickness distributions obtained by using different spray gun settings and painting parameters. Besides the technical specifications of the spray gun, air and paint nozzles, and paint needle; basic settings like paint tank pressure and gun needle-valve position affect paint cone angle and paint flowrate, which finally characterize the spray painting process. After deciding on an acceptable paint thickness distribution, the robot program is generated in the robot’s programming language. While moving along the path, the Cartesian position, velocity and orientation of the gun are necessary. Since there is no Cartesian input function available in the controller of the industrial robot used in this study, an inverse kinematic solution algorithm is developed to obtain the generated data in the joint space of the manipulator.