Point-driven robot selective grinding method based on region growing for turbine blade

Abstract

The complex geometric characteristics and the uneven allowance distribution of turbine blades restrict the grinding accuracy of robots. A novel point-driven robot selective grinding method based on region growing is proposed to enhance the surface accuracy of the turbine blade. First, this method calculates the curvature of every surface point among the turbine blade point clouds located at the slicing plane. Then, all surface points are segmented into intake edge points, exhaust edge points, convex points, and concave points. Moreover, the ideal normal grinding force (INGF) of every surface point at blade edges and profile is calculated based on the allowance distribution and material removal rate of belt grinding. INGF values, as the main characteristics of these surface points, are used in the voxel-based region growing to obtain multiple grinding regions in the blade surface, and their corresponding INGF values are calculated. Finally, the planned robotic grinding trajectories are modified based on the INGF values of these grinding regions. Robotic grinding experiments with the blade point clouds are conducted. The surface accuracy of the turbine blade with the proposed method is improved by 46.49% compared to that with the traditional grinding method.