A vision-based calibration method for aero-engine blade-robotic grinding system
Chen Chen,Zhenhua Cai,Tingyang Chen,Zifan Li,Fan Yang,Xufeng Liang
DOI: https://doi.org/10.1007/s00170-023-10822-6
IF: 3.563
2023-01-18
The International Journal of Advanced Manufacturing Technology
Abstract:Robotic grinding has been gradually applied in the field of blade grinding due to its high precision and high flexibility. One important task in robotic grinding is to determine the position of the workpiece coordinate system. In addition, the mounting errors that exist during robotic grinding operations can lead to machining deviations. To solve above problems, this paper presents a vision-based robot workpiece posture error calibration method to fulfill the precision grinding requirements of aero-engine blades. Firstly, the 3D laser scanner is used to obtain point cloud information of aero-engine blades, where the blades are fixed to the end flange of the robot. Due to environmental interference and the limitation of the scanning angle, the initial point cloud is large and redundant; thus, this study applies statistical filtering, voxel filtering, and cluster segmentation for pre-processing. Therefore, to solve the problem of low efficiency of traditional ICP for aligning residual point clouds, this paper proposes to use the trimmed ICP algorithm to improve the alignment speed and accuracy between the obtained point cloud and the discrete CAD model point cloud to calibrate the pose errors of aero-engine blades. The experiments verified that the calibration method proposed in this paper has high precision and good surface machining consistency. The overall surface roughness of the blade is reduced from over Ra8μm to about Ra0.4 μm, which meets the technological requirements.
engineering, manufacturing,automation & control systems