Accurate localization and girth weld grinding planning for an in-pipe machining robot of thin-walled conical pipe

In-pipe robots have been applied to detection, cleaning, welding, grinding, drilling, etc., which realizes the narrow space operation efficiently and economically. However, intelligent operation is still a problem, especially for high-precision machining of inner welds in thin-walled conical pipes, due to the low stiffness and precision of the robot machining system, uncertain robot position, and poor consistency of target characteristics. To address this problem, an intelligent girth weld grinding method of an in-pipe machining robot for thin-walled conical pipes is proposed. An in-pipe machining robot (named IPMR-I) with adaptive motion ability, controllable stiffness, and high machining precision is designed, which benefits by the designs of controllable contact forces and a high-precision three-axis machining mechanism. A weld locating method based on time-series point clouds, combined with an analytical pose model, is used for robot self-localization, obtaining the accurate pose of the robot relative to the weld region. Furthermore, an intelligent machining path planning method is proposed with the abilities of the weld boundary recognition, machining path generation, and optimization, which adaptively realizes the high machining quality and safety facing the welding irregularity (e.g., deformation and misalignment) inside thin-walled conical pipes. Several weld bead grinding experiments were conducted inside thin-walled conical pipes to verify the proposed methods’ validity. The results proved that IPMR-I with the proposed intelligent girth weld grinding method completed autonomous high-quality machining without manual intervention. Without damaging the base material, the maximum residual weld height is 0.18 mm, with the average residual height controlled at approximately 0.08 mm.