An Intelligent Path Generation Method of Robotic Grinding for Large Forging Parts

Abstract

Large forging parts widely adopted in oil rigs, wind mills, large vessels and other complex equipment often carry random forging defects such as parting lines, burrs and high islands, which are traditionally removed through manual grinding by skilled operators. These random defects pose a big challenge to researchers interested in large forging parts grinding path generation by a CAD/CAM system. This paper proposes a new path generation method based on intelligent defect recognition of robotic grinding for large forging parts. A point cloud is first constructed of random defects on the surfaces of the to-be-ground parts from the unmatched points that emerge from the matching of the point clouds captured by a laser camera from the parts against those from the standard parts, a process employing both a random sample consensus algorithm and a modified iterative closest point algorithm. Then, the grinding path generation strategy is established by sorting the random defects according to a law of area on the fitting surface and robotic grinding motion programs are generated by transferring the coordinates of the random defects from the laser camera frame into the robot base frame. Finally, robotic grinding tests are conducted to verify the identification accuracy of the proposed new method. Results of the tests indicate that the method has accurately identified all random defects on a 10-m long forging part and intelligently generated subsequent robotic grinding paths according to the identified random feature categories. This study therefore provides an intelligent tool for finishing large forging parts.

S. Yan, Q. Wang—Authors with equal contributions.