Development and Control of an Automated Robotic Weld Bead Grinding System

Abstract

In a robotic weld bead grinding system there are large interaction forces between the robot and the workpiece. Jigging errors and robot compliance make position control unreliable. Instead, metal removal can be controlled by controlling the power applied to the workpiece. The approach used in this work has been to develop a nonlinear force control law for the PUMA 560 robot and implement a structured light vision system that measures material volume in real time. A power trajectory planner has also been incorporated into the system permitting the delivery of a power (the product of force and wheel speed) trajectory to the weld bead. The resulting system is able to plan and execute a multi-pass grinding sequence for removing a weld bead.