SUMMARY

Machining thin-walled, compliant parts is a cost-efficient way to manufacture lightweight and structurally sound parts as used extensively in the aerospace industry. Such parts are difficult to machine using traditional CNC machines due to part compliance, increased susceptibility to chatter, and the need for specialized tooling or fixturing devices. These challenges are heightened while machining with a robotic manipulator due to its lower stiffness and easily excited dynamics. However, due to the unique benefits of industrial robotic manipulators such as low cost and a large workspace to footprint ratio, there has been extensive research to maximize the accuracy and path compensation of robotic manipulators. This thesis introduces a methodology to compensate the path of a robotic manipulator to increase the accuracy of peripherally milled compliant parts. The research purpose is to develop an offline path compensation methodology as a solution to the part inaccuracies that occur during machining due to part compliance arising from the forces involved in machining. Two approaches to the compensation methodology are pursued in this thesis. The first approach utilizes experimentally determined dimensional errors to iteratively compensate a nominal path. In the second approach, milling force and part deflection models are used to predict the path compensation needed to compensate the part compliance induced errors. Experiments are performed on a 6-DOF industrial robotic manipulator with a laser-tracker based real-time closed-loop feedback control system. The experiments demonstrate the effectiveness of the iterative robot path compensation strategy in improving part accuracy. The benefits and implications of such a model-based compensation strategy are discussed and future improvements to the methodology are recommended.