Title: Trajectory Optimization Based on Thermal Data in a Robotic Wire-Arc Additive Manufacturing System
Date: Tuesday, July 14th, 2026
Time: 12PM to 2PM ET
Location: GTMI 114 or via Teams
Kathryn Kelly
Robotics Ph.D. Student
Woodruff School of Mechanical Engineering
Georgia Institute of Technology
Committee:
Dr. Christopher Saldaña (advisor) -- George W. Woodruff School of Mechanical Engineering
Dr. Thomas Kurfess -- George W. Woodruff School of Mechanical Engineering
Dr. Shreyes Melkote -- George W. Woodruff School of Mechanical Engineering
Dr. Kyle Saleeby -- Georgia Tech Manufacturing Institute
Dr. Sean Wilson -- Georgia Tech Research Institute
Abstract:
Wire-arc additive manufacturing has been widely studied in the pursuit of large-scale metal additive processes. Compared to other additive methods that deposit material at grams-per-hour, wire-arc can deposit material at kilograms-per-hour. The addition of a robot manipulator allows for the development of more complex and much larger geometries than traditional machining platforms. When depositing large amounts of material, many variables become more critical to the process effecting final product quality. Many of these variables, such as heat input, are directly controlled by the thermal properties fo the material selected. While trajectories for robotic additive manufacturing have been studied, the design of optimal trajectories that prioritize local thermomechanical conditions are not as well understood. This dissertation aims to address this gap through: (1) exploration of the efficacy of intersecting path-plan strategies for the infill of wire-arc parts, (2) investigating the impact of various intersecting path-plan strategies on thermomechanical response and part quality, and (3) studying how these strategies might be implemented for closed-loop online control algorithms. This work is impactful in providing new tools and methods for manufacturing process developers to optimize trajectories that enable deposition during the cooling process, while reducing thermal gradients and improving geometric accuracy and surface finish. Most importantly, it reduces the time needed to print large-scale components over current robotic additive methods.