Machine Learning to Improve Friction Stir Welding Quality

Project Objective

Apply ML.AI to defect detection and mitigation in Friction Stir Welding as performed by NASA robotic welding machines as guided by human operators.

Project Description

We will apply ML.AI machine learning methods to large data sets of logged welding (robotic) machine feedback data of rpm, travel speed, forces, etc., and will correlate these with post-welds test results of Yield Strength, Tensile Stress, elongation at fracture, hardness, NDE and voids detection as well as X-ray. We will seek to uncover the root causes for the formation of a defect and will generate protocols for the avoidance of these problematic area and therefore that steer away from these and towards a high-quality consistent weld via real time feedback control.

Project Results and Conclusions

Improved quantified understanding of defect formation in FSW friction stir welding. This to be incorporated into real time feedback controllers of the robotic welding machines at NASA via Reinforcement Learning methods and via MPCs model predictive controllers (ongoing separate project EM32/ES11) and as additional quantified insight into welding machine variable data logged and tested post weld samples results. The goal is ML.AI real time feedback controlled robotic FS welding that produces consistently high-quality defect-free welds for Space Launch System and beyond.

This is early/mid-stage (TRL 3) — the most realistic path in is NASA SBIR/STTR, which funds small businesses and research institutions to develop technology aligned with NASA's needs (equity-free, phased funding). Check whether a current SBIR/STTR solicitation topic overlaps with this project's technology area, or contact the project directly (above) to ask.

None of these are guaranteed paths for this specific project — TechPort itself doesn't have an "apply" button. Reaching out to the contact(s) above with a specific question is usually the fastest way to find out what's actually open.