Closing the loop in Battery Can Manufacturing
AI-based defect prediction for spin-grooving
Download PDF (~19 MB)The problem
In spin-grooving, defects emerge from non-linear interactions between machine parameters, tool positioning, and material response. It is a multivariate chain that is hard to debug from final-part outcomes alone.
"The causal chain from parameter settings to force evolution to defect formation is managed through iterative adjustment and post-process inspection and not structured evaluation during setup."
Result
During validation, this approach reduced the observed defect rate from 30% to less than 3%. More broadly, physics-informed AI makes complex forming processes more predictable, less dependent on tacit expert knowledge, and more scalable across production steps in advanced manufacturing.
Want the full write-up, charts and all?
Download PDF (~19 MB)Stay in the loop
Get a short note when we publish the next case study. No marketing fluff, unsubscribe anytime.
Please enable marketing cookies so we can process your email.
Exploring a similar project?
Tell us about your process and the quality signals you're trying to act on. We'll show you where a physics-informed approach can close the loop.
Discuss your use case