Major project
Manufacturing, Robotics & Industry 4.0
Machine learning stand-in for a slow stress analysis
Train a model that predicts a part's peak stress from its dimensions in milliseconds instead of minutes.
The brief
A designer trying 500 dimensions cannot wait ten minutes for each finite-element run. Run a batch of analyses on a parametric part, train a model on the results, and use it to explore designs quickly, while measuring where it can and cannot be trusted. Good looks like: a sampling plan that covers the design space, errors reported against held-out analyses, and a clear warning outside the trained range.
What you will build
- Parametrise a simple part such as a plate with a hole
- Plan the sample designs and run the analyses
- Train a regression model on the results
- Test on held-out designs and extreme values
- Add an uncertainty or out-of-range warning
- Use it to find the best design
Free, open data to build it with
Sign up free to get the data and links
It takes a minute with your email. AiTeky Pass (₹999/month or ₹9,999/year) is free for your first 30 days.