Major project
Automotive, EV & Aerospace
Battery cell ageing model from lab cycling data
Predict how a lithium-ion cell loses capacity with each charge cycle and when it reaches end of life.
The brief
Battery makers test cells for months to learn how long they last, which is slow and costly. Using laboratory cycling data, build a model that predicts remaining capacity and end-of-life from the first few hundred cycles, and compare a physics-based approach with machine learning. Good looks like: end-of-life predicted with an error band, tests on cells the model has not seen, and an honest view of when the early prediction fails.
What you will build
- Load the cycling data for several cells
- Build capacity-fade curves
- Create early-cycle features
- Train a model to predict end of life
- Compare with a simple physics-based model
- Report error on unseen cells
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.