Predicting Battery Capacity Loss on New Data Sets
About the Project
In a joint project with the Toyota Research Institute (TRI), we are working on expanding the research and work done in the paper, Data-driven prediction of battery cycle life before capacity degradation. Our first goal was to validate the models discovered in the paper with the original dataset, the second to run that same model on the Battery Archive dataset, and last to expand on the models built in the original paper.
A battery is considered degraded when it can only hold 80% of its maximum capacity when it was new. Using the BEEP (Battery Evaluation and Early Prediction) library developed by TRI, we featurize electrochemical measurements to predict the number of cycles in these batteries until degradation. We aim to accurately predict when the battery will degrade early in its lifetime, which enables faster research into new battery technologies.
Student Team
- David Barnett - Statistics major and data science minor
- Steven Taruc - Statistics major and data science minor
- Matthew Mazzagatte - Math major and data science minor
- Anish Yakkala - Statistics major and data science minor
Mentors
- Dr. Chirranjeevi Gopal, Client
- Dr. Patrick Herring, Client
- Dr. Joseph Montoya, Client
