(251d) BEEP: Battery Estimation for Early Prediction of Long-Cycle Lifetime

Authors: 
Montoya, J., Toyota Research Institute
Herring, P., Toyota Research Institute
Aykol, M., TRI
Gopal, C., Toyota Research Institute
We present BEEP (Battery Estimation and Early Prediction), an open-source framework for management and processing of high-throughput battery cycling data streams. This software enables ingestion of raw cycling data and metadata from cell testing equipment, validation to ensure data integrity, and structuring of cell cycles into analytics-ready data structures. BEEP further enables featurization of structured cycling data to serve as input to machine learning, and include a number of ready-made models from existing data as well as a framework for creating new user-generated models. We demonstrate this pipeline's performance for training early-prediction models for cycle life. Thus, BEEP is shown to bridge the software and expertise gap between cell-level battery testing and data-driven battery development.