Searching for a Pulse: Evaluating the Use of Rapid DC Pulses for Diagnosing Battery Health, State-of-Charge, and Safety
Description
This is the data and code associated with the manuscript 'Searching for a pulse: Evaluating the Use of Rapid DC Pulses for Diagnosing Battery Health, State-of-Charge, and Safety', link: https://iopscience.iop.org/article/10.1149/1945-7111/addd50/meta. A collection of about 50,000 independent DC pulse sequences, collected from ~80 cells of 4 different types of commercially produced Li-ion batteries, with varying state-of-health (lab-aged, field-aged, and both), pulse state-of-charge, pulse design, pulse current, battery temperature, and battery hysteresis, while also measuring battery performance metrics relevant to a variety of battery applications (rate capability, mobility application, and stationary energy storage application). To diagnose safety, we have attempted to quantify potential safety and reliability issues using both electrochemical (post-charge voltage relaxation) and physical methods (rapid X-ray CT screening). We then use machine-learning methods to predict different target metrics using rapid pulse data and analyze results. Code with potential future updates to analysis is also at https://github.com/NREL/battery_pulse_diagnostics
Files
Institutions
- National Renewable Energy Laboratory
- University of Colorado Boulder
- Carnegie Mellon University
Categories
Funders
- United States Department of EnergyUnited States