| About this Abstract |
| Meeting |
2022 TMS Annual Meeting & Exhibition
|
| Symposium
|
Advanced Materials for Energy Conversion and Storage 2022
|
| Presentation Title |
AI BMS Design with Sensor and ML Integration |
| Author(s) |
Alexey Y. Serov, Meghana Sudarshan, Surya Ayalasomayajula, Casey M. Jones, Vikas Tomar |
| On-Site Speaker (Planned) |
Alexey Y. Serov |
| Abstract Scope |
This work focuses on taking robust machine learning models built off of public datasets and applying the resulting predicted values and outputs directly on top of real-time operating battery management systems. A parasitic sensor network composed of a main microcontroller, a host CPU, resistance temperature detection devices, voltage and current measurements was created. The resulting network was integrated with a commercial battery management system. Real-time data for a simple 18650 battery pack with four cells in series was gathered. This real-time data stream was then actively used alongside public data-trained neural network algorithms for a robust and predictive “AI BMS”. Using the power of non-linear models to include battery health impacts not normally considered in battery management systems nor accounted for in common linear models, the sensor system in conjunction with a model trained on public datasets made predictive decisions about battery health characteristics on top of normal BMS operations. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Machine Learning, Modeling and Simulation, Other |