A Comparative Study of Recurrent Neural Network Architectures for Battery Voltage Prediction 2021-01-1252
Electrification is the well-accepted solution to address carbon emissions and modernize vehicle controls. Batteries play a critical in the journey of electrification and modernization with battery voltage prediction as the foundation for safe and efficient operation. Due to its strong dependency on prior information, battery voltage was estimated with recurrent neural network methods in the recent literatures exploring a variety of deep learning techniques to estimate battery behaviors. In these studies, standard recurrent neural networks, gated recurrent units, and long-short term memory are popular neural network architectures under review. However, in most cases, each neural network architecture is individually assessed and therefore the knowledge about comparative study among three neural network architecture is limited. In addition, many literatures only studied either the dynamic voltage response or the voltage relaxation. This paper presents a comparative study on the battery voltage predication using all three neural network architectures. In this study, all neural network architectures use common pulse data for training and validation. The selected pulse data covers the voltage response in not only dynamic events but also during relaxation. Then the predictions are made in severe battery operation conditions such as high current, low temperature, and long duration. The results indicate the LSTM has the best prediction accuracy across various temperatures and current pulse sizes. All three neural network structures show the robustness at -30°C. But the standard RNN and GRU show the prediction error increases when the pulse size is higher.
Citation: Cho, G., ZHU, D., and Campbell, J., "A Comparative Study of Recurrent Neural Network Architectures for Battery Voltage Prediction," SAE Technical Paper 2021-01-1252, 2021, https://doi.org/10.4271/2021-01-1252. Download Citation
Author(s):
Gyouho Cho, DI ZHU, Jeffrey Campbell
Affiliated:
Ford Motor Co.
Pages: 8
Event:
SAE Powertrains, Fuels & Lubricants Digital Summit
ISSN:
0148-7191
e-ISSN:
2688-3627
Related Topics:
Neural networks
Machine learning
Batteries
Architecture
Research and development
SAE MOBILUS
Subscribers can view annotate, and download all of SAE's content.
Learn More »