Abstract:Lithium-ion batteries are pivotal to technological advancements in transportation, electronics, and clean energy storage. The optimal operation and safety of these batteries require proper and reliable estimation of battery capacities to monitor the state of health. Current methods for estimating the capacities fail to adequately account for long-term temporal dependencies of key variables (e.g., voltage, current, and temperature) associated with battery aging and degradation. In this study, we explore the usage of transformer networks to enhance the estimation of battery capacity. We develop a transformer-based battery capacity prediction model that accounts for both long-term and short-term patterns in battery data. Further, to tackle the data scarcity issue, data augmentation is used to increase the data size, which helps to improve the performance of the model. Our proposed method is validated with benchmark datasets. Simulation results show the effectiveness of data augmentation and the transformer network in improving the accuracy and robustness of battery capacity prediction.

What problem does this paper attempt to address?

The paper mainly targets the issue of lithium-ion battery capacity prediction, especially how to improve the prediction accuracy and the robustness of the model. Specifically, the research attempts to address the following key issues: 1. **Limitations of existing methods**: The current methods used for lithium-ion battery capacity estimation (including physics-based and data-driven approaches) are insufficient in handling the long-term temporal dependencies of key variables (such as voltage, current, and temperature) during the battery aging and degradation process. 2. **Data scarcity issue**: Data-driven approaches typically require a large amount of training data to achieve good performance, but obtaining a large volume of high-quality battery data is challenging in practice. To address the aforementioned issues, the paper proposes the following solutions: - **Adopting Transformer networks**: Utilizing the self-attention mechanism of Transformer networks to capture both long and short-term patterns in battery data, which helps to predict battery capacity more accurately. Compared to traditional recurrent neural networks (such as LSTM), Transformer networks are more effective in capturing long-term dependencies when processing sequence data. - **Data augmentation techniques**: To tackle the data scarcity issue, researchers have employed data augmentation techniques, generating additional data samples by adding Gaussian noise to the original dataset, thereby increasing the volume of data. This method helps to enhance the flexibility and robustness of the model, enabling it to better handle various changes and uncertainties that may occur in practical applications. Experimental validation on two standard datasets (NASA Group 1 and University of Michigan datasets) shows that the proposed method, which combines Transformer networks and data augmentation techniques, significantly improves the accuracy of lithium-ion battery capacity prediction, outperforming other common methods (such as LSTM and CNN).

Transformer-based Capacity Prediction for Lithium-ion Batteries with Data Augmentation

A Hybrid Battery Degradation Model Combining Arrhenius Equation and Neural Network for Capacity Prediction under Time-varying Operating Conditions

A Data Augmentation Method for Lithium-Ion Battery Capacity Estimation Based on Wassertein Time Generative Adversarial Network

Capacity estimation of Li-ion battery based on transformer-adversarial discriminative domain adaptation

Battery Health State Prediction Based on Singular Spectrum Analysis and Transformer Network

An Improved Transformer Model for Remaining Useful Life Prediction of Lithium-Ion Batteries under Random Charging and Discharging

Online State-of-Health Estimation for Fast-Charging Lithium-Ion Batteries Based on a Transformer–Long Short-Term Memory Neural Network

Opportunities and Challenges in Transformer Neural Networks for Battery State Estimation: Charge, Health, Lifetime, and Safety

Early prediction of remaining useful life for lithium-ion batteries based on CEEMDAN-transformer-DNN hybrid model

Enhancing lithium-ion battery lifespan early prediction using a multi-branch vision transformer model

A hybrid data-driven method for rapid prediction of lithium-ion battery capacity

Data-driven capacity estimation for lithium-ion batteries with feature matching based transfer learning method

A Data-Driven Hybrid Approach for Capacity Estimation on Lithium-ion Battery

A Battery Prognostics and Health Management Technique Based on Knee Critical Interval and Linear Complexity Self-Attention Transformer in Electric Vehicles

Fast Capacity Prediction of Lithium-Ion Batteries Using Aging Mechanism-Informed Bidirectional Long Short-Term Memory Network.

Capacity prediction of lithium-ion batteries with fusing aging information

Temporal Attention Mechanism Based Indirect Battery Capacity Prediction Combined with Health Feature Extraction

Prognosis of Multivariate Battery State of Performance and Health via Transformers

Predicting Capacity Fading Behaviors of Lithium Ion Batteries: an Electrochemical Protocol-Integrated Digital-Twin Solution

A Novel Fine-Tuning Model Based on Transfer Learning for Future Capacity Prediction of Lithium-Ion Batteries

A Hybrid Drive Method for Capacity Prediction of Lithium-Ion Batteries