Research on precise lithium battery state of charge estimation method based on CALSE-LSTM model and pelican algorithm
Zujun Ding1, Daiming Hu1, Yang Jing1
1Huaiyin Institute of Technology, Huaiyin, Jiangsu, 223002, China.
Heliyon
|September 10, 2024
Summary
This study introduces CALSE-LSTM, a novel model for accurate lithium battery State of Charge (SoC) estimation. It significantly reduces estimation errors compared to existing methods.
Area of Science:
- Electrical Engineering
- Computer Science
- Materials Science
Background:
- Accurate State of Charge (SoC) estimation is crucial for lithium battery performance and safety.
- Existing models often struggle with complex battery dynamics and feature extraction.
Purpose of the Study:
- To develop an advanced fusion model, CALSE-LSTM, for optimizing lithium battery SoC estimation accuracy.
- To leverage dual-attention mechanisms and parameter fine-tuning for enhanced feature learning.
Main Methods:
- Integration of Convolutional Neural Networks (CNNs), Long Short-Term Memory Networks (LSTMs), and dual-attention mechanisms (self-attention and squeeze-excitation).
- Utilizing battery historical data and power consumption as input features.
- Parameter fine-tuning with the Pelican algorithm.
Main Results:
- CALSE-LSTM achieved a Root Mean Squared Error (RMSE) of 1.73% under Urban Dynamometer Driving Schedule (UDDS) conditions.
- Significant error reduction compared to GRU (31.9%), LSTM (31.3%), and CNN-LSTM (15%).
- Ablation studies confirmed the effectiveness of the dual-attention mechanism.
Conclusions:
- CALSE-LSTM demonstrates superior accuracy and efficiency in lithium battery SoC estimation.
- The dual-attention mechanism and inclusion of power consumption as a feature notably improve performance.
- The model shows strong potential for practical applications in battery management systems.
