Related Experiment Video
Updated: Jul 5, 2026
06:37
Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
Published on: December 15, 2023
5.3K
State of Health for Lithium-Ion Batteries Based on Explainable Feature Fragments via Graph Attention Network and
Wenpeng Luan1, Hanju Cai1, Xiaohui Wang1
1School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, China.
Sensors (Basel, Switzerland)
|October 16, 2025
Summary
This study introduces a new method for accurately estimating lithium-ion battery health using interpretable features and advanced deep learning. The approach enhances safety and reliability for electric vehicle applications.
Area of Science:
- Materials Science
- Electrical Engineering
- Computer Science
Background:
- Accurate state of health (SoH) estimation is crucial for lithium-ion battery safety and performance.
- Current methods struggle with interpretable feature extraction and modeling spatial correlations.
- Addressing these limitations is key for reliable battery management systems.
Purpose of the Study:
- To develop a novel framework for interpretable lithium-ion battery SoH estimation.
- To improve the accuracy and reliability of battery degradation modeling.
- To enable practical deployment on embedded systems.
Main Methods:
- Utilized incremental capacity (IC) analysis to extract physically meaningful voltage and capacity fragments from charging data.
- Transformed features into graph-structured data for analysis.
- Employed a graph attention network (GAT) and bi-directional gated recurrent unit (Bi-GRU) with residual connections for spatial-temporal learning.
Main Results:
- Achieved superior performance on two benchmark datasets with an average Mean Absolute Error (MAE) of 0.561% and Root Mean Square Error (RMSE) of 0.783%.
- Demonstrated a low computational footprint (1.68 MFLOPs per inference).
- Confirmed fast inference time (17.55 ms) suitable for embedded platforms.
Conclusions:
- The proposed framework offers a highly accurate and interpretable approach to lithium-ion battery SoH estimation.
- The synergistic spatial-temporal learning effectively captures degradation trends.
- The model's efficiency and speed make it suitable for real-world applications, enhancing battery safety and longevity.
Related Concept Videos
01:19Ligand-gated Ion Channels
Ligand-gated ion channels are transmembrane proteins with a channel for ions to pass through and a binding site for a ligand. The channel opens only when a ligand attaches to the binding site.
Three Subfamilies of Ligand-gated Ion Channels
Ligand-gated ion channels fall into three subfamilies. The 'Cys-loop' includes the nicotinic acetylcholine receptors, γ-aminobutyric acid (GABA), glycine, and 5-hydroxytryptamine receptors. The second one is the 'Pore-loop' channels that include the...
Three Subfamilies of Ligand-gated Ion Channels
Ligand-gated ion channels fall into three subfamilies. The 'Cys-loop' includes the nicotinic acetylcholine receptors, γ-aminobutyric acid (GABA), glycine, and 5-hydroxytryptamine receptors. The second one is the 'Pore-loop' channels that include the...
01:30Ligand-Gated Ion Channel Receptor: Gating Mechanism
Ligand-gated ion channels are transmembrane proteins that play a vital role in intercellular communication and functions of the nervous system. They allow the influx of ions across the membrane once the neurotransmitter binds, allowing the subsequent transmission of electrical excitation across the neurons. Other ligand-gated ion channels, like the γ-aminobutyric acid (GABA) receptor, permit anions like chloride into the cells on the binding of the GABA molecule. Their entry into the cell...