E2F2: End-to-End Feature-Free Network for Lithium-Ion Battery State of Health Prediction

With the development of new energy technologies, such as solar and wind power, lithium-ion batteries are being widely used for energy storage and distribution. The State-of-Health (SOH) is very important for the safety and reliability of battery applications. However, existing SOH prediction methods...

Full description

Saved in:
Bibliographic Details
Published in2024 6th International Conference on Electronic Engineering and Informatics (EEI) pp. 362 - 366
Main Authors Huang, Ziyue, Dang, Lujuan, Xing, Lei
Format Conference Proceeding
LanguageEnglish
Published IEEE 28.06.2024
Subjects
Accuracy
Adaptation models
battery
convolution
Data models
Feature extraction
Gated Recurrent Unit (GRU)
Lithium-ion batteries
Logic gates
Neural networks
Safety
state of health (SOH)
Training
Wind power generation
Online AccessGet full text

Cover

More Information
Summary:With the development of new energy technologies, such as solar and wind power, lithium-ion batteries are being widely used for energy storage and distribution. The State-of-Health (SOH) is very important for the safety and reliability of battery applications. However, existing SOH prediction methods usually require health feature extraction for model training, which limits the coherence and reliability of SOH prediction. To deal with this issue, we propose an End-to-End Feature-Free network (E2F2). The proposed model is an end-to-end neural network model that combines convolutional and Gated Recurrent Unit (GRU) architectures to process raw battery data, effectively capturing both spatial and temporal dependencies without relying on pre-defined features, thus simplifying modeling and enhancing adaptability. Experimental results on a real dataset illustrate the high accuracy and reliability of the proposed method.
DOI:10.1109/EEI63073.2024.10696695