Newswise — Researchers have developed a novel state-of-health (SOH) estimation model for lithium-ion batteries that integrates advanced signal processing and optimized machine learning techniques. The study demonstrates significant improvements in the accuracy and reliability of battery health assessments, crucial for various industries relying on lithium-ion technology.
Full text:
Researchers have developed a novel state-of-health (SOH) estimation model for lithium-ion batteries that integrates advanced signal processing and optimized machine learning techniques. The study, published in Frontiers in Energy, demonstrates significant improvements in the accuracy and reliability of battery health assessments, crucial for various industries relying on lithium-ion technology.
Lithium-ion batteries are central to modern technology, powering everything from smartphones to electric vehicles. Accurate SOH estimation is crucial for ensuring the longevity and safe operation of these batteries. However, the presence of cyclic aging noise in data poses significant challenges to current estimation models.
The research introduces a model that combines a whale optimization algorithm with variational modal decomposition (WOA-VMD) to efficiently process data and mitigate noise interference. The model further employs convolutional neural networks (CNN) for feature extraction and a support vector machine (SVM) for precise SOH estimation. Tested on a publicly available dataset, the model outperformed traditional techniques, enhancing estimation accuracy and generalization capabilities.
The model was validated using a dataset that includes different temperatures, discharge rates, and depths. This comprehensive testing ensures the model’s robustness across various real-world conditions, making it a versatile tool for industries using lithium-ion batteries.
This advancement is poised to significantly impact industries such as consumer electronics, automotive, and energy storage by improving battery life management and reliability. The model not only proposes a new standard for SOH estimation but also sets the stage for future research in optimizing battery performance through machine learning.
The research was supported by the Action Programme for Cultivation of Young and Middle-aged Teachers in Universities in Anhui Province, China; the Supporting Programme for Outstanding Young Talents in Colleges and Universities of Anhui Provincial Department of Education, China; the Huainan Normal University Scientific Research Project, China and the Key Projects of Huainan Normal University, China. For more detailed insights, the full study is available in Frontiers in Energy: https://journal.hep.com.cn/fie/EN/1157270154645213769.
link