Estimation of lithium-ion battery lifespan with deep learning
Loading...
Date
2026-07-10
Authors
Νέζης, Σταύρος
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
In the 21st century an increasing number of appliances, machines and critical applications rely on battery technology. Thus, it is very important to have methods that can be used for the prognostics of batteries. Accurate estimation of lithium-ion batteries is critical for ensuring reliability, safety, and cost-efficiency in electrical appliances, electric vehicles and energy storage systems. This thesis presents a deep learning framework for the early prediction of lithium-ion battery lifespan that can also be used online as the battery degrades. The proposed approach uses the well-known MATR dataset and leverages time-series data from different sensors to learn degradation patterns directly from cycle data. Different model architectures and learning paradigms were examined to test their performance on this task. The proposed approach achieves comparable performance as state-of-the art deep learning models while utilizing a simpler model with a comparatively small parameter count.
Description
Keywords
Machine learning, Lithium ion batteries, Rul estimation, Battery prognostics