Lithium Battery Enhancement Through Electrical Characterization and Optimization Using Deep Learning
Year: 2025
Type: article
Abstract: Currently, research on lithium-ion batteries has been driven by increased demand for electric vehicles to mitigate greenhouse gas emissions. Despite advances, batteries still face significant challenges in efficiency, lifetime, safety, and material optimization for their production. In this context, we present a Transformer-based deep learning model to predict lithium-ion batteries' charging and discharging behavior from their electrochemical parameters. A database of experiments characterized by their electrochemistry and their charge-discharge response was performed to achieve this goal. The database was generated by experimentation on a set of commercial batteries and new approaches in materials science. Subsequently, a model based on the Deep-Learning Transformers methodology was designed and trained from the database measurements. The results show that the deep-learning model can accurately predict lithium-ion batteries' charging and discharging behavior, even in complex scenarios such as abrupt changes in charging capacity. In conclusion, developing a Transformer deep-learning-based model to predict lithium-ion batteries' charging and discharging behavior is effective. These results significantly affect the battery industry and their optimization in electric vehicles.
Source: World Electric Vehicle Journal
DOI: 10.3390/wevj16030167
Language: English
Subfield: Automotive Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,689
Award: G2818852300
OpenAlex ID: W4408413622