Computational method for optimal electrolyte screening using Bayesian Optimization and physics based battery model
Year: 2023
Type: preprint
Abstract: Lithium-ion batteries (LIBs) are widely adopted in various applications, including electric vehicles and large-scale stationary energy storage, due to their high energy density and specific power. The performance of LIBs is significantly influenced by the composition of the liquid electrolyte. This study focuses on developing a framework to expedite the design and screening of electrolytes for LIBs by integrating physics-based electrochemical models with Bayesian optimization algorithms. The primary objective is to tune the transport properties of the electrolyte to simultaneously maximize cycle life and minimize internal heating, thereby enhancing battery performance. To support the framework, two electrolyte databases are created, for matching optimized conductivity profiles with existing electrolyte compositions. Additionally, a novel metric known as the deviation index is introduced to quantify the similarity between conductivity profiles. The effectiveness of the framework is validated through multiple case studies, showcasing its ability to optimize electrolyte properties. The proposed electrolyte composition optimization framework provides a systematic and efficient approach to screen and identify electrolytes with optimal properties, driving advancements in battery technology. Moreover, the framework’s versatility enables its adaptation and extension to different battery chemistries, facilitating broader applications and paving the way for sustainable and efficient energy storage systems.
DOI: 10.1149/osf.io/nuh87
Institution: Tata Consultancy Services (India)
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4386050870