About this Abstract |
Meeting |
2021 TMS Annual Meeting & Exhibition
|
Symposium
|
Advanced Materials for Energy Conversion and Storage VII
|
Presentation Title |
Bio-inspired, Machine Learning-enabled Vascular Structures for Fast-Charging Lithium-ion Batteries |
Author(s) |
Po-Chun Hsu |
On-Site Speaker (Planned) |
Po-Chun Hsu |
Abstract Scope |
Vascular structures are ubiquitous in nature. Examples such as roots, blood vessels, and lung alveoli are the outcomes of millions of years of evolution to balance between surface area and mass transport. For lithium-ion batteries, vascular channels in the electrodes also enhance the kinetics by ensuring the fresh electrolyte supply to achieve fast charging. However, it is challenging to find the right parameters out of the immense parameter hyperspace. In this talk, I will explain how to use deep learning and finite element modeling to predict the battery behavior and, more importantly, to inverse-design the parameters of vascular electrode structure. We can arbitrarily choose the performance criteria among charge capacity, cycle life, maximum local temperature, and maximum local stress. We envision this research will significantly broaden our design hyperspace of high-rate energy storage devices and provide an effective approach for solving multiscale and multiphysics problems in reactive flow systems. |
Proceedings Inclusion? |
Planned: |
Keywords |
Energy Conversion and Storage, Machine Learning, Modeling and Simulation |