The electric vehicle (EV) revolution has prompted automotive manufacturers to develop new battery cell and pack technology at an extremely rapid pace--dictating the need to employ modeling methods to become resilient to material delays and remove strain from experimental capability. Automotive manufacturers are concerned with how material, electrode, cell, and pack design decisions impact each other, some of which are hard to measure until after development. Therefore, multi-scale models are of particular interest to help improve the EV and battery pack development process. One phenomenon that can greatly impact design decisions is the volume change of active material upon lithiation, especially considering the promising materials with high volume change, like silicon. The active material volume change can lead to changes in the electrode's porosity and volume, impacting performance. The volume change of the electrode can lead to swelling of the full pouch cell (or pressure generated in a prismatic cell). And finally, the cell volume and pressure change can impact design decisions at the full module or pack scales. To address this, the authors have launched an industrially-led project focused on the development of a multi-scale, mechano-electrochemical battery model, with particular focus on automotive applications. Initially, theoretical models were developed to capture how volume change may impact performance in a single electrode[1] and dual insertion electrodes.[2] Then, applied models were developed to account for volume change impacts on the cell and pack scales[3], incorporate realistic, thermodynamically non-ideal active material volume change[4], and accurately simulate electrodes with blended active materials to account for preferential lithiation and unique volume change behavior.[5] In this presentation, the development of these models will be briefly summarized. Throughout the presentation, the discussion will have a particular emphasis on how each modeling topic may be useful to automotive manufacturers. Additionally, brief discussion will be provided on concepts that would improve the usefulness of this type of model, such as accounting for the impact of battery aging.[6,7] The authors thank the GM Global Propulsion System's team for computing resources and partial funding. This work was partially funded by the IGERT Program at University of South Carolina under NSF Award #1250052.
Read full abstract