Application note
Validating simulation-assisted cell design
eVTOL cell design, tested on 57 cells built by Avrion Battery Labs
A physics-based model screened thousands of electrode designs for two eVTOL missions with opposing requirements, and Avrion Battery Labs built 19 of them (57 cells) to test it. Before any cell was built, the model chose the direction of change for each mission, and the go-around design gained 21% in stack-specific energy over the baseline. One build loop corrected the model, which then predicted held-out designs to a 3.5% mean error and the gain between electrode loadings more accurately than the design spreadsheet.

What you'll learn
- How a design program changes when the builds validate a model instead of searching the design space
- Which model corrections one build loop found, and which of them transfer to designs the model was not fitted to
- How the corrected model compares with a design spreadsheet and data-driven predictors on held-out designs
- Where the model stops being accurate, and what the parameterization data must cover
- What one to two fewer build loops are worth per program
Who it's for
- Cell design engineers
- Modeling leads
- R&D and program managers

Dr. Stephen Glazier, CTO, Avrion Battery Labs
Collaborating with Ionworks enables our customers with tools and insight that support faster development and more predictable outcomes.Read the collaboration announcement