Industrial Applications of Granular Simulations

Abstract

The transition of lithium-ion battery manufacturing from empirical trial-and-error methods to theory-driven predictive modeling is essential for achieving the performance improvements and cost reductions required for next-generation energy storage. This thesis develops a granular simulation framework within the BIOVIA Materials Studio to investigate the evolution of electrode microstructures during the manufacturing process. By implementing multisphere rigid body approximations of prolate and oblate ellipsoids, the model achieves a high-fidelity representation of realistic active material particles observed in µCT and SEM imaging. Utilizing a geometry optimization-based approach instead of traditional Discrete Element Methods (DEM), the simulation workflow comprising slurry generation, drying, and calendering, demonstrates significant computational efficiency and scalability. Key results indicate that the predicted porosity values for uncalendered (0.469 ±0.002) and calendered (0.232 ±0.003) LIB cathode structures align closely with experimental benchmarks in existing literature. Furthermore, tortuosity analysis reveals that mechanical compaction induces structural anisotropy, significantly increasing the tortuosity factor for ionic transport in the calendering direction. In a secondary phase of research, this work successfully addressed inaccuracies in the COMPASS III forcefield for predicting phosphoric acid (H3PO4) properties. Through the selective adjustment of van der Waals parameters for double-bonded oxygen and central phosphorus atoms, the predicted crystal density and cell parameters were brought into excellent agreement with experimental values.