This dissertation presents advancements in high-definition (HD) simulation methods for packed-bed chromatography, focusing on simulation accuracy, workflow efficiency, and applicability to calibration of reduced-order models. Comprehensive updates to simulation infrastructure enhance simulation accuracy through targeted improvements in geometry representation, mesh generation, and numerical stabilization. Local and global strategies for contact-point modeling were explored, revealing that minimal global particle size reduction offered the best trade-off between mesh generation success and simulation reliability. A novel mesh generation technique leveraging geometric similarity among particles achieved an order-of-magnitude speedup with respect to generic techniques. A new stabilization method based on the contravariant metric tensor was applied to reduce relative holdup volume error by up to a factor of three. Anisotropic mesh refinement near particle and wall interfaces yielded significant reductions in element count while maintaining or improving accuracy. A key contribution is the development and application of a calibration framework for reduced-order models (ROMs), specifically 1D and 2D general rate models. HD simulation data for thin columns were spatially averaged and used to estimate dispersion and mass transfer coefficients. Results demonstrate excellent fits for polydisperse packings, while discrepancies in monodisperse packings highlight the limitations of ROM assumptions under strong geometric wall effects. The framework underscores the necessity of using internal state variables such as solute mass distributions for robust parameter estimation, especially in models with radial resolution. This establishes a foundation for ROM calibration using high-resolution simulations across diverse geometries. The simulation infrastructure was extended to support laterally periodic (unconfined) geometries, with new meshing capabilities and updates to the codebase of the numerical solver, XNS. Preliminary results from periodic HD simulations show a lack of wall-effect artifacts. However, particle loading was observed to be asymmetric owing to local nonlinearities in flow and transport within the interstitial domain. While full implementation of double-periodic boundary conditions remains ongoing, the current system demonstrates the potential to bridge the gap between HD and reduced-order modeling for realistic, large-scale chromatographic systems. This work lays the foundation for future efforts to accurately resolve dispersion phenomena via calibrated ROMs and scalable HD simulations of confined and unconfined columns.
Jayghosh Subodh Rao (Wed,) studied this question.