CFD-based studies of scaled-up adsorption columns

Adsorption is an actively researched process in chemical engineering with advances in various realms. Regarding numerical studies, most research has been focused on lab-scale and pilot-scale packed beds, leaving a gap in the understanding of flow dynamics within industry-scale systems. The present study attempts to address this gap by applying a multidimensional CFD model to industry-sized 2D geometries of adsorption columns. The model is validated using published experimental data of a pilot-scale column containing 41 kg of adsorbent. A column of 1.055 m in diameter and 2.395 m in length is considered along with three inflow distribution modes: plug, jet, and conical inflow. The effect of heat transfer through the wall is also studied by comparing the usage of a constant heat transfer coefficient ( h w ) against a physics-based h w formulation. The evolution across time and space of relevant adsorption variables like composition, temperature, and loading is analysed, where aspects such as irregular spatial distributions, front morphing, radial gradients, and the influence of heat transfer are explored. It is determined that the inflow mode does not have a permanent effect on species distribution at large scales, as the composition fronts revert to a plug-flow configuration at 42% and 32% of the column’s length for the jet and conical cases, respectively. However, the lack of proper radial distribution, along with convective cooling, induces irreversible asymmetries in the temperature fronts that affect adsorption at the latter stages of a DCB experiment. Finally, a constant h w is proven to be an acceptable modelling assumption, as it differs from the physics-based h w approach in less than 6.87% and has negligible impact on the overall adsorption dynamics. • Flow dynamics within three scaled-up geometries of adsorption columns are studied. • The validated CFD model shows accuracy and scalability across small and large setups. • The inflow mode does not permanently affect species distribution. • Convective cooling can irreversibly morph the posterior side of thermal fronts. • The fixed heat transfer approach has ∼ 6% discrepancy but maintains overall accuracy.