Multi-cell model for pressure swing adsorption process

Pressure Swing Adsorption process is a discrete–continuous system by nature and it is extremely time consuming to simulate steady state performance for a given set of design and operating parameters. A multitude of design variations is offered by the configuration of Pressure Swing Adsorption cycle in terms of choice, sequence, and durations of various possible component steps implemented on two or more adsorber beds. Often, simplifying assumptions are made to speed up each simulation. These assumptions erode the quality of match between reality and simulation and make the resultant design approximate. Use of assumptions like no adsorption/desorption during the pressurization and blowdown steps, constancy in volumetric flow during the adsorption and purge steps makes the model computationally lighter but raises questions on its predictive power. A new modeling approach, namely Multi-cell Model is presented in this work. It is shown to avoid the extensive time taken with equation-based simulations and to have better predictive power. The model is used to study a representative Pressure Swing Adsorption process for nitrogen enrichment from the air. Numerical convergence with respect to the spatial and temporal step sizes and mass balance closure is verified. The model is generic in nature and is valid for any multi-bed, multi-adsorbent, multi-component Pressure Swing Adsorption process executing any combination of component steps.