Non-experimental methodology for developing pressure drop correlations for structured adsorbents with parallel channels

• Published in: Adsorption : journal of the International Adsorption Society Vol. 29; no. 1; pp. 29 - 43
• Main Authors: Mohammadi, Nima, Sanders, Ryan T., Holland, Charles E., Ebner, Armin D., Ritter, James A.
• Format: Journal Article
• Language: English
• Published: New York Springer US 2023; Springer Nature B.V
• Subjects: Adsorbents; Boussinesq equations; Channels; Chemistry; Chemistry and Materials Science; Computational fluid dynamics; Correlation; Engineering Thermodynamics; Experimental methods; Experiments; Friction factor; Heat and Mass Transfer; Industrial Chemistry/Chemical Engineering; Mathematical models; Methodology; Pressure drop; Standard deviation; Surfaces and Interfaces; Thin Films; Three dimensional models; Packed bed; Computational fluid dynamics; Fixed bed adsorption; Structured adsorbent; Pressure drop
• ISSN: 0929-5607; 1572-8757
• DOI: 10.1007/s10450-023-00374-2

A new non-experimental methodology based on a 3D Navier–Stokes (NS) computational fluid dynamics (CFD) model in lieu of experiments, is proposed for developing 1D axial pressure drop correlations for structured adsorbents with parallel channels. To demonstrate the methodology, a 1D correlation was developed by fitting and validating a Darcy–Weisbach (DW) type expression against the 3D-NS model. The methodology was further validated by contrasting the 1D-DW correlation against bench-scale experimental data obtained from a Catacel structured adsorbent with parallel triangular channels and against two 1D pressure drop correlations in the literature for parallel triangular channels. A wide range of velocities, pressures, channel dimensions and gas molecular weights were explored. To resolve the 1D-DW correlation, expressed in terms of a Darcy friction factor involving two fitting parameters f 1 and f 2 , an analytic expression derived from the differential 1D-DW model was simultaneously regressed with all the results obtained from the 3D-NS model using air at 25 °C. Then predictions from the differential 1D-DW correlation, now solved numerically in COMSOL 5.2, were contrasted against those from the 3D-NS model for the same conditions using CO 2 and He in addition to air at 25 °C. The 1D-DW correlation agreed well with the 3D-NS model, with the average and standard deviation of the average relative errors (AREs) being 1.79% ± 1.89%. The 1D-DW correlation also showed good agreement with experiment for all outlet pressures and gases. The average and standard deviation of the AREs were 10.96% ± 5.59%. Additionally, the 1D-DW correlation agreed well with the Boussinesq, and Shah and London correlations, with the average of the AREs relative to the 3D-NS model respectively being 2.00, 2.47 and 6.17%. These results validated the new non-experimental procedure for developing 1D axial pressure drop correlations from 3D CFD modeling in lieu of experiments. This new methodology is applicable to virtually any structured adsorbent shape with parallel channels and is especially advantageous where experiments might be problematic.