<p>A 1D axial pressure drop correlation was developed for structured adsorbents with parallel channels of any cross-sectional shape simply in terms of its hydraulic diameter. It is based on the Darcy-Weisbach (DW) equation, which includes both laminar and inertial contributions to the pressure drop. It was created from pressure drop data generated from the new, non-experimental methodology developed recently by the authors that uses 3D Navier–Stokes (NS) computational fluid dynamics (CFD). To resolve the 1D-DW correlation, expressed in terms of a Darcy friction factor involving just two fitting parameters, <i>f</i><sub><i>1</i></sub> and <i>f</i><sub><i>2</i></sub>, an analytic expression derived from the differential 1D-DW model was regressed simultaneously with 300 3D-NS simulations of four different channel cross-sectional shapes (equilateral triangle, isosceles triangle, square, and rectangle) in air at 1&#xa0;atm and 25&#xa0;°C. Different hydraulic diameters created 30 different structures of these four shapes with each one evaluated at 10 different velocities. The values of <i>f</i><sub><i>1</i></sub> and <i>f</i><sub><i>2</i></sub> were determined to be 1.73e-2 and 47.68, respectively. Using the values of <i>f</i><sub><i>1</i></sub> and <i>f</i><sub><i>2</i></sub>, pressure drop predictions from the differential 1D-DW correlation, solved numerically, were contrasted against those from bench-scale experiments carried out with the triangular channel Catacel and trapezoidal channel Gore structured adsorbents at 0.5, 1 and 3&#xa0;atm for air, CO<sub>2</sub> and He at 25&#xa0;°C. The 1D-DW correlation showed very good agreement with experiment for all three outlet pressures and all three gases, and it provided much better predictions compared to two correlations in the literature, i.e., the Shah and London, and Muzychka and Yovanovich correlations. Overall averages of the respective percent average relative errors and the corresponding standard deviations were 4.3 ± 2.3, 14.1 ± 10.8 and 22.9 ± 11.1. These results demonstrated the 1D-DW correlation could be used with confidence in an adsorption process simulator to predict the pressure drop in a structured adsorbent with parallel channels of any cross-sectional shape of known hydraulic diameter.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

New universal pressure drop correlation for structured adsorbents with parallel channels of any cross-sectional shape

  • Ryan T. Sanders,
  • Armin D. Ebner,
  • Charles E. Holland,
  • Marjorie A. Nicholson,
  • James A. Ritter

摘要

A 1D axial pressure drop correlation was developed for structured adsorbents with parallel channels of any cross-sectional shape simply in terms of its hydraulic diameter. It is based on the Darcy-Weisbach (DW) equation, which includes both laminar and inertial contributions to the pressure drop. It was created from pressure drop data generated from the new, non-experimental methodology developed recently by the authors that uses 3D Navier–Stokes (NS) computational fluid dynamics (CFD). To resolve the 1D-DW correlation, expressed in terms of a Darcy friction factor involving just two fitting parameters, f1 and f2, an analytic expression derived from the differential 1D-DW model was regressed simultaneously with 300 3D-NS simulations of four different channel cross-sectional shapes (equilateral triangle, isosceles triangle, square, and rectangle) in air at 1 atm and 25 °C. Different hydraulic diameters created 30 different structures of these four shapes with each one evaluated at 10 different velocities. The values of f1 and f2 were determined to be 1.73e-2 and 47.68, respectively. Using the values of f1 and f2, pressure drop predictions from the differential 1D-DW correlation, solved numerically, were contrasted against those from bench-scale experiments carried out with the triangular channel Catacel and trapezoidal channel Gore structured adsorbents at 0.5, 1 and 3 atm for air, CO2 and He at 25 °C. The 1D-DW correlation showed very good agreement with experiment for all three outlet pressures and all three gases, and it provided much better predictions compared to two correlations in the literature, i.e., the Shah and London, and Muzychka and Yovanovich correlations. Overall averages of the respective percent average relative errors and the corresponding standard deviations were 4.3 ± 2.3, 14.1 ± 10.8 and 22.9 ± 11.1. These results demonstrated the 1D-DW correlation could be used with confidence in an adsorption process simulator to predict the pressure drop in a structured adsorbent with parallel channels of any cross-sectional shape of known hydraulic diameter.