(476f) Numerical Investigation of Packed Bed Reactors With Non-Spherical Particles

Introduction

Packed bed reactors are widely used in the chemical and process industry, amongst others for highly exothermic or endothermic catalytic reactions. Such reactors are characterized by a small tube to particle diameter ratio (D/d-ratio) to ensure a safe thermal management. For the design of such apparatuses the well-known correlations for packed bed reactors can not be used, because these reactors are dominated by the influence of the confining wall, which affects the porosity and the velocity field and as a result also the species and temperature distribution within the bed.

Due to the increasing computational power and the possibility of a detailed analysis of the transport phenomena in such reactors CFD (computational fluid dynamics) becomes more and more an indispensable tool. In the last years several authors have investigated mass, energy and species transport in spatially resolved packed beds mostly consisting of spheres, although non-spherical particles are much more common in industry.

Aim of this contribution

The aim of this contribution is to show firstly a fully computational work flow for the generation of packings consisting of non-spherical particles, meshing and calculation of the flow, temperature and species field and secondly to compare these packings in terms of radial porosity, velocity and temperature distribution.

Methods

In our work we have investigated packings of more complex particle shapes like cylinders and Raschig rings, which are often used in industry. The whole numerical process includes the modeling of these particles, the generation of the packed bed with the discrete element method (DEM), the meshing of the complex packing geometry as well as the CFD calculation which is done with the commercial tool STAR-CCM+ by CD-adapco.

For the generation of the packed bed DEM is used. DEM is based on solving Newton's Law for spherical particles. More complex particles like cylinders or Raschig rings can be modeled by combining several spheres to approximate the geometrical shape.

For the meshing of the packed bed for the CFD calculation the approximated particles are replaced with the original exact shapes. The meshing difficulties in the vicinity of the contact points of two particles or between the wall and the particle is handled by a method which is already published by the authors and was presented at AIChE annual meeting 2012: If the distance between two surfaces fall below a predefined value, the surface is modified locally resulting in a small gap. This gap can then be filled with cells of a reasonable quality and does not significantly affect the results.

Results

With the described methods we have investigated three different particle shapes (cylinder, Raschig ring, and 4 hole cylinder) with a D/d-ratio between 5.0 and 10.0 and a particle Reynolds number Re_p= 5 - 5000. All packings consists of up to 1500 single particles which is equal to a packing height of approximately 20d. Heat transport is simulated by applying a constant wall temperature.

The resulting packings are validated in terms of global bed porosity, radial and axial porosity profile with experimental results from literature. The results agree very well with these data. Where possible the CFD calculations are validated against literature data, too. The results are in a reasonable agreement for pressure drop and the radial velocity distribution.

Further investigations show, that the well-known dependence between the radial porosity profile and the radial velocity profile is not valid for Raschig rings. Additionally, the radial velocity is higher for all investigated particle shapes with a maximum for the cylindrical shapes compared to a packing of spheres. Furthermore the higher radial velocity leads to a higher convective radial heat transport which is beneficial for the thermal management of the reactor.