A Novel Computational Approach in Modelling Tubular Fixed Bed Reactors: Aspects of Hydrodynamics and Heat Transfer Elyas M. Moghaddam 1 , Esmaeel A. Foumeny 2 , Andrzej Stankiewicz 1 , Johan T. Padding 1* 1 Process & Energy Department, TU Delft, Leeghwaterstraat 39, Delft, The Netherlands; 2 Petrochemical Commercial Company, No.2499, Vanak Sq. Valiasr Ave., Tehran, Iran *Corresponding author: j.t.padding@tudelft.nl Highlights • In narrow tubular fixed bed reactors, lateral heterogeneities emerge due to topological constraints imposed by the confining wall. • A novel approach based on Rigid Body Dynamics (RBD) is adopted to synthesize realistic random packing structures of spherical and cylindrical pellets with 3 < dt/dp < 8. • CFD simulations of hydrodynamics and heat transfer were performed and validated using literature data. • The CFD results demonstrate a remarkable influence of local flow maldistribution on the temperature field across the entire radius of narrow beds. 1. Introduction Fixed bed arrangements find wide applications, particularly in reaction engineering where they are employed as tubular catalytic reactors for the transformation of reactants into desired products. The design of such systems is predominantly rooted in pseudo-homogeneous models with effective parameters extracted from averaged semi-empirical correlations. This prevailing design procedure is inadequate for tubular fixed beds of low to moderate tube-to-particle ratios, say dt/dp ≤8, where the role of bed hydrodynamics in propagation of transport scalars is considerable [1,2]. This has persuaded researchers to seek for reliable design procedures, taking into account details of hydrodynamics, as well as the behavior of transport scalars at the pellet scale, in such narrow tubular fixed beds [3,4]. However, the majority of researchers have concentrated on rather simple packing arrangements of spherical pellets, based on inhouse codes, e.g. [5], or even commercial packages such as PFC 3D which is a commercial Discrete Element Method (DEM) package, coupling them with supplementary simulation tools such as Computational Fluid Dynamics (CFD) or Lattice Boltzmann (LB). Hence very few works have been devoted to packing of non-spherical pellets and the problem of heat transfer in such packings, e.g. [6]. This stems essentially from the intrinsic complexities connected to packing simulations of non-spherical catalyst pellets and very high computational expenses imposed in heat transfer studies as it is required to create a mesh inside the catalyst objects as well. The main emphasis of this contribution is centered on the behavior of flow hydrodynamics and lateral heat transfer at the pellet scale for random packing structures of spheres and equilateral solid cylinders. 2. Methods A novel approach based on Rigid Body Dynamics (RBD) is adopted to generate realistic random packing structures. RBD is an analytical hard-body scheme, capable of simulating the dynamic behaviour of assemblies of objects based on Newton’s laws of motion and Lagrangian mechanics. In this work, an inhouse code was developed to generate random packings of spheres and equilateral solid cylinders with 3 < dt/dpv < 8. CFD simulations of the flow field and heat transfer were then performed for some of the packing models in laminar, transitional and turbulent flow regimes, for 5 ≤ Rep ≤ 3,000, in which the problem of wall-to-bed heat transfer, viz. the wall-heated fixed bed problem, is resolved. The flow is assumed to be compressible and non-isothermal with physical properties of air, and the catalyst pellets are considered as alumina and glass with thermal conductivities of 40 and 1.01 W/mK, repectively. The commercial CFD code adopted here is ANSYS FLUENT V.14.5 in which the governing equations, including the equations of conservation