Gunther Brenner

Numerical simulations of single phase reacting flows in randomly packed fixed-bed reactors and experimental validation

Abstract Randomly packed fixed-bed reactors are widely used in the chemical process industries. Their design is usually based on pseudo-homogeneous model equations with averaged semi-empirical parameters. However, this design concept fails for low tube-to-particle diameter ratios (=aspect ratios) where local phenomena dominate. The complete three-dimensional (3D) structure of the packing has therefore to be considered in order to resolve the local inhomogeneities. New numerical methods and the increase of computational power allow us to simulate in detail single phase reacting flows in such reactors, exclusively based on material properties and the 3D description of the geometry, thus without the use of semi-empirical data. The successive simulation steps (packing generation, fluid flow and species calculation) and their validation with experimental data are described in this paper. In order to synthetically generate realistic random packings of spherical particles, we apply a Monte-Carlo method. The subsequent numerical simulation of the 3D flow field and coupled mass transport of reacting species is done by means of lattice Boltzmann methods. The simulation results reveal that not only the local behaviour but also integral quantities like the pressure drop depend remarkably on the local structural properties of the packings, a feature which is neglected when using correlations with averaged values.

CFD-calculation of flow, dispersion and reaction in a catalyst filled tube by the lattice Boltzmann method

The behaviour of a reacting, viscous flow inside the complex geometry of a fixed-bed reactor has been studied by means of a lattice Boltzmann automata (LBA). In particular, two tasks have been investigated in detail. The geometrical structures of the fixed bed have been generated with a Monte-Carlo method. This allows to simulate very efficiently the placement of randomly packed spheres in a cylinder and to obtain detailed information of statistical properties, such as the voidage distribution. This geometrical information is the base for the subsequent numerical flow simulation using a lattice Boltzmann automata. This approach allows to predict the local fluid velocity distribution in the bed as well as the transport and reactions of chemical species taking into account the effect of heterogeneous catalysis. Thus, detailed information can be provided for the diffusion/dispersion effects around catalytic particles and ensembles of spheres. It is shown that this approach allows to improve the understanding of the underlying transport and reaction phenomena in these reactors and to obtain a reliable database for their operating behaviour and design.

Numerical analysis of the pressure drop in porous media flow with lattice Boltzmann (BGK) automata

The lattice Boltzmann (LB) method is used for a detailed study on the origins of the pressure drop in porous media flow. In agreement with the experimental results [Durst et al., J. Non-Newtonian Fluid Mech. 22 (1987) 169] it is shown, that the elongation and the contraction of fluid elements is an important factor for the pressure loss in porous media flow, and that a significant error is made, when only shear forces are taken into account. To obtain the geometry information of the porous media for our simulations, we used the 3D computer tomography technique.

Analysis of the flow field and pressure drop in fixed-bed reactors with the help of lattice Boltzmann simulations

The pressure drop of technical devices is a crucial property for their design and operation. In this paper, we show how the results of lattice Boltzmann simulations can be used in science and engineering to improve the physical understanding of the pressure drop and the flow inhomogeneities in porous media, especially in sphere-packed fixed-bed reactors with low aspect ratios. Commonly used pressure drop correlations are based on simplified assumptions such as the capillary or tortuosity model, which do not reflect all hydrodynamic effects. Consequently, empirical correlations for certain classes of media have been introduced in the past to bridge the gap between the models and the experimental findings. As is shown in this paper by the detailed analysis of the velocity field in the void space of packed beds, the pressure drop is due to more complex hydrodynamics than considered in the above-mentioned models. With the help of lattice Boltzmann simulations, we were able to analyse the different contributions to the total dissipation, namely shear and deformation of the fluid, for different geometries over a wide range of Reynolds numbers. We further show that the actual length of the flow paths changes considerably with the radial and circumferential position.

Numerical Simulation and Evaluation of Velocity Fluctuations During Rotating Stall of a Centrifugal Pump

Flow instabilities like rotating stall can lead to severe vibrations in turbomachines if the eigenfrequency of the rotor is equal to the stall frequency. The goal of the present work is to shed some light on the origin of the rotating stall phenomenon in a centrifugal pump. The resulting fluctuating loads are quantified using numerical computations. For the chosen configuration transient PIV data are available for validation. In addition to measuring the stall frequency in the stationary frame, the CFD data is analyzed in the rotating frame. A Fourier analysis is done for a large number of sample points. This enables us to determine the local variation of amplitudes for a given frequency. Together with eigenfrequencies and eigenmodes of the rotor determined from modal analysis, it is possible to evaluate the risk of resonance vibration excited from fluctuating fluid forces.

Combustion modeling of turbulent jet diffusion H2/air flame with detailed chemistry

Abstract In the present paper, turbulent jet diffusion flames are investigated numerically using a finite volume method for the solution of the Navier–Stokes and reaction equations governing the problem. The method is based on a finite volume discretization and the SIMPLE approach for velocity and pressure coupling. For validation of the modeling of turbulence and numerical method, results are shown for an inert turbulent jet flow. Different versions of the standard k–e turbulence model including the Rodi correction are compared with experimental results by Panchapakesan and Lumley. The focus is on the investigation of an axisymmetric turbulent hydrogen/air diffusion flame using a time-dependent numerical model with a detailed chemical mechanism. The chemical reactions are described by nine species and 16 or 17 pairs of elementary steps. The transport and thermodynamic physical properties for each species and gas mixture are obtained from the CHEMKIN-II package. An algebraic correlation closure (ACC) model is used for the coupling of turbulence and chemistry. The temperature and major species (H2, O2, H2O, N2) distributions are in good agreement with the experimental measurements. The numerical results obtained from the detailed chemistry calculations depend on how the turbulent diffusion coefficients are selected for the species and energy equations.

COMPARATIVE STUDY OF THERMAL FLOWS WITH DIFFERENT FINITE VOLUME AND LATTICE BOLTZMANN SCHEMES

In the present paper, a comparative study of numerical solutions for steady flows with heat transfer based on the finite volume method (FVM) and the relatively new lattice Boltzmann method (LBM) is presented. In the last years, the LB methods have challenged the classical FV methods to solve the Navier–Stokes equations and have proven to be superior in accuracy and efficiency for certain applications. Most of these studies were related to the transport of mass and momentum. In the meantime, significant effort has been invested in the application of the LBM to simulate flows including heat transfer. The studies in the present paper are the analysis of performance and accuracy aspects of LBM applied to the prediction of these flows. For a fully developed laminar flow between parallel plates, analytical solutions for the heat transfer in fully developed thermal boundary layers are available and may be compared with the respective numerical results. Finally, a hybrid approach is proposed to circumvent numerical problems of the thermal LB methods.

Monte Carlo simulation of a nephelometric experiment

Abstract A Monte Carlo method is applied to simulate the extinction of a laser beam collimated on a suspension of fly-ash particles. The objectives are to evaluate the method by comparing predicted angular distributions of the scattered intensities with measurements available in the literature, to investigate the influence of multiple scattering on these measurements and to estimate the sensitivity of the results on the complex refractive index of the medium. The Lorenz–Mie theory is used to model the radiative properties of the particles. The examined cases showed that their scattering behaviour is sensitive on the imaginary part of their refractive index, for which widely varying values are available in the literature. The proportion of multiple scattering in the total scattered intensity, as received in various angular positions, is calculated. That allows the estimation of angular regions where this proportion is minimised.

Numerical simulation of turbulent jet flow and combustion

Abstract The work applies the κ − ϵ turbulent model, with pressure boundary condition for the entrainment atmosphere surface, to calculate the steady free jet flow. It is solved by SIMPLE method using multi-grid solver to accelerate the computational speed. Based on the fulfillment of the above isothermal jet flow, combusting jet flows of diffusion flame and partial premixed flame are simulated, using the assumption of fast chemical reaction and Eddy-Dissipation-Concept (EDC) model, respectively. The numerical results are compared with experimental and theoretical results. The numerical results of isothermal jet and diffusion jet flame agree well with tests by Panchapakesan and Lumley, Lockwood and Moneib. It is found that the EDC model has some errors in modeling partial premixed jet flames.

Simulating fluid flow over sinusoidal surfaces using the lattice Boltzmann method

The objective of this paper is to analyze the laminar flow of a Newtonian fluid, pressurized and sheared, through a two-dimensional channel with one sinusoidal wall. The motivation of this work is the investigation of shear flows in lubrication. The effects of Reynolds number and the geometrical dimensions on the velocity distributions and on the flow factors are studied numerically using the lattice Boltzmann method (LBM). The results are verified with analytical and other numerical solutions.