Effect of drainage channel dimensions on the performance of wave-plate mist eliminators

We investigated the effects of drainage channel dimensions on droplet removal efficiency and pressure drop of the gas droplet flow in a wave-plate mist eliminator. Droplet dispersion in turbulent gas flows is numerically simulated using eddy interaction model (EIM) and Eulerian-Lagrangian method. Reynolds stress transport model (RSTM) with enhanced wall treatment and shear stress transport (SST) k-ω model are used for simulating the turbulent airflow. Comparison between the numerical simulations and available experimental data shows that eddy lifetime constant (C _L ) can affect the results significantly, and by selecting suitable values of the eddy lifetime constant, both turbulence models give reasonable predictions of droplet removal efficiency. Simulations of gas droplet flow in the eliminators with various drainage channel dimensions show that the drainage channel length (L _DC ) has a greater effect on droplet removal efficiency than the drainage channel width (W _DC ).     

Modeling and simulation of the injection of urea-water-solution for automotive SCR DeNOx-systems

The evaporation of water from a single droplet of urea water solution is investigated theoretically by a Rapid Mixing model and a Diffusion Limit model, which also considers droplet motion and variable properties of the solution. The Rapid Mixing model is then implemented into the commercial CFD code Fire 8.3 from AVL Corp. Therein, the urea water droplets are treated with Lagrangian particle tracking. The evaporation model is extended for droplet boiling and thermal decomposition of urea. CFD simulations of a SCR DeNOx-system are compared to experimental data to determine the kinetic parameters of the urea decomposition. The numerical model allows to simulate SCR exhaust system configurations to predict conversion and local distribution of the reducing agent.     

The influence of an anisotropic Langevin dispersion model on the prediction of micro- and nanoparticle deposition in wall-bounded turbulent flows

This paper deals with the issues of stochastic dispersion models and associated best practice responses for the investigation of micro- and nanoparticle deposition in turbulent flows. For such applications, Reynolds averaged turbulence models are widely used in combination with particle Lagrangian tracking, due to their relative simplicity and computational efficiency. Such approaches imply to generate the instantaneous velocity of the fluid at particle location to reproduce the effect of turbulence on particle transport. The default dispersion model used in most CFD codes is an eddy lifetime model, which frequently overestimates the deposition rates. In this work, a simple method is proposed to implement a three-dimensional stochastic dispersion model based on the Langevin equation in the Fluent^® commercial code. Comparisons are provided between this model, complemented by the simulation of Brownian effects, and available numerical data obtained using either an eddy lifetime model or a simple Langevin model. Computations are carried out in horizontal and vertical channel flows and in circular pipe flows as well. The use of the proposed anisotropic Langevin model is shown to improve the accuracy of deposition prediction in the whole range of particle inertia.     

Optimal design of drainage channel geometry parameters in vane demister liquid–gas separators

Vane liquid–gas demisters are widely used as one of the most efficient separators. To achieve higher liquid disposal and to avoid flooding, vanes are enhanced with drainage channels. In this research, the effects of drainage channel geometry parameters on the droplet removal efficiency have been investigated applying CFD techniques. The observed parameters are channel angle, channel height and channel length. The gas phase flow field was determined by the Eulerian method and the droplet flow field and trajectories were computed applying the Lagrangian method. The turbulent dispersion of the droplets was modeled using the discrete random walk (DRW) approach. The CFD simulation results indicate that by applying DRW model, the droplet separation efficiency predictions for small droplets are closer to the corresponding experimental data. The CFD simulation results showed that in the vane, enhanced with drainage channels, fewer low velocity sectors were observed in the gas flow field due to more turbulence. Consequently, the droplets had a higher chance of hitting the vane walls leading to higher separation efficiency. On the other hand, the parameters affect the liquid droplet trajectory leading to the changes in separation efficiency and hydrodynamic characteristic of the vane. To attain the overall optimum geometry of the drainage channel, all three geometry parameters were simultaneously studied employing 27 CFD simulation cases. To interpolate the overall optimal geometry a surface methodology method was used to fit the achieved CFD simulation data and finally a polynomial equation was proposed.     

CFD simulations of bubble column reactors: 1D, 2D and 3D approach

CFD simulations have been carried out for the predictions of flow pattern in bubble column reactors using 1D, 2D and 3D k-ε models. An attempt has been made to develop a complete correspondence between the operation of a real column and the simulation. Attention has been focused on the cylindrical bubble columns because of their widespread applications in the industry. All the models showed good agreement with the experimental data for axial liquid velocity and the fractional gas hold-up profiles. However, as regards to eddy diffusivity, only the 3D model predictions agree closely with the experimental data.The CFD model has been extended for the estimation of an axial dispersion coefficient (D_L) using 1D, 2D and 3D models. Excellent agreement was found only between the experimental values and the 3D predictions. The 1D and 2D simulations, however, yielded D_L values, which were lower by 25-50%. For this, a mechanistic explanation has been provided.     

Numerical simulations of airflow and droplet transport in a wave-plate mist eliminator

In this paper, the droplet transport and deposition in the turbulent airflow inside a wave-plate mist eliminator was studied using an Eulerian–Lagrangian computational method. The Reynolds Stress Transport Model (RSTM) with standard wall functions and with enhanced wall treatment was used for simulating the airflow field. A computer code for solving the Reynolds-averaged Navier–Stokes (RANS) equations in conjunction with the RSTM on two-dimensional collocated unstructured meshes was developed. For droplet trajectory analysis, another computer code was developed that accounts for the drag and lift forces action on the droplets. The Eddy Interaction Model (EIM) was used to model the droplet dispersion in turbulent airflow. The gas flow code was validated by comparing the computational model results for a fully developed asymmetric channel turbulent flow with the experimental data. Then the airflow and droplet trajectory analysis were performed for a mist eliminator with smooth walls and the resultant removal efficiency curves were evaluated and compared with the available experimental data. The results showed that the enhanced wall treatment improved the predictions of the droplet removal efficiency especially for small droplets in which the removal efficiency was lower than 50%. On the other hand, the Reynolds Stress Transport Model (RSTM) with standard wall functions cannot predict the removal efficiency correctly, especially for low gas velocities.     

Axial mixing in pipe flows: turbulent and transition regions

In the present work, the flow pattern in pipe flow has been simulated using low Reynolds number k-ε model. The CFD model has been extended to simulate the axial dispersion phenomena in both the transition and turbulent regions. An extensive comparison of the predicted axial dispersion coefficient with the experimental data has been presented along with the predictions of various models published in the literature. The proposed CFD model for axial mixing was found to give an excellent agreement with the experimental measurements.     

Challenges of Simulating Droplet Coalescence within a Spray

Abstract

This article reports various challenges that have been encountered in the process of developing validated Lagrangian and Eulerian models for simulating particle agglomeration within a spray dryer. These have included the challenges of accurately measuring droplet coalescence rates within a spray, and modeling properly the gas–droplet and droplet-droplet turbulence interactions. We have demonstrated the relative versatility and ease of implementation of the Lagrangian model compared with the Eulerian model, and the accuracy of both models for predicting turbulent dispersion of droplets and the turbulent flow-field within a simple jet system. The Lagrangian and Eulerian droplet coalescence predictions are consistent with each other, which implies that the numerical aspects of each simulation are handled properly, suggesting that either approach can be used with confidence for future spray modeling. However, it is clear that considerable research must be done in the area of particle turbulence modeling and accurate measurement of particle agglomeration rates before any Computational Fluid Dynamics tool can be employed to accurately predict particle agglomeration within a spray dryer.     

CFD simulation of large-scale bubble plumes: Comparisons against experiments

Computational fluid dynamics (CFD) simulations have been carried out to simulate a turbulent, bubble plume in a liquid pool. A two-fluid enhanced k-? model has been used, with the extra source terms introduced to account for the interaction between the bubbles and the liquid and transient calculations have been performed to study the plume growth, the acceleration of the liquid due to viscous drag, and the approach to steady-state conditions. In order to obtain correct spreading of the plume observed experimentally, it was observed that interfacial forces like lift and turbulent dispersion plays important role. The sensitivity analysis for drag coefficient and two different turbulent dispersion forces is presented. The development of the flow variables has been compared with the experimental data. From the CFD predictions obtained in the present work, it can be concluded that a two-dimensional (2D) axisymmetric assumption in this case is justified, with 2D model predictions in good agreement with the experimental data and those of a three-dimensional (3D) model, except for the shear stresses and turbulent kinetic energy. In general, quantitative comparison with the experimental data has revealed that, by applying proper models of inter-phase momentum transfer, and performing simulations based on the two-fluid model, satisfactory predictions of mean flow quantities can be obtained for this application away from the injector.     

Liquid phase axial mixing in bubble columns operated at high pressures

Liquid phase axial mixing was measured in a 100 mm i.d. bubble column operated in the pressure range of 0.1-0.5 MPa. Water, ethanol and 1-butanol were used as the liquid phase and nitrogen as the gas phase. The temperature and superficial gas velocity were varied in the range of 298-323 K and 0.01-0.21 m/s, respectively. The axial dispersion coefficient increased with an increase in the gas density due to pressure. The temperature had surprisingly a small effect. A CFD model was developed for the prediction of flow pattern in terms of mean velocity and eddy diffusivity profiles. The model was further extended for the prediction of residence time distribution and hence the axial dispersion coefficient (D_L). The predictions of axial dispersion coefficient agree favorably with all the experimental data collected in this work as well as published in the literature. The model was extended for different gas-liquid systems. The predicted values of axial dispersion coefficient were found to agree very well with all the experimental data.     

Challenges of Simulating Droplet Coalescence within a Spray

This article reports various challenges that have been encountered in the process of developing validated Lagrangian and Eulerian models for simulating particle agglomeration within a spray dryer. These have included the challenges of accurately measuring droplet coalescence rates within a spray, and modeling properly the gas-droplet and droplet-droplet turbulence interactions. We have demonstrated the relative versatility and ease of implementation of the Lagrangian model compared with the Eulerian model, and the accuracy of both models for predicting turbulent dispersion of droplets and the turbulent flow-field within a simple jet system. The Lagrangian and Eulerian droplet coalescence predictions are consistent with each other, which implies that the numerical aspects of each simulation are handled properly, suggesting that either approach can be used with confidence for future spray modeling. However, it is clear that considerable research must be done in the area of particle turbulence modeling and accurate measurement of particle agglomeration rates before any Computational Fluid Dynamics tool can be employed to accurately predict particle agglomeration within a spray dryer.     

Numerical investigation of bend and torus flows, part I : effect of swirl motion on flow structure in U-bend

As a preliminary study to flow modeling in torus reactors, simulations are carried out in well-known standard geometries, namely 90° and 180° bends. For the latter, two flow configurations are considered, with and without initial swirl motion so as to approach torus reactor conditions. Efficiency of the commercial CFD code FLUENT is investigated by comparing predicted results with experimental measurements available in the literature for both bended configurations. Different turbulent models and near-wall considerations are considered, including k-ε and high Reynolds-stresses models, with the standard wall-function approach as well as the two-layer zonal model and low-Rek-ε models.After validation of the numerical strategy, a parametric study is made to better understand the interactions between Dean vortices, involved by the bend curvature, and the main rotating motion generated by the swirl motion. Simulations are achieved for various values of the initial swirl intensity applied at the bend entry. Numerical simulations show different flow structures, resulting from the progressive Dean vortices perturbations with the increase of swirl intensity.     

Models for fuel droplet heating and evaporation: Comparative analysis

The results of comparative analysis of liquid and gas phase models for fuel droplets heating and evaporation, suitable for implementation into computational fluid dynamics (CFD) codes, are presented. Among liquid phase models, the analysis is focused on the model based on the assumption that the liquid thermal conductivity is infinitely large, and the so-called effective thermal conductivity model. Seven gas phase models are compared. These are six semi-theoretical models based on various assumptions and a model based merely on the approximation of experimental data. It is pointed out that the gas phase model, taking into account the finite thickness of the thermal boundary layer around the droplet, predicts the evaporation time closest to the one based on the approximation of experimental data. In most cases, the droplet evaporation time depends strongly on the choice of the gas phase model. The droplet surface temperature at the initial stage of heating and evaporation does not practically depend on the choice of the gas phase model, while the dependence of this temperature on the choice of the liquid phase model is strong. The direct comparison of the predictions of various gas models, with available experimental data referring to droplet heating and evaporation without break-up, leads to inconclusive results. The comparison of predictions of various liquid and gas phase models with the experimentally observed total ignition delay of n-heptane droplets without break-up, has shown that this delay depends strongly on the choice of the liquid phase model, but practically does not depend on the choice of the gas phase model. In the presence of droplet break-up processes, the evaporation time and the total ignition delay depend strongly on the choice of both gas and liquid phase models.     

Modelling of a RDC using a combined CFD-population balance approach

In the present study we propose an extension of the Euler/Lagrangian approach for liquid-liquid two phase flows when the volume fraction of the dispersed phase is not small. The continuous phase velocity is obtained by solving the Reynolds-averaged Navier-Stokes equations augmented with the k-ε turbulence model. The motion of the dispersed phase is calculated by solving the equations of motion taking into account inertia, drag and buoyancy forces. The coupling between the phases is described by momentum source terms and the terms that account for turbulence generation by the droplets’ motion. Collision and breakage of the droplets are treated by a single particle Monte-Carlo stochastic simulation method. This method is based on a mass flow formulation and operator splitting technique. For validation of the numerical procedure the droplet size distribution and flow fields in a rotating disc contactor are calculated and compared with the existing experimental results.     

Mixing effects on particle formation in supercritical fluids

The process termed solution enhanced dispersion by supercritical fluids (SEDS™) is investigated. In the process particles are created in the rapid antisolvent process using a twin-fluid nozzle to co-introduce the SCF antisolvent and solution. Results of experimental and numerical studies are presented for two regions of pressure: above the mixture critical pressure where a single-phase exists for all solvent–antisolvent compositions, and below the mixture critical pressure where the two-phase region is observed. In experimental studies paracetamol (in the single-phase system) and nicotinic acid (in the two-phase system) were precipitated from ethanol solution using supercritical CO₂ as an antisolvent. To interpret the phenomena affecting creation of the supersaturation and to predict suprsaturation distribution, balances of momentum (flow), species (mixing), energy (heating and cooling) and population (droplet and crystal size distributions) are applied. The Favre averaged k–? model of the CFD code FLUENT is applied together with specific models for precipitation subprocesses and Peng–Robinson equation of state. This includes application of the PDF closure procedure for precipitation and the drop breakage kernel that is based on multifractal theory of turbulence for modelling drop dispersion. Thermodynamic effects of mixing and decompression are included as well. Predicted values not always agree with experimental data but anyhow simulations predict all trends observed in experiments.     

Modeling fluid behavior and droplet interactions during liquid-liquid separation in hydrocyclones

A mechanical separation process in a hydrocyclone is described in which disperse water droplets are separated from a continuous diesel fuel phase. This separation process is influenced by droplet-droplet interaction effects like droplet breakup and coalescence resulting in a change of droplet size distribution. A simulation model is developed coupling the numerical solution of the flow field in the hydrocyclone based on computational fluid dynamics with population balances. The droplet size distribution is discretized and each discrete droplet size fraction is assumed to be an individual phase within a multiphase-mixture model. The droplet breakup and coalescence rates are defined as mass transfer rates between the discrete phases by the aid of user-defined functions. All model equations are solved with the CFD software package FLUENT™. The investigations show the impact of the cyclone geometry on the coupled population and separation dynamics. Cyclone separators with an optimized geometry show less steep velocity gradients increasing the coalescence rates and improving the separation efficiency. The calculated droplet size distributions at the cyclone overflow and at the underflow show good accordance with experimental data. The basic modeling approach can be extended and adapted to other disperse multiphase flow systems.     

CFD modeling of radial spreading of flow in trickle-bed reactors due to mechanical and capillary dispersion

CFD simulations of trickle-bed reactors are presented with radial spreading of the liquid due to mechanical and capillary dispersion. Simulations are performed with various particle sizes and the significance of the dispersion mechanisms at the industrially relevant particle size range is analyzed. The effect of the bed porosity distribution and particle size to the simulation results is also discussed. The choice of the radial porosity profile is found to have a significant impact to the simulation results, especially when the column to particle diameter ratio, D/d_p, is small, in which case the wall flow is important. The dependence of the standard deviation of porosity on the sample size is determined experimentally. Introducing just random variation of porosity to the model is found to describe inadequately the dispersive flow behavior. The presented hydrodynamic model with proper capillary and mechanical dispersion terms succeeds in capturing the features of the two independent physical phenomena. Separate models are presented for each dispersion mechanisms and it is shown that they both can have a significant contribution to the overall dispersion of liquid flowing through a packed bed. The hydrodynamic model is validated against the experimental dispersion profiles from Herskowitz and Smith [1978. Liquid distribution in trickle-bed reactors. A.I.Ch.E Journal 24, 739-454], Boyer et al. [2005. Study of liquid spreading from a point source in trickle-bed via gamma-ray tomography and CFD simulation. Chemical Engineering Science 60, 6279-6288] and Ravindra et al. [1997. Liquid flow texture in trickle-bed reactors: an experimental study. Industrial & Engineering Chemistry Research, 36, 5133-5145]. The extent of liquid dispersion predicted by the presented hydrodynamic model is in excellent agreement with the experiments.     

CFD based extraction column design——Chances and challenges

This paper shows that one-dimensional (1-D) [and three-dimensional (3-D) computational fluid dynamics (CFD)] simulations can replace the state-of-the-art usage of pseudo-homogeneous dispersion or back mixing models. This is based on standardized lab-scale cel experiments for the determination of droplet rise, breakage, coalescence and mass transfer parameters in addition to a limited number of additional mini-plant experiments with original fluids. Alternatively, the hydrodynamic parameters can also be derived using more sophisticated 3-D CFD simulations. Computational 1-D modeling served as a basis to replace pilot-plant experiments in any column geometry. The combination of 3-D CFD simulations with droplet population balance models (DPBM) increased the accuracy of the hydrodynamic simulations and gave information about the local droplet size. The high computational costs can be reduced by open source CFD codes when using a flexible mesh generation. First combined simulations using a three way coupled CFD/DPBM/mass-transfer solver pave the way for a safer design of industrial-sized columns, where no correlations are available.     

Experimental and computational study of the “end of the vortex” phenomenon in reverse‐flow centrifugal separators

The “end of the vortex” (EoV) phenomenon, a flow instability that plays a crucial role in cyclone design and operation is studied in this article. In the experimental part of the study, tests were carried out to understand the origin and nature of the EoV and to study the effects of the flowrate through, and the length of, the cylindrical cyclone on the EoV. In the theoretical part, computational fluid dynamics (CFD) models, in agreement with the geometrical configurations and operating conditions used in the present and earlier experimental studies, were constructed and investigated. Three‐dimensional simulations were carried out using the large eddy turbulence model with the commercial CFD package Star‐CD. Bending of the vortex core to the wall of the vessel and its precessional motion, constituting the phenomenon of the EoV, was observed in the simulations in most geometries. The results are in a good agreement not only with the present experimental results but also with previous experimental observations. © 2011 American Institute of Chemical Engineers AIChE J, 2012