# Publications

Either browse through the list of published material or use the search below to limit by, year, or author or document type.

## 2009 |

Mohanarangam, Krishna; Stephens, Darrin W CFD Modeling of floating and settling phases in settling tanks Conference Seventh International Conference on CFD in the Minerals and Process Industries, 2009. Abstract | Links | BibTeX | Tags: ASM, Cenospheres, Floating, Modelling, Multiphase, Oil, Phases, PVC, SST, Tanks @conference{mohanarangam2009cfd, title = {CFD Modeling of floating and settling phases in settling tanks}, author = {Krishna Mohanarangam and Darrin W Stephens}, doi = {10.13140/RG.2.1.5078.7686}, year = {2009}, date = {2009-01-01}, booktitle = {Seventh International Conference on CFD in the Minerals and Process Industries}, journal = {Seventh International Conference on CFD in the Minerals and Process Industries}, pages = {9--11}, abstract = {A Computational Fluid Dynamics (CFD) model for modelling a floating phase has been developed and tested on a settling tank. The current model used for settling tanks is able to predict the settling of solids and the formation of a higher density layer of solids at the bottom of the vessel. Due to the widespread use of settling tanks in water and other chemical industries, floating phases (cenospheres, oil, PVC, etc) form a major part of the separation process. With this in mind, a model has been developed to incorporate both the settling as well as the floating of the secondary phases. The simulations were performed by customizing the commercially available software ANSYS-CFX (release 10.0). Multi-phase simulations were performed with clay, sand and a floating solid (density less than the continuous phase) as the secondary phases. Numerical instability was encountered in the volume fraction of the floating phase at the top boundary, where the floating phase collected, when using the unmodified version of ANSYS-CFX. This was mainly due to the volume fraction tending towards unity without any gradient at the top boundary. To prevent this happening, an extra term that is ignored in the CFX implementation was included in the slip velocity calculation. This essentially sets up a volume fraction gradient of the floating phase. Two variants of particle sizes for the floating phase were used to access this phenomenon. Contour plots of the floating phase volume fraction are presented within the feedwell as well in the cross-section of the tank to depict the preferential concentration of the phase. Further results are also shown for the settling solids. NOMENCLATURE C D drag co-efficient C µ k-ε turbulence model constant C ε1-2 k-ε turbulence model constant d diameter g acceleration due to gravity k Turbulence Kinetic Energy (T.K.E) p pressure r solids fraction Re Reynolds number Sc t Turbulent Schmidt number S1-S2 Diameter of floating species t time U velocity Y mass fraction x,y,z cartesion co-ordinate system ε turbulence dissipation rate μ dynamic viscosity eff}, keywords = {ASM, Cenospheres, Floating, Modelling, Multiphase, Oil, Phases, PVC, SST, Tanks}, pubstate = {published}, tppubtype = {conference} } A Computational Fluid Dynamics (CFD) model for modelling a floating phase has been developed and tested on a settling tank. The current model used for settling tanks is able to predict the settling of solids and the formation of a higher density layer of solids at the bottom of the vessel. Due to the widespread use of settling tanks in water and other chemical industries, floating phases (cenospheres, oil, PVC, etc) form a major part of the separation process. With this in mind, a model has been developed to incorporate both the settling as well as the floating of the secondary phases. The simulations were performed by customizing the commercially available software ANSYS-CFX (release 10.0). Multi-phase simulations were performed with clay, sand and a floating solid (density less than the continuous phase) as the secondary phases. Numerical instability was encountered in the volume fraction of the floating phase at the top boundary, where the floating phase collected, when using the unmodified version of ANSYS-CFX. This was mainly due to the volume fraction tending towards unity without any gradient at the top boundary. To prevent this happening, an extra term that is ignored in the CFX implementation was included in the slip velocity calculation. This essentially sets up a volume fraction gradient of the floating phase. Two variants of particle sizes for the floating phase were used to access this phenomenon. Contour plots of the floating phase volume fraction are presented within the feedwell as well in the cross-section of the tank to depict the preferential concentration of the phase. Further results are also shown for the settling solids. NOMENCLATURE C D drag co-efficient C µ k-ε turbulence model constant C ε1-2 k-ε turbulence model constant d diameter g acceleration due to gravity k Turbulence Kinetic Energy (T.K.E) p pressure r solids fraction Re Reynolds number Sc t Turbulent Schmidt number S1-S2 Diameter of floating species t time U velocity Y mass fraction x,y,z cartesion co-ordinate system ε turbulence dissipation rate μ dynamic viscosity eff |

Mohanarangam, Krishna; Nguyen, Tuan V; Stephens, Darrin W Evaluation of two equation turbulence models in a laboratory-scale thickener feedwell Conference Seventh International Conference on CFD in the Minerals and Process Industries, 2009. Abstract | Links | BibTeX | Tags: Feedwell, Model, Models, SST, Turbulence @conference{mohanarangam2009evaluation, title = {Evaluation of two equation turbulence models in a laboratory-scale thickener feedwell}, author = {Krishna Mohanarangam and Tuan V Nguyen and Darrin W Stephens}, doi = {10.13140/RG.2.1.1933.0407}, year = {2009}, date = {2009-01-01}, booktitle = {Seventh International Conference on CFD in the Minerals and Process Industries}, journal = {Seventh International Conference on Computational Fluid Dynamics in the Minerals and Process Industries}, pages = {9--11}, abstract = {Single phase modelling studies have been carried out using commercially available software ANSYS-CFX (release 11.0) on a laboratory scale thickener feedwell geometry. With the increase in complexity of feedwell and thickener geometries, meshing with a hexahedral mesh is time-consuming and sometimes impossible. The first objective of this study is to test the effectiveness of using tetrahedral/prism meshes in thickener feedwell geometries. Experimental results from a previously published lab-scale thickener feedwell geometry has been compared against the numerical predictions to verify the accuracy of these meshes towards replicating the flow structure. Mesh independency studies were also carried with these tetrahedral/prism meshes. The second objective is to test the suitability of four currently available two-equation turbulence models in our thickener feedwell geometry and their resulting flow structure. These turbulence models have been tested for open feedwell geometries with and without a shelf. NOMENCLATURE 1 a SST k-ω turbulence model constant B body forces c r1-3 curvature correction constant scale C curvature correction constant C ε1-2 k-ε turbulence model constant C μ k-ε turbulence model constant D rate of deformation 1 F First SST blending function 2 F Second SST blending function r f modified streamline curvature strength rotation f streamline curvature strength k turbulence kinetic energy k P shear production of turbulence kb P buoyancy production of turbulence p pressure p ' modified pressure * r curvature correction function r% curvature correction function S strain rate t time U velocity 3 α SST k-ω turbulence model constant β ′ SST k-ω turbulence model constant 3 β SST k-ω turbulence model constant ε turbulence dissipation rate μ dynamic viscosity eff μ effective viscosity t μ turbulent viscosity ρ density k σ k-ε turbulence model constant 3 k σ SST k-ω turbulence model constant ε σ k-ε turbulence model constant 2 ω σ SST k-ω turbulence model constant 3 ω σ SST k-ω turbulence model constant t υ kinematic turbulent viscosity Ω vorticity ω turbulence frequency Subscripts i, j, k velocity components INTRODUCTION Thickeners, as the name dictates, are used to concentrate fine particles from a slurry feed. Thickeners usually consist of a cylindrical feedwell surrounded concentrically by a large tank which forms the main body of the thickener. Slurry is fed into the feedwell along with a flocculant to induce the aggregation process under the turbulent conditions within the feedwell. Aggregates settle under gravity to produce a clear liquor collected from the outer edge of the upper surface of the thickener (overflow) and a concentrated underflow suspension of solids at the bottom of the tank. A slowly rotating rake is usually positioned at the base of the thickener to help move sediment out of the thickener for disposal or further processing. Industrial thickeners may be up to 100m in diameter, with feedwells up to 15m. The feedwell is core to the overall operational performance of a thickener. Feedwell use as a flocculation reactor is a relatively recent innovation, with the introduction of synthetic polymer flocculants in the 1960s. Feedwells also aid in dissipating the kinetic energy of the feed stream, helping to achieve uniform settling with minimum turbulence, and thereby reducing/eliminating short-circuiting in the thickener.}, keywords = {Feedwell, Model, Models, SST, Turbulence}, pubstate = {published}, tppubtype = {conference} } Single phase modelling studies have been carried out using commercially available software ANSYS-CFX (release 11.0) on a laboratory scale thickener feedwell geometry. With the increase in complexity of feedwell and thickener geometries, meshing with a hexahedral mesh is time-consuming and sometimes impossible. The first objective of this study is to test the effectiveness of using tetrahedral/prism meshes in thickener feedwell geometries. Experimental results from a previously published lab-scale thickener feedwell geometry has been compared against the numerical predictions to verify the accuracy of these meshes towards replicating the flow structure. Mesh independency studies were also carried with these tetrahedral/prism meshes. The second objective is to test the suitability of four currently available two-equation turbulence models in our thickener feedwell geometry and their resulting flow structure. These turbulence models have been tested for open feedwell geometries with and without a shelf. NOMENCLATURE 1 a SST k-ω turbulence model constant B body forces c r1-3 curvature correction constant scale C curvature correction constant C ε1-2 k-ε turbulence model constant C μ k-ε turbulence model constant D rate of deformation 1 F First SST blending function 2 F Second SST blending function r f modified streamline curvature strength rotation f streamline curvature strength k turbulence kinetic energy k P shear production of turbulence kb P buoyancy production of turbulence p pressure p ' modified pressure * r curvature correction function r% curvature correction function S strain rate t time U velocity 3 α SST k-ω turbulence model constant β ′ SST k-ω turbulence model constant 3 β SST k-ω turbulence model constant ε turbulence dissipation rate μ dynamic viscosity eff μ effective viscosity t μ turbulent viscosity ρ density k σ k-ε turbulence model constant 3 k σ SST k-ω turbulence model constant ε σ k-ε turbulence model constant 2 ω σ SST k-ω turbulence model constant 3 ω σ SST k-ω turbulence model constant t υ kinematic turbulent viscosity Ω vorticity ω turbulence frequency Subscripts i, j, k velocity components INTRODUCTION Thickeners, as the name dictates, are used to concentrate fine particles from a slurry feed. Thickeners usually consist of a cylindrical feedwell surrounded concentrically by a large tank which forms the main body of the thickener. Slurry is fed into the feedwell along with a flocculant to induce the aggregation process under the turbulent conditions within the feedwell. Aggregates settle under gravity to produce a clear liquor collected from the outer edge of the upper surface of the thickener (overflow) and a concentrated underflow suspension of solids at the bottom of the tank. A slowly rotating rake is usually positioned at the base of the thickener to help move sediment out of the thickener for disposal or further processing. Industrial thickeners may be up to 100m in diameter, with feedwells up to 15m. The feedwell is core to the overall operational performance of a thickener. Feedwell use as a flocculation reactor is a relatively recent innovation, with the introduction of synthetic polymer flocculants in the 1960s. Feedwells also aid in dissipating the kinetic energy of the feed stream, helping to achieve uniform settling with minimum turbulence, and thereby reducing/eliminating short-circuiting in the thickener. |