t e c h n i ca l

P A P ER

enhanced turbulence modeling in SOLIDWORKS FLOW SIMULATION

OVERVIEW
SolidWorks Flow Simulation is a powerful, intelligent, and easy-to-use computational fluid dynamics (CFD) program that facilitates the work of product engineers who use SolidWorks 3D CAD software for 3D design creation. This paper details the approach and theory behind turbulence modeling as used in SolidWorks Flow Simulation to simulate complex industrial turbulent flows with heat and mass transfer. Detailed explanations and examples will support how SolidWorks Flow Simulations turbulence model coupled with the immersed boundary Cartesian meshing techniques allow for accurate flow field resolution and enable Flow Simulation to offer a best-in-class CFD solution.

INTRODUCTION
Today, product lifecycle management concepts (PLM) are widely deployed by engineers in many industries as the means by which 3D manufactured product data are used and maintained consistently during an entire products lifecycle and across all its design changes. The basis of a PLM concept is the availability of high-quality, complete, detailed, and accurate 3D product model data within a CAD system as the central element. 3D product model data are therefore both the foundation and starting point for all virtual prototyping and physical simulations today. The use of fluid flow simulations (or Computational Fluid Dynamics, CFD) in such a CADembedded context is obviously very attractive as it can not only accelerate the design process, but it can make these processes more predictable and reliable, against a background of increasing design complexity and dependence on external development partners. The systematic search for the best solution for a design is the objective of most CFD simulations and CFD software packages. The main criterion for flow and heat transfer simulation as an integral part of a PLM concept is efficient turnaround of high-quality CFD solutions, from geometry changes to resultant engineering interpretation in order to keep pace with design changes. To make CFD usable for mechanical designers and design engineers from other engineering disciplines, SolidWorks Flow Simulation is the unique SolidWorks-embedded general purpose concurrent CFD software package largely automated to minimize the specialist expertise required to operate traditional CFD software. The capabilities required for CAD-embedded CFD to not only handle very complex geometries without simplification, but also to simulate complex industrial turbulent flows with heat and mass transfer is very important, together with benchmarking SolidWorks Flow Simulations turbulence capabilities against some classic industrial CFD validation cases. SolidWorks Flow Simulation is a mature code with over 10 years of commercial presence and a thousand man-years of development effort behind it (Gavriliouk, 1993). It utilizes a modified k-e two-equation turbulence model designed to simulate accurately a wide range of turbulence scenarios in association with its pioneering immersed boundary Cartesian meshing techniques that allow for accurate flow field resolution with low cell mesh densities. For more information on immersed boundary meshes see for instance Kalitzin and Iaccarino (2002) and their work at Stanford University with General Motors on this approach for generating fast turnaround times for CFD simulations in upfront engineering design studies. Depending on the fluid being examined and the flow conditions being considered, any fluid flow situation can be viewed (Schlichting, 1959) as one of the following: Laminar (a smooth flow without any disturbances) Turbulent (a flow regime characterized by random three-dimensional vorticity and intensive mixing) Transitional between laminar and turbulent (an alternation between laminar and turbulent flow regions) There are usually no difficulties involved with CFD codes in simulating laminar flows which have clear unique solutions. However, direct simulations of turbulent flows taking into account fluid volume fluctuations are practically impossible for industrial situations because of the small physical sizes involved and the wide spectra of velocity fluctuations that would require extremely fine computation meshes to resolve them, long CPU times to simulate them, and large computer memory to store the data produced. Hence, industrial turbulent fluid flows are simulated today
Enhanced Turbulence Modeling in SolidWorks Flow Simulation 2

usually by considering their effect on the time-average fluid flow characteristics in the volume being considered via semiempirical models of turbulence that close the fundamental Navier-Stokes equations being solved (Wilcox, 1994). The classical two-equation k-e empirical model for simulating turbulence effects in fluid flow CFD simulation (Launder & Spalding, 1972 and Wilcox, 1994) is still widely used and considered reliable for most industrial CFD simulations and it requires the minimum amount of additional information to calculate the flow field. In SolidWorks Flow Simulation the k-e model is used with a range of additional empirical enhancements added to cover a wide range of industrial turbulent flow scenarios (such as shear flows, rotational flows). For instance, damping functions proposed by Lam and Bremhorst (1981) for better boundary layer profile fit when resolving boundary layers with computational meshes have been added (the LB k-e model). In addition to turbulence modeling, when simulating flows it is also necessary to simulate fluid boundary layer effects near solid bodies or walls that can be difficult to resolve due to high velocity and temperature gradients across these near-wall layers. To solve the Navier-Stokes equations with a two-equation k-e turbulence model without resolving the near-wall fluid boundary layer would require a very fine computational mesh, hence a wall function approach had been proposed by Launder and Spalding (1972, 1974) to reduce mesh sizes. According to this now classic approach, the fluid wall frictional resistance and heat fluxes from the fluid to the wall are used to calculate the wall boundary conditions for solving the NavierStokes equations. Naturally, the main domain flows physical properties will be those of the boundary layers external boundary conditions. In SolidWorks Flow Simulation, Van Driests (1956) universal profiles are employed to describe turbulent boundary layers and two approaches (called Two-Scale Wall Functions, 2SWF) have been devised to fit a fluids boundary layer profile relative to the main flows properties: 1) When the fluid mass centers of the near-wall mesh cells are located inside the boundary layer, i.e., the physical fluid flow boundary layer is thick 2) When the fluid mass centers of the near-wall mesh cells are located outside the boundary layer, i.e., the physical fluid flow boundary layer is thin These two approaches allow SolidWorks Flow Simulation to overcome the traditional CFD code restriction of having to employ a very fine mesh density near the walls in the calculation domain and to use immersed boundary Cartesian meshes for all geometries (Kalitzin and Iaccarino, 2002).

Figure 1. Structure of the ETM approach used in SolidWorks Flow Simulation CFD software

Both the aforementioned SolidWorks Flow Simulation modification to the classical k-e turbulence model and the SolidWorks Flow Simulation modifications to the Launder-Spalding wall function approach for specifying the wall boundary conditions for the Navier-Stokes equations are called Enhanced Turbulence Modeling (ETM). Its structure is shown schematically in Figure1.

Enhanced Turbulence Modeling in SolidWorks Flow Simulation

1.The SolidWorks Flow Simulation Modified k- Turbulence Model

The modified k-e turbulence model with damping functions proposed by Lam and Bremhorst (1981) describes laminar, turbulent, and transitional flows of homogeneous fluids consisting of the following turbulence conservation laws:

Lam and Bremhorsts damping functions , 1, 2 decrease turbulent viscosity and turbulence energy and increase the turbulence dissipation rate when the Reynolds number Ry based on the average velocity of fluctuations and distance from the wall becomes too small. When = 1, 1 = 1, 2 = 1 the approach obtains the original k-e model.

2.Two-Scale Wall Functions

The Two-Scale Wall Functions (2SWF) in SolidWorks Flow Simulation consist of two approaches to coupling the boundary layer calculation with the main flow properties: 1)The thick-boundary-layer approach when d > A y 2)The thin-boundary-layer approach when d A y where is the boundary layer thickness, y is the distance from the near-wall computational mesh cells fluid mass center to the wall (Figure 2), and A 1 is a coefficient depending on flow conditions.

Enhanced Turbulence Modeling in SolidWorks Flow Simulation

Figure 2. Computational mesh cells near a wall, their fluid mass centers, and the distances from them to the wall (dashed lines)

2.1The Thick-Boundary-Layer Approach When a fluids boundary layer thickness, d > A y, where y is the distance from the near-wall computational mesh cells fluid mass center to the wall (Figure 2), and A 1 is a coefficient depending on flow conditions, that is, the near-wall computational mesh cells fluid mass center is located inside the boundary layer. Instead of y, SolidWorks Flow Simulation uses the dimensionless value, . Since the

computational mesh used in SolidWorks Flow Simulation is always an immersive boundary non-body-fitted Cartesian mesh, y+ of some near-wall cells could be very small (Figure 2). Hence, the corresponding dimensionless distance from the turbulent equilibrium regions outer boundary to the wall follows from an analysis of the experimental data presented by Wilcox (1994) and is equal to y+UP = 300. With this approach SolidWorks Flow Simulation obtains the momentum, heat flux, and turbulent boundary conditions for the Navier-Stokes equations which are described below. 2.1.1The Momentum Boundary Condition Since y+ < y+UP , and proceeding from the Van Driest mixing length (1956), the dimensionless longitudinal velocity in the boundary layer depends

Enhanced Turbulence Modeling in SolidWorks Flow Simulation

on the dimensionless y+ in the following manner:

, (2.1) where K = 0.4054 is the Karman constant, Av = 26 is the Van Driest coefficient. The wall shear stress being the momentum boundary condition governing the fluid velocity u at the near-wall computational mesh cells center is determined as ,(2.2) where K = K(y+) is a correction coefficient determined from equation (2.1) and the function u+ = u+ K(y+) . In SolidWorks Flow Simulation an equivalent sand wall roughness ks is taken into account by adding u+def(ks+) function in equation (2.1):

,(2.3) where u+def is a dimensionless function of the dimensionless roughness height , where ks is the physical roughness height. The u+def(ks+) function has been determined from the experimental data obtained in flow over a flat plate (Wilcox, 1994) and the boundary condition u+def(ks = 0) = 0 . Correspondingly, K = K(y+, ks+ ) . If , < 4000, where ue is the fluid flow velocity at the boundary layers

fluid boundary, i.e., at the distance of the boundary layer thickness d from the wall, the boundary layer is considered to be laminar, so K = 1. 2.1.2The Heat Flux Boundary Condition The heat flux across the boundary layer can be determined from

.(2.4)

Enhanced Turbulence Modeling in SolidWorks Flow Simulation

Since it is constant along the y direction, the heat flux from the fluid to the wall is the heat transfer boundary condition governing the fluid temperature T at the nearwall computational mesh cells center and it can be determined from

,(2.5) where K is described in the previous section, and from equation (2.4). is determined

In the same manner as in the previous section, if Red, the boundary layer is considered to be laminar, so K = 1, f = 1. 2.1.3Turbulence Boundary Conditions The turbulence boundary conditions for k and e depend on where the near-wall computational mesh cells fluid mass center is located: in the turbulence equilibrium region or nearer to the wall. If it is located in the turbulence equilibrium region, i.e., y+ > 30, then at the near-wall computational mesh cells fluid mass center

.(2.6) If y+ 30, i.e., the near-wall computational mesh cells fluid mass center is located in the turbulence nonequilibrium region, then at this cells fluid mass center , (2.7)

which had been determined from experimental data obtained on a plate (Lapin, 1982). 2.2The Thin-Boundary-Layer Approach In the thin-boundary-layer approach the Prandtl boundary layer equations already integrated along the normal-to-the-wall (i.e., along the y ordinate) from 0 (at the wall) to the boundary layer thickness d are solved along a fluid streamline near the wall. If the boundary layer is laminar, these equations are solved in SolidWorks Flow Simulation with a method of successive approximations based on the Shvetz trial functions technology (Ginzburg, 1970). If the boundary layer is turbulent or transitional (between laminar and turbulent), SolidWorks Flow Simulation uses a generalization of this method to such boundary layers employing the Van Driest hypothesis about the mixing length in turbulent boundary layers (1956). Three-dimensional effects of fluid flow over concave and convex surface walls, i.e., fluid spilling or collecting near such walls, are taken into account inside SolidWorks Flow Simulation through corresponding corrections for the wall curvature. Flow and boundary layer separations are also determined with a special method in

Enhanced Turbulence Modeling in SolidWorks Flow Simulation

SolidWorks Flow Simulation taking the near-zero wall shear stress into account. The equivalent sand wall roughness and the external flows turbulence on the boundary layer are taken into account through semiempirical coefficients correcting the wall shear stress and the heat flux from the fluid to the wall in SolidWorks Flow Simulation. Fluid compressibility, turbulence kinetic energy dissipation, and various mass forces are also taken into account through corresponding empirical and semiempirical models. From the boundary layer calculation SolidWorks Flow Simulation obtains the boundary layer thickness d, the wall shear stress te , and the heat flux from the w e fluid to the wall qw , which are used as boundary conditions for the Navier-Stokes equations, which are described below. 2.2.1 The Momentum and Heat Flux Boundary Conditions The momentum and heat flux boundary conditions are written as ,(2.8) 2.2.2 Turbulence Boundary Conditions Turbulence boundary conditions for k and e are determined from the condition of turbulence equilibrium in the near-wall computational mesh cell: (2.9) .

3. B  enchmark CFD Turbulence Validations with SolidWorks Flow Simulation

Several classical CFD turbulence benchmark examples are presented here to demonstrate the following: SolidWorks Flow Simulations accuracy in calculating a wide region of Reynolds numbers on the immersed boundary Cartesian computational meshes used in SolidWorks Flow Simulation SolidWorks Flow Simulation simulation results with refinement of the immersed boundary computational mesh 3.1Flow over a Flat Plate Consider an airflow of u=20 m/s velocity over a smooth flat plate of 0.5 m length (Figure 3).

Enhanced Turbulence Modeling in SolidWorks Flow Simulation

Figure 3. Schematic representation of boundary layer development on a flat plate

SolidWorks Flow Simulation predictions for a computational mesh of 20050 cells of the obtained fluid velocity field are shown in Figure 4.

Figure 4. Computational mesh and the resultant velocity distribution at z=0 (zoom-view)

CFD simulation results obtained in SolidWorks Flow Simulation on a similar computational mesh of 200100 cells with a varying incoming flows turbulence intensities of Tu=0.5%, 0.8%, 1.0%, 1.25%, 1.5%, 2.0% are presented in Figures 5, 6, and 7. In these calculations y+6 in the turbulent region, so the boundary layers laminar sublayer is not being resolved by the computational mesh. Taking the SolidWorks Flow Simulation-calculated local flat plate friction coefficient: (4.1) for these cases and plotting it versus the Reynolds number in

comparison with the Blasius semiempirical law (Schlichting, 1979) for the laminar boundary layer (on the plates inlet section): (4.2)

Enhanced Turbulence Modeling in SolidWorks Flow Simulation

and the Prandtl-Schlichting semiempirical law (Schlichting, 1979) for the turbulent boundary layer (on the plates following section). The results are presented in Figure 5: (4.3)

Figure 5. The smooth flat plates local friction coefficient calculated with SolidWorks Flow Simulation and for comparison with the Blasius and Prandtl-Schlichting formulae (Schlichting 1979)

In the SolidWorks Flow Simulation calculations the minimum Rex, at which the boundary layer is laminar for any incoming flow turbulence intensity, is close to the experimental value of Rex= 6104 (Schlichting, 1979), and the SolidWorks Flow Simulation-calculated maximum Rex , at which the boundary layer is turbulent at any incoming flow turbulence intensity, is close to the experimental value of Rex = 3106 (Schlichting, 1979).

Figure 6. The flat plate boundary layer velocity profiles calculated with SolidWorks Flow Simulation on the flat plate inlet section with laminar boundary layer (x=0.125 m), on the flat plate exit section with turbulent boundary layer (x=0.4375 m) in comparison with experimental data (Ref. 2) and semiempirical theory (Schlichting, 1979)

Enhanced Turbulence Modeling in SolidWorks Flow Simulation

10

The SolidWorks Flow Simulation-calculated local Nusselt number dependences Nux(Rex) at the different incoming flow turbulence intensities with the semiempirical dependences for airflows over flat plates (Pr=0.72) with laminar (4.4) and turbulent (4.5) boundary layers is shown in Figure 7: , (4.4) , (4.5)

Figure 7. The local Nusselt number of airflow over a smooth flat plate vs. calculated with SolidWorks Flow Simulation and, for comparison, with the semiempirical equations (4.4) and (4.5) (Lienhard and Lienhard, 2004)

3.2Couette flow between Two Parallel Flat Plates at Re=3.4 x104 A classical plane flow is one between two parallel infinite flat plates spaced at a distance h from one another and moving at velocity U in opposite directions (Figure 8), resulting in the flow Reynolds number, Re=rUh . =3.4x104

Figure 8. A turbulent Couette flow between two parallel plates moving in opposite directions

Enhanced Turbulence Modeling in SolidWorks Flow Simulation

11

Dimensionless velocity profiles calculated within SolidWorks Flow Simulation for different computational meshes (10, 20, 40, 80 mesh cells across the channel) in comparison with experimental data (Schlichting, 1979) are shown in Figure 9.

Figure 9. SolidWorks Flow Simulation-calculated dimensionless velocity profiles of turbulent Couette flow between two parallel plates moving in opposite directions in comparison with the experimental data (Schlichting, 1979)

We can see good agreement between the SolidWorks Flow Simulation calculations with the experimental data: y+ =75 for the 10-cells mesh, y+ =37.5 for the 20-cells mesh, y+ =19 for the 40-cells mesh, and y+ =9.5 for the 80 cells mesh showing how robust the solver is with coarse meshes. 3.3Flow over a Backward-Facing Step at Re=5,000 Consider airflow in a rectangular channel having parallel walls (Hirsch, 1988), 1.0 m length 15.1 cm width Y0=10.1 cm height as the inlet section, and an H=1.27 cm height backward-facing step on the bottom of the domain (Figure 10). Due to the channels large aspect ratio (the tunnel-width to step-height ratio is equal to 12), three-dimensional effects in the flow separation region downstream of the backward-facing step are minimal, and due to the channels small expansion ratio (Y0 + H)/Y0 = 1.125, the pressure gradient downstream of the sudden expansion is also at a minimum.

Enhanced Turbulence Modeling in SolidWorks Flow Simulation

12

At a distance 4H upstream of the backward-facing step, the inlet airflow has a velocity of 44.2 m/s, the atmospheric total pressure and temperature this flows Mach number is equal to 0.128and the fully turbulent boundary layer of 1.9 cm thickness provided by the high Reynolds number of 5,000 (based on the boundary layer momentum thickness).

Figure 10. Flow in a rectangular channel with parallel walls and a backward-facing step on the bottom

SolidWorks Flow Simulation calculations of this flow were performed on uniform computational meshes having 10, 20, and 40 cells per the step height (Figure 11).

Figure 11. The SolidWorks Flow Simulation computational meshes of different density used: the number of cells traversing the step height: a) 10 cells, b) 20 cells, c) 40 cells

Enhanced Turbulence Modeling in SolidWorks Flow Simulation

13

SolidWorks Flow Simulation-calculated flow velocity profiles, wall friction downstream of the backward-facing step, and the separated flow reattachments distance from the backward-facing step are presented in Figs.12 and 13 and in Table 1. In these calculations y+ = 0 up to 54 in the computational mesh of 10 cells per step height, H, y+ = 0 up to 27 at the computational mesh of 20 cells per H, and y+ = 0 up to 13 at the computational mesh of 40 cells per H. Good convergence of these SolidWorks Flow Simulation calculation results with the computational mesh refinements chosen and good agreement of these results with the experimental data (Driver and Seegmiller, 1985) and calculations (Wilcox, 1994).

Figure 12. SolidWorks Flow Simulation-calculated flow velocity profiles obtained with the computational mesh of 10 cells per step height in comparison with experimental data (Driver and Seegmiller, 1985)

Enhanced Turbulence Modeling in SolidWorks Flow Simulation

14

Figure 13. The SolidWorks Flow Simulation-calculated wall friction coefficient of the flow past the back step in comparison with experimental data (Driver and Seegmiller, 1985)

Table 1 SolidWorks Flow Simulation-calculated distance of the separated flow reattachment point downstream of the backward facing step in comparison with experimental data (Driver and Seegmiller 1985) and calculations (Wilcox 1994).

x/H

Deviation from experimental data (%) 0 -13 -3 1 -16 3

Experiment (Driver and Seegmiller 1985) SolidWorks Flow Simulation, 10 cells per H SolidWorks Flow Simulation, 20 cells per H SolidWorks Flow Simulation, 40 cells per H Calculation with the k-e model (Wilcox 1994) Calculation with the k-w model (Wilcox 1994)

6.2 5.37 6.00 6.27 5.2 6.4

3.4Flow over a Cylinder at Re=3.7x105 Consider unsteady airflow over a circular cylinder at the incoming airflow velocity of 153 m/s under normal pressure and temperature at the cylinder-diameter-based Reynolds number of Re=3.7x105. SolidWorks Flow Simulation calculations were then carried out with computational mesh densities having 20, 40, and 80 cells per cylinder diameter. In all these calculations the thin-boundary-layer turbulence model was employed. The computational mesh and results for 20 cells per cylinder diameter is shown in Figure 14a. SolidWorks Flow Simulation-calculated velocity fields and an experimental shadowgraph (Driver and Seegmiller 1985) of real flow over a circular cylinder at nearly the same Reynolds number are shown in Figure 14.

Enhanced Turbulence Modeling in SolidWorks Flow Simulation

15

Figure 14. Predicted flow velocity fields over a circular cylinder, calculated with SolidWorks Flow Simulation for different computational meshes having a) 20 cells per diameter, b) 40 cells per diameter, c) 80 cells per diameter, and, d) similar real flow shadowgraph from Driver and Seegmiller (1985)

The drag coefficient, Cd, of the cylinder calculated with SolidWorks Flow Simulation for the different computational meshes in comparison with experimental data (Driver and Seegmiller 1985), as well as the corresponding y+ values, are presented in Table 2.

Table 2 Circular cylinder drag coefficients calculated with SolidWorks Flow Simulation in comparison with experimental data (Driver and Seegmiller 1985).

Mesh density (cells per cylinder diameter) SolidWorks Flow Simulation, 20 cells per diameter SolidWorks Flow Simulation, 40 cells per diameter SolidWorks Flow Simulation, 80 cells per diameter Experiment (Driver and Seegmiller 1985)

Cd

Deviation from the Ref.12 experimental data (%) -18 -5 2 0

y+max

0.82 0.95 1.02 1.0

650 330 170 n/a

The SolidWorks Flow Simulation simulation predictions for varying computational mesh densities shows very good agreement with the experimental data from Driver and Seegmiller (1985).

Enhanced Turbulence Modeling in SolidWorks Flow Simulation

16

3.5 Flow over a Generic Car Body Shape (the Ahmed Body) A classical automotive external aerodynamics wind tunnel test case is the so-called Ahmed Body (Lienhart, Stoots, and Becker, 2000) which has a curved, chamfered front end, box-like main body and sloping rear section. Using SolidWorks Flow Simulation, an approaching airflow of 40 m/s on the generic model car body of 1,044 mm length, 389 mm width, and 288 mm height and mounted in a wind tunnel of 1,870 mm 1,400 mm cross-section at a 50 mm height above the floor on 4 stilts of 30 mm diameter was createdsee Figure 16. The model car bodys slanting rear section of 222 mm length was modeled with different (35 and 25) slant angles. In the experimental study, flow had started to separate from the bodys slanting surface at an angle of 30. The flows Reynolds number based on the body height and the incoming flow velocity is equal to Re = 7.68x105. SolidWorks Flow Simulation calculations were performed with a computational mesh of 209 cells in length, 58 cells in height, and 78 cells in width to resolve the car body (Figure 15).

Figure 15. The SolidWorks Flow Simulation computational mesh over the model car body: a) the 25 rear slant, b) the 35 rear slant

SolidWorks Flow Simulation-calculated flow streamlines and velocity contours upstream, over, and downstream of the model car body are shown in Figure 16 for the two sloping rear slant angles. The SolidWorks Flow Simulation-calculated flow velocity profiles and body drag coefficients in comparison with the experimental ones (Lienhart, Stoots, and Becker, 2000) are shown in Figure 17 and Table 3. It can be seen from Figs.16 and 17 that in the SolidWorks Flow Simulation calculations the flow over the body is attached to the bodys slanting rear surface if it has the 25 rear slant angle and separates from this surface if it has the 35 angle, and the calculated flow velocity profiles are close to the experimental ones. From Table 3 it is observed that the SolidWorks Flow Simulation-calculated body drag coefficients agree well with the experimental ones.

Enhanced Turbulence Modeling in SolidWorks Flow Simulation

17

Figure 16. SolidWorks Flow Simulation-calculated flow streamlines and velocity contours upstream, over, and downstream of the model car body: a) the 25 body slant, b) the 35 body slant

Figure 17. SolidWorks Flow Simulation-calculated flow velocity profiles in the bodys symmetry plane in comparison with experimental data at the bodys slant angles of 25 and 35 degrees: SolidWorks Flow Simulation calculation, experiment (Lienhart, Stoots, and Becker, 2000)

Enhanced Turbulence Modeling in SolidWorks Flow Simulation

18

Table 3 The model car bodys drag coefficient calculated with SolidWorks Flow Simulation and obtained in wind tunnel experiments (Lienhart, Stoots, and Becker, 2000)

Ahmed bodys slant angle

Cd, exp

Cd, SolidWorks

Deviation from the Ref. 13 experimental data (%) -4.8 6.6

y+max on the slanted surface 1401 1402

25 35

0.298 0.257

0.284 0.274

Conclusions
The general purpose SolidWorks-embedded CFD solver in the SolidWorks Flow Simulation software package has been benchmarked against a wide range of CFD turbulence cases and its two-equation modified k - e turbulence model with its unique two-scale wall functions approach and immersed boundary Cartesian meshes leads to good predictions for spatial laminar, turbulent, and transitional flows over a range of compressible and anisotropic flows. Boundary layer resolution was good in all cases even though mesh densities were varied and coarse by traditional CFD approaches and the determination of wall friction and heat fluxes from the fluid flow to the wall interface over a wide range of Reynolds numbers was excellent.

References
Driver, D.M. and Seegmiller, H.L., 1985. Features of a Reattaching Turbulent Shear Layer in Divergent Channel Flow. AIAA Journal, Vol. 23, p. 163 Gavriliouk, V.N., Denisov, O.P., Nakonechny, V.P., Odintsov, E.V., Sergienko, A.A., Sobachkin, A.A., 1993. Numerical Simulation of Working Processes in Rocket Engine Combustion Chamber. 44th Congress of the International Astronautical Federation, IAF-93-S.2.463, October 16-22, Graz, Austria Ginzburg, I. P., 1970. Theory of Drag and Heat Transfer. Leningrad, LGU (in Russian) Griffiths, W.D., and Boysan, F., 1996. Computational fluid dynamics (CFD) and empirical modelling of the performance of a number of cyclone samplers, Journal of Aerosol Science, Vol. 27, No. 2, pp. 281-304 Hirsch, C., 1988. Numerical Computation of Internal and External Flows, Volume I, Fundamentals of Numerical Discretization. New York: John Wiley & Sons Kalitzin, G., and Iaccarino, G., 2002. Turbulence Modeling in an Immersed-Boundary RANS-Method, Center for Turbulence Research Annual Research Briefs, Stanford University, California, pp. 415 - 426 Lam, C.K.G. and Bremhorst, K.A., 1981. Modified Form of Model for Predicting Wall Turbulence. ASME Journal of Fluids Engineering, Vol. 103, pp. 456-460 Lapin, Y.V., 1982. Turbulent Boundary Layer in Supersonic Gas Flows. Moscow, Nauka, (in Russian) Launder, B.E. and Spalding, D.B., 1972. Lectures in Mathematical Models of Turbulence. Academic Press, London, England. Launder, B. E. and Spalding, D. B. The Numerical Computation of Turbulent Flows. Computer Methods in Applied Mechanics and Engineering, Vol. 3, 1974, pp. 269-289

Enhanced Turbulence Modeling in SolidWorks Flow Simulation

19

Lienhard IV, J.H. and Lienhard V, J.H., 2004. A Heat Transfer Textbook. 3rd ed., Cambridge, MA: Phlogiston Press Lienhart, H., Stoots, C., Becker, S., 2000. Flow and turbulence structures in the wake of a simplified car model (Ahmed model). DGLR Fach Symp. der AG STAB, Stuttgart University Mentor Graphics Corp., 2011. Advanced Immersed Boundary Cartesian Meshing Technology in SolidWorks Flow Simulation Schlichting, H., 1959. Entstehung der Turbulenz. Berlin (in German) Schlichting, H., 1979. Boundary-Layer Theory. McGraw-Hill, New York Van Driest, E.R., 1956. On Turbulent Flow Near a Wall. Journal of the Aeronautical Sciences, Vol. 23, No. 10, p. 1007 Wilcox, D.C., 1994. Turbulence Modeling for CFD. DCW Industries

Nomenclature
ui i-th component of the fluid velocity vector r fluid density k turbulence energy e dissipation rate of turbulence energy fluid viscosity t fluid turbulent viscosity tij ij-th component of the laminar stress tensor tRij ij-th component of the Reynolds stress tensor Pr the dimensionless Prandtl number Prt the dimensionless turbulent Prandtl number Cp fluid-specific heat capacity under constant pressure T temperature xi i-th component of the Cartesian coordinate system ni i-th component of the normal-to-the-wall in the fluid region y distance from the wall along the normal to it y+ dimensionless distance from the wall along the normal to it tw wall shear stress qw heat flux from the wall to the fluid

Enhanced Turbulence Modeling in SolidWorks Flow Simulation

20

k the Karman factor Av the Van Driest dumping factor ks equivalent sand roughness height d boundary layer thickness calculated by the integral method ue fluid velocity at the fluid boundary of the boundary layer tew wall friction calculated by the integral method qew heat flux from the wall to the fluid, calculated by the integral method

Subscripts
l, j, k directions of the Cartesian coordinate system w at the wall + parameter is dimensionless on the basis of wall friction, fluid density, and fluid laminar viscosity To learn more about how SolidWorks Flow Simulation solutions can help you develop better products, please contact your local SolidWorks reseller or visit www.solidworks.com.

Dassault Systmes SolidWorks Corporation 175 Wyman Street Waltham, MA 02451 USA Phone: 1 800 693 9000 Outside the US: +1 781 810 5011 Email: [email protected] www.solidworks.com

SolidWorks is a registered trademark of Dassault Systmes SolidWorks Corporation in the US and other countries. Other brand and product names are trademarks or registered trademarks of their respective owners. 2013 Dassault Systmes. All rights reserved. MKTURBMODwpENG0313