Sie sind auf Seite 1von 35

Numerical Study using FLUENT of the Separation and Reattachment Points for Backwards-Facing Step Flow

by

Luke Jongebloed

An Engineering Project Submitted to the Graduate

Faculty of Rensselaer Polytechnic Institute

in Partial Fulfillment of the

Requirements for the degree of

Master of Engineering

Major Subject: Mechanical Engineering

Approved:

Ernesto Gutierrez-Miravete, Project Adviser

Rensselaer Polytechnic Institute Hartford, Connecticut

December, 2008

CONTENTS

Numerical Study using FLUENT of the Separation and Reattachment Points for

Backwards-Facing Step Flow

i

LIST

OF

SYMBOLS

iii

LIST OF TABLES

v

LIST

OF

FIGURES

vi

ACKNOWLEDGMENT

vii

ABSTRACT

 

viii

1. Background

1

1.1 Introduction

1

1.2 Previous research

3

2. Methodology

 

5

2.1 Theory

5

2.2 Approach using FLUENT

8

3. Discussion

 

10

3.1 Laminar

12

3.2 Turbulent

15

4. Conclusion

 

18

5. References

19

6. Appendix

21

6.1

FLUENT Input

21

ii

LIST OF SYMBOLS

A

0

Model constant

A

s

Model variable

C

2

Model constant

C

µ

Model variable

D

Hydraulic diameter of backwards step

ER

Expansion ratio

G k

Turbulent generation term

h

Height of inlet channel

H

Height of outlet

I

Identity matrix

i

Sub index

j

Sub index

k

Turbulent kinetic energy

k

Sub index

Re

Reynolds number

S

Step height

S

Magnitude of mean strain

S ij

Mean strain tensor

t

Time

u

Fluid velocity

U

Characteristic velocity scale

W

Model variable

x

Direction vector

x1

Reattachment point for 1st bottom recirculation zone

x2

Separation point for 2st bottom recirculation zone

x3

Reattachment point for 2nd bottom recirculation zone

x4

Reattachment point for 1st top recirculation zone

x5

Separation point for 2st top recirculation zone

Xe

Inlet channel length

Xo

Outlet channel length

iii

ε

Dissipation rate

µ

Eddy viscosity

σ

k

Model constant

σ

ε

Epsilon model constant

σ

Model variable

τ

Stress tensor

υ

Kinematic viscosity

µ

Dynamic viscosity

ρ

Density

LIST OF TABLES

Table 1 – Backward-facing step dimensions (all in meters)

2

Table 2 – Miscellaneous reference values used in this

3

Table 3 – Number of nodes for grid reference number used to indicate amount of mesh

refinement in discussion

9

Table 4 – Effect of mesh refinement for

14

Table 5 – Comparison of reattachment and separation points for Re=800 and ER=1.942

for various numerical

14

Table 6 – Separation points obtained for turbulent

15

Table 7 – Comparison of methods used to obtain solution for

17

v

LIST OF FIGURES

Figure 1 – Schematic of backward-facing step turbulent-flow

1

Figure 2 – Three recirculation zones for laminar

2

Figure 3 – Schematic of backward-facing step geometry (not to scale)

2

Figure 4 – Schematic showing region of grid refinement, 200m downstream from step

9

Figure 5– Comparison of separation and reattachment points for present analysis with

10

Figure 6 – Comparison of separation and reattachment points between present analysis

(to scale)

experimental data collected by Armaly et al

and experimental data collected by Armaly et al. for Re<1000

12

Figure 7 – Streamlines for Re=800; colored by velocity

15

Figure 8 – Plot of streamlines for Re=8000; colored by velocity magnitude

16

Figure 9 – Scaled residuals for Re=7470

17

vi

ACKNOWLEDGMENT

I would like to thank my cat for sitting with me and providing support while I completed my school work.

vii

ABSTRACT

A numerical investigation is conducted on the affect of Reynolds number on the separation and attachment points for backward-facing step flow. Both turbulent and laminar flow is considered for two-dimensional viscous flow, neglecting compressibility, heat generation, and external body forces. A steady-state coupled pressure and velocity algorithm is used for laminar flow and a steady-state segregated pressure-velocity algorithm is used with a realizable k-ε wall-enhanced turbulence model. The expansion ratio of inlet height to outlet height is a 1.942. The results are compared to published experimental and numerical data. The present study agrees with published data for low Reynolds numbers (Re<400) and high Reynolds numbers (Re>15000). Results exhibit behavior of published data, but are slightly lower in magnitude for 400<Re<1200 and

6600<Re<15000.

viii

1. Background

1.1

Introduction

A numerical analysis is performed using FLUENT to investigate backward-facing step

flow for Reynolds numbers in the laminar and turbulent regions. Separation and

reattachment lengths are determined for each Reynolds number and the results are

compared to experimental data and numerical analyses found in literature.

Flow over a backward-facing step produces recirculation zones where the fluid

separates and forms vortices. For turbulent flow, the fluid separates at the step and

reattaches downstream, as show below in Figure 1. Only a single recirculation zone

develops for turbulent flow and the reattachment point is believed to be independent of

the Reynolds number and depend only on the ratio of inlet height to outlet height.

depend only on the ratio of inlet height to outlet height. Figure 1 – Schematic of

Figure 1 – Schematic of backward-facing step turbulent-flow. 1 For laminar flow, various recirculation zones occur downstream from the step, as

shown below in Figure 2. Separation occurs when adverse pressure gradients act on the

fluid. As the Reynolds number increases from zero, the first region of separation occurs

at the step to x1 on the bottom wall. Next, the second region of separation occurs

between x4 and x5 on the top wall. As the Reynolds number increases into the transition

zone, a third separation region occurs between x2 and x3 on the bottom wall.

1 Figure from R.L. Simpson.

1

Theoretically, recirculation zones will continue to develop downstream as the Reynolds

number increases and the flow remains laminar; however, this has not been observed

experimentally and the flow will eventually become turbulent.

and the flow will eventually become turbulent. Figure 2 – Three recirculation zones for laminar flow.

Figure 2 – Three recirculation zones for laminar flow.

The geometry for the backward-facing step used in this analysis is similar to that used by

Armaly et al. Figure 3 and Table 1 provide the dimensions of the geometry.

Figure 3 and Table 1 provide the dimensions of the geometry. Figure 3 – Schematic of

Figure 3 – Schematic of backward-facing step geometry (not to scale).

Significant length is provided for the inlet channel to ensure that the flow is fully

developed and does not contain any additional effects created by the flow source. The

significant length of the outlet channel ensures that the outlet condition does not affect

the flow near the step. The expansion ration, ER, is ratio of the outlet height over the

inlet height. For this case, ER = 1.942.

Table 1 – Backward-facing step dimensions (all in meters).

Height of inlet channel

h

5.2

Height of outlet

H

10.1

Step height

S

4.9

Inlet channel length

Xe

200

Outlet channel length

Xo

500

The Reynolds number is defined as,

Re =

uD

ν

, where u is the inlet velocity, υ is the

kinematic viscosity, and D is the hydraulic diameter. The Reynolds number has been

2

expressed differently throughout literature; D can be based on the inlet height, the outlet height, or the step height. In agreement with Armaly et al., this study will use D=2h. It is important to know how the Reynolds number was calculated when comparing data. Also of importance is the method used to calculate the inlet velocity. The average velocity can be used or one can use functions of the measured velocity profile, e.g., Armaly et al. used 2/3 maximum measured inlet velocity. Another factor that may affect the comparison of results is the turbulent intensity of the inlet velocity. Although inlet velocity parameters have significant effect on the reattachment points (Badran and Bruun), a relatively long inlet channel length should dissipate the discrepancies. Table 1 lists various constants used in this study.

Table 2 – Miscellaneous reference values used in this study.

Density

Dynamic viscosity

Expansion Ratio

ρ

µ

ER

1.225

1.7894

× 10

1.942

5

3 kg m kg m ⋅ s
3
kg m
kg m ⋅ s

-

1.2 Previous research

The in-depth experimental data analysis performed by Armaly et al has provided the majority of data used for comparison in the present study. Others, including Driver and D. M., Seegmiller; D.E. Abbott and S.J. Kline; Denham. M. K. & Patrick; Etheridge, D.W. & Kemp, have performed similar experiments and yielded similar results. These experiments have provided useful data to compare with and validate numerical schemes. In the following we can summarize the relationship between the location of the separation point and Reynolds number. The two dimensional approximation is only valid for Re <400 and Re>6600, for 400<Re<6600 non-symmetric side-wall separation occurs and three-dimensional effects become important. As summarized by Lima et al., flow can be considered laminar for Re<1200 and turbulent for Re>6600. Within the laminar region the reattachment points will increase with Re. Within the turbulent region the

will increase with Re. Within the turbulent region the reattachment point is constant ( x 1

reattachment point is constant ( x1 S 8 for ER=2). As show in Figure 1 and Figure 2

above, there may be multiple separation and reattachments points for laminar flow and

3

there is only one reattachment point for turbulent flow. For all flows, as ER increases the distance that separation occurs increases. The single reattachment point for turbulent flow has been observed to be independent of the Reynolds number and depends only on geometry. Both Armaly et al. and Abbot and Kline have determined experimentally that for turbulent flow (Re>6600)

experimentally that for turbulent flow (Re>6600) the reattachment point x 1 S ≈ 8 at ER=1.94.

the reattachment point x1 S 8 at ER=1.94. The reattachment length decrease for

decreasing step heights; e.g., De Brederode and Bradshaw found the reattachment point

e.g., De Brederode and Bradshaw found the reattachment point x 1 S ≈ 6 at ER=1.2
e.g., De Brederode and Bradshaw found the reattachment point x 1 S ≈ 6 at ER=1.2

x1 S 6 at ER=1.2 and Moss et al. found the reattachment point x1 S 5.5 at ER=1.1.

Backward-facing step flow research continues to be pursued as analysis methods evolve. Lima et al. investigated two-dimensional laminar flow with Reynolds number varying between 100 and 2500. Convergence could not be obtained using a steady state segregated finite volume method (FVM); instead, an unsteady flow was analyzed for very large time. Good agreement with Armaly et al. was found for x1 with Re<689 and x5 with 648<Re<900; however, good agreement was not found for the upper separation point, x4. Lee and Mateescu performed an experimental and numerical analysis for 400<Re*<2800 and ER=1.17 and 2.0 (* Re based on outlet height). Data was collected using multiple hot-film sensors (MHFS) arrays and, although good agreement found with numerical prediction, it was recommended to further develop MHFS arrays for detection of turbulent flow separation. Kim et al. provides a validation study of turbulent near-wall treatment methods for Re=38000 and ER=1.125. Standard wall functions, non- equilibrium wall functions, and two-layer models were studied for various turbulence models and compared with data collected by Driver and Seegmiller. Results indicate that the non-equilibrium wall function with renormalized (RNG) and realizable (RKE) k-ε turbulence models compare best with experimental data. Bidwas et al. provided a numerical study of 2D and 3D laminar flow with ER= 1.9423, 2.5 and 3.0. LESOCC and FASTEST-2D FVM were used and good agreement was found with various sources of data.

4

2. Methodology

Flow over a backward-facing step is studied based on the numerical methods discussed in Section 2.1 and the approach in Section 2.2. Compressibility and energy terms are neglected. The following boundary conditions are used: non-slip walls, zero gauge pressure outlet, and constant normal inlet velocity that does not vary along the height of inlet. The Reynolds number is varied from 50 to 1250 and the separation and reattachment points are determined from minimum values of the coefficient of friction. Reattachment points are also determined for various Reynolds numbers in the turbulent region (Re>6600).

2.1

Theory

The governing equations for computational fluid dynamics (CFD) are based on conservation of mass, momentum, and energy. FLUENT uses a finite volume method (FVM) to solve the governing equations. The FVM involves descretization and integration of the governing equation over the control volume. The following is a summary of the theory involved in the FLUENT analysis and is based on the FLUENT User’s Manual, Bardina et al., and Anderson. The basic equations for steady-state laminar flow are conservation of mass and momentum. When heat transfer or compressibility is involved the energy equation is also required. The governing equations are, Continuity Equation:

required. The governing equations are, Continuity Equation: Momentum equation: where, τ, the stress tensor is,

Momentum equation:

equations are, Continuity Equation: Momentum equation: where, τ, the stress tensor is, Turbulent flow can be

where, τ, the stress tensor is,

Momentum equation: where, τ, the stress tensor is, Turbulent flow can be modeled using mean and

Turbulent flow can be modeled using mean and fluctuating values for components,

such as velocity,

u i = u + u. Substituting the mean and fluctuating value equations into

i

i

5

the

equations:

Navier-Stokes

equations

yields

the

Reynolds-averaged

Navier-Stokes

(RANS)

equations yields the Reynolds-averaged Navier-Stokes (RANS) The k-ε model is semi-empirical two-equation turbulence

The k-ε model is semi-empirical two-equation turbulence model that is based on an exact solution for the turbulent kinetic energy (k) and a model of the dissipation rate (ε).

kinetic energy (k) and a model of the dissipation rate (ε). To model the Reynolds stress,

To model the Reynolds stress, , in the RANS equations, the κ-ε model uses the Bousinesq approximation to relate the Reynolds stresses to the mean velocity gradients.

relate the Reynolds stresses to the mean velocity gradients. Along with the Bousinesq approximation above, the

Along with the Bousinesq approximation above, the following definition of the eddy viscosity is used,

the following definition of the eddy viscosity is used, The realizable portion of the k-ε model

The realizable portion of the k-ε model is based on the following relationship, which can be obtained by determining the point that the average normal stress becomes

negative.

C , is determined by equilibrium

The realizable k-ε model coefficient,

µ

analysis at high Reynolds numbers.

model coefficient, µ analysis at high Reynolds numbers. The realizable k-ε model is defined by the

The realizable k-ε model is defined by the following two equations,

k-ε model is defined by the following two equations, and G k represents the generation of

and

k-ε model is defined by the following two equations, and G k represents the generation of

G k represents the generation of turbulence kinetic energy due to the mean velocity gradients and relies on the Boussinesq approximation.

generation of turbulence kinetic energy due to the mean velocity gradients and relies on the Boussinesq

6

where the modulus of the mean rate-of-strain tensor,

where the modulus of the mean rate-of-strain tensor, and S = ij    

and

S =

ij

u

i

u

u

1

2

u

j

i

+

j

The variable in the eddy viscosity is,

1 2 ∂ u j i + j The variable in the eddy viscosity is, where,

where,

∂ u j i + j The variable in the eddy viscosity is, where, and the

and the model constants are,

the eddy viscosity is, where, and the model constants are, where, The following values are used

where,

eddy viscosity is, where, and the model constants are, where, The following values are used for

The following values are used for the remaining constants,

eddy viscosity is, where, and the model constants are, where, The following values are used for

7

2.2

Approach using FLUENT

The continuity and momentum equations, along with the realizable k-ε model with wall

enhancements and pressure gradients effects for turbulent flows, are solved using the FVM in FLUENT. A pressure based solver is used since the flow is incompressible and separation is caused by adverse pressure gradients. As demonstrated by Kim et al. the realizable k-ε model with wall treatment performs well for boundary layers subject to separation and is used. A coupled pressure and velocity algorithm is used for laminar flows, which solves the continuity and momentum equations in a simultaneous fashion and removes the approximations associated with segregated algorithms where the momentum and continuity equations are solved separately. The coupled algorithm is employed because of convergence issues with segregated solvers on backward-facing step flows. The coupled algorithm does not offer solution accuracy improvement over segregated solvers; rather it provides improvement in stability and ability to converge. The semi- implicit method for pressure-linked equations SIMPLE algorithm is a segregated algorithm and is used for turbulent flow analysis in this study. The SIMPLE algorithm solves the momentum equation first, then solves for pressure, and later corrects the descretized solutions. The SIMPLE algorithm can offer increased convergence time due

to the smaller memory requirement over the coupled algorithm.

A convergence criterion of 1

× 10

5

is used for continuity, x-velocity and y-velocity.

is used for k and ε. All solutions converged with

second order pressure and third order MUSCL (Monotone Upstream-centered Schemes for Conservation Laws) momentum interpolation schemes for laminar flow. All solutions converged with second order pressure, momentum, turbulent intensity, and turbulent dissipation interpolation schemes for turbulent flow. Third order MUSCL schemes did not provide significant accuracy for turbulent flow. Adequate grid independence is satisfied with a quadrilateral mesh of 59251 nodes. The entire surface is meshed with 30151 nodes then adapted to 59251 nodes with refinement only in the region of recirculation from the step (x=200) to 200 m downstream from the step. Further adaptation to 174645 nodes in this region does not

A convergence criterion of

1

×

10

3

8

provide significant increase in the accuracy of the results. Shown below, Table 3 and

Figure 4 summarize the amount of grid adaptation used and the area of refinement.

Table 3 – Number of nodes for grid reference number used to indicate amount of mesh refinement in discussion section.

grid

number of

region

number

nodes

refinement

0

30151

no

1

59251

yes

2

174645

yes

30151 no 1 59251 yes 2 174645 yes Figure 4 – Schematic showing region of grid

Figure 4 – Schematic showing region of grid refinement, 200m downstream from step (to scale).

9

3. Discussion

Laminar flow exists for Re<1200, transitional flow occurs for 1200<Re<6600, and

devloped turbulent flow occurs for Re>6600, as reported by Armaly et al. The separation

and reattachments zones vary almost linearly with the Reynolds number for laminar flow

and are constant for turbulent flow. Below, Figure 5 shows the present results for

Re<8000 with experimental data obtained by Armaly et al.

Re<8000 with experimental data obtained by Armaly et al. Figure 5– Comparison of separation and reattachment

Figure 5– Comparison of separation and reattachment points for present analysis with experimental data collected by Armaly et al. 2

Present results compare well for the first bottom reattachment point, x1, when

Re<400, then the predicted x1 becomes increasingly lower as the Reynolds number

increases. The top separation point, x4, develops shortly after Re=400; at this point

results are comparable with data. As the Reynolds number increases from 400, x4 does

2 Figure adapted from Armaly et al. (1983).

10

not compare well with experimental data. The top reattachment point, x5, compares well with data for Re<750; above this values x5 continues to increase as experimental data begins to decline when turbulent transition begins to occur. The reattachment point, x1, is shown to be independent of Reynolds number for turbulent flow, but is slightly lower the experimental for Reynolds numbers near 8000. The second bottom recirculation region was observed in the present analysis; however convergence was not obtained for this region (Re>1250). Oscillating residuals were experienced and can be attributed to the use of a steady-state method. Numerical unsteady analyses have obtained convergence for flows with three or more recirculation zones, e.g., Lima et al. In this region, however, the flow is 3-D and numerical and measured values do not agree.

11

3.1

Laminar

The plot in Figure 6 below shows a closer view for Re<1000 to compare presents results

with the experimental results of Armaly et al. Generally, good agreement is found with

experiment, although the present analysis obtained values slightly lower– especially for

second bottom reattachment point, x5.

especially for second bottom reattachment point, x5. Figure 6 – Comparison of separation and reattachment

Figure 6 – Comparison of separation and reattachment points between present analysis and experimental data collected by Armaly et al. for Re<1000. 3

To demonstrate grid independence, the grid was adapted multiple times for

in

Re=800

and

the

summary

is

provided

3 Plot adapted from Barber and Fonty.

12

Table 4 below. The first adaptation of the initial mesh 59251 nodes provided reasonable accurate results and is used for all laminar Reynolds number. Adaptation was only performed in the region between the step and 200 meters downstream (see Figure 4).

13

Table 4 – Effect of mesh refinement for Re=800.

grid

number 4

x1/S

x4/S

x5/S

0 10.83

8.44

19.26

1 11.67

9.24

19.92

2 11.88

9.42

20.02

Table 5 below summarizes the values for separation and reattachment points at

Re=800 determined by various authors. The results of the present study are lower than

the average of values obtained by the various authors, but are still within the range of

data. The largest difference between present results and average literature value is the

upper reattachment point, x5/S=19.92; average is 20.62. This present result of

x5/S=19.92 is closer to the experimental value (Armly et al.) 19.33 than the average

20.62.

Table 5 – Comparison of reattachment and separation points for Re=800 and ER=1.942 for various numerical studies.

 

x1/S

x4/S

x5/S

(x5-x4)/S

Presentt study Lima Gartling† Lee and Mateescu† Barton Kim and Moin Guj and Stella Gresho et al. Keskar and Lyn Grigoriev and Dargush Rogers and Kwak Erturk

11.67

9.24

19.92

10.68

11.97

9.51

20.40

10.89

12.20

9.70

20.96

11.26

12.00

9.60

20.60

11.00

11.51

9.14

20.66

11.52

11.90

-

-

-

12.05

9.70

20.20

10.50

12.20

9.72

20.98

11.26

12.19

9.71

20.96

11.25

12.18

9.70

20.94

11.24

11.48

-

-

-

11.83

9.48

20.55

11.07

Average

11.93

9.55

20.62

11.07

Armaly et al.*

14.00

11.11

19.33

8.22

† ER=2.0, * Experimental

A plot of the stream lines over the step for Re=800 is shown in Figure 7. The

fluid velocity at the step dissipates as slower recirculation region absorb some of the

momentum.

4 See Table 2.

14

Figure 7 – Streamlines for Re=800; colored by velocity magnitude. 3.2 Turbulent For turbulent flow

Figure 7 – Streamlines for Re=800; colored by velocity magnitude.

3.2

Turbulent

For turbulent flow (Re>6600) Armaly et al. and Abbot and Kline have determined

Armaly et al. and Abbot and Kline have determined experimentally that the reattachment point, x 1

experimentally that the reattachment point, x1 S 8 at ER=1.94. Table 6 below

summarizes the range of values obtained in the present study for the reattachment point at various turbulent Reynolds numbers. The average value, x1/S=7.21, for the separation point is lower than x1/S=8, the accepted value. The higher Reynolds numbers studied are closer to the accepted value. For Re= 17799, x1/S=8.0, which is in agreement with the accepted value.

Table 6 – Separation points obtained for turbulent flow.

Re

x1/S

7000

6.92

7476

6.61

7830

6.54

8000

6.80

11400

7.01

17799

8.00

24480

8.60

average

7.21

A plot of the streamlines over the step is displayed in Figure 8 below. A second eddy near the step corner is observed. The velocity of the recirculation zone is on the order of magnitude lower than the velocity at the step.

15

Figure 8 – Plot of streamlines for Re=8000; colored by velocity magnitude. Solving for turbulent

Figure 8 – Plot of streamlines for Re=8000; colored by velocity magnitude.

Solving for turbulent flow required multiple levels of refinement to obtain an accurate solution. Figure 9 shows a plot of the scaled residuals for the solution at Re=7470. Table 7 provides the type of methods used to achieve each level of convergence displayed in Figure 9.

16

Figure 9 – Scaled residuals for Re=7470. RKE with mesh refinement provides significant accuracy over

Figure 9 – Scaled residuals for Re=7470.

RKE with mesh refinement provides significant accuracy over SKE with out wall

treatment. The third order MUSCL RKE with number 2 mesh refinement does not offer

significant increase in accuracy over second order RKE with number 1 mesh refinement.

Table 7 – Comparison of methods used to obtain solution for Re=7470.

 

Momentum, Turbulent Kinetic Energy, and turbulent dissipation rate

wall

enhancements

and pressure

Mesh

Method

Pressure

gradient effects

refinement 5

x1/S

SKE

1st order

1st order 1st order 1st order 2nd order 3rd order MUSCL 3rd order MUSCL

no

0

5.37

RKE

1st order

yes

0

5.74

RKE

1st order

yes

1

6.26

RKE

2nd order

yes

1

6.54

RKE

2nd order

yes

1

6.59

RKE

2nd order

yes

2

6.61

5 See Table 2.

17

4. Conclusion

The values for the separation and reattachment points obtained in this study compare fairly well with published numerical data. The present results begin to differ from experimental data at Reynolds numbers (Re>400) where three-dimensional effects become important. The normalized values (x1/S~6.5) for the turbulent reattachment points are less than accepted value (x1/S~8) for low turbulent Reynolds numbers (Re~8000); however for higher Reynolds numbers (Re>15000), good agreement is found. A general trend in the laminar results of this analysis is slightly lower values for separation and reattachment points than compared with other numerical studies. This difference with present results for laminar flow can be attributed to the range of methods and grids used to perform the numerical calculations. Unsteady methods iterated over a large time span are typically used for the laminar case because of convergence issues; however, this study used a steady-state method with a coupled pressure-velocity algorithm.

18

5. References

Anderson, J.D. Jr., Computational Fluid Dynamics: The Basics with Applications, McGraw Hill, 1995

Armaly, B.F., Durst, F., Pereira, J.C.F., and Schonung, B., Experimental and theoretical investigation of backward-facing step flow, J. Fluid. Mech. 127 (1983), pp. 473–

496.

Badran, O.O., Bruun, H.H., Effect of inlet conditions on flow over backward facing step, Journal of Wind Engineering and Industrial Aerodynamics, v 74-76, Apr-Aug, 1998, p 495-509.

Barber, B.W., Fonty, A., A numerical study of laminar flow over a confined backward- facing step using a novel viscous-splitting vortex algorithm, 4th GRACM Congress on Computational Mechanics, Patras, 27-29 June 2002.

Bardina, J. E., Huang, P. G., Coakley, T. J., Turbulence Modeling Validation, Testing, and Development, AIAA-1997-2121, NASA Technical Memorandum 110446

Barton I.E., The entrance effect of laminar flow over a backward-facing step geometry, Int J Numer Methods Fluids 1997, 25:633–44.

Biswas, G., Breuer, M., Durst F., Backward-facing step flows for various expansion reatios at low and moderate Reynolds numbers, Journal of Fluid Engineering Vol. 126, May 2004, 363-374.

De Brederode, V., Bradshaw, P., Three-dimensional flow in nominally two-dimensional separation bubbles. I. Flow behind a rearward-facing step, Aero Report 72-19, Imperial College of Science and Technology (1972), London, England.

Denham. M. K. & Patrick, M. A. Laminar flow over a downstream-facing step in a two- dimensional flow channel. Trans. Inst. Chem. Engrs 52 (1974), 361.

Driver, D. M., Seegmiller, H. L. and Marvin, J., Time-dependent behavior of a reattaching shear layer, AIAA J. 25 (1987), 914-919.

Erturk E., Numerical solutions of 2-D steady incompressible flow over backward-facing step, Part I: High Reynolds number solution, Computers & Fluids 37 (2008),

633-655.

Etheridge, D.W. & Kemp, P.H., Measurements of turbulent flow downstream of a backward-facing step, J. Fluid Mech. 86 (1978), 545.

Fluent Inc, User’s Guide, 6.3.26 version, 2006

Gartling D.K., A test problem for outflow boundary conditions – flow over a backward- facing step, Int J Numer Methods Fluids 1990, 11:953–67.

19

Gresho P.M., Gartling D.K., Torczynski JR, Cliffe KA, Winters KH, Garratt TJ, et al. Is the steady viscous incompressible two-dimensional flow over a backward-facing step at Re = 800 stable?, Int J Numer Methods Fluids 1993, 17:501–41.

Grigoriev

M.M.,

Dargush

G.F.,

A

poly-region

boundary

element

method

for

incompressible viscous fluid flows, Int J Numer Methods Eng 1999, 46:1127–58.

Guj G., Stella F., Numerical solutions of high-Re recirculating flows in vorticity– velocity form, Int J Numer Methods Fluids 1988, 8:405–16.

Keskar J., Lyn D.A., Computations of a laminar backward-facing step flow at Re = 800 with a spectral domain decomposition method, Int J Numer Methods Fluids 1999, 29:411–27.

Kim J., Moin P., Application of a fractional-step method to incompressible Navier– Stokes equations, J Comp Phys 1985, 59:308–23.

Kim, Ghajar, Tang, Foutchm Comparison of near-wall treatment methods for high Reynolds number backward-facing step flow, International Journal of Computational Fluid Dynamics, Vol. 19, No. 7 (2005), 493-500.

Lee, T. and Mateescu, D., Experimental and numerical investigation of 2-D backward- facing step flow, Journal of Fluids and Structures (1998) 12, 703-716.

Lima, R.C., Andrade, C.R., and Zaparoli, E.L., Numerical study of three recirculation zones in the unilateral sudden expansion flow, International Communications in Heat and Mass Transfer, Volume 35, Issue 9, November 2008, Pages 1053-1060.

Moss, Bakers, Bradburly, 1979 Measurements of mean velocity and Reynolds stresses in some regions of recirculation flows, In Turbulent Shear Flows 1 (ed. F. Durst, B. C. Launder, F. W. Schmidt & J. H. Whitelaw). Springer.

Rogers S.E., Kwak D., An upwind differencing scheme for the incompressible Navier– Stokes equations, Appl Numer Math 1991, 8:43–64.

Simpson, R.L., Aspects of turbulent boundary-layer separation, Prog. Aerospace Sci. Vol 32 (1996), 457-521.

20

6. Appendix

6.1 FLUENT Input

FLUENT Version: 2d, dp, pbns, lam (2d, double precision, pressure-based, laminar) Release: 6.3.26 Title:

Models

------

Model

Settings

-------------------------------------

Space

2D

Time

Steady

Viscous

Laminar

Heat Transfer

Disabled

Solidification and Melting

Species Transport Coupled Dispersed Phase

Disabled

Disabled

Disabled

Pollutants

Disabled

Pollutants

Disabled

Soot

Disabled

Boundary Conditions

-------------------

Zones

name

id

type

---------------------------------------

fluid

2

fluid

outlet

3

pressure-outlet

inlet

4

velocity-inlet

top_wall

5

wall

bottom_wall

6

wall

default-interior

8

interior

21

Boundary Conditions

fluid

Condition

---------------------------------------------------------------

Material Name Specify source terms? Source Terms Specify fixed values? Fixed Values Motion Type X-Velocity Of Zone (m/s) Y-Velocity Of Zone (m/s) Rotation speed (rad/s) X-Origin of Rotation-Axis (m) Y-Origin of Rotation-Axis (m) Deactivated Thread Porous zone? X-Component of Direction-1 Vector Y-Component of Direction-1 Vector Relative Velocity Resistance Formulation? Direction-1 Viscous Resistance (1/m2) Direction-2 Viscous Resistance (1/m2)

Choose alternative formulation for inertial resistance?

Direction-1 Inertial Resistance (1/m) Direction-2 Inertial Resistance (1/m) C0 Coefficient for Power-Law C1 Coefficient for Power-Law Porosity

Value

air

()

()

0

no

no

0

0

0

no

no

0

0

1

0

0

0

0

0

0

0

yes

no

1

outlet

Condition

-----------------------------------------------

Gauge Pressure (pascal)

Value

0

Backflow Direction Specification Method

1

X-Component of Flow Direction

1

Y-Component of Flow Direction

0

X-Component of Axis Direction

1

Y-Component of Axis Direction

0

Z-Component of Axis Direction

0

X-Coordinate of Axis Origin (m)

0

Y-Coordinate of Axis Origin (m)

0

Z-Coordinate of Axis Origin (m)

0

is zone used in mixing-plane model?

no

Specify targeted mass flow rate

no

Targeted mass flow (kg/s)

1

inlet

Condition

---------------------------------------------------

Velocity Specification Method Reference Frame Velocity Magnitude (m/s) X-Velocity (m/s) Y-Velocity (m/s)

Value

2

0.00028099999

0

0

0

X-Component of Flow Direction

1

Y-Component of Flow Direction

0

X-Component of Axis Direction

1

Y-Component of Axis Direction

0

Z-Component of Axis Direction

0

X-Coordinate of Axis Origin (m)

0

Y-Coordinate of Axis Origin (m)

0

Z-Coordinate of Axis Origin (m)

0

Angular velocity (rad/s)

0

is zone used in mixing-plane model?

top_wall

no

Condition

----------------------------------------------------------

Value

Wall Motion Shear Boundary Condition

0

0

Define wall motion relative to adjacent cell zone?

 

yes

Apply a rotational velocity to this wall?

 

no

Velocity Magnitude (m/s)

 

0

X-Component of Wall Translation

 

1

Y-Component of Wall Translation

0

Define wall velocity components?

no

X-Component of Wall Translation (m/s)

 

0

Y-Component of Wall Translation (m/s)

0

Rotation Speed (rad/s)

0

X-Position of Rotation-Axis Origin (m)

 

0

Y-Position of Rotation-Axis Origin (m)

0

X-component of shear stress (pascal)

0

Y-component of shear stress (pascal)

0

Specularity Coefficient

0

bottom_wall

Condition

Value

 

----------------------------------------------------------

 

Wall Motion Shear Boundary Condition

0

0

Define wall motion relative to adjacent cell zone?

 

yes

Apply a rotational velocity to this wall?

 

no

Velocity Magnitude (m/s)

 

0

X-Component of Wall Translation

 

1

Y-Component of Wall Translation

0

Define wall velocity components?

no

X-Component of Wall Translation (m/s)

 

0

Y-Component of Wall Translation (m/s)

0

Rotation Speed (rad/s)

0

X-Position of Rotation-Axis Origin (m)

 

0

Y-Position of Rotation-Axis Origin (m)

0

X-component of shear stress (pascal)

0

Y-component of shear stress (pascal)

0

Specularity Coefficient

0

default-interior

Condition

-----------------

Value

Solver Controls

---------------

Equations

Equation

-----------------

Flow

Solved

yes

Numerics

Numeric

---------------------------------------

Enabled

Absolute Velocity Formulation

yes

Relaxation

Variable

Relaxation Factor

-------------------------------

Density

1

Body Forces

1

Linear Solver

Solver

Termination

Residual Reduction

Variable

Type

Criterion

Tolerance

-----------------------------------------------------

Flow

F-Cycle

0.1

Pressure-Velocity Coupling

Parameter

---------------------------------------------

Value

Type

Coupled

Courant Number

200

Explicit Momentum Relaxation Factor Explicit Pressure Relaxation Factor

0.75

0.75

Discretization Scheme

Variable

----------------------------

Pressure

Momentum

Scheme

Second Order Third-Order MUSCL

Solution Limits

Quantity Limit

---------------------------------

Minimum Absolute Pressure

1

Maximum Absolute Pressure

5e+10

Minimum Temperature

1

Maximum Temperature

5000

Material Properties

-------------------

Material: air (fluid)

Property

----------------------------------------------------------------

Density Cp (Specific Heat) Thermal Conductivity Viscosity

kg/m3

Units

Method

constant

Value(s)

1.225

j/kg-k

w/m-k

constant

1006.43

0.0242

constant

kg/m-s

constant

1.7894e-05

Molecular Weight

kg/kgmol

constant

28.966

L-J Characteristic Length

angstrom

constant

3.711

L-J Energy Parameter

k

constant

78.6

Thermal Expansion Coefficient

1/k

constant

0

Degrees of Freedom

constant

0

Speed of Sound

m/s

none

#f

Material: aluminum (solid)

 

Property

Units

Method

Value(s)

 

---------------------------------------------------

Density

Cp (Specific Heat) j/kg-k constant 871

Thermal Conductivity w/m-k

kg/m3

constant

2719

constant

202.4