Numerical Study using FLUENT of the Separation and Reattachment Points for BackwardsFacing 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 GutierrezMiravete, Project Adviser
Rensselaer Polytechnic Institute Hartford, Connecticut
December, 2008
CONTENTS
Numerical Study using FLUENT of the Separation and Reattachment Points for
BackwardsFacing 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 _{i}_{j} 
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 
iv
LIST OF TABLES
Table 1 – Backwardfacing 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 backwardfacing step turbulentflow 
1 
Figure 2 – Three recirculation zones for laminar 
2 
Figure 3 – Schematic of backwardfacing 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 backwardfacing step flow. Both turbulent and laminar flow is considered for twodimensional viscous flow, neglecting compressibility, heat generation, and external body forces. A steadystate coupled pressure and velocity algorithm is used for laminar flow and a steadystate segregated pressurevelocity algorithm is used with a realizable kε wallenhanced 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 backwardfacing 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 backwardfacing 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.
Figure 1 – Schematic of backwardfacing step turbulentflow. ^{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.
Figure 2 – Three recirculation zones for laminar flow.
The geometry for the backwardfacing 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 – Schematic of backwardfacing 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 – Backwardfacing 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 =
u⋅D
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}_{.}_{7}_{8}_{9}_{4}
× 10 ^{−}
1.942
5

1.2 Previous research
The indepth 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 nonsymmetric sidewall separation occurs and threedimensional 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
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)
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
x1 S ≈ 6 at ER=1.2 and Moss et al. found the reattachment point x1 S ≈ 5.5 at ER=1.1.
Backwardfacing step flow research continues to be pursued as analysis methods evolve. Lima et al. investigated twodimensional 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 hotfilm 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 nearwall treatment methods for Re=38000 and ER=1.125. Standard wall functions, non equilibrium wall functions, and twolayer models were studied for various turbulence models and compared with data collected by Driver and Seegmiller. Results indicate that the nonequilibrium wall function with renormalized (RNG) and realizable (RKE) kε 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 FASTEST2D FVM were used and good agreement was found with various sources of data.
4
2. Methodology
Flow over a backwardfacing 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: nonslip 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 steadystate 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:
Momentum equation:
where, τ, the stress tensor is,
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:
NavierStokes
equations
yields
the
Reynoldsaveraged
NavierStokes
(RANS)
The kε model is semiempirical twoequation turbulence model that is based on an exact solution for the turbulent kinetic energy (k) and a model of the dissipation rate (ε).
To model the Reynolds stress, , in the RANS equations, the κε model uses the Bousinesq approximation to relate the Reynolds stresses to the mean velocity gradients.
Along with the Bousinesq approximation above, the following definition of the eddy viscosity is used,
The realizable portion of the kε 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 kε model coefficient,
µ
analysis at high Reynolds numbers.
The realizable kε model is defined by the following two equations,
and
G _{k} represents the generation of turbulence kinetic energy due to the mean velocity gradients and relies on the Boussinesq approximation.
6
where the modulus of the mean rateofstrain tensor,
and
S =
ij
∂ u
i
∂ u
∂ u
1
2
∂ u
j
i
+
j
The variable in the eddy viscosity is,
where,
and the model constants are,
where,
The following values are used for the remaining constants,
7
2.2
Approach using FLUENT
The continuity and momentum equations, along with the realizable kε 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 kε 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 backwardfacing 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 pressurelinked 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, xvelocity and yvelocity.
is used for k and ε. All solutions converged with
second order pressure and third order MUSCL (Monotone Upstreamcentered 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 
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.
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 steadystate 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 3D 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.
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
_{n}_{u}_{m}_{b}_{e}_{r} _{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 
(x5x4)/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 (Re>6600) Armaly et al. and Abbot and Kline have determined
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 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 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 threedimensional 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 steadystate method with a coupled pressurevelocity 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 backwardfacing 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 7476, AprAug, 1998, p 495509.
Barber, B.W., Fonty, A., A numerical study of laminar flow over a confined backward facing step using a novel viscoussplitting vortex algorithm, 4th GRACM Congress on Computational Mechanics, Patras, 2729 June 2002.
Bardina, J. E., Huang, P. G., Coakley, T. J., Turbulence Modeling Validation, Testing, and Development, AIAA19972121, NASA Technical Memorandum 110446
Barton I.E., The entrance effect of laminar flow over a backwardfacing step geometry, Int J Numer Methods Fluids 1997, 25:633–44.
Biswas, G., Breuer, M., Durst F., Backwardfacing step flows for various expansion reatios at low and moderate Reynolds numbers, Journal of Fluid Engineering Vol. 126, May 2004, 363374.
De Brederode, V., Bradshaw, P., Threedimensional flow in nominally twodimensional separation bubbles. I. Flow behind a rearwardfacing step, Aero Report 7219, Imperial College of Science and Technology (1972), London, England.
Denham. M. K. & Patrick, M. A. Laminar flow over a downstreamfacing step in a two dimensional flow channel. Trans. Inst. Chem. Engrs 52 (1974), 361.
Driver, D. M., Seegmiller, H. L. and Marvin, J., Timedependent behavior of a reattaching shear layer, AIAA J. 25 (1987), 914919.
Erturk E., Numerical solutions of 2D steady incompressible flow over backwardfacing step, Part I: High Reynolds number solution, Computers & Fluids 37 (2008),
633655.
Etheridge, D.W. & Kemp, P.H., Measurements of turbulent flow downstream of a backwardfacing 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 twodimensional flow over a backwardfacing step at Re = 800 stable?, Int J Numer Methods Fluids 1993, 17:501–41.
Grigoriev
M.M.,
Dargush
G.F.,
A
polyregion
boundary
element
method
for
incompressible viscous fluid flows, Int J Numer Methods Eng 1999, 46:1127–58.
Guj G., Stella F., Numerical solutions of highRe recirculating flows in vorticity– velocity form, Int J Numer Methods Fluids 1988, 8:405–16.
Keskar J., Lyn D.A., Computations of a laminar backwardfacing 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 fractionalstep method to incompressible Navier– Stokes equations, J Comp Phys 1985, 59:308–23.
Kim, Ghajar, Tang, Foutchm Comparison of nearwall treatment methods for high Reynolds number backwardfacing step flow, International Journal of Computational Fluid Dynamics, Vol. 19, No. 7 (2005), 493500.
Lee, T. and Mateescu, D., Experimental and numerical investigation of 2D backward facing step flow, Journal of Fluids and Structures (1998) 12, 703716.
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 10531060.
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 boundarylayer separation, Prog. Aerospace Sci. Vol 32 (1996), 457521.
20
6. Appendix
6.1 FLUENT Input
FLUENT Version: 2d, dp, pbns, lam (2d, double precision, pressurebased, 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 
pressureoutlet 

inlet 
4 
velocityinlet 

top_wall 
5 
wall 

bottom_wall 
6 
wall 

defaultinterior 
8 
interior 
21
Boundary Conditions
fluid
Condition

Material Name Specify source terms? Source Terms Specify fixed values? Fixed Values Motion Type XVelocity Of Zone (m/s) YVelocity Of Zone (m/s) Rotation speed (rad/s) XOrigin of RotationAxis (m) YOrigin of RotationAxis (m) Deactivated Thread Porous zone? XComponent of Direction1 Vector YComponent of Direction1 Vector Relative Velocity Resistance Formulation? Direction1 Viscous Resistance (1/m2) Direction2 Viscous Resistance (1/m2)
Choose alternative formulation for inertial resistance?
Direction1 Inertial Resistance (1/m) Direction2 Inertial Resistance (1/m) C0 Coefficient for PowerLaw C1 Coefficient for PowerLaw 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
22
Backflow Direction Specification Method
1
XComponent of Flow Direction 
1 

YComponent of Flow Direction 
0 

XComponent of Axis Direction 
1 

YComponent of Axis Direction 
0 

ZComponent of Axis Direction 
0 

XCoordinate of Axis Origin (m) 
0 

YCoordinate of Axis Origin (m) 
0 

ZCoordinate of Axis Origin (m) 
0 

is zone used in mixingplane 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) XVelocity (m/s) YVelocity (m/s)
Value
2
0.00028099999
0
0
0
XComponent of Flow Direction 
1 

YComponent of Flow Direction 
0 

XComponent of Axis Direction 
1 

YComponent of Axis Direction 
0 

ZComponent of Axis Direction 
0 

XCoordinate of Axis Origin (m) 
0 

YCoordinate of Axis Origin (m) 
0 

ZCoordinate of Axis Origin (m) 
0 

Angular velocity (rad/s) 
0 
is zone used in mixingplane model?
top_wall
no
Condition

Value
23
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 

XComponent of Wall Translation 
1 

YComponent of Wall Translation 
0 

Define wall velocity components? 
no 

XComponent of Wall Translation (m/s) 
0 

YComponent of Wall Translation (m/s) 
0 

Rotation Speed (rad/s) 
0 

XPosition of RotationAxis Origin (m) 
0 

YPosition of RotationAxis Origin (m) 
0 

Xcomponent of shear stress (pascal) 
0 

Ycomponent 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 

XComponent of Wall Translation 
1 

YComponent of Wall Translation 
0 

Define wall velocity components? 
no 

XComponent of Wall Translation (m/s) 
0 

YComponent of Wall Translation (m/s) 
0 

Rotation Speed (rad/s) 
0 

XPosition of RotationAxis Origin (m) 
0 

YPosition of RotationAxis Origin (m) 
0 

Xcomponent of shear stress (pascal) 
0 

Ycomponent of shear stress (pascal) 
0 

Specularity Coefficient 
0 
24
defaultinterior
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
FCycle
0.1
PressureVelocity Coupling
25
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 ThirdOrder 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/kgk
w/mk
constant
1006.43
0.0242
constant
kg/ms
constant
1.7894e05
Molecular Weight 
kg/kgmol constant 
28.966 
LJ Characteristic Length 
angstrom constant 
3.711 
LJ Energy Parameter 
k constant 78.6 
26
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/kgk constant 871
Thermal Conductivity w/mk
kg/m3
constant
2719
constant
202.4
27
Viel mehr als nur Dokumente.
Entdecken, was Scribd alles zu bieten hat, inklusive Bücher und Hörbücher von großen Verlagen.
Jederzeit kündbar.