Numerical Simulation of

Two-Dimensional Laminar
Incompressible Wall Jet Flow
P. Rajesh Kanna
Research Scholar
Under Backward-Facing Step
Two-dimensional laminar incompressible wall jet flow over a backward-facing step is
Manab Kumar Das1 solved numerically to gain insight into the expansion and recirculation of flow processes.
Associate Professor
e-mail: [email protected] Transient streamfunction vorticity formulation of the Navier-Stokes equation is solved
with clustered grids on the physical domain. The behavior of the jet has been studied for
Department of Mechanical Engineering, different step geometry (step length, l, step height, s) and Reynolds number (Re). It is
Indian Institute of Technology Guwahati, found that the presence of a step in the wall jet flow creates recirculation and the
North Guwahati, Guwahati-781039, Assam, reattachment length follows an almost linear trend within the range considered for both
India parameters Re and step geometry. Simulations are made to show the effect of entrainment
on recirculation eddy. Detailed study of u velocity decay is reported. The velocity profile
in the wall jet region shows good agreement with experimental as well as similarity
results. The distance where the similarity profile forms is reduced by increasing the step
geometry whereas an increment in Re increases this distance. The effects of Re, step
length, and step height on wall vorticity are presented. The parametric study is helpful to
predict the reattachment location for wall jet flows over step. 关DOI: 10.1115/1.2243298兴

Keywords: wall jet, recirculation, wall vorticity, similarity, backward-facing step

1 Introduction Wall jet flow over the step has many practical applications. Wall
jet flow is commonly used in the windshield defroster system in
Depending upon the distance of the confining boundaries from
automobiles. A jet of hot air is blown along the windshield to
the discharge, a jet can be analyzed as a free jet or a bounded one.
protect the surface from its surroundings. Introducing a recircula-
If the boundaries 共parallel to inlet axis兲 are sufficiently away from
tion by means of a step to enhance the heat transfer is an interest-
the origin of the jet, the flow is termed as a free jet. However, a
bounded jet will occur when it interacts with a parallel wall. ing study on this situation. The recirculation is used to shift the
Bounded jet flows occur in many engineering applications such as peak Nusselt number location in the downstream direction. An-
environmental discharges, heat exchangers, fluid injection sys- other example can be found from the cooling of electronic com-
tems, cooling of combustion chamber wall in a gas turbine, auto- ponents. Steplike geometries are very often found in multi-
mobile demister, and others. Bounded jets can be classified into processor electronic components. For better performance, they are
three types: 共a兲 impinging jet aimed towards the boundary, 共b兲 designed to maintain at constant low temperature. Wall jets are
wall jet where fluid is discharged at the boundary, and 共c兲 offset used to cool such hot surfaces. The knowledge of recirculation
jet from a vertical wall of a stagnant pool issuing parallel to a eddy size is technically important from the design point of view in
horizontal solid wall. connection with Re and the step length and height. It is a common
The flow emanating from a two-dimensional 共2D兲 plane wall jet belief that the peak Nusselt number will occur near the reattach-
is shown in Fig. 1共a兲 where the main features and regions of ment location. So, it is very important to find the reattachment
interest are depicted. Fluid is discharged from a slot along the location in connection with Re and geometry. Wall jet flow over a
vertical wall into the ambient near a horizontal solid boundary step has several similar applications like a refrigerated air curtain,
parallel to the inlet jet direction. The jet flow features are different the paper industry, etc. The understanding of the flow behavior of
in various regions. In the near-field up to the step from the point the wall jet over a step is important in engineering practices. If
of discharge, the jet behaves like a plane wall jet. Further the jet attachment is not desired for more mixing, a knowledge of calcu-
expands across the step. Due to entrainment between the solid lation and design is required to prevent it. For a case where at-
wall and the jet there is a reduction of pressure in this region, tachment is desirable, study of the involved variables is required
forcing the jet to deflect towards the boundary and eventually so that a precise location of the attachment point and the contain-
attach with it. This is called the Coanda effect. In the region ment of the flow can be established. Moreover, this simulation
around the attachment point, that is, the impingement region and will be helpful to set approximate guidelines for carrying out ex-
part of the recirculation region, the jet can be partly characterized periments.
as an impingement jet. The jet becomes a wall jet in the far field. Analytical solutions of the wall jet are available, based on the
Other factors like free-stream velocity, ambient stratification, self-similarity of the velocity field 关1,2兴. However, these solutions
buoyancy 共density difference兲, discharge orientation, etc., further are valid only far away from the jet inlet and, in most applications,
complicate the jet-boundary interaction and the behavior of a wall the near-field development holds the key to important features of
jet. the jet flow and recirculation. Therefore, the near-field develop-
ment of a wall jet has been the subject of a lot of research in
1
recent years.
Author to whom all correspondence should be addressed.
Contributed by the Fluids Engineering Division of ASME for publication in the
Amitay and Cohen 关3兴 have reported the effects of wall blowing
JOURNAL OF FLUIDS ENGINEERING. Manuscript received January 27, 2005; final manu- and suction on the stability characteristics of a laminar incom-
script received January 25, 2006. Review conducted by Joseph Katz. pressible two-dimensional plane wall jet. Cohen et al. 关4兴 did

Journal of Fluids Engineering Copyright © 2006 by ASME SEPTEMBER 2006, Vol. 128 / 1023

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Fig. 1 Schematic diagram and boundary conditions in a wall jet under backward-facing step problem

work on transition of wall jets subjected to blowing and suction. 2 Mathematical Formulation
Quintana et al. 关5兴 experimentally investigated the mean and fluc-
An incompressible two-dimensional laminar wall jet under
tuating characteristics of a plane unsteady laminar wall jet for
backward-facing step jet flow is considered. For the sake of sim-
constant wall temperature. Seidel 关6兴 has done numerical work to
plicity, the jet is assumed to be isothermal and have the same
find the effect of high amplitude forcing on laminar and turbulent
density as the ambient fluid. Also, the velocity profile at the jet
wall jets over a heated flat plate. Seidel has used DNS for laminar
inlet is taken as parabolic.
case and RANS for turbulent wall jet. Recently, Bhattacharjee and
The governing equations for incompressible laminar flow are
Loth 关7兴 simulated laminar and transitional cold wall jets. They
solved by stream-function-vorticity formulation. The transient
have investigated the significance of three different inlet profiles,
nondimensional governing equations in the conservative form are
viz., parabolic, uniform, and ramp, and presented the detailed re-
Stream function equation
sults of time-averaged wall jet thickness and temperature distribu-
tion. They used the RANS approach for higher Reynolds number ⵜ 2␺ = − ␻ 共1兲
and DNS approach for three-dimensional wall jet.
Vorticity equation
The stability of 2D free jets was studied experimentally by Sato
关8兴 and Sato and Sakao 关9兴. Theoretical analyses for axisymmetric ⳵␻ ⳵共u␻兲 ⳵共v␻兲 1 2
free jets were carried out by Batchelor and Gill 关10兴 and Cohen + + = ⵜ␻ 共2兲
⳵t ⳵x ⳵y Re
and Wygnanski 关11兴. Sarma et al. 关12兴 have studied 2D incom-
pressible jet development inside a duct in the laminar flow regime where ␺ is the stream function, u = ⳵␺ / ⳵y, v = −⳵␺ / ⳵x, and ␻
for cases with and without entrainment of ambient fluid. The ef- = ⳵v / ⳵x − ⳵u / ⳵y.
fect of channel length in the backward-facing step is studied by The variables are scaled as
Barton 关13兴. For similar geometry for the pulsatile inlet condition
heat transfer, results were reported by Valencia and Hinojosa 关14兴. ū v̄ x̄ ȳ ␻
¯ t̄
u= ; v= ; x= ; y= ; ␻= ; t=
Jacob et al. 关15兴 investigated the sound radiation by a backward- Ū Ū h h Ū/h h/Ū
facing step under a plane wall jet.
Though many studies have been conducted on wall jets, the with the overbar indicating a dimensional variable and Ū, h de-
available literature on theoretical simulation of plane laminar wall noting the average jet velocity at the nozzle exit and the jet width,
jets over step jet flow is not carried by anybody. The case of an respectively.
entraining jet located near the jet discharge, which occurs in sev- The boundary conditions needed for the numerical simulation
eral practical applications, has not been studied. In the present have been prescribed. For an offset jet with entrainment, the fol-
study, a time marching incompressible flow solver has been ap- lowing dimensionless conditions have been enforced as shown in
plied for simulating the flow features of wall jet for a range of step Fig. 1共b兲. The inlet slot height is assumed as h = 0.05 m.
geometry 共step length, l, step height, s兲 and Reynolds number. At the jet inlet, along AG 共Fig. 1共b兲兲,

1024 / Vol. 128, SEPTEMBER 2006 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 u共y兲 = 120y − 2400y 2 ; ␻共y兲 = 4800y − 120;

␺共y兲 = 60y 2 − 800y 3 共3a兲

Along FG, AB, BC, and CD due to no-slip condition,
u=v=0 共3b兲
Along EF,
⳵u
=0 共3c兲
⳵x
At the downstream boundary, the condition of zero first-
derivative has been applied for velocity components. This condi-
tion implies that the flow has reached a developed condition.
Thus, at 共DE兲,
⳵u ⳵v
= =0 共3d兲
⳵x ⳵x

3 Numerical Procedure
The unsteady vorticity transport equation 共2兲 in time is solved
by alternate direction implicit scheme 共ADI兲. The central differ-
encing scheme is followed for both the convective as well as the Fig. 2 Schematic diagram and boundary conditions of
diffusive terms 关16兴. It consists of two half time-steps. L-shaped lid driven cavity flow problem †30‡
The first half time-step:

Fig. 3 Velocity profiles along CL1 and CL2

Journal of Fluids Engineering SEPTEMBER 2006, Vol. 128 / 1025

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Fig. 4 Backward-facing step flow with upstream channel problem. „Domain shown near the step.…

␻i,j
n+1/2
− ␻i,j
n
− 共Cyvi−1,j
n
+ Sy兲␻i,1−j
n+1
+ 共1 + 2Sy兲␻i,j
n+1
− 共− Cyvi,j+1
n
+ Sy兲␻i,j+1
n+1
+ Lx共u␻兲i,j
n+1/2
+ Ly共v␻兲i,j
n
⌬t/2 = 共Cxui−1,j
n
+ Sx兲␻i−1,j
n+1/2
+ 共1 − 2Sx兲␻i,j
n+1/2

1 + 共− Cxui+1,j
n
+ Sx兲␻i+1,j
n+1/2
共6b兲
− 关Lxx共␻兲i,j
n+1/2
+ Lyy共␻兲i,j
n
兴=0 共4a兲
Re where
The second half time-step:

␻i,j
n+1
− ␻i,j
n+1/2
+ Lx共u␻兲i,j
n+1/2
+ Ly共v␻兲i,j
n+1
⌬t/2
1
− 关Lxx共␻兲i,j
n+1/2
+ Lyy共␻兲i,j
n+1
兴=0 共4b兲
Re
where,
共u␻兲i+1,j − 共u␻兲i−1,j 共v␻兲i,j+1 − 共v␻兲i,j−1
Lx共u␻兲i,j = , Ly共v␻兲i,j =
⌬xi + ⌬xi−1 ⌬y j + ⌬y j−1
共5a兲

␻i−1,j − 2␻i,j + ␻i+1,j

Lxx共␻兲i,j = , 共5b兲
⌬xi ⴱ ⌬xi−1

␻i,j−1 − 2␻i,j + ␻i,j+1

Lyy共␻兲i,j =
⌬y j ⴱ ⌬y j−1
Equations 共4a兲 and 共4b兲 are rearranged to give the following equa-
tions 共6a兲 and 共6b兲.

− 共Cxui−1,j
n
+ Sx兲␻i−1,j
n+1/2
+ 共1 + 2Sx兲␻i,j
n+1/2
− 共− Cxui+1,j
n
+ Sx兲␻i+1,j
n+1/2

= 共Cyvi,j−1
n
+ Sy兲␻i,j−1
n
+ 共1 − 2Sy兲␻i,j
n
+ 共− Cyvi,j+1
n
+ Sy兲␻i,j+1
n
Fig. 5 Reattachment length for different Reynolds number:
共6a兲 parabolic inlet profile

1026 / Vol. 128, SEPTEMBER 2006 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Table 1 Primary vortex reattachment length „ER= 0.25, Re Table 2 Primary vortex reattachment length „ER= 0.75, Re
= 500… = 200…

关32兴 关14兴 Present 关32兴 关14兴 Present

0.84628 1.1158 1.0707 6.4868 7.0261 7.1469

⌬t ⌬t
Cx = , Cy = , ⳵␺ ␺i,j+1 − ␺i,j−1
2共⌬xi + ⌬xi−1兲 2共⌬y j + ⌬y j−1兲 u= = 共8a兲
⳵y ⌬y j + ⌬y j−1
⌬t 1 ⌬t 1
Sx = , Sy = ⳵␺ ␺i+1,j − ␺i−1,j
Re ⌬xi ⴱ 共⌬xi + ⌬xi−1兲 Re ⌬y j ⴱ 共⌬y j + ⌬y j−1兲 v=− = 共8b兲
⳵x ⌬xi + ⌬xi−1
The discretization of Eq. 共1兲 is given by:
The velocities 共Eqs. 共6a兲 and 共6b兲兲 are calculated at the nth time
Lxx共␺兲 + Lyy共␺兲 = − ␻i,j 共7兲 level while advancing to the 共n + 1兲th time level. Because of this
The velocity components are updated by the following equations: approximation in the nonlinear terms, the second-order accuracy

Fig. 6 Streamline contour

Journal of Fluids Engineering SEPTEMBER 2006, Vol. 128 / 1027

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Fig. 7 Vorticity contour

of the method is somewhat lost. However, something of the merical tests verify that this first-order form is more accurate than
second-order accuracy of the linearized system is retained if the second-order form.
velocity field is slowly varying 关16兴. The wall jet consists of an inner region and an outer region. It
It is first-order accurate in time and second-order accurate in is a combination of boundary layer flow over a flat plate at the
space O共⌬t , ⌬x2 , ⌬y 2兲 and is unconditionally stable. The Poisson inner region and plane free jet at the outer region 关19兴. The veloc-
equation 共7兲 is solved explicitly by five-point Gauss-Seidel meth- ity profile has a point of inflexion. The surrounding medium of the
ods. Thom’s vorticity condition has been used to obtain the wall
wall jet may be quiescent or co-flow or counter-flow depending
vorticity as given below.
upon the applications. In the present case, the surrounding is con-
2共␺w+1 − ␺w兲 sidered to be quiescent.
␻w = − 共9兲
⌬n2 At the bottom wall and the left side wall, constant stream lines
where ⌬n is the grid space normal to the wall. It has been shown are assumed based on inlet flow. At the outlet in the downstream
by Napolitano et al. 关17兴 and Huang and Wetton 关18兴 that conver- direction, stream-wise gradients are assumed to be zero. At the
gence in the boundary vorticity is actually second order for steady entrainment boundary, the normal velocity gradient is zero 关20兴.
problems and for time-dependent problems when t ⬎ 0. Roache The detailed boundary conditions are,
关16兴 has reported that for a Blausius boundary-layer profile, nu- along FE 共Fig. 1共b兲兲,

1028 / Vol. 128, SEPTEMBER 2006 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Fig. 8 Effect of Re on wall vorticity: l = 2h, s = 1h

2共␺w − ␺w+1兲 For the computation, time step 0.0001 is used. It has been ob-
␻共y兲 = ; ␺ = 0.05 共10a兲 served that for coarse grids, a larger time step can be used,
⌬xi ⴱ ⌬xi−1
whereas, for fine grids, the solution diverges with large time step.
along BA, While selecting ␬, a time step also needs to be considered. The
2共0 − ␺w+1兲 maximum vorticity error behavior is complicated as explained by
␻共y兲 = ; ␺=0 共10b兲 Roache 关16兴. While marching in time for the solution, it has been
⌬xi ⴱ ⌬xi−1 observed that the maximum vorticity error gradually decreases. It
along BC, then increases drastically and finally decreases asymptotically,
leading to a steady-state solution. The convergence criterion is to
2共0 − ␺w+1兲 be set in such a way that it should not terminate at false stage. At
␻共x兲 = ; ␺=0 共10c兲
⌬y j ⴱ ⌬y j−1 steady state, the error reaches the asymptotic behavior. Here it is
set as the sum of the vorticity error reduced to either the conver-
The rationality of the entrainment and exit boundary conditions
are given in detail in 关21兴. gence criteria ␧ or large total time.
The solution approaches steady state asymptotically while the imax,j max

time reaches infinity. The computational domain considered here

is clustered Cartesian grids. For unit length, the grid space at ith
兺
i,j=1
共␻i,j
t+⌬t
− ␻i,j
t
兲⬍␧ 共12兲

node is 关22兴, Comini et al. 关23兴 used low Re steady-state solution as an initial

xi = 冋 ␬i
− sin
imax ␽
i␽
imax
冉 冊册 共11兲
guess value for high Re flow for stream function and vorticity.
The same practice is followed here also.

where ␽ is the angle and ␬ is the clustering parameter. ␽ = 2␲

stretches both ends of the domain whereas ␽ = ␲ clusters more 4 Validation of the Code
grid points near one end of the domain. ␬ varies between 0 and 1. To validate the developed code, the two-dimensional lid-driven
When it approaches 1, more points fall near the end. square-cavity flow problem 关24兴 and the backward-facing flow

Journal of Fluids Engineering SEPTEMBER 2006, Vol. 128 / 1029

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Fig. 9 Effect of step length on wall vorticity: Re= 400, s = 1h

problem 关25,26兴 have been solved. Excellent agreement has been v-velocity components at these locations are compared with 关30兴
obtained with the benchmark solutions and reported elsewhere for Re= 100 and Re= 1000 共Fig. 3兲 and excellent agreement is
关27兴. The laminar plane wall jet problem then has been solved and obtained.
the computed velocity profiles are compared with the similarity
solutions of Glauert 关1兴 and the experimental results of Quintana 4.2 Backward-Facing Step Flow With Upstream Channel.
et al. 关5兴 in a similar way as represented by Seidel 关6兴. Details are Backward-facing step with upstream channel flow is the problem
given by Kanna and Das 关27兴. of interest for the study of sudden expansion flows. Armaly et al.
To validate the present numerical procedure, the sudden expan- 关25兴 have reported the experimental results in which the geometry
sion flow problem and split domain problems are solved. The has a long inlet channel and fully developed flow is expected at
computational domain reported by Durst et al. 关28兴 is considered the expansion. To simulate the experimental situation, the inlet
for sudden expansion flow. The details of the comparison are channel is considered in the numerical computation also. Kondoh
given by Kanna and Das 关29兴. et al. 关31兴 reported heat transfer study on backward-facing step
Due to scarcity of experimental results of the present geometry, flow with upstream channel geometry. They considered various
the following two split domain problems are solved and compared expansion ratios. However, their reattachment length 共ER= 2.0,
with the published results. Re= 100, x1 / h = 4.8兲 is larger than the other published work
共x1 / h = 2.8 关25,14兴兲. Barton studied in detail the effect of the inlet
1. L-shaped lid driven cavity problem.
channel 关13兴. He has reported the reattachment length by setting
2. Backward-facing step flow with upstream channel.
the inlet at different upstream locations in the channel for various
4.1 L-shaped Lid Driven Cavity Problem. The schematic Reynolds numbers. Thangam and Knight 关32兴 studied the effect of
diagram and boundary conditions of the problem are shown in step height in the sudden expansion flow.
Fig. 2. The geometry considered is similar to that of Oosterlee et The geometry considered here is similar to the one considered
al. 关30兴. The top lid is moving in the positive x direction. Along by Barton 关13兴, which is having 10⫻ h as the upstream channel
the solid walls a no-slip condition is applied. Results are com- length. The schematic diagram of the problem and boundary con-
pared at two locations named CL1 and CL2 共Fig. 2兲. u- and ditions are shown in Fig. 4共a兲. Clustered grids are used near the

1030 / Vol. 128, SEPTEMBER 2006 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Fig. 10 Effect of step height on wall vorticity: Re= 400, l = 2h

step, which is shown in Fig. 4共b兲. At the inlet, the parabolic ve- at this length. To understand the flow physics of the present situ-
locity profile is assumed and at exit the fully developed condition ation, three parameters considered here are Re共=Ūh / ␯兲, step
is assumed. The expansion ratio 共ER兲 is defined as h / H. The length 共l兲, and step height 共s兲. Results are presented here for l
streamline contour and velocity vector are shown in Figs. 4共c兲 and
= 1h, 2h, 3h and s = 0.5h, 1h, 2h. Since at low Re the jet has
4共d兲, respectively for Re= 500 and ER= 0.5. The ER= 0.75 case is
difficulty overcoming the downstream friction 关33兴, Re has been
considered to visualize the effect of ER. Results are shown for
chosen between 300 and 600. The experimental study of the lami-
Re= 200, ER= 0.75 in Figs. 4共e兲 and 4共f兲. For large values of Re nar plane wall jet is presented in 关19兴. Based on the jet exit Rey-
the primary vortex reattachment length is compared with 关25,13兴 nolds number, they reported that the laminar wall jet results up to
共Fig. 5兲. The present computations show excellent agreement with
Re= 770. Bhattacharjee and Loth 关7兴 reported that early transition
computational results and up to Re= 400 with experimental re-
begins at about Re= 700 for the plane wall jet. Kanna and Das
sults. Backward-facing steps with different expansion ratios are
关29兴 have shown that the laminar wall attaching the offset jet can
tested and the reattachment length is compared with the published
be possible up to Re= 600. Based on these results, the maximum
results and good agreement is obtained 共Tables 1 and 2兲.
Re for the wall jet flow over a step is set as Re= 600.
The clustered grids are used for the computations. The domain
5 Results and Discussion has been chosen as 25⫻ h in the streamwise direction from the
The parabolic profile, issuing from the inlet slot, spreads over a step and 20⫻ h in normal direction. For the grid independence
plane wall and further expands across a backward-facing step. It study the reference case considered is Re= 400, l = 2h, and s = 1h.
reattaches on the bottom wall due to the Coanda effect and Three grid systems 71⫻ 61, 97⫻ 85, and 127⫻ 125 are consid-
spreads along the wall 共Fig. 1共a兲兲. The length between the step ered. The variation in the reattachment length is less than 2.5%
and the reattachment point is called the reattachment length 共x1, and the 97⫻ 85 grid system is used for the entire computation.
Fig. 1共b兲兲. The flow in the domain has the nature of the wall jet, The grids are clustered near the step and solid walls. Uniform
recirculation, developing, and the wall jet at far downstream. Be- grids are considered for the inlet to ensure that the parabolic ve-
fore flow expands from the step, it behaves like a plane wall jet. locity profile falls exactly at the inlet slot. Typical grids are shown
However, due to entrainment, self-similarity nature is not obtained in Fig. 1共c兲.

Journal of Fluids Engineering SEPTEMBER 2006, Vol. 128 / 1031

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Fig. 11 Similarity profile Re= 400, l = 2h, s = 1h

Figure 6 shows the streamline contours for a range of the pa- the walls. The effect of these parameters on vorticity is shown in
rameters Re, l, and s. The streamline contours illustrate the main Fig. 7. Near the bottom wall the u velocity gradient is greater in
jet flow, recirculation, and entrainment flow. When Re increases the normal direction than the v velocity gradient in the streamwise
from 300 to 600, the main flow is confined towards the bottom direction. It is noticed that when Re increases this is magnified
wall. Particularly after the step, the jet deflects towards the bottom 共Figs. 7共a兲 and 7共b兲兲. For longer step, vorticity values are reduced
wall 共Figs. 6共a兲 and 6共b兲兲. The size of the recirculation eddy is 共Figs. 7共c兲 and 7共d兲兲. However, when height of the step is in-
increased considerably. At low Re the jet spreads more in the creased, the u-velocity normal direction gradient is increased
normal direction, which causes the reduction in ambient fluid 共Figs. 7共e兲 and 7共f兲兲.
flow. It is noticed that the increase in the length of the step reduces Walls are the sources of vorticity. Here solid walls are assumed
the reattachment length 共Figs. 6共c兲 and 6共d兲兲. When the height of as impermeable which exist along AB, BC, CD, and FG 共Fig.
the step is increased, the main flow deflects towards the bottom 1共b兲兲. It may have either positive or negative values depending
wall. Recirculation eddy size is increased 共Figs. 6共e兲 and 6共f兲兲. At upon the velocity gradients. Since walls are a source of vorticity,
exit, the streamlines become parallel and thus the assumption for the magnitude of vorticity along the wall will be higher than other
fully developed flow is satisfied. Near the bottom wall, stream- parts of the domain. Thus the distribution of vorticity along the
lines are less dense in order to satisfy the no-slip conditions along

Table 4 Effect of step length on wall jet similarity region dis-

Table 3 Effect of Re on wall jet similarity region distance. l tance. Re= 300, s = 1h.
= 2h, s = 1h.
Length of the step x/h
Re x/h
l = 1h 8.84
300 7.72 7.72
400 7.74 l = 2h
500 7.99 l = 3h 6.42
600 8.33 l = 0h 关29兴 14.24

1032 / Vol. 128, SEPTEMBER 2006 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Table 5 Effect of step height on wall jet similarity region dis-
tance. Re= 300, l = 2h.

Height of the step x/h

s = 0.5h 9.65
s = 1h 7.74
s = 2h 7.56

solid walls is important for understanding the physics of the flow.

The effect of Re and step geometry on wall vorticity is presented
in detail in Figs. 8–10. Vorticity distributions are shown for ver-
tical walls followed by horizontal walls. The effect of Re on vor-
ticity along BC is shown in Fig. 8共a兲. Along step height the vor-
ticity is in the positive direction and its value is increased when
Re is increased. However, the maximum vorticity occurred at the
same height 共y = 0.833h兲. It is important to note that along BC
there exist two singular corner points. Roache 关16兴 reported in
detail about how to treat these corner singular points. However, at
high Re, he concluded that those forms are less efficient. Above
the inlet, the left wall causes vorticity mainly due to the entrain-
ment of the ambient fluid. Vorticity is generated by the left wall up Fig. 12 Effect of step height on upstream velocity at x = 1h:
to a maximum of y = 2h above the inlet and further it approaches Re= 400, l = 2h
zero 共Fig. 8共b兲兲. At higher Re it increases in the negative direction.
This is attributed to the high entrainment towards the inlet. The
vorticity is generated due to negative v velocity gradient. The is reduced for longer step length. However, far downstream, the
vorticity variation along step length is shown in Fig. 8共c兲. When effect of step length is less significant. The effect of step height is
Re increases it tends towards negative direction because the nor- shown in Fig. 10. Along BC the vorticity is increased in the posi-
mal direction u gradient is increased. Along CD the wall vorticity tive direction to a maximum value. As the top corner is ap-
varies from positive to negative in the downstream direction 共Fig. proached, it tends towards the negative direction. The magnitude
8共d兲兲. From the step it increases in the positive direction and then is increased when step height is increased 共Fig. 10共a兲兲 due to large
decreases. After reattachment its direction is changed to the nega- recirculation, which causes more positive v velocity gradient.
tive direction. At far from the reattachment it becomes invariant at Along FG, the magnitude of the vorticity is less affected by the
high Re. variation in height. However, due to more entrainment at high step
The effect of step length on vorticity is shown in Fig. 9. It is height, negative vorticity is generated far away from the inlet 共Fig.
noticed that when the length of the step is increased, wall vorticity 10共b兲兲. The variation in step height does not have significant effect
on BC is decreased 共Fig. 9共a兲兲. However, the maximum values on the vorticity along the step length 共Fig. 10共c兲兲. Thus the sepa-
occur at the same height. When length is increased, the left wall ration in the flow is less significant for the upstream flow wall
vorticity is decreased 共Fig. 9共b兲兲 due to the reduction of entrain- vorticity. It is noticed that the positive direction peak value is
ment. Figure 9共c兲 shows the effect of change in step length on higher for s = 2h and the negative direction peak value is higher
vorticity. It is observed that when length increases, the wall vor- for s = 0.5h 共Fig. 10共d兲兲.
ticity is decreased for the same distance. Along the length of the The wall jet boundary layer thickness 共␦兲 is the normal distance
step, the vorticity value tends towards positive direction. It is no- where u = 0.5umax. The similarity u velocity profile at different
ticed that the vorticity along CD is affected by the upstream step downstream locations is compared with experimental 关5兴 and
length 共Fig. 9共d兲兲. The peak value of the vorticity in any direction similarity 关1兴 results 共Fig. 11兲 for Re= 400, l = 2h, and s = 1h. The

Fig. 13 Local maximum u velocity: l = 2h, s = 1h

Journal of Fluids Engineering SEPTEMBER 2006, Vol. 128 / 1033

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Table 7 Effect of step height on primary vortex reattachment
length. Percentage of variation from s = 1h: s = 1h.

% of increment % of reduction
Re x1 / h for s = 1h if s = 2h if s = 0.5h

300 2.128 118.65 58.51

400 2.239 113.31 60.87
500 2.386 104.48 57.88
600 2.507 100.03 55.08

共i.e., y / h兲 is shown in Fig. 13共b兲. It is noticed that when Re in-

creases, the jet umax location 共y / h兲 shifts down. The reattachment
lengths of all the cases considered here are presented in Fig. 14. It
is observed that the increase in step length reduces the reattach-
ment length. However, it is noticed that the reattachment length is
reduced when the step length is absent 关29兴. This is mainly due to
the large entrainment near the expansion of the jet. When step
height increases, increase in recirculation length is observed.
However, the variation in x1 is more sensitive to step height than
step length 共Tables 6 and 7兲.
Fig. 14 Reattachment length for different geometry and Rey-
nolds number: l = 0h, s = 1h †29‡ 6 Conclusions
The two-dimensional incompressible nonbuoyant laminar wall
u velocity profiles at locations in the middle of step length 共x jet over backward-facing step is solved by stream function and
= 1h兲, reattachment point 共x = 2.874h兲, and inlet are shown in Fig. vorticity formulation considering the problem as an asymptotic
11共a兲. Due to entrainment, a negative direction velocity value is solution to the transient equation. The flow characteristics like
present. The same is shown at a few downstream locations 共Fig. strength of recirculation eddy, impingement length, decay of local
11共b兲兲. After reattachment, flow is under development to a certain velocity, and wall jet region are studied in detail for a range of
flow property and physical property. The entrainment is increased
distance. Location x = 5h corresponds to the developing region
for high Re and it is less affected with change in step length. The
velocity profile. Similarity profiles in the wall jet region are
reattachment length follows almost a linear trend with Re and step
shown in Fig. 11共c兲. Present computations show excellent agree-
geometry for the range considered. However, it is more sensitive
ment with experimental and similarity results for the inner region
to step height than length. The upstream flow properties are not
as well as the outer region of the wall jet. However, at further
affected by expansion in the geometry. Vorticity from the solid
downstream locations, the jet expands in the normal direction and
wall is affected by Re. The maximum vorticity location is inde-
thus it does not have agreement with the similarity solution at the
pendent of Re and step geometry. Vorticity from the step height is
entrainment location, i.e., free shear layer. The experimental result
increased for higher Re and decreased if step length is increased.
shown here is chosen from one particular location where it satis-
From the left wall, vorticity is generated to a few inlet slot heights
fies the similarity solution. The length from the step to the wall jet
and further it is independent of Re and step geometry. At the
region, i.e., where the similarity profiles begin and closed-form
bottom wall, vorticity becomes a negative value from a positive
solutions of the wall jet can be applicable, is measured for differ-
ent Re and step geometry. It is observed that when Re is in- value after reattachment point. The umax location shifts down
creased, due to more entrainment, distance where the similarity when Re is increased. Similarity is not observed along step length
principle can be applied is also increased 共Table 3兲. When length due to entrainment. Similarity exists for the wall jet far away from
increases, this distance is reduced due to reduction of entrainment. the reattachment. Similarity profile formation distance is reduced
However, this distance is larger for the zero step length case by increasing the step geometry whereas an increment in Re in-
共Table 4兲. This is caused by the entrainment. It is observed that creases this distance. The bottom wall vorticity is considerably
when step height increases this distance is reduced 共Table 5兲. affected near the impingement region.
The effect of the step height on the upstream velocity is shown
in Fig. 12. The local u velocity at the x = 1h location is compared Nomenclature
with different step height. It is noticed that the downstream step
h ⫽
inlet slot height, m
height variation does not affect the local u velocity upstream. The
i ⫽
x-direction grid point
maximum u velocity in the normal direction from the entire
downstream direction is presented for different Re and shown in
j ⫽
y-direction grid point
Fig. 13. The local normalized 共i.e., umax / Umax兲 umax decays mono- l ⫽
step length, m
tonically in the downstream direction, finally reaching an n ⫽
normal direction
asymptotic condition. The corresponding normalized y location Re ⫽
Reynolds number for the fluid
s ⫽
step height, m
t̄ ⫽
dimensional time, s
Table 6 Effect of step length on primary vortex reattachment t ⫽
nondimensional time
length. Percentage of reduction from l = 1h: s = 1h. ū , v̄ ⫽
dimensional velocity components along 共x , y兲
axes, m/s
Re x1 / h for l = 1h % of reduction if l = 2h % of reduction if l = 3h
u , v ⫽ dimensionless velocity components along 共x , y兲
300 2.378 10.51 15.39 axes
400 2.389 6.28 11.01 Ū ⫽ inlet mean velocity, m/s
500 2.514 5.09 10.02
600 2.643 5.15 9.72 x̄ , ȳ ⫽ dimensional Cartesian coordinates along and
normal to the plate, m

1034 / Vol. 128, SEPTEMBER 2006 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 x , y ⫽ dimensionless Cartesian coordinates along and Study of Sound Generated by Backward-Facing Steps Under Wall Jet,” AIAA
J., 39共7兲, pp. 1254–1260.
normal to the plate 关16兴 Roache, P. J., 1998, Fundamentals of Computational Fluid Dynamics, Her-
mosa, Albuquerque, NM, Chap. 3.
Greek Symbols
关17兴 Napolitano, M., Pascazio, G., and Quartapelle, L., 1999, “A Review of Vortic-
␦ ⫽ wall jet boundary layer thickness ity Conditions in the Numerical Solution of the ␨-␺ Equations,” Comput.
␧ ⫽ convergence criterion Fluids, 28, pp. 139–185.
␩ ⫽ similarity variable 关18兴 Huang, H., and Wetton, B. R., 1996, “Discrete Compatibility in Finite Differ-
␬ ⫽ clustering parameter ence Methods for Viscous Incompressible Fluid Flow,” J. Comput. Phys., 126,
pp. 468–478.
␺ ⫽ dimensionless stream function 关19兴 Bajura, R. A., and Szewczyk, A. A., 1970, “Experimental Investigation of a
␻ ⫽ dimensionless vorticity Laminar Two-Dimensional Plane Wall Jet,” Phys. Fluids, 13, pp. 1653–1664.
关20兴 Kang, S. H., and Greif, R., 1992, “Flow and Heat Transfer to a Circular
Subscripts Cylinder With a Hot Impinging Air Jet,” Int. J. Heat Mass Transfer, 35共9兲, pp.
max ⫽ maximum 2173–2183.
w ⫽ wall 关21兴 Kanna, P. R., and Das, M. K., 2006, “A Short Note on the Entrainment and
Exit Boundary Conditions,” Int. J. Numer. Methods Fluids, 50共8兲, pp. 973–
985.
References 关22兴 Kuyper, R. A., Van Der Meer, T. H., Hoogendoorn, C. J., and Henkes, R. A.
W. M., 1993, “Numerical Study of Laminar and Turbulent Natural Convection
关1兴 Glauert, M. B., 1956, “The Wall Jet,” J. Fluid Mech., 1共1兲, pp. 1–10.
in an Inclined Square Cavity,” Int. J. Heat Mass Transfer, 36共11兲, pp. 2899–
关2兴 Schlichting, H., and Gersten, K., 2000, Boundary Layer Theory, 8th ed.
2911.
Springer, New York.
关23兴 Comini, G., Manzan, M., and Nonino, C., 1994, “Finite Element Solution of
关3兴 Amitay, M., and Cohen, J., 1997, “Instability of a Two-Dimensional Plane
the Streamfunction-Vorticity Equations for Incompressible Two-Dimensional
Wall Jet,” J. Fluid Mech., 344, pp. 67–94.
关4兴 Cohen, J., Amitay, M., and Bayly, B. J., 1992, “Laminar-Turbulent Transition Flows,” Int. J. Numer. Methods Fluids, 19, pp. 513–525.
of Wall-Jet Flows Subjected to Blowing and Suction,” Phys. Fluids A, 4, pp. 关24兴 Ghia, U., Ghia, K. N., and Shin, C. T., 1982, “High Re Solutions for Incom-
283–289. pressible Flow Using the Navier-Stokes Equations and Multigrid Method,” J.
关5兴 Quintana, D. L., Amitay, M., Ortega, A., and Wygnanski, I. J., 1997, “Heat Comput. Phys., 48, pp. 387–411.
Transfer in the Forced Laminar Wall Jet,” ASME J. Heat Transfer, 119, pp. 关25兴 Armaly, B. F., Durst, F., Pereira, J. C. F., and Schonung, B., 1983, “Experi-
451–459. mental and Theoretical Investigation of Backward-Facing Step Flow,” J. Fluid
关6兴 Seidel, J., 2001, “Numerical Investigations of Forced Laminar and Turbulent Mech., 127, pp. 473–496.
Wall Jets Over a Heated Surface,” Ph.D. thesis, Faculty of the Department of 关26兴 Gartling, D. K., 1990, “A Test Problem for Outflow Boundary Conditions-
Aerospace and Mechanical Engineering. The Graduate College, The Univer- Flow Over a Backward-Facing Step,” Int. J. Numer. Methods Fluids, 11, pp.
sity of Arizona, Tucson, AZ. 953–967.
关7兴 Bhattacharjee, P., and Loth, E., 2004, “Simulations of Laminar and Transi- 关27兴 Kanna, P. R., and Das, M. K., 2005, “Conjugate Forced Convection Heat
tional Cold Wall Jets,” Int. J. Heat Fluid Flow, 25, pp. 32–43. Transfer From a Flat Plate by Laminar Plane Wall Jet Flow,” Int. J. Heat Mass
关8兴 Sato, H., 1960, “The Stability and Transition of a Two-Dimensional Jet,” J. Transfer, 48, pp. 2896–2910.
Fluid Mech., 7, pp. 53–80. 关28兴 Durst, F., Pereira, J. C. F., and Tropea, C., 1993, “The Plane Symmetric
关9兴 Sato, H., and Sakao, F., 1964, “An Experimental Investigation of the Instabil- Sudden-Expansion Flow at Low Reynolds Numbers,” J. Fluid Mech., 248, pp.
ity of a Two-Dimensional Jet at Low Reynolds Numbers,” J. Fluid Mech., 20, 567–581.
pp. 337–352. 关29兴 Kanna, P. R., and Das, M. K., 2005, “Numerical Simulation of Two-
关10兴 Batchelor, G. K., and Gill, A. E., 1962, “Analysis of the Stability of Axisym- Dimensional Laminar Incompressible Offset Jet Flows,” Int. J. Numer. Meth-
metric Jets,” J. Fluid Mech., 14, pp. 529–551. ods Fluids, 49共4兲, pp. 439–464.
关11兴 Cohen, J., and Wygnanski, I., 1987, “The Evolution Instabilities in the Axi- 关30兴 Oosterlee, C. W., Wesseling, P., Segal, A., and Brakkee, E., 1993, “Benchmark
symmetric Jet. Part 1. The Linear Growth of Disturbances Near the Nozzle,” J. Solutions for the Incompressible Navies-Stokes Equations in General Co-
Fluid Mech., 176, pp. 191–219. ordinates on Staggered Grids,” Int. J. Numer. Methods Fluids, 17, pp. 301–
关12兴 Sarma, A. S. R., Sundararajan, T., and Ramjee, V., 2000, “Numerical Simula- 321.
tion of Confined Laminar Jet Flows,” Int. J. Numer. Methods Fluids, 33, pp. 关31兴 Kondoh, T., Nagano, Y., and Tsuji, T., 1993, “Computational Study of Laminar
609–626. Heat Transfer Downstream of a Backward-Facing Step,” Int. J. Heat Mass
关13兴 Barton, I. E., 1997, “The Entrance Effect of Laminar Flow Over a Backward- Transfer, 36共3兲, pp. 577–591.
Facing Step Geometry,” Int. J. Numer. Methods Fluids, 25, pp. 633–644. 关32兴 Thangam, S., and Knight, D. D., 1989, “Effect of Step Height on the Separated
关14兴 Valencia, A., and Hinojosa, L., 1997, “Numerical Solution of Pulsating Flow Flow Past to Backward Facing Step,” Phys. Fluids A, 1, pp. 604–606.
and Heat Transfer Characteristics in a Channel With Backward-Facing Step,” 关33兴 Chiriac, V. A., and Ortega, A., 2002, “A Numerical Study of the Unsteady
Heat Mass Transfer, 32, pp. 143–148. Flow and Heat Transfer in a Transitional Confined Slot Jet Impinging on an
关15兴 Jacob, M. C., Louisot, A., Juve, D., and Guerrand, S., 2001, “Experimental Isothermal Surface,” Int. J. Heat Mass Transfer, 45, pp. 1237–1248.

Journal of Fluids Engineering SEPTEMBER 2006, Vol. 128 / 1035

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use