9-4 MOMENTUM INTEGRAL EQUATION 415

JIld ary-layer
:knesses are
am speed is
INTRODUCTION
@, located TO
tic pressure FLUID
1m dynamic
MECHANICS
SIXTH EDITION

Jisplace-
ROBERT W. FOX
Purdue University
ALAN T. McDONALD
-eam dy- Purdue University
PHILIP J PRITCHARD
Manhattan College

the area
rea than
) will be
I CD.

@
:ement . , JOHN WILEY & SONS, INC.

9-4 MOMENTUM INTEGRAL EQUATION

Blasiu s' exact solution involved performing a rather subtle mathem atical transforma­
tion of two differential equations based on the insight that the lamin ar boundary layer
velocity profile is self- similar-only its sca le changes as we move along the plate.
Even with this transformation, we note that numerical integration was necessary to
obtain results for the boundary-layer thickness 8(x), velocity profile u/U versus y/8,
and wall shear stress T.ix). Furthermore, the analysis is limited to laminar bound ary
layers only (Eq. 9.4 does not include the turbulent Reynolds stresses discussed in
Chapter 8), and for a fl at plate only (no pressure variations).
416 CHAPTER 9 / EXTERNAL INCO MPRESSIBLEVISCOUS FLOW

u(x) _

, .. "
, }.~~:- - -
I
I
- --.::~
]
c

I
I
---
---r o(x)
I I
CV - - I
.r
l
I
I I

Lx
I I
I
1/ L ~d

l-dx ~1
Fig. 9.4 Differential control volume in a boundary layer.

To avoid these difficulties and limitations, we now consider a method for de­
riving an algebraic equation that can be used to obtain approximate information on
boundary-layer growth for the general case (laminar or turbulent boundary layers,
with or without a pressure gradient). The approach is one in which we will again
apply the basic equations to a control vol ume. The derivation, from the mass con­
servation (OT continuity) equation and the momentum eq uation, will take several
pages.
Consider incompressible, stead y, two-dimensional flow over a so lid sUIface. The
boundary -layer thickness, 8, grows in some manner with increasing distance, x. For
our analysis we choose a difielential control volume, of length dx, width w, and
height 8(x), as shown in Fig. 9.4. The freestream ve locity is U(x).
We wish to determine the boundary-layer thickness, 8, as a function of x. There
wi II be ma ss flow across surfaces ab and cd of differential control volume abed.
What about surface be? Will there be a mass flow across this suIt'ace? In Example
Problem 9.2, (on the CD), we showed th at the edge of the boundary layer is not a
st reamline. Thus there will be mass flow across surface be. Since control surt'ace ad
is adjacent to a so lid boundary, there will not be flow across ad. Before considering
the forces acting on the co ntrol volume and the momentum fluxes throug h the control
surt'ace, let us apply the continuity equa tion to determine the ma ss flux through each
portion of the control surface.

a. Continuity Equation
Basic equation:
= 0(1)

;at if cv
p dll + 1 V.
cs
p dA =0 (4.12)

Ass umptions: (1) Stead flow.

(2) Two-dimensional flow.

Then

Jrcs pV· dA = 0

or
. . .
m.lx = -m.ab - med
 9-4 MOMENTUM INTEGRAL EQUATION 417

Now let us evaluate these terms for the differential control volume of width w:

Surface Mass Flux

ab SUiface ab is located at x. Since the flow is two-dimen sional (no variation with z),
the mass flux through ab is

lilah = -{J: PUdY}W

cd Surface cd is located at x + dx. Expanding In in a Taylor series about location x,
Dd for de­ we obtain
mati on on
ry layers,
i11 x +dx
. +a
I11 x
din] dx
Niff again x x
nass con­
:e several and hence

face . The
~e, x. For
h W, and
li1 ctl = {J: pu dy + :x [J:pu dy Jdx }W

be Thus for surface be we obtain

' x. There
ne abed.
Example
is not a
Inbc = -{ :x [J: pu dy JdX}W
wface ad
lsidering Now let us consider the momentum fluxes and forces associated with control
~ control
volume abed. These are related by the momentum equation.
Igh each
b. Momentum Equation
Apply the x component of the momentum equation to control volume abed:
Basic equation:
= 0(3) = 0(1)

(4. I 2)
Fsx + {= $rIJ
;8> CV
up dV + J upV . i4
cs
(4.18a)

Assumption: (3) Fax = O.

Then

Fs x = mfab + mfbc + mfcd

where mf represents the x component of momentum flux .

To apply this equation to differential control volume abed, we must obtain ex­
pressions for the x momentum flux through the control surface and also the surface
forces acting on the control volume in the x direction. Let us consider the momentum
flux first and again consider each segment of the control surface.
418 CHAPTER 9 I EXTERNAL INCOMPRESSIBLE VISCOUS FLOW

Surface Momentum Flux (mO

ab Surface ab is located at x. Since the flow is two-dimensional, the x momentum flux

through ab is

mf"b = -U;UPUdY}W

ed Surface ed is located at x + dx. Expanding the x momentum flux (mf) in a Taylor

series about location x, we obtain

mfx+tLt = mfx + - ­ amf] dx

ox x

or

mfed = {f: u pu dy + aax [f; upu dY}X}W

be Since the mass crossing slllface be has velocity component U in the x direction, the
x momentum flux across be is given by
mfbc = Umbc

mfbc = -U{ aax[J; pu dy }tx}w

From the above we can evaluate the net x momentum flux through the control
surface as

fcs u pV . dA = -{J: u pu dV}W + {J: u pu ct.V}W

+ {:x [J: upu dy JdX}W - u{ dd [J: pu dy Jd X}W
x
Collecting terms, we find that

fcs u pV . dA = + {:X [J:u pu dY] dx - U:x [J: pu elyJdX}W

Now that we have a suitable expression for the x momentum flu x through the control
surface, let us consider the surface forces acting on the control volume in the x direc­
tion. (For convenience the differential control volume has been redrawn in Fig. 9.5.)
We recognize that normal forces having nonzero components in the x direction act on
three surfaces of the control surface. In addition, a shear force acts on sllIface ad.
c
_----' --1
I do

h !~ I I I

iI 0\ I:
I I
a IL _ __ _ _ _ __ _ _ _ JI d

I-dx-I
Fig. 9.5 Differential
control volume
 9-4 MOMENTUM INTEGRAL EQUATION 419

Since the velocity gradient goes to zero at the edge of the boundary layer, the shear
force acting along surface be is negligible.
omentum flux
Surface Force

ah If the press ure at x is P, then the force acting on surface ah is given by

) in a Taylor I F ob = PWO
(The boundary layer is very thin ; its thickness has been greatly exaggerated in all
the sketches we have made. Because it is thin, pressure variations in the y direction
may be neglected , and we ass ume that within the boundary la yer, p = p(x).)

ed Expanding in a Taylor series, the pressure at x + dx is given by

P
Px+dx = P + -d ] dx
dx x
direction, the
The force on surface cd is then given by

0:d = - [p+ ddxP ] x dx JW(O + do)

he The average press ure acting over surface he is

+ ~ dP ] dx
the control P 2 dx x

Then the x component of the normal force acting over he is given by

Fhc = (p+ ~ :1 dXJWdO

v ad The average shear force acting on ad is given by

Fad = -(Tw + 1dT w)wdx

Summing the x components of all forces acting on the control volume, we obtain
=0 =0
the control
l: , I }
F:,:, + { - dx 0 dx - 2: dx d1dO - Tw dx - 2:1 d7'dx
dp 1 dp .
he x direc­ W
I Fig. 9.5.)
tion act on where we note that dx do « 0 dx and dTw« T "" and so neglect the second and
Uiface ad. fourth terms.
Substituting the expressions for
. ,
equatIOn, we obtam
1
~
U pv . dA and Fs x
into the x momentum

{-:OdX-Twdx}W = {:x[J:UPUdY]dt-U :x[J: PUdY]dX}W

Dividing this equ ation by dxgives
W

-0 -dp - Tw= - a 1
° a 1°
dx ax 0 pu dy - U -ax 0 pu dy
U (9.16)

Equation 9.16 is a "momentum integral" equation that gives a relation between the x
components of the forces acting in a boundary layer and the x momentum flux .
420 CHAPTER 9 I EXTERNAL INCOMPRESSIBLE VISCOUS FLOW

The pressure gradient, dp/dx, can be determined by applying the Bernoulli

equation to the inviscid flow outside the boundary layer: dp/dx = - pU dU/dx. If we
recognize that 0 = J: dy, then Eq. 9.16 can be written as

'T w = - ~ rDu pu dy + U ~ rD pu dy + dU rD pU dy
ax Jo ax Jo dx Jo
Since
U- a iD pu dy = - a iD puU dy - -dU i pu dy
D

ax 0 ax 0 dx 0

we have
'T w =- a i D pu(U-u)dy+-
dU i p(U-u)dy
D

ax 0 dx 0

and
'Tw = ~U2 ro P~ (I- ~)dY + U dU r
ax J
D

U U dx Jo
D
P(l- ~)dY
U
Using the definitions of displacement thickness, 0* (Eq. 9.1), and momentum thick­
ness, () (Eq. 9.2), we obtain

(9.17)

Equation 9.17 is the momentum integral equation. This equation will yield an
ordinary differential equation for boundary-layer thickness, provided that a suit­
able form is assumed for the velocity profile and that the wall shear stress can
be related to other variables. Once the boundary-layer thickness is determined, the
momentum thickness, displacement thickness, and wall shear stress can then be
calculated.
Equation 9.17 was obtained by applying the basic equations (continuity and x
momentum) to a differential control volume. Reviewing the assumptions we made in
the derivation, we see that the equation is restricted to steady, incompressible, two­
dimensional flow with no body forces parallel to the surface.
We have not made any specific assumption relating the wall shear stress, 'Tw, to
the velocity field. Thus Eq. 9.17 is valid for either a laminar or turbulent boundary­
layer flow. In order to use this equation to estimate the boundary-layer thickness as a
function of x, we must:
1. Obtain a first approximation to the freestream velocity distribution, Vex). This is deter­
mined from inv iscid flow theory (the ve locity that would exist in lhe ahsence of a bound­
ary layer) and J epentC on ody ~ h a pe. The pressure in the boundary layer is related to the
freestrcil m \ cloci l_" U( -), using the Bernoulli equation.
2. Assume a reasonable velocity-profile shape inside the boundary layer.
3. Derive an expression for 'Tw using the results obtained from item 2.
To illustrate the application of Eq. 9.17 to boundary-layer flows, we consider
first the case of flow with zero pressure gradient over a flat plate (Section 9-5)-the
results we obtain for a laminar boundary layer can then be compared to the exact Bla­
sius results. The effects of pressure gradients in boundary-layer flow are then dis­
cussed in Section 9-6.