BOUNDARY LAYER

Ref: Frank M. White, Fluid Mechanics, McGraw-

Hill, Inc, 4th edition
Boundary Layer

Ludwig Prandtl introduced

concept of Boundary Layer
in 1904 for the first time.
• For high Reynolds Number flow, two length scales
1. Far from surface, viscous forces are unimportant
and inertial forces dominate (Bernoulli eq may be
applied).
2. Near the surface, viscous forces are comparable
to inertial forces (Navier-Stokes eq can be applied)
Examples boundary layers (layers
where there is velocity gradient) in
Chemical Engineering
1. Development of fully developed laminar flow
(a) and turbulent flow (b) in circular pipes

Momentum
boundary
layer of flow
in pipes

Higher flow induces higher du/dy close to the wall →

higher shear stress → higher pump power in case of
power of pump.
2. Development of boundary layer around
spherical particles when Re is increased

.

Boundary layer in front

part of the particle
Separation point of
boundary layer
7.1. Reynolds-Number and
Geometry Effects
 The technique of boundary-layer (BL) analysis can
be used to compute viscous effects near solid walls
and to “patch” (to join) these onto the outer inviscid
motion.
 In Fig. 7.1 a uniform stream U moves parallel to a
sharp flat plate of length L. If the Reynolds number
UL/ is low (Fig. 7.1a), the viscous region is very
broad and extends far ahead and to the sides of the
plate due to retardation of the oncoming stream
greatly by the plate.
Fig. 7.1 Comparison of flow past a sharp flat plate at low and high
Reynolds numbers: (a) laminar, low-Re flow; (b) high-Re flow
 At a high-Reynolds-number flow (Fig. 7.1b) the
viscous layers, either laminar or turbulent, are very
thin, thinner even than the drawing shows.
 We define the boundary layer thickness  as the
locus of points where the velocity u parallel to the
plate reaches 99 % of the external velocity U.
 As we shall see in Sec. 7.4, the accepted formulas
for flat-plate flow are
 where Rex = Ux/ is called the local Reynolds number
of the flow along the plate surface. The turbulent-flow
formula applies for Rex > approximately 106.

 The blanks indicate that the formula is not applicable.

In all cases these boundary layers are so thin that
their displacement effect on the outer inviscid layer is
negligible.
 Thus the pressure distribution along the plate can be
computed from inviscid theory as if the boundary layer
were not even there.
 This external pressure field acts as a forcing function
in the momentum equation along the surface.
 For slender bodies (streamlined bodies), such as
plates and airfoils parallel to the oncoming stream,
the assumption of negligible interaction between the
boundary layer and the outer pressure is an excellent
approximation because pressure along plate is
constant.
 For a blunt-body (bluff body) flow, however, there is a
pressure distribution over the surface body, so the
pressure must be taken into account.
streamlined body
 Figure 7.2 shows two sketches of flow past a two-
or three-dimensional blunt body.
 In the idealized sketch (7.2a), there is a thin film of
boundary layer about the body and a narrow sheet
of viscous wake in the rear.
 In a actual flow (Fig. 7.2b), the boundary layer is
thin on the front, or windward, side of the body,
where the pressure decreases along the surface
(favorable pressure gradient).
.

Fig. 7.2 Illustration of the strong interaction between

viscous and inviscid regions in the rear of blunt body
flow: (a) idealized and definitely false picture of
blunt-body flow (according to Bernoulli’s law); (b)
actual picture of blunt body flow.
 But in the rear the boundary layer encounters
increasing pressure (adverse pressure gradient)
and breaks off, or separates, into a broad, pulsating
wake.
 The mainstream is deflected by this wake, so that
the external flow is quite different from the
prediction from inviscid theory with the addition of a
thin boundary layer
7.2. von Karman’s Momentum-
Integral (theoretical work)
 A boundary layer of unknown thickness grows
along the sharp flat plate in Fig. 7.3.
 The no-slip wall condition retards the flow, making
it into a rounded profile u(y), which merges into the
external velocity U constant at a “thickness” y (x).
 By utilizing the control volume of Fig. 7.3, we found
(without making any assumptions about laminar
versus turbulent flow) that the drag force on the
plate = the momentum integral across the exit
plane due to the change of velocity from U to
u(x,y)
 Fig. 7.3 Growth of a boundary layer on a flat plate.
Drag force = rate of momentum change = V x mass rate =  ((U –
u) x  b (width) x u x y).
This represents how much the momentum is lost due to friction
 where b is the plate width into the paper and the
integration is carried out along a vertical plane at a
constant x (von Kármán, 1921).
 Imagine that if the flow remains at uniform velocity
U, but the surface of the plate is moved upwards 
to consider that momentum flux reduction =
momentum flux loss boundary layer actually does,
then momentum thickness is defined as the loss of
momentum (=drag force) per unit width b divided
by U2 up to x concerned.
 (von Kármán, 1921).

Covers area from

0 to x

 By comparing this with Eq. (7.4) Kármán arrived at

what is now called the momentum integral relation for
flat-plate boundary-layer flow (w  )
Vertical variable
horisontal variable
 It is valid for either laminar or turbulent flat-plate flow.
vertical- variable

horisontal variable Variables in

momentum
integral relation
(7.5)
 Eq. 7.9 is the desired thickness estimate. It is all
approximate, of course, part of Kármán’s
momentum-integral theory [7], but it is startlingly
accurate, being only 10 percent higher than the
known exact solution for laminar flat-plate flow,
which we gave as Eq. (7.1a).
 By combining Eqs. (7.9) and (7.7) we also obtain a
shear-stress estimate along the plate
 Again this estimate, in spite of the crudeness of the
profile assumption (7.6) is only 10 % > the known
exact laminar-plate-flow solution cf = 0.664/Rex1/2,
treated in Sec. 7.4.
 The dimensionless quantity cf, called the skin-friction
coefficient, is analogous to the friction factor f in
ducts.
 A boundary layer can be judged as “thin” if, say, the
ratio /x < about 0.1. /x = 0.1 = 5.0/Rex1/2 or Rex >
2500.
 For Rex < 2500 we can estimate that boundary-
layer theory fails because the thick layer has a
significant effect on the outer inviscid flow
(thickness creates pressure distribution across the
boundary layer).
 The upper limit on Rex for laminar flow is about 3 x
106, where measurements on a smooth flat plate
[8] show that the flow undergoes transition to a
turbulent boundary layer.
 From 3 x 106 upward the turbulent Reynolds
number may be arbitrarily large, and a practical
limit at present is 5 x 109 for oil supertankers.
 Displacement Thickness
 Another interesting effect of a boundary layer is
displacement of the outer streamlines.
 As shown in Fig. 7.4, outer streamlines must deflect
outward a distance *(x) to satisfy conservation of
mass between the inlet and outlet as a result of
fluid entrainment from fluid flow to boundary layer
 .

 The quantity * is called the displacement thickness of

the boundary layer.
 To relate it to u(y), cancel  and b from Eq. (7.11),
evaluate the left integral, and add and subtract U from
the right integrand:
 Bernoulli
eq. applies

Navier Stokes
eq. applies

Fluid
entrainment

Fig. 7.4 Displacement effect of a boundary

layer. Fluid entrainment occurs from free fluid
flow to the boundary layer, so mass rate at 0 =
mass rate at 1
Hypothetical thickness attributed to fluid
entrainment
 Mass loss/(.w)

δ
ρUδ* w   ρ U  u  dy w
0

Imagine that if the flow remains at

uniform velocity U, but the surface of the
plate (wall) is moved upwards
(displaced) * to consider mass flux
reduction = mass flux loss of the main
flow the BL generates (as if above the
plate, flow is uniform), then displacement
thickness is the loss of mass from main
flow per unit width divided by U2 .
 Introducing von Karman’s profile approximation (7.6)
into (7.12), we obtain by integration the approximate
result

 These estimates are only 6% away from the exact

solutions for laminar flat-plate flow given in Sec. 7.4:
* = 0.344  = 1.721x/Rex1/2.
 Since * << x for large Rex and the outer streamline
slope is proportional to *, we conclude that the
velocity normal for entrainment to the wall << the
velocity parallel to the wall. This is a key assumption
in boundary-layer theory (Sec. 7.3).
 EXAMPLE 7.2
 Are low-speed, small-scale air and water boundary
layers really thin? Consider flow at U = 1 ft/s past a
flat plate 1 ft long. Compute the boundary-layer
thickness at the trailing edge for (a) air and (b)
water at 20°C.
(ReL > 2500, then BL is thin)
7.4 The Flat-Plate Boundary
Layer: Blasius’s Laminar flow
(experimental work)
 The classic and most often used solution of
boundary-layer theory is for flat-plate flow, as in
Fig. 7.3, which can represent either laminar or
turbulent flow.
 For laminar flow past the plate, the boundary-
layer equations can be solved exactly for u and v,
assuming that the free-stream velocity U is
constant (dU/dx = 0).
 The solution was given by Prandtl’s student
Blasius, in 1908.
 With a coordinate transformation, Blasius showed
experimentally (exactly) that the dimensionless
velocity profile u/U is a function only of the single
composite dimensionless variable (y)[U/( x)]1/2:

 The boundary conditions are (see next slide)

Figure 1. The result of experiment where Vx/V∞ = function
of , which will be used to derive formula of , 0, dan 
(from deNevers) Vx/V∞ = u/U in Frank White’s book
 Some tabulated values of the velocity-profile
shape f()= u/U are given in Table 7.1.

 Since u/U → 1.0 only as y → ∞, it is customary to
select the boundary layer thickness at that point
where u/U = 0.99. From the table, this occurs at 
 5.0:

 With the profile known, Blasius could also compute

the wall shear and displacement thickness from
the fact that the slope at y = 0 is d  u / U 
 0.332
d
 At a constant x and y  0, it gives
 d   U   dy   U  d u / U 
0,5 0,5
 du 
  du     x   du  or   U 
 x    x  dy   x  d

 or w(x) = 0.332  U (U/(x))0.5

 Therefore,

 Notice how close these are to Karman’s integral

estimates, Eqs. (7.9), (7.10), and (7.13).

 In the equation of shear stress
 w(x) = 0.332  U (U/(x))0.5
 The wall shear drops off with x1/2 because of
boundary-layer growth and varies as velocity U to
the 1.5 power. This is in contrast to laminar pipe
flow, where w is proportional to U and is
independent of x.
 Covers area from
0 to x

 If w(x) is substituted into Eq. (7.4), we compute

the total drag force

 The drag increases only as x1/2 . The non

dimensional drag coefficient is defined as
 .
 Thus, for laminar plate flow, CD = 2 x the value of
the skin-friction coefficient at the trailing edge. This
is the drag on one side of the plate.
 Kármán pointed out that the drag could also be
computed from the momentum relation (7.2). In
dimensionless form, Eq. (7.2) becomes

 This can be rewritten in terms of the momentum

thickness at the trailing edge (at x = L)
 Computation of  from the Blasius’s profile u/U or
from CD gives 

 The ratio of displacement to momentum thicknesses,

called the dimensionless-profile shape factor, is also
useful in integral theories. For laminar flat-plate flow

 A large shape factor then implies that boundary-layer

separation is about to occur (low shear stress tends
to separate boundary layer).
 If we plot the Blasius velocity profile from Table 7.1
in the form of u/U vs y/, we can see why the
simple integral-theory guess from von Karman, Eq.
(7.6), was such a great success. This is done in
Fig. 7.5.

 The simple parabolic approximation of von

Karman’s estimate is not far from true Blasius’s
profile (based on experiments); hence its
momentum thickness is within 10 percent of the
true value.
 EXAMPLE 7.3
 A sharp flat plate with L =1 m and b = 3 m is
immersed parallel to a stream of velocity 2 m/s.
 Find the drag on one side of the plate, and at the
trailing edge find the thicknesses , *, and  for
(a) air,  =1.23 kg/m3 and  =1.46x10-5 m2/s, and
(b) water,  =1000 kg/m3 and  =1.02 x 10-6 m2/s.
 Part a (for air).
 Part b (for water).
 The drag is 215 x more for water in spite of the
higher Reynolds number and lower drag coefficient
because water is 57 x more viscous and 813 x
denser than air.
 From Eq. (7.26), in laminar flow, it should have
(57)1/2(813)1/2 = 7.53(28.5) = 215 x more drag.
The water layer is 3.8 times thinner than the air
layer, which reflects the square root of ratio of air to
water kinematic viscosity = 14.3 = 3.8. Or, the
higher the kinematic viscosity, the thicker is the
boundary layer,