Advanced heat transfer

SENOUCI Mohammed
2023-2024

Chapter II
BOUNDARY LAYER FLOW

Higher School of Electrical and Energetic

Engineering of Oran
 Boundary layer concept
Theoretical solutions to convective heat exchange problems are based on
boundary layer theory. The term boundary layer was introduced in 1904 by
the German Ludwig Prandtl.
Navier-Stokes equations and the energy equation are simplified using the
boundary layer concept.
The hydrodynamic (or velocity) boundary is the layer (or thickness), near
the wall (solid surface), in which the velocity varies from the value 0 on
the wall to the value that the velocity would have for an ideal fluid. The
thickness of the boundary layer is defined as the distance y at which the
velocity reaches 99% of the value it would have in an ideal fluid (the
velocity of the external fluid, uinf or u∞), (u/u∞)=0,99. In the boundary layer,
the velocity gradients are high and the viscosity forces involved are
significant. The equations governing boundary layer flows were introduced
in 1904 by the German physicist Ludwig Prandtl.
la couche limite est une zone de forts cisaillements induits par des
gradients de vitesse intenses. Son importance est déterminante car c’est là
que se font l’esentiel des transferts de quantité de mouvement et de chaleur.
 The thermal boundary layer will only develop if the temperature of the fluid
and that of the solid surface are different, The fluid particles in adjacent layer
to the plate get the same temperature, Ts (or TW), that of surface . The particles
exchange heat energy with particles in adjoining fluid layers and so on. As a
result, the temperature gradients are developed in the fluid layers and a
temperature profile is developed in the fluid flow, which ranges from Ts (or
TW) at the surface to fluid temperature Tinf (or T∞) sufficiently far from the
surface in y direction. The region where this gradient develops is called the
thermal boundary layer and its thickness δt is defined as the distance for which
the ratio (T −Tinf)/(Ts−Tinf)=0,99. As with the hydrodynamic boundary layer,
the thickness of the thermal boundary layer increases with distance x.

Fig. 1 : Hydrodynamic and thermal boundary layer over a flat plate

 Given the link between mass transfer and heat transfer, it is necessary to
take the flow regime into account. The transition between flow regimes was
demonstrated by the Irishman Osborne Reynolds (Fig. 2). An example of
boundary layer development on a flat plate is given in Figure 3,

Fig. 2

Fig. 3

Viscous sublayer
 To predict the flow regime it is first useful to determine the critical
distance for which the transition regime occurs. For this a dimensionless
number, called local Reynolds number, is defined:

uinf x u inf x
Re x = =
 

The flow is a laminar flow if: Rex  Recrit = 5  10 5

and a turbulent flow if: Re x  Re crit = 5  10 5

 Laminar boundary layer governing equations
The conservation of mass (continuity equation), momentum
(momentum equation) and energy (energy equation) are not easy to
solve analytically. Boundary-layer approximations, as shown in Figure
4, can be employed.

Fig. 4 : Boundary-layer approximations

 Prandtl stated that some of the terms in the Navier–Stokes equations can be
neglected for the boundary-layer flows. As a result of this assumption, well-
known boundary-layer equations arose. By introducing the consequences of
the Prandtl hypothesis [1]: velocity in the x-direction is greater than that in
the y-direction, streamwise velocity change in the y-direction is greater than
that in the x-direction; temperature change in the y-direction is greater than
that in the x-direction:
v u v u
v  u d' ou  et  (1)
x x y y
   
V V  2V  2V
 d' ou  (2)
x y x 2
y 2
We find, for two-dimensional, steady, incompressible flow and Neglecting
the gravity term, the following equations:
u u 1 p  2u
u +v = − . + 2 (3)
x y  x y
v v 1 p  2v
u +v = − . + 2
x 
y



y
 
y
 (4)
a b c d
 By comparing the two equations term by term, we observe that according to
the hypotheses (1) and (2) the terms (a), (b) and (d) of (4) are very small
compared to the corresponding terms of (3), It must therefore be the same for
the last term (c).
p p (5)

y x
We are led to write:
p
=0 (6)
y
The continuity and momentum equations therefore take the following final
form: u v (7)
+ =0
x y
u u 1 p  2u
u +v =− + 2 (8)
x y  x y
p
=0 (9)
y
These last three equations are called boundary layer equations (or Prandtl
equations).
  2T  2T
Analogous reasoning in the thermal boundary layer ( 2 
2
) provides a
x y
simplified equation for temperature:
T T   2T 
u +v = a  (10)
x y  y 2 
 
a is the thermal diffusivity
 Laminar flat-plate boundary layer: exact solution
The solution for the laminar boundary layer on a horizontal flat plate was
obtained by Prandtl’s student H. Blasius [2] in 1908.
It is essential to specify the conditions of validity of the calculation that
follows, i.e. the experimental conditions under which we are working. We
therefore consider the following problem:
• isothermal flat plate at an imposed temperature TP (TS or TW)
• external flow parallel to the plate (in the x direction), with uniform
velocity U∞ = Cte. The V∞ component is therefore zero (V∞ = 0) (Fig. 5).
• external flow at uniform temperature T∞.
p
It follows from equation (8) that in this case we have :  = 0
x
L’équation de conservation de la quantité de mouvement selon 𝑥 reste
valable même hors de la couche limite et donc sur la frontière où 𝑢 = 𝑈∞. À
cet endroit, le champ de vitesse ne varie plus avec la distance 𝑦 et il n’y a
pas de gradient important dans la direction transversale. On obtient donc :
dU  1 dP
U =−
dx  dx
 Fig. 5 : Flow over a flat plate: problem data
p
according to (eq. 9) we have in the boundary layer = 0 therefore p = p(x).
P y P P p
It followsthat is independent of y. We have therefore =  and so = 0
x x x x
The boundary layer equations are finally written as follows :
u v
+ =0 (11)
x y
u u  2u
u + v = 2 (12)
x y y

T T   2T 
u +v = a  (13)
x y  2 
 y 
 With boundary conditions :
• at y = 0, u = 0, v = 0
• at y = ∞, u = U∞ =U, u = 0 (14)
y
Equations (11) and (12), with boundary conditions equation (14) are a set of
nonlinear, coupled partial differential equations for the unknown velocity
field u and v. To solve them, Blasius reasoned that the velocity profile, u/U∞,
should be similar for all values of x when plotted versus a nondimensional
distance from the wall. Use similarity profile assumption to turn 2 PDE into 1
ODF. The boundary layer thickness, was a natural choice for
nondimensionalizing the distance from the wall. Thus the solution is of the
form:
u y
= g ( ) = g ( ) (15)
where : U 

y y U
= = =y
 x x (16)
U
η is called the similarity variable
We now introduce the stream function, ψ, where
 
u= and v = − (17)
y x
satisfies the continuity equation (eq. 11) identically. Substituting for u and v
into (eq. 12) reduces the equation to one in which ψ is the single dependent
variable. Defining a dimensionless stream function as:

f ( ) = (18)
xU 
With ψ defined by (eq. 18) and η defined by (eq. 16), we can evaluate each of
the terms in (eq. 12). The velocity components are given by:

∞ (19)
∞ ∞

and
∞ ∞
∞ ∞ (20)
Or:
∞ (21)

By differentiating the velocity components, it also can be shown that:

∞
(22)

∞ ∞ (23)
and:
∞
(24)

Substituting these expressions into (eq. 12), we obtain:

∞ (25)

with boundary conditions:

(26)
 The second-order partial differential equations governing the growth of
the laminar boundary layer on a flat plate (eqs. 11 and 12) have been
transformed to a nonlinear, third-order ordinary differential equation (eq.
25) with boundary conditions given by (eq. 26).
It is not possible to solve (eq. 25) in closed form; Blasius solved it using a
power series expansion about with the above mentioned BCs (eq. 26). The
same equation later was solved more precisely -again using numerical
methods- by L, Howarth (1938), who reported results to 5 decimal places.
The numerical values of f, df/dη, and d2f/dη2 in Table 1.
 ∞
∞

Table 1: The Function f(η) for the Laminar Boundary Layer along a
Flat Plate at Zero Incidence.
Fig. 6 : Velocity profile in laminar boundary layer
 The energy equation is solved in a similar maner by first defining a
dimensionless temperature variable as :
T (  ) − TW
 ( ) = (27)
T − TW
Where it is also assumed that θ and T may be expressed as function of the
similarity variable η. Equation (13) becomes:

(28)

substituting the u (eq. 19) and v (Eq. 21) expressions into the energy
equation (eq. 28) gives:

(29)
With the boundary conditions:

(30)
The solution is given by:

(31)

This solution was given by Pohlhausen (1921).

Fig. 7: Pohlhausen's solution for temperature distribution

for laminar flow over a semi-infinte isothermal flat plate
 One important consequence of this solution is that, for 0,6 < Pr < 15, it
was found that:
 d (  ) 
  = 0.332 Pr 1 / 3
 d  =0 (32)

The slope of the temperature profile at the surface, y = 0, is well presented

by:
 dT  U  d (  )  U
  = ( T − Tw )    = ( T − Tw )  ( 0.332 Pr 1 / 3 ) (33)
 dy  y =0 x  d  =0 x

Local convection heat transfer coefficient is given by:

( −k ( T / y )w U
hx = = 0.332k Pr 1 / 3 (34)
Tw − T ) x
Nusselt number is then:
hx x
Nu = = 0.332 Re 1 / 2 Pr 1 / 3
k (35)
The average value of h and Nu , are: hL
Nu =
h = 2 hx x = L k = 0.664 Re 1/ 2
Pr 1/ 3
The average value of h and Nu , are:

h = 2hx x= L (36)

hL
Nu = = 0.664 Re 1 / 2 Pr 1 / 3 (37)
k
 Characterization of the boundary layer
Three different parameters are typically used to quantify how thick a
boundary layer is at any given position.
• The Boundary Layer Thickness:
The distance from the wall for which the flow velocity has essentially
reached the 'asymptotic' velocity U∞.

u y = = 0.99U  (38)

• Displacement thickness δ1(or δ*)

An imaginary displacement of fluid from the surface to account for "lost"
mass flow in boundary layer.
(39)

• Momentum thickness δ2 (or θ):
The equivalent thickness of a fluid layer (velocity u) with momentum equal
to the momentum lost from friction.

(40)
 
 Other quantities related to these boundary layer thickness can be defined.
• Wall shear stress (La contrainte de cisaillement à la paroi)

(41)

• Wall shear stress coefficient (Coecient de frottement)

(42)


• Totale friction force (or drag force), per unit width, for one side of the
plate

(43)
 For the Blasius similarity solution for a two-dimensional boundary layer, we
can compute the the quantities defined above:
• The Boundary Layer Thickness:

(44)
• Displacement thickness:
(45)

• Momentum thickness:
(46)

• Thermal boundary layer thickness

(47)
• Wall shear stress:

(48)

• Local wall shear stress coefficient:

(49)
• Average wall shear stress coefficient:

(50)

• Totale friction force

(51)
 The von Karman and Pohlhausen Approximate Solution
Let us return to the boundary-layer equations (The Prandtl equations).
These equations, for a steady two-dimensional incompressible flow, are:
u v
+ =0 (52)
x y
u u  2u (53)
u + v = 2
x y y
With the boundary conditions are:
y = 0; u = v = 0
(54)
y =  ;u = U

Due to equations (46) and (48) equation (47) can be reduced to the form:
d 2 dU   w
U 2 + ( 2 2 +  1 )U  = (55)
dx dx 
This equation is known as the "Karman momentum integral equation" for
two-dimensional steady incompressible boundary layer [1].
 One of the earliest and, until recently, most widely used approximate
methods for the solution of the Karman equation is that developed by
Pohlhausen (1921). This method consists of giving a polynomial
approximation of the nondimensional velocity profile (u/U∞) in the
boundary layer as a function of the nondimensional coordinate normal to
the wall (η=δ/y), the constants (ai) and the degree of the polynomial (n) are
determined from the boundary conditions of the problem studied
(geometric shape of the obstacle, potential flow, etc.).
u n
= f ( ) =  ai i ; 0    1 (56)
U i =0

Pohlhausen, suggested that the velocity profile is a polynomial of forth

order: u
= 2 − 2 3 +  4 (57)
U

solving this equation with the appropriate boundary conditions, gives:

 5.84 x
= (58)
Re x
 1 = 0.3 (59)
 2 = 0.117 (60)
2 U 
w = (61)

Best results are obtained with a 3rd order polynomial:
u 2 1
=  − 3 (62)
U 3 2
We found:
4.64 x
= (63)
Re x
 1 = 0.373 (64)
 2 = 0.143 (65)
3U 
w = (66)
2
Table 2: Comparison of the Blasius and von Karman Approximations
Turbulent boundary layer
‫شكرا‬