Chapter 1 - Viscous Incompressible Flow (Rev200) (ANSWER)

CHAPTER ONE [BTME2223 APPLIED FLUID MECHANICS]
Chapter One: Viscous Incompressible Flow
Motivation
The flow of real fluid differs from that of an ideal fluid due to its viscosity. The ideal fluid experiences
no resistance to its motion and hence, uniform velocity distribution is obtained. In case of real fluid,
the velocity remains zero at the stationary wall and increases asymptotically to free stream velocity
away from the wall. The viscosity of fluid, thus, resists the motion of fluid and modifies the
configuration over and around the boundary. The concept of boundary layer is most significant
contribution to the development of fluid mechanics. The theory has been successful in giving the
physical picture of the role of viscosity in the fluid flow problems and also in predicting the drag
force exerted on various shaped bodies, surface vehicles, aerofoil and etcetera.
Chapter Objectives:
 To develop the Navier-Stokes equations, which apply to a differential element of viscous
incompressible fluid.
 To obtain analytical solutions of the equations of motion for simple flow fields.
 To predict boundary layer thickness and other boundary layer properties.
 To show how to determine the shear stress or shear drag created by both laminar and turbulent
boundary layers that form on a flat surface.
 To discuss the effect of flow separation and methods for reducing drag.
Chapter Outline:
1.1 Introduction
1.2 Navier-Stokes Equations
1.3 Boundary Layer Flow over a Flat Plate and the Equations
1.4 Blasius Flow over a Flat Plate for Boundary Layer
1.5 Momentum Integral Equation
1.6 Turbulent Boundary Layer
1.7 Boundary Layer Separation & Wake Formation
1.1 Introduction
In the previous level of fluid mechanics studies (BTME2213 Fluid Mechanics) we considered most
of the applications of the equations of motion to ideal fluids, where only the forces of gravity and
pressure influence the flow.
Incompressible flow is the flow with constant density. Viscous flow means that flow of viscous fluid
with viscosity. All real fluids have viscosity. Viscosity is defined as the property of a fluid which
offers resistance to the movement of one layer fluid over another adjacent layer of the fluid. As
consequences the fluid velocity which is on the surface is equal to zero.
When you have an inviscid flow, then the Navier-Stokes equations reduces to the Euler Equation.
The Euler equation is essentially Newton’s second law applied on a flowing infinitesimal volume
element and it addresses conservation of mass, momentum, and energy absent the effect of viscosity.
The Navier-Stokes equations describes the motion of fluid substance based on Newton’s second law
together with the assumptions that stress in the fluid is the sum of a diffusing viscous term (shear
stress) and a pressure term.
May2014[Rev2:May2017]-pkq\dmee\afm\c1 1-1
For NON-COMMERCIAL use only.
1.2 Navier-Stokes Equations

Generalized equations of motion of a real flow are named after the
inventors of them (Claude-Louis Navier, a French engineer, and
George Gabriel Stokes, an English mathematician) and they are
known as Navier-Stokes equations. These are the exact equations
that govern the motion of real fluids and generally valid for
compressible or incompressible laminar flows. When a motion
becomes turbulent, these equations are generally not able to provide Claude-Louis George Gabriel
Navier Stokes
with a complete solution.
However, they are derived from the Newton’s second law which states that the product of mass and
acceleration is equal to sum of the external forces acting on the body. The external forces can be
classified into two kinds – body force (due to gravitational or electromagnetic force) and surface
forces (pressure and skin friction or known as friction force). Let the body force per unit mass be
⃗⃗⃗⃗
𝐹𝐵 = 𝑓𝑥 𝑖 + 𝑓𝑦 𝑗 + 𝑓𝑍 𝑘 (1.1)
Consider a differential fluid element in the flow field as shown in Figure 1.1a. We wish to evaluate
the surface forces acting on the boundary of this rectangular parallelepiped. The net force on the body
due to the surface forces in x-, y- and z-directions are depicted in Figure 1.1b, 1.1c and 1.1d
respectively.
Figure 1.1a An infinitesimal fluid element Figure 1.1b Net force in x-direction
Figure 1.1c Net force in y-direction
Figure 1.1d Net force in z-direction
Net surface force in 𝑥-direction:

𝜕𝜎𝑥 𝜕𝜏𝑦𝑥 𝜕𝜏𝑧𝑥
d𝐹𝑥 = d𝑥 ∙ d𝑦d𝑧 + d𝑦 ∙ d𝑥d𝑧 + d𝑧 ∙ d𝑥d𝑦
𝜕𝑥 𝜕𝑦 𝜕𝑧
𝜕𝜎𝑥 𝜕𝜏𝑦𝑥 𝜕𝜏𝑧𝑥

d𝐹𝑦 = ( + + ) d𝑥d𝑦d𝑧 (1.2a)
𝜕𝑥 𝜕𝑦 𝜕𝑧
Net surface force in 𝑦-direction:

𝜕𝜏𝑥𝑦 𝜕𝜎𝑦 𝜕𝜏𝑧𝑦
d𝐹𝑦 = d𝑥 ∙ d𝑦d𝑧 + d𝑦 ∙ d𝑥d𝑧 + d𝑧 ∙ d𝑥d𝑦
𝜕𝑥 𝜕𝑦 𝜕𝑧
𝜕𝜏𝑥𝑦 𝜕𝜎𝑦 𝜕𝜏𝑧𝑦

d𝐹𝑦 = ( + + ) d𝑥d𝑦d𝑧 (1.2b)
𝜕𝑥 𝜕𝑦 𝜕𝑧
Net surface force in 𝑧-direction:

𝜕𝜏𝑥𝑧 𝜕𝜏𝑦𝑧 𝜕𝜎𝑧
d𝐹𝑧 = d𝑥 ∙ d𝑦d𝑧 + d𝑦 ∙ d𝑥d𝑧 + d𝑧 ∙ d𝑥d𝑦
𝜕𝑥 𝜕𝑦 𝜕𝑧
𝜕𝜏𝑥𝑧 𝜕𝜏𝑦𝑧 𝜕𝜎𝑧

d𝐹𝑧 = ( + + ) d𝑥d𝑦d𝑧 (1.2c)
𝜕𝑥 𝜕𝑦 𝜕𝑧
Normal stresses and shearing stresses are denoted by 𝜎 and 𝜏 respectively. The stress system is
having nine scalar quantities. These nine quantities form a stress tensor. The set of nine components
of stress tensor can be described as
𝜎𝑥 𝜏𝑦𝑥 𝜏𝑧𝑥
π = [𝜏𝑥𝑦 𝜎𝑦 𝜏𝑧𝑦 ] (1.3)
𝜏𝑥𝑧 𝜏𝑦𝑧 𝜎𝑧
The above stress tensor is symmetric, which means that two shearing stresses with subscripts which
differ only in their sequence are equal. Hence,
𝜏𝑥𝑦 = 𝜏𝑦𝑥 ; 𝜏𝑥𝑧 = 𝜏𝑧𝑥 ; 𝜏𝑦𝑧 = 𝜏𝑧𝑦
Total force in 𝑥-direction:

D𝑢
𝜌d𝑥d𝑦d𝑧 = d𝐹𝑥 + 𝜌d𝑥d𝑦d𝑧(𝑓𝑥 )
D𝑡
D𝑢 𝜕𝜎𝑥 𝜕𝜏𝑦𝑥 𝜕𝜏𝑧𝑥
𝜌 = 𝜌𝑓𝑥 + + + (1.4)
D𝑡 𝜕𝑥 𝜕𝑦 𝜕𝑧
Total force in 𝑦-direction:

D𝑣
𝜌d𝑥d𝑦d𝑧 = d𝐹𝑦 + 𝜌d𝑥d𝑦d𝑧(𝑓𝑦 )
D𝑡
D𝑣 𝜕𝜏𝑥𝑦 𝜕𝜎𝑦 𝜕𝜏𝑧𝑦
𝜌 = 𝜌𝑓𝑦 + + + (1.5)
Total force in 𝑧-direction:

D𝑤
𝜌d𝑥d𝑦d𝑧 = d𝐹𝑧 + 𝜌d𝑥d𝑦d𝑧(𝑓𝑧 )
D𝑡
D𝑤 𝜕𝜏𝑥𝑧 𝜕𝜏𝑦𝑧 𝜕𝜎𝑧
𝜌 = 𝜌𝑓𝑧 + + + (1.6)
The well-known Newton’s viscosity law is

𝜕𝑉
𝜏=𝜇 (1.7)
𝜕𝑛
where 𝑛 is the coordinate direction normal to the solid-fluid interface, 𝜇 is the coefficient of viscosity
and 𝑉 is velocity. This law is valid for parallel flows. There are more generalized relations which can
relate stress field and velocity field for any kind of flow. Such relations are called constitutive
equations. We shall consider here the Stokes’ viscosity law (the derivation is far beyond the syllabus).
According to Stokes’ law of viscosity, shear stress is proportional to rate of shear strain so that
𝜕𝑣 𝜕𝑢
𝜏𝑥𝑦 = 𝜏𝑦𝑥 = 𝜇 ( + ) (1.8a)
𝜕𝑥 𝜕𝑦
𝜕𝑤 𝜕𝑢
𝜏𝑥𝑧 = 𝜏𝑧𝑥 = 𝜇 ( + ) (1.8b)
𝜕𝑥 𝜕𝑧
𝜕𝑣 𝜕𝑤
𝜏𝑦𝑧 = 𝜏𝑧𝑦 = 𝜇 ( + ) (1.8c)
𝜕𝑧 𝜕𝑦
Also, the expressions of Stokes’ law of viscosity for normal stresses are
𝜕𝑢 2 𝜕𝑢 𝜕𝑣 𝜕𝑤
𝜎𝑥 = −𝑃 + 2𝜇 − 𝜇( + + ) (1.9a)
𝜕𝑥 3 𝜕𝑥 𝜕𝑦 𝜕𝑧
𝜕𝑣 2
𝜎𝑦 = −𝑃 + 2𝜇 ⃗)
− 𝜇(∇ ∙ 𝑉 (1.9b)
𝜕𝑦 3
𝜕𝑤 2
𝜎𝑧 = −𝑃 + 2𝜇 ⃗)
− 𝜇(∇ ∙ 𝑉 (1.9c)
𝜕𝑧 3
D𝑢 D𝑣 D𝑤
In order to express , and in terms of field derivatives, Eqs (1.8) and (1.9) are introduced into
D𝑡 D𝑡 D𝑡
Eqs (1.4), (1.5) and (1.6) and we obtain
D𝑢 𝜕 𝜕𝑢 2 𝜕 𝜕𝑣 𝜕𝑢 𝜕 𝜕𝑤 𝜕𝑢
𝜌 = 𝜌𝑓𝑥 + [−𝑃 + 2𝜇 ⃗ )] +
− 𝜇(∇ ∙ 𝑉 [𝜇 ( + )] + [𝜇 ( + )] (1.10a)
D𝑡 𝜕𝑥 𝜕𝑥 3 𝜕𝑦 𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑥 𝜕𝑧
D𝑣 𝜕 𝜕𝑣 𝜕𝑢 𝜕 𝜕𝑣 2 𝜕 𝜕𝑣 𝜕𝑤
𝜌 = 𝜌𝑓𝑦 + [𝜇 ( + )] + [−𝑃 + 2𝜇 ⃗ )] + [𝜇 ( +
− 𝜇(∇ ∙ 𝑉 )] (1.10b)
D𝑡 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑦 3 𝜕𝑧 𝜕𝑧 𝜕𝑦
D𝑤 𝜕 𝜕𝑤 𝜕𝑢 𝜕 𝜕𝑣 𝜕𝑤 𝜕 𝜕𝑤 2
𝜌 = 𝜌𝑓𝑧 + [𝜇 ( + )] + [𝜇 ( + )] + [−𝑃 + 2𝜇 ⃗ )]
− 𝜇(∇ ∙ 𝑉 (1.10c)
D𝑡 𝜕𝑥 𝜕𝑥 𝜕𝑧 𝜕𝑦 𝜕𝑧 𝜕𝑦 𝜕𝑧 𝜕𝑧 3
These differential equations are known as Navier-Stokes equations. However, the general form of
continuity equation is simplified in case of incompressible flow where 𝜌 = constant. Therefore,
𝜕𝑢 𝜕𝑣 𝜕𝑤
+ + =0 OR ∇∙𝑉 ⃗ =0 (1.11)
𝜕𝑥 𝜕𝑦 𝜕𝑧
Invoking the incompressible continuity equation into the Navier-Stokes equations

D𝑢 𝜕𝑃 𝜕 2𝑢 𝜕 2𝑢 𝜕 2𝑣 𝜕 2𝑤 𝜕 2𝑢
𝜌 = 𝜌𝑓𝑥 − + 2𝜇 2 + 𝜇 [ 2 + + + ]
D𝑡 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑥𝜕𝑦 𝜕𝑥𝜕𝑧 𝜕𝑧 2
D𝑢 𝜕𝑃 𝜕 2𝑢 𝜕 2𝑢 𝜕 2𝑢 𝜕 𝜕𝑢 𝜕𝑣 𝜕𝑤
𝜌 = 𝜌𝑓𝑥 − + 𝜇 [ 2 + 2 + 2] + 𝜇 [ + + ]
D𝑡 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑧
D𝑢 𝜕𝑃
𝜌 = 𝜌𝑓𝑥 − + 𝜇∇2 𝑢 (1.12)
D𝑡 𝜕𝑥
where D𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑥 𝜕𝑢 𝜕𝑦 𝜕𝑢 𝜕𝑧
= + + +
D𝑡 𝜕𝑡 𝜕𝑥 𝜕𝑡 𝜕𝑦 𝜕𝑡 𝜕𝑧 𝜕𝑡
D𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢
= +𝑢 +𝑣 +𝑤 = ⃗ ∙ ∇𝑢
+𝑉
D𝑡 𝜕𝑡 𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑡
For 𝑦-direction:
D𝑣 𝜕𝑃 𝜕 2𝑣 𝜕 2𝑣 𝜕 2𝑣 𝜕 𝜕𝑢 𝜕𝑣 𝜕𝑤
𝜌 = 𝜌𝐹𝑦 − + 𝜇 [ 2 + 2 + 2] + 𝜇 [ + + ]
D𝑡 𝜕𝑦 𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑦 𝜕𝑥 𝜕𝑦 𝜕𝑧
D𝑣 𝜕𝑃
𝜌
= 𝜌𝑓𝑦 − + 𝜇∇2 𝑣 (1.13)
D𝑡 𝜕𝑦
where D𝑣 𝜕𝑣 𝜕𝑣 𝜕𝑣 𝜕𝑣
= +𝑢 +𝑣 +𝑤
D𝑡 𝜕𝑡 𝜕𝑥 𝜕𝑦 𝜕𝑧
and 𝑧-direction:
D𝑤 𝜕𝑃
𝜌 = 𝜌𝑓𝑧 − + 𝜇∇2 𝑤 (1.14)
D𝑡 𝜕𝑧
where D𝑤 𝜕𝑤 𝜕𝑤 𝜕𝑤 𝜕𝑤
= +𝑢 +𝑣 +𝑤
D𝑡 𝜕𝑡 𝜕𝑥 𝜕𝑦 𝜕𝑧
In short, vector notation maybe used to write Navier-Stokes and continuity equations for
incompressible flow as
⃗
D𝑉
𝜌 ⃗⃗⃗⃗𝐵 − ∇𝑃 + 𝜇∇2 𝑉
= 𝜌𝐹 ⃗ (1.15)
D𝑡
∇∙𝑉 ⃗ =0
where 𝑉 ⃗ = 𝑢𝑖 + 𝑣𝑗 + 𝑤𝑘
We observe that we have four unknown quantities, 𝑢, 𝑣, 𝑤 and 𝑃, and four equations: equation of
motion in three directions (three equations) and the continuity equation. In principle, these equations
are solvable but to date generalized solution is not available due to the complex nature of the set of
these equations. The highest order terms, which come from the viscos forces, are linear and of second
order (the Laplacian term). The first order convective terms are non-linear and hence, the set is termed
as quasi-linear.
Boundary Condition: Since boundary conditions are so critical to

a proper solution, we discuss the types of boundary conditions that are
commonly encountered in fluid flow analyses. The most-used boundary
condition is the no-slip condition, which states that for a fluid in contact
with a solid wall, the velocity of the fluid must equal that of the wall,
⃗ fluid = 𝑉
𝑉 ⃗ wall . In other words, as its name implies, there is no “slip”
between the fluid and the wall. Fluid particles adjacent to the wall
adhere to the surface of the wall and move at the same velocity as the
wall. For a stationary wall with 𝑉 ⃗ wall = 0 the fluid adjacent to a
stationary wall has zero velocity.
When two fluids (fluid A and fluid B) meet at an interface, the interface
⃗A = 𝑉
boundary conditions are 𝑉 ⃗ B and 𝜏𝐴 = 𝜏𝐵 where, in addition to the
condition that the velocities of the two fluids must be equal, the shear
stress 𝜏 acting on a fluid particle adjacent to the interface in the
direction parallel to the interface must also match between the two
fluids.
A degenerate form of the interface boundary condition occurs at the free

surface of a liquid, meaning that fluid A is a liquid and fluid B is a gas
(usually air). The interface is flat, surface tension effects are negligible,
but the water is moving horizontally (like water flowing in a calm river).
In this case, the air and water velocities must match at the surface and
the shear stress acting on a water particle on the surface of the water
must equal that acting on an air particle just above the surface. According to the previous case,
boundary conditions at water–air interface:
∂𝑢 ∂𝑢
𝑢water = 𝑢air and 𝜏water = 𝜇water ) = 𝜏air = 𝜇air )
𝜕y water 𝜕y air
A quick glance at the fluid property tables reveals that 𝜇water is over 50 times greater than 𝜇air . In
order for the shear stresses to be equal the velocity gradient of air must be 50 times greater than
velocity gradient for water. Thus, it is reasonable to approximate the shear stress acting at the surface
of the water as negligibly small compared to shear stresses elsewhere in the water. Another way to
say this is that the moving water drags air along with it with little resistance from the air; in contrast,
the air doesn’t slow down the water by any significant amount. In summary, for the case of a liquid
in contact with a gas, and with negligible surface tension effects, the free surface boundary condition
is 𝜏 = 0.
Example 1.1
Water of 60℃ with viscosity of 4.7 × 10−4 Ns/m2 flows steadily between two large flat plates. The
lower plate moves to the left at a speed of 0.3 m/s while the upper plate is stationary. The plate spacing
is 3 mm and the flow is laminar. Determine the pressure gradient required to produce zero net flow
at the cross section.
The plate moves in x-direction and hence, Eq. (1.12) is applied.
D𝑢 𝜕𝑃
𝜌 = 𝜌𝐹𝑥 − + 𝜇∇2 𝑢
D𝑡 𝜕𝑥
𝜕𝑢 𝜕𝑢 𝜕𝑣 𝜕𝑤 𝜕𝑃 𝜕2𝑢 𝜕2𝑢 𝜕2𝑢
+𝑢 +𝑣 +𝑤 = 𝜌𝐹𝑥 − + 𝜇 [ 2 + 2 + 2]
𝜕𝑡 𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑧
The equation can be simplified further based on the flow

direction, no gravitational force in x-direction and type of flow,
hence,
𝜕𝑃 𝜕2𝑢
=𝜇 2
𝜕𝑥 𝜕𝑦
Integrate the function,
1 𝜕𝑃 2
𝑢= 𝑦 + 𝐶1 𝑦 + 𝐶2
2𝜇 𝜕𝑥
The boundary conditions:
At 𝑦 = 0, 𝑢 = −𝑈: 𝐶2 = −𝑈
𝑈 1 𝜕𝑃
At 𝑦 = 𝑏, 𝑢 = 0: 𝐶1 = − 𝑏
𝑏 2𝜇 𝜕𝑥
Therefore,
1 𝜕𝑃 2 𝑦
𝑢= (𝑦 − 𝑏𝑦) + 𝑈 ( − 1)
2𝜇 𝜕𝑥 𝑏
Now,
𝑏 𝑏
1 𝜕𝑃 2 𝑦
𝑄 = ∫ 𝑢 d𝑦 = ∫ [ (𝑦 − 𝑏𝑦) + 𝑈 ( − 1)] d𝑦
0 0 2𝜇 𝜕𝑥 𝑏
1 𝜕𝑃 3 𝑈𝑏
𝑄=− 𝑏 −
12𝜇 𝜕𝑥 2
For zero net flow, 𝑄 = 0.
1 𝜕𝑃 3 𝑈𝑏
𝑏 =−
12𝜇 𝜕𝑥 2
𝜕𝑃 6𝑈
=− 2
𝜕𝑥 𝑏
𝜕𝑃 6 × 0.3 × 4.7 × 10−4
=−
𝜕𝑥 0.0032
𝜕𝑃
= −94.0 Pa/m
𝜕𝑥
Example 1.2
When the supply valve 𝐴 is slightly opened, a very
viscous Newtonian liquid in the rectangular tank
overflows as shown in the figure. Determine the
velocity profile of the liquid as it slowly spills over
the sides if the flow is steady and fully developed.
The plate moves in x-direction and hence, Eq. (1.12) is applied. The Navier-Stokes equation is now become
D𝑢 𝜕𝑃 𝜕2𝑢 𝜕2𝑢 𝜌𝑔
𝜌 = 𝜌𝐹𝑥 − + 𝜇∇2 𝑢 0 = 𝜌𝑔 + 𝜇 2 → =−
D𝑡 𝜕𝑥 𝜕𝑦 𝜕𝑦 2 𝜇
𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑃 𝜕2𝑢 𝜕2𝑢 𝜕2𝑢 Integrate the function,
+𝑢 +𝑣 +𝑤 = 𝜌𝐹𝑥 − + 𝜇 [ 2 + 2 + 2]
𝜕𝑡 𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑢 𝜌𝑔
=− 𝑦 + 𝐶1
𝜕𝑦 𝜇
So,
𝑣=𝑤=0 𝜌𝑔 2
𝑢=− 𝑦 + 𝐶1 𝑦 + 𝐶2
2𝜇
For a general velocity equation in 𝑥-direction is given by
𝑢 = 𝑢(𝑥, 𝑦, 𝑧, 𝑡) The boundary conditions:
For a steady flow, At 𝑦 = 0, 𝑢 = 0: 𝐶2 = 0
𝜕𝑢
=0 𝜕𝑢
𝜕𝑡 At 𝑦 = 𝑎, = 0:
From the diagram, only 𝑥-𝑦 axis is involved. Thus, the velocity 𝜕𝑦
𝜌𝑔
equation can further simplify to 0=− 𝑎 + 𝐶1
𝑢 = 𝑢(𝑥, 𝑦) 𝜇
From the continuity equation, 𝜌𝑔
d𝑢 d𝑣 d𝑤 d𝑢 𝐶1 = 𝑎
+ + =0 → =0 𝜇
d𝑥 d𝑦 d𝑧 d𝑥 Therefore,
Therefore, we can make a conclusion to say that the velocity is a 𝜌𝑔 2 𝜌𝑔
function of 𝑦 only which is 𝑢=− 𝑦 + 𝑎𝑦
2𝜇 𝜇
𝑢 = 𝑢(𝑦)
𝜌𝑔
From the diagram give, the surface of the liquid is exposed to 𝑢= (2𝑎𝑦 − 𝑦 2 )
2𝜇
atmosphere and thus,
𝜕𝑃
=0
𝜕𝑥
The gravitational force in x-direction is parallel and follow the
gravity direction, thus,
𝐹𝑥 = 𝑔
1.3 Boundary Layer Flow over a Flat Plate and The Equations
In 1904, Ludwig Prandtl, a well-known Germen
scientist, introduced the concept of boundary layer and
derived the equations for the boundary layer flow by
correction reduction of Navier-Stokes equations. He
hypothesized that for fluid having relatively small 𝑈∞
viscosity, the effect of internal friction in the fluid is
significant only in a narrow region surrounding solid
boundaries or bodies over which the fluid flows. Thus,
close to the body is the boundary layer where shear Figure 1.2 Prandtl’s boundary layer concept
splits the flow into an outer flow region and a thin
stresses exert an increasingly larger effect on the fluid
boundary layer region (not to scale).
as one moves from free stream towards the solid
boundary. However, outside the boundary layer where the effect of the shear stresses on the flow
is small compared to values inside the boundary layer (since the velocity gradient 𝜕𝑢⁄𝜕𝑦 is
negligible), the fluid particles experience no vorticity (irrotational flow, ∇ × 𝑉 ⃗ = 0), and therefore,
the flow is similar to a potential flow. Hence, the surface at the boundary layer interface is a rather
fictitious (imaginary) one dividing rotational and irrotational flow.
The effect of the viscous flow over the solid

𝜕𝑢
surface, there is friction in the flow very near 𝑈∞ =0
𝜕𝑦
to the surface which signifies that the fluid is
retarded until it adheres to the surface. The
transition of the mainstream velocity from zero
𝑈∞
at the surface to full magnitude takes place
across the boundary layer. The thickness (𝛿)
of the boundary layer which is a function of the
coordinate direction x and is considered to be
very small compared to the characteristic
length 𝐿 of the domain. As we move down the Figure 1.3 Growth of a boundary layer on a flat plate.
plate to larger and larger values of x, Reynolds
number increases linearly with x. At some point, infinitesimal disturbance in the flow begin to grow
and the boundary layer cannot remain laminar. It begins a transition process toward turbulent flow.
For a smooth plate with a uniform free stream, the transition process begins at a critical Reynolds
number, 𝑅𝑒𝑥,critical ≈ 2 × 105 , and continues until the boundary layer is fully turbulent at transition
Reynolds number, 𝑅𝑒𝑥,critical ≈ 3 × 106 . The transition process is quite complicated, and details are
beyond the scope.
In the normal direction, within the thin layer, the gradient 𝜕𝑢⁄𝜕𝑦 is very large compared to the
gradient in the flow direction 𝜕𝑢⁄𝜕𝑥. Next step is to simplify the Navier-Stokes equations for steady
two-dimensional laminar incompressible flows and gravity effect is neglected. A schematic of the
general boundary-layer problem is given in Figure 1.3. The x-coordinate is in the main flow direction
along the body surface; the y-coordinate extends upward normal to the surface; the z-coordinate
points out of the page. Considering the Navier-Stokes equations together with the equation of
continuity, the following dimensional form is obtained.
𝜕𝑢 𝜕𝑢 1 𝜕𝑃 𝜇 𝜕 2 𝑢 𝜕 2 𝑢
𝑢 +𝑣 =− + [ + ] (1.16)
𝜕𝑥 𝜕𝑦 𝜌 𝜕𝑥 𝜌 𝜕𝑥 2 𝜕𝑦 2
𝜕𝑣 𝜕𝑣 1 𝜕𝑃 𝜇 𝜕 2 𝑣 𝜕 2 𝑣
𝑢 +𝑣 =− + [ + ] (1.17)
𝜕𝑥 𝜕𝑦 𝜌 𝜕𝑦 𝜌 𝜕𝑥 2 𝜕𝑦 2
𝜕𝑢 𝜕𝑣
+ =0 (1.18)
𝜕𝑥 𝜕𝑦
Now we will perform what is known as an order of magnitude analysis. Here the velocity
components 𝑢 and 𝑣 are acting along the streamwise 𝑥 and normal 𝑦 direction respectively. The
static pressure is 𝑃, while 𝜌 is the density and 𝜇 is the dynamic viscosity of the fluid. The equations
are not non-dimensionalised. The length and the velocity scales are chosen as 𝐿 and 𝑈∞ respectively.
The non-dimensional variables are:
𝑢 𝑣 𝑃 𝑥 𝑦
𝑢∗ = ; 𝑣∗ = ; 𝑃∗ = ; 𝑥∗ = ; 𝑦∗ =
𝑈∞ 𝑈∞ 𝜌𝑈∞ 𝐿 𝐿
The boundary layer is thin; thus, non-dimensional variables for velocity and thickness in the
y-direction have orders of magnitude of 𝜀. In equation form, we would write, for example,
𝑣 ∗ ~𝑂(𝜀) ; 𝑦 ∗ ~𝑂(𝜀)
This states that the non-dimensional variables for velocity in the y-direction have an order of
magnitude of 𝜀. Relative to the y-directed factors, the non-dimensional variables for velocity and
distances in the x-direction have orders of magnitude of unity. For example,
𝑢∗ ~𝑂(1) ; 𝑥 ∗ ~𝑂(1)
The values of 𝛿 and unity are important in a comparative rather than an absolute sense. In this regard,
we will determine the order of magnitude of each term in the continuity and Navier–Stokes equations
and discard the terms of order 𝜀 as being negligible. The continuity equation in differential form in
Eq. (1.18), with each term’s order of magnitude term written underneath, is
𝜕𝑢 𝜕𝑣
+ =0
𝜕𝑥 𝜕𝑦
𝜕(𝑢∗ 𝑈∞ ) 𝜕(𝑣 ∗ 𝑈∞ )
+ =0
𝜕(𝑥 ∗ 𝐿) 𝜕(𝑦 ∗ 𝐿)
𝑈∞ 𝜕𝑢∗ 𝜕𝑣 ∗
[ + ]
𝐿 𝜕𝑥 ∗ 𝜕𝑦 ∗
𝜕𝑢∗ 𝜕𝑣 ∗
+
𝜕𝑥 ∗ 𝜕𝑦 ∗
(1.19)
𝑂(1) 𝑂(𝜀)
+
𝑂(1) 𝑂(𝜀)
Thus, both terms are of order 1, and either can be neglected.
The x-component of the Navier–Stokes equation is, neglecting gravity,

𝜕𝑢 𝜕𝑢 1 𝜕𝑃 𝜇 𝜕 2 𝑢 𝜕 2 𝑢
𝑢 +𝑣 =− + [ + ]
𝜕𝑥 𝜕𝑦 𝜌 𝜕𝑥 𝜌 𝜕𝑥 2 𝜕𝑦 2
𝜕(𝑢∗ 𝑈∞ ) 𝜕(𝑢∗ 𝑈∞ ) 1 𝜕(𝑃∗ 𝜌𝑈∞

2)
𝜇 𝜕 2 (𝑢∗ 𝑈∞ ) 𝜕 2 (𝑢∗ 𝑈∞ )
𝑢∗ 𝑈∞ + 𝑣 ∗
𝑈∞ = − + [ + ]
𝜕(𝑥 ∗ 𝐿) 𝜕(𝑦 ∗ 𝐿) 𝜌 𝜕(𝑥 ∗ 𝐿) 𝜌 𝜕(𝑥 ∗ 𝐿)2 𝜕(𝑦 ∗ 𝐿)2
2
𝑈∞ ∗
𝜕𝑢∗ ∗
𝜕𝑢∗ 𝑈∞2
𝜕𝑃∗ 𝜇 𝑈∞ 𝜕 2 𝑢∗ 𝜕 2 𝑢∗
[𝑢 +𝑣 ]=− + ∙ [ + ]
𝐿 𝜕𝑥 ∗ 𝜕𝑦 ∗ 𝐿 𝜕𝑥 ∗ 𝜌 𝐿2 𝜕𝑥 ∗ 2 𝜕𝑦 ∗ 2
𝜕𝑢∗
∗ ∗
𝜕𝑢∗ 𝜕𝑃∗ 𝜇 𝜕 2 𝑢∗ 𝜕 2 𝑢∗
𝑢 +𝑣 =− ∗+ [ + ]
𝜕𝑥 ∗ 𝜕𝑦 ∗ 𝜕𝑥 𝜌𝑈∞ 𝐿 𝜕𝑥 ∗ 2 𝜕𝑦 ∗ 2
𝜕𝑢∗
∗ ∗
𝜕𝑢∗ 𝜕𝑃∗ 𝜇 𝜕 2 𝑢∗ 𝜕 2 𝑢∗
𝑢 +𝑣 =− ∗+ [ + ]
𝜕𝑥 ∗ 𝜕𝑦 ∗ 𝜕𝑥 𝜌𝑈∞ 𝐿 𝜕𝑥 ∗ 2 𝜕𝑦 ∗ 2
(1.20)
𝑂(1) 𝑂(1) 𝑂(1) 𝑂(1) 𝑂(1)
𝑂(1) + 𝑂(𝜀) = + 𝑂(𝜀 2 ) [ + ]
𝑂(1) 𝑂(𝜀) 𝑂(1) 𝑂(1) 𝑂(𝜀 2 )
The y-component of the Navier–Stokes equations is

𝜕𝑣 𝜕𝑣 1 𝜕𝑃 𝜇 𝜕 2 𝑣 𝜕 2 𝑣
𝑢 +𝑣 =− + [ + ]
𝜕𝑥 𝜕𝑦 𝜌 𝜕𝑦 𝜌 𝜕𝑥 2 𝜕𝑦 2
∗
𝜕(𝑣 ∗ 𝑈∞ ) ∗
𝜕(𝑣 ∗ 𝑈∞ ) 1 𝜕(𝑃∗ 𝜌𝑈∞
2)
𝜇 𝜕 2 (𝑣 ∗ 𝑈∞ ) 𝜕 2 (𝑣 ∗ 𝑈∞ )
𝑢 𝑈∞ + 𝑣 𝑈∞ =− + [ + ]
𝜕(𝑥 ∗ 𝐿) 𝜕(𝑦 ∗ 𝐿) 𝜌 𝜕(𝑦 ∗ 𝐿) 𝜌 𝜕(𝑥 ∗ 𝐿)2 𝜕(𝑦 ∗ 𝐿)2
2
𝑈∞ ∗
𝜕𝑣 ∗ ∗
𝜕𝑣 ∗ 𝑈∞2
𝜕𝑃∗ 𝜇 𝑈∞ 𝜕 2 𝑣 ∗ 𝜕 2 𝑣 ∗
[𝑢 +𝑣 ]=− + ∙ [ + ]
𝐿 𝜕𝑥 ∗ 𝜕𝑦 ∗ 𝐿 𝜕𝑦 ∗ 𝜌 𝐿2 𝜕𝑥 ∗ 2 𝜕𝑦 ∗ 2
𝜕𝑣 ∗
∗ ∗
𝜕𝑣 ∗ 𝜕𝑃∗ 𝜇 𝜕 2𝑣 ∗ 𝜕 2𝑣 ∗
𝑢 +𝑣 =− ∗+ [ + ]
𝜕𝑥 ∗ 𝜕𝑦 ∗ 𝜕𝑦 𝜌𝑈∞ 𝐿 𝜕𝑥 ∗ 2 𝜕𝑦 ∗ 2
𝜕𝑣 ∗ 𝜕𝑣 ∗ 𝜕𝑃∗ 𝜇 𝜕 2𝑣 ∗ 𝜕 2𝑣 ∗
𝑢∗ + 𝑣 ∗
= − + [ + ]
𝜕𝑥 ∗ 𝜕𝑦 ∗ 𝜕𝑦 ∗ 𝜌𝑈∞ 𝐿 𝜕𝑥 ∗ 2 𝜕𝑦 ∗ 2
(1.21)
𝑂(𝜀) 𝑂(𝜀) 𝑂(𝜀) 𝑂(𝜀)
𝑂(1) + 𝑂(𝜀) = 𝑂(𝜀) + 𝑂(𝜀 2 ) [ + ]
𝑂(1) 𝑂(𝜀) 𝑂(1) 𝑂(𝜀 2 )
where 𝛿 ≪ 𝐿. Therefore, 𝑂(𝜀) ≪ 1⁄𝑂(𝜀) and the term 𝜕 2 𝑣 ∗ ⁄𝜕𝑥 ∗ 2 is neglected.
Compare the term

𝜇
∝ 𝜀2
𝜌𝑈∞ 𝐿
𝛿 2 1 𝛿 1
( ) ∝ → ∝ (1.22)
𝐿 𝑅𝑒𝑥 𝐿 √𝑅𝑒𝑥
The boundary layer equation is a function of Reynolds number as derived in Eq. (1.22). Some of
the characteristics of boundary layer are discussed below.
a. The boundary layer grows thicker in the direction of flow, for example, 𝛿 ∝ 𝑥.
b. The flow of viscous fluid results in a thicker boundary layer, for example, 𝛿 ∝ 𝜇.
c. The boundary layer remains thin in a high velocity stream, for example, 𝛿 ∝ 1⁄𝑈∞ .
d. As the boundary layer grows thicker the wall shear stress decreases in the direction of flow,
for example,
𝜕𝑢
𝜏=𝜇
𝜕𝑦
e. It is known that the fluid in the boundary layer is subjected to a pressure gradient from outside
the layer. If the pressure decreases in the direction of flow (as happens in contracting channel
or pipe), it increases the momentum of the fluid in the boundary layer which results in a
thinner boundary layer. If the pressure increases in the flow direction (flow through a
diverging channel, draft tube, etc.), the positive or adverse pressure gradient hastens the
growth of boundary layer and a thicker boundary is obtained at the given section.
The variation in velocity takes place asymptotically in vertical direction and hence the distance
required will be too large to acquire velocity 𝑈∞ . Therefore, the boundary thickness is defined in
many ways.
a. Normal boundary layer thickness, 𝛿0.99 . It is defined as the distance normal (as shown in
Figure 1.4a) to the wall where the velocity differs by 1% from the free stream velocity 𝑈∞ .
In other words, 𝑢 = 0.99𝑈∞ at 𝑦 = 𝛿0.99 .
b. Displacement thickness, 𝛿 ∗ . It is defined as the distance normal to the wall by which the
actual boundary (or wall) should be shifted in order that actual discharge (volume rate
of flow per unit time) would be the same as that of ideal fluid past the displaced boundary
(as shown in Figure 1.4b). In fact, the displacement boundary layer thickness indicates the
distance by which the ambient flow (of velocity 𝑈∞ ) is displaced from the wall due to the
growth of boundary layer.
𝑈∞ 𝑈∞
𝑈∞ − 𝑢
Figure 1.4 Boundary layer thickness: (a) standard boundary layer thickness, (b) boundary layer displacement
thickness.
Because of the velocity deficit, 𝑈∞ − 𝑢, within the boundary layer, the flowrate across
section b–b is less than that across section a–a. However, if we displace the plate at section
a–a by an appropriate amount 𝛿 ∗ , the boundary layer displacement thickness, the flowrates
across each section will be identical. This is true if
𝛿0.99
∗ (𝑈∞ − 𝑢) 𝑏d𝑦
𝛿 𝑏𝑈∞ = ∫
0
where b is the plate width. Thus,
𝛿0.99
𝑢
𝛿∗ = ∫ (1 − ) d𝑦 (1.23)
0 𝑈∞
c. Momentum thickness, 𝛿 ∗∗ . It is defined as the distance by which the boundary should be

displaced to compensate for the reduction in momentum of the flowing fluid on account of
boundary layer formation. Thus,
𝛿0.99
2 ∗∗
𝜌𝑈∞ 𝛿 =∫ 𝜌𝑢(𝑈∞ − 𝑢) d𝑦
0
𝛿0.99
∗∗
𝑢 𝑢
𝛿 =∫ (1 − ) d𝑦 (1.24)
0 𝑈∞ 𝑈∞
1.4 Blasius Flow over a Flat Plate

The classical problem considered by H. Blasius was a two-dimensional, steady, incompressible flow
over a flat plate at zero angle of incidence with respect to the uniform stream of velocity 𝑈∞ . The
fluid extends to infinity in all directions from the plate. The physical problem is already illustrated in
Figure 1.3. Blasius wanted to determine:
a. The velocity field solely within the boundary layer,
b. The boundary layer thickness, 𝛿0.99 ,
c. The shear stress distribution on the plate, and
d. The drag force on the plate.
The Prandtl boundary layer equations in the case under consideration are
𝜕𝑢 𝜕𝑢 𝜇 𝜕 2 𝑢
𝑢 +𝑣 = (1.25)
𝜕𝑥 𝜕𝑦 𝜌 𝜕𝑦 2
𝜕𝑢 𝜕𝑣
+ =0 (1.18)
𝜕𝑥 𝜕𝑦
The boundary conditions are
at 𝑦 = 0, 𝑢=𝑣=0 (Condition 1 & 2)
𝑦 = ∞, 𝑢 = 𝑈∞ (Condition 3)
The solution to the Navier–Stokes equation for this flow begins with an order-of-magnitude analysis
to determine what terms are important. Within the boundary layer the usual balance between viscosity
and convective inertia is struck, resulting in the scaling argument
𝛿 2 𝜇
( ) ∝ (1.22)
𝐿 𝜌𝑈∞ 𝐿
Furthermore, from the scaling argument it is apparent that the boundary layer grows with the
downstream coordinate x,
𝛿 2 𝜇
( ) ∝
𝑥 𝜌𝑈∞ 𝑥
𝜈𝑥
𝛿~√ (1.26)
𝑈∞
Condition 3 is the patching condition between the inviscid and boundary-layer regions. The problem
was solved in 1908 by Blasius by using a coordinate transformation, the mathematical details of
which are beyond the scope of this discussion. The solution form is
𝑢
= 𝑓 ′ (𝜂) (1.27)
𝑈∞
where 𝜂 = 𝑦⁄𝛿 (1.28)
Eq. (1.29) is the final form of Blasius Equation which is a third order nonlinear differential equation.
Blasius obtained the solution of this equation in the form of series expansion through analytical
techniques which is beyond the scope of this syllabus.
2𝑓 ′′′ (𝜂) + 𝑓(𝜂)𝑓 ′′ (𝜂) = 0 (1.29)
The boundary conditions are
at 𝜂 = 0: 𝑓(𝜂) = 0 ; 𝑓 ′ (𝜂) = 0
𝜂 = ∞: 𝑓 ′ (𝜂) = 1
η F F' 0.5(ηF' - F) F" η F F' 0.5(ηF' - F) F"

0 0.0000 0.0000 0.0000 0.3320 3.6 1.9294 0.9233 0.6972 0.0981
0.1 0.0017 0.0332 0.0008 0.3320 3.7 2.0222 0.9326 0.7142 0.0889
0.2 0.0066 0.0664 0.0033 0.3320 3.8 2.1159 0.9411 0.7301 0.0801
0.3 0.0149 0.0996 0.0075 0.3318 3.9 2.2104 0.9487 0.7447 0.0719
0.4 0.0266 0.1328 0.0133 0.3314 4 2.3056 0.9555 0.7581 0.0642
0.5 0.0415 0.1659 0.0207 0.3309 4.1 2.4014 0.9615 0.7704 0.0571
0.6 0.0597 0.1989 0.0298 0.3301 4.2 2.4979 0.9669 0.7816 0.0505
0.7 0.0813 0.2319 0.0405 0.3289 4.3 2.5948 0.9716 0.7916 0.0445
0.8 0.1061 0.2647 0.0528 0.3274 4.4 2.6922 0.9758 0.8007 0.0390
0.9 0.1342 0.2973 0.0667 0.3254 4.5 2.7899 0.9795 0.8088 0.0340
1 0.1656 0.3298 0.0821 0.3230 4.6 2.8881 0.9826 0.8160 0.0295
1.1 0.2001 0.3619 0.0990 0.3200 4.7 2.9865 0.9854 0.8224 0.0255
1.2 0.2379 0.3937 0.1173 0.3166 4.8 3.0851 0.9877 0.8280 0.0219
1.3 0.2789 0.4252 0.1369 0.3125 4.9 3.1840 0.9898 0.8329 0.0187
1.4 0.3230 0.4562 0.1579 0.3078 5 3.2831 0.9915 0.8372 0.0159
1.5 0.3701 0.4868 0.1800 0.3026 5.1 3.3823 0.9930 0.8409 0.0135
1.6 0.4203 0.5167 0.2032 0.2966 5.2 3.4816 0.9942 0.8441 0.0113
1.7 0.4734 0.5461 0.2274 0.2901 5.3 3.5811 0.9952 0.8468 0.0095
1.8 0.5295 0.5747 0.2525 0.2829 5.4 3.6807 0.9961 0.8491 0.0079
1.9 0.5884 0.6026 0.2783 0.2751 5.5 3.7803 0.9968 0.8511 0.0066
2 0.6500 0.6297 0.3047 0.2667 5.6 3.8800 0.9974 0.8528 0.0054
2.1 0.7143 0.6560 0.3316 0.2578 5.7 3.9798 0.9979 0.8542 0.0045
2.2 0.7811 0.6813 0.3588 0.2483 5.8 4.0796 0.9983 0.8553 0.0036
2.3 0.8505 0.7056 0.3862 0.2384 5.9 4.1795 0.9987 0.8563 0.0030
2.4 0.9222 0.7289 0.4136 0.2281 6 4.2794 0.9989 0.8571 0.0024
2.5 0.9962 0.7512 0.4409 0.2174 6.1 4.3793 0.9991 0.8577 0.0019
2.6 1.0724 0.7724 0.4679 0.2064 6.2 4.4792 0.9993 0.8583 0.0016
2.7 1.1507 0.7925 0.4945 0.1953 6.3 4.5791 0.9994 0.8587 0.0012
2.8 1.2309 0.8115 0.5206 0.1840 6.4 4.6791 0.9996 0.8591 0.0010
2.9 1.3129 0.8293 0.5460 0.1727 6.5 4.7790 0.9996 0.8593 0.0008
3 1.3967 0.8460 0.5706 0.1614 6.6 4.8790 0.9997 0.8596 0.0006
3.1 1.4821 0.8616 0.5944 0.1502 6.7 4.9790 0.9998 0.8597 0.0005
3.2 1.5690 0.8760 0.6171 0.1391 6.8 5.0790 0.9998 0.8599 0.0004
3.3 1.6573 0.8894 0.6389 0.1283 6.9 5.1789 0.9998 0.8600 0.0003
3.4 1.7468 0.9017 0.6595 0.1179 7 5.2789 0.9999 0.8601 0.0002
3.5 1.8376 0.9130 0.6789 0.1078
Table 1.1 Blasius solution for laminar flow over a flat plate
Blasius Boundary Layer

2
f
df/d
1.8
d2f/d2
1.6
1.4
1.2
𝑓 1
f
0.8
0.6
0.4
0.2
0
0 1 2 3 4 5 6 7 8
𝜂
Figure 1.5 Blasius boundary layer result from MATLAB
With the profile known, wall shear stress can be evaluated as

𝜕𝑢
𝜏𝑤 = 𝜇 | (1.30)
𝜕𝑦 𝑦=0
Invoking Eqs (1.27) and (1.28) and the values from Table 1.1 into Eq. (1.30), we obtain
𝜕(𝑈∞ 𝑓 ′ (𝜂)) 𝜕𝜂
𝜏𝑤 = 𝜇 ∙ |
𝜕𝜂 𝜕𝑦 𝜂=0
𝜕 ′ 𝜕𝜂
𝜏𝑤 = 𝜇𝑈∞ 𝑓 (𝜂) ∙ |
𝜕𝜂 𝜕𝑦 𝜂=0
1
𝜏𝑤 = 𝜇𝑈∞ × 0.3321 ×
𝛿
2
0.3321𝜌𝑈∞
𝜏𝑤 = (1.31)
√𝑅𝑒𝑥
and the local skin friction coefficient is
𝜏𝑤 0.664
𝐶𝑑 = = (1.32)
1 2 √𝑅𝑒𝑥
2 𝜌𝑈∞
In 1951, Liepmann and Dhawan, measured the shearing stress on a flat plate directly. Their results
showed a striking confirmation of Eqs (1.31) and (1.32). Total frictional force per unit width for
the plate of length 𝐿 is
𝐿
𝐹 = ∫ 𝜏𝑤 d𝑥
0
𝐿
2
𝜇
𝐹 = ∫ 0.3321𝜌𝑈∞ ∙√ d𝑥
0 𝜌𝑈∞ 𝑥
𝐿
2
0.3321𝜌𝑈∞ √𝑥
𝐹= ∙ |
1 |
√𝑈∞ 2
𝜈 0
2
2
𝜈 0.664𝜌𝑈∞ 𝐿
𝐹 = 0.664𝜌𝑈∞ 𝐿√ =
𝑈∞ 𝐿 √𝑅𝑒𝐿
And the average skin friction coefficient (the total skin friction drag coefficient) is
𝐹 1.328
𝐶𝐷 = = (1.33)
1 2 √𝑅𝑒𝐿
2 𝜌𝑈∞ 𝐿
where 𝑅𝑒𝐿 = 𝑈∞ 𝐿⁄𝜈 and 𝜈 is the kinematic viscosity.
Since 𝑢⁄𝑈∞ approaches 1.0 as 𝑦 → ∞, it is customary to select the boundary layer thickness 𝛿 as
that point where 𝑢⁄𝑈∞ approaches 0.99. From Table 1.1, 𝑢⁄𝑈∞ reaches 0.99 at 𝜂 = 5.0 and we can
write
𝑈∞
𝜂 = 5.0 = 𝛿√
𝜈𝑥
5.0𝑥 (1.34)
𝛿0.99 =
√𝑅𝑒𝑥
For displacement thickness, substitute the Eqs (1.27) and (1.28) into Eq. (1.23), we obtain
∞
𝜈𝑥
𝛿 ∗ = ∫ (1 − 𝑓 ′ ) d (√ 𝜂)
0 𝑈∞
∞
𝑥
𝛿∗ = ∫ (1 − 𝑓 ′ ) d𝜂
√𝑅𝑒𝑥 0
𝑥
𝛿∗ = [𝜂 − 𝑓(𝜂)]∞
0
√𝑅𝑒𝑥
When we consider the numerical result of 𝜂 − 𝑓(𝜂), we see that for all values above 𝜂 = 4.9 the
result of 𝜂 − 𝑓(𝜂) is remain constant which is about 1.7206. Therefore, this results in an equation for
the displacement thickness is
1.7206𝑥
𝛿∗ = (1.35a)
√𝑅𝑒𝑥
5.0𝑥 (1.35b)
𝛿 ∗ = 0.344 × ≈ 0.344𝛿
√𝑅𝑒𝑥
With substitution of Eqs (1.27) and (1.28) into Eq. (1.24), we can evaluate the momentum thickness
numerically for a flat plate as
∞
∗∗
𝜈𝑥
𝛿 = ∫ 𝑓 ′ (1 − 𝑓 ′ ) d (√ 𝜂)
0 𝑈∞
∞
𝑥
𝛿 ∗∗ = ∫ 𝑓 ′ (1 − 𝑓 ′ ) d𝜂
√𝑅𝑒𝑥 0
The equation in the integral 𝑓 ′ (1 − 𝑓 ′ ) cannot be integrated analytically. However, for 𝜂 > 4.9 the
𝑓 ′ will not change anymore and will be equal to 1.0. The numerical result of the integral is 0.664. So
we obtain the momentum thickness in our case
0.644𝑥
𝛿 ∗∗ = (1.36a)
√𝑅𝑒𝑥
5.0𝑥 (1.36b)
𝛿 ∗∗ = 0.1288 × ≈ 0.13𝛿
√𝑅𝑒𝑥
Note that these results are obtained for laminar flow over flat plate for 𝑹𝒆 < 𝟓 × 𝟏𝟎𝟓 .
Example 1.3
Air at 30°C flows over a flat plate at a free stream velocity of 5 m/s. Determine the boundary layer
thickness (𝛿) , displacement thickness (𝛿 ∗ ) and momentum thickness (𝛿 ∗∗ ) at distances 0.2 m,
0.5 m and 0.8 m. Also determine the skin friction coefficients, both local and average, at these
locations. The property values for air at 30°C are 𝜌 = 1.165 kg/m3 and 𝜈 = 16 × 10−6 m2 /s.
Eqs (1.32) to (1.36) will be applied to solve the problem.
𝜌𝑈∞ 𝑥 5.0𝑥 1.7206𝑥 0.644𝑥 0.664 1.328

𝑅𝑒𝑥 = ; 𝛿= ; 𝛿∗ = ; 𝛿 ∗∗ = ; 𝐶𝑑 = ; 𝐶𝐷 =
𝜇 √𝑅𝑒𝑥 √𝑅𝑒𝑥 √𝑅𝑒𝑥 √𝑅𝑒𝑥 √𝑅𝑒𝐿
Distance, m Reynolds number 𝛿, mm 𝛿 ∗ , mm 𝛿 ∗∗ , mm 𝐶𝑑 𝐶𝐷

0.2 62,500 4 × 10−3 1.376 × 10−3 5.152 × 10−4 2.656 × 10−3 5.312 × 10−3
0.5 156,250 6.325 × 10−3 2.176 × 10−3 8.146 × 10−4 1.68 × 10−3 3.36 × 10−3
0.8 250,000 8 × 10−3 2.753 × 10−3 1.03 × 10−3 1.328 × 10−3 2.656 × 10−3
Comparing the result of local drag coefficient and total drag coefficient, we note that the total drag coefficient is twice the
value of the friction coefficient at 𝑥 = 𝐿; that is 𝐶𝐷 = 2𝐶𝑑 .
1.5 Momentum Integral Equation for Boundary Layer

The boundary layer equations are a set of partial differential equations which we have to solve to
obtain results like the Blasius equation. In this section we will not obtain the exact (Blasius) solution,
but we will approximate the solution of the boundary layer equations.
Formulation of integral equations describing the flow and solving them using an assumed velocity
variation satisfying the boundary conditions. This method is more versatile and results in easier
solution of problems. The difference between the results obtained by the exact method and by the
integral method is found to be within acceptable limits. The
momentum integral technique utilizes a control volume
approach to obtain such quantitative approximations of
boundary layer properties along surfaces with zero or
nonzero pressure gradients. The momentum integral
technique is straightforward, and in some applications does
not require use of a computer. It is valid for both laminar
and turbulent boundary layers.
We begin with the control volume sketched in Figure 1.6.

The bottom of the control volume is the wall at 𝑦 = 0, and
the top is at 𝑦 = 𝑌, high enough to enclose the entire height
of the boundary layer. The control volume is an
infinitesimally thin slice of width d𝑥 in the x-direction. In Figure 1.6 Control volume used in derivation of
accordance with the boundary layer approximation, the momentum integral equation.
𝜕𝑃⁄𝜕𝑦 = 0, so we assume that pressure 𝑃 acts along the
entire left face of the control volume,
𝑃left face = 𝑃
In the general case with nonzero pressure gradient, the pressure on the right face of the control volume
differs from that on the left face.
d𝑃
𝑃right face = 𝑃 + d𝑥
d𝑥
In a similar manner we write the incoming mass flow rate through the left face as
𝑌
𝑚̇left side = 𝜌𝑏 ∫ 𝑢 d𝑦 (1.37)
0
and the outgoing mass through the right face as

𝑌 𝑌
d
𝑚̇right side = 𝜌𝑏 [∫ 𝑢 d𝑦 + (∫ 𝑢 d𝑦) d𝑥] (1.38)
0 d𝑥 0
where 𝑏 is the width of the control volume in Figure 1.6. Since Eq. (1.38) differs from Eq. (1.37),
and since no flow can cross the bottom of the control volume (the wall), mass must flow into or out
of the top face of the control volume. We illustrate this in Figure 1.7 for the case of a growing
boundary layer in which 𝑚̇right side < 𝑚̇left side and 𝑚̇top is positive (mass flows out). Conservation of
mass over the control volume yields
𝑌
d
𝑚̇top = −𝜌𝑏 (∫ 𝑢 d𝑦) d𝑥 (1.39)
d𝑥 0
We now apply conservation of x-momentum for the chosen control volume.

The x-momentum is brought in through the left face and is removed through
the right and top faces of the control volume. The net momentum flux out
of the control volume must be balanced by the force due to the shear stress
acting on the control volume by the wall and the net pressure force on the
control surface, as shown in Figure 1.6. Assume no gravity force exerts on
the system and thus, the steady control volume x-momentum equation is
⃗⃗⃗ Figure 1.7 Mass flow balance

𝐹𝑆 = ∬ 𝑛⃗ 𝑃d𝐴
on the control volume
𝑆
𝐹𝑆 = ∫ ⃗ ∙ 𝑛⃗d𝐴 + ∫
𝜌𝑢𝑉 ⃗ ∙ 𝑛⃗d𝐴 + ∫ 𝜌𝑢𝑉
𝜌𝑢𝑉 ⃗ ∙ 𝑛⃗d𝐴
left side right side top
𝑌 𝑌 𝑌 𝑌
2
d 2 2
d
𝐹𝑆 = −𝜌𝑏 ∫ 𝑢 d𝑦 + 𝜌𝑏 [∫ 𝑢 d𝑦 + (∫ 𝑢 d𝑦) d𝑥] − 𝜌𝑏𝑈∞ (∫ 𝑢 d𝑦) d𝑥
0 0 d𝑥 0 d𝑥 0
d𝑃
where 𝐹𝑆 = 𝑃(𝑏𝑌) − 𝑏𝑌 [𝑃 + d𝑥] − 𝜏𝑤 𝑏 ∙ d𝑥
d𝑥
where the momentum flux through the top surface of the control volume is taken as the mass flow
rate through that surface times 𝑈∞ . Some of the terms cancel, and we rewrite the equation as
𝑌 𝑌
d𝑃 d 2
d
−𝑏𝑌 d𝑥 − 𝜏𝑤 𝑏 ∙ d𝑥 = 𝜌𝑏 (∫ 𝑢 d𝑦) d𝑥 − 𝜌𝑏 𝑈∞ (∫ 𝑢 d𝑦) d𝑥
d𝑥 d𝑥 0 d𝑥 0
𝑌 𝑌
d𝑃 d 2
d
−𝑌 − 𝜏𝑤 = 𝜌 (∫ 𝑢 d𝑦) − 𝜌𝑈∞ (∫ 𝑢 d𝑦) (1.40)
d𝑥 d𝑥 0 d𝑥 0
The Bernoulli’s equation applied to the outer inviscid flow is given by
d𝑈∞ 1 d𝑃
𝑈∞ =− (1.41)
d𝑥 𝜌 d𝑥
𝑌
For convenience we note that 𝑌 = ∫0 d𝑦. Invoking the Eq (1.41) into Eq. (1.40) reverse the produce
in the second term on the right hand of Eq. (1.40).
d𝑈∞ 𝑌 d 𝑌 𝑑 𝑌 𝑑𝑈∞ 𝑌
𝜌𝑈∞ ∫ d𝑦 − 𝜏𝑤 = 𝜌 (∫ 𝑢2 d𝑦) − 𝜌 [ (𝑈∞ ∫ 𝑢 𝑑𝑦) − ∫ 𝑢 𝑑𝑦] (1.42)
d𝑥 0 d𝑥 0 𝑑𝑥 0 𝑑𝑥 0
d𝑈∞ 𝑌 𝑑𝑈∞ 𝑌 𝑑 𝑌
𝜌𝑈∞ ∫ d𝑦 − 𝜌 ∫ 𝑢 𝑑𝑦 − 𝜏𝑤 = 𝜌 (∫ 𝑢(𝑢 − 𝑈∞ ) 𝑑𝑦)
d𝑥 0 𝑑𝑥 0 𝑑𝑥 0
𝑌
d𝑈∞ 𝑌 𝑑 𝜏𝑤
(
∫ 𝑈∞ − 𝑢)d𝑦 + (∫ 𝑢(𝑈∞ − 𝑢) 𝑑𝑦) =
d𝑥 0 𝑑𝑥 0 𝜌
∞
𝑑 2
𝑢 𝑢 d𝑈∞ ∞ 𝑢 𝜏𝑤
(𝑈∞ ∫ (1 − ) 𝑑𝑦) + 𝑈∞ ∫ (1 − ) d𝑦 = (1.43)
𝑑𝑥 0 𝑈∞ 𝑈∞ d𝑥 0 𝑈∞ 𝜌
We previously defined displacement thickness 𝛿 ∗ in Eq. (1.23) and momentum thickness 𝛿 ∗∗ in

Eq. (1.24) for a flat plate boundary layer. Eq. (1.43) can thus be written in more compact form as
𝑑 2 ∗∗ )
d𝑈∞ 𝜏𝑤
(𝑈∞ 𝛿 + 𝑈∞ 𝛿 ∗ = (1.44)
𝑑𝑥 d𝑥 𝜌
Equation (1.44) is known as momentum integral equation or Von Kármán integral equation for
two-dimensional incompressible laminar and turbulent boundary layer. This equation can be used to
determine the wall shear stress at any section or wall. The effect of pressure gradient is described by
d𝑈∞
the second term on the left hand side in Eq. (1.44) For external flow, the existence of 𝑈∞
d𝑥
d𝑈∞
depends on the shape of the body. During the flow over a flat plate, 𝑈∞ = 0 and Eq. (1.44) is
d𝑥
reduced to
𝑑 𝜏𝑤
(𝑈∞2 𝛿 ∗∗ ) = (1.45a)
𝑑𝑥 𝜌
𝑑𝛿 ∗∗ 𝜏𝑤
= (1.45b)
𝑑𝑥 𝜌𝑈∞2
Equation (1.45) can be solved if a velocity profile satisfying the boundary conditions is assumed.
However, we assume a velocity profile which a polynomial of 𝜂 = 𝑦⁄𝛿 . As it has been seen earlier,
𝜂 is a form of similarity variable. This implies that with the growth of boundary layer as distance 𝑥
varies from the leading edge, the velocity profile 𝑢⁄𝑈∞ remains geometrically similar. Out of the
popularly used profiles the results obtained from a cubic profile given below is in closer agreement
with the exact solution.
𝑢 𝑦 𝑦 2 𝑦 3
= 𝑎0 + 𝑎1 + 𝑎2 ( ) + 𝑎3 ( ) (1.46)
𝑈∞ 𝛿 𝛿 𝛿
In order to determine the constants 𝑎0 , 𝑎1 , 𝑎2 and 𝑎3 we shall prescribe the following boundary
conditions
at 𝑦 = 0, 𝑢=0 --- 1st Boundary Condition
𝑦 = 𝛿, 𝑢 = 𝑈∞ --- 2nd Boundary Condition
𝜕𝑢
𝑦 = 𝛿, =0 --- 3rd Boundary Condition
𝜕𝑦
𝜕 2𝑢
𝑦 = 0, =0 --- 4th Boundary Condition
𝜕𝑦 2
These requirements will yield

3 1
𝑎0 = 0 ; 𝑎1 = ; 𝑎2 = 0 ; 𝑎3 = −
2 2
and the velocity profile becomes
𝑢 3𝑦 1 𝑦 3
= − ( ) (1.47)
𝑈∞ 2 𝛿 2 𝛿
Substituting Eq. (1.47) into the momentum thickness equation in Eq. (1.24)
𝛿
∗∗
𝑢 𝑢
𝛿 =∫ (1 − ) d𝑦
0 𝑈∞ 𝑈∞
𝛿
∗∗
3𝑦 1 𝑦 3 3𝑦 1 𝑦 3
𝛿 =∫ ( − ( ) ) (1 − + ( ) ) d𝑦
0 2𝛿 2 𝛿 2𝛿 2 𝛿
39
𝛿 ∗∗ = 𝛿 (1.48)
280
The wall shear stress is given by
𝜕𝑢
𝜏𝑤 = 𝜇 |
𝜕𝑦 𝑦=0
𝜕 3𝑦 1 𝑦 3
𝜏𝑤 = 𝜇 [𝑈∞ ( − ( ) )]|
𝜕𝑦 2𝛿 2 𝛿 𝑦=0
3𝜇𝑈∞
𝜏𝑤 = (1.49)
2𝛿
Substituting the Eqs (1.48) and (1.49) into Eq. (1.45b) we get
d 39 1 3𝜇𝑈∞
( 𝛿) = ∙
d𝑥 280 𝜌𝑈∞2 2𝛿
39 d𝛿 3𝜇
=
280 d𝑥 2𝛿𝜌𝑈∞
140𝜇
∫ 𝛿d𝛿 = ∫ d𝑥
13𝜌𝑈∞
𝛿 2 140 𝜇
= 𝑥+𝐶
2 13 𝜌𝑈∞
where 𝐶 is any arbitrary unknown constant. The condition at the leading edge (at 𝑥 = 0, 𝛿 = 0)
yields 𝐶 = 0. Therefore, the equation is reduced to
280 𝜇
𝛿2 = 𝑥
13 𝜌𝑈∞
4.64𝑥
𝛿0.99 = (1.50)
√𝑅𝑒𝑥
This is the value of boundary layer thickness on a flat plate. Although, the method is an approximation
one, the result is found to be reasonably accurate. The value is slightly lower than the exact (Blasius)
solution of laminar flow over a flat plate given by Eq. (1.34). As such, the accuracy depends on the
order of the velocity profile as mentioned in Eq. (1.46).
𝑢
𝑈∞
𝑢
𝑈∞
𝑢
𝑈∞
𝑢
𝑈∞
Table 1.2 Comparisons of Approximate Results to Exact Solution for Flow over a Flat Plate
Example 1.4
Water flows over a flat plate at a free stream velocity of 0.15 m/s. There is no pressure gradient and
laminar boundary layer is 6 mm thick. Assume a sinusoidal velocity profile is
𝑢 𝜋 𝑦
= sin ( ∙ )
𝑈∞ 2 𝛿
For the flow conditions stated above, calculate the local wall shear stress and skin friction coefficient.
Assume the dynamic viscosity and the density of the medium is 1.02 × 10−3 Ns/m and 1000 kg/m3.
From Eq. (1.30), And the skin coefficient at this point (local),
𝜕𝑢 𝜏𝑤
𝜏𝑤 = 𝜇 | 𝐶𝑑 =
𝜕𝑦 𝑦=0 1 2
2 𝜌𝑈∞
𝜋𝑈∞ 𝜋 𝑦 0.04
𝜏𝑤 = 𝜇 ∙ cos ( ∙ )| 𝐶𝑑 =
2𝛿 2 𝛿 𝑦=0 1 2
2 × 1000 × 0.15
𝜋𝑈∞
𝜏𝑤 = 𝜇 𝐶𝑑 = 0.0035
2𝛿
The wall shear stress for this scenario is derived and
shown as above. Therefore, the wall shear stress is
1.02 × 10−3 × 𝜋 × 0.15

𝜏𝑤 = = 0.04 N/m2
2 × 0.006
1.6 Turbulent Boundary Layer

There are no precise definitions of turbulence. This is similar to diseases in a sense that the diseases
can only be diagnosed by their symptoms. Then, how do we know that a flow is turbulent? We have
to look at the characteristics of the flow in order to find out whether it is turbulent or not. The
characteristics are:
i. Irregularity – The turbulence is irregular and random. Because of this, a

deterministic approach to turbulence problems is impossible. We
have to use the statistical approach instead.
ii. Diffusivity – The turbulence is diffusive. In fact, this is a key feature of

turbulence in terms of industrial applications, such as mixing of
fluids. On the contrary, non-turbulent flows are not very diffusive.
We can see this difference when cream is poured into a coffee cup
– mixing takes place quickly if we stir coffee with a spoon but it
is a very slow process if not stirred.
iii. Large Reynolds – As the Reynolds number of the flow increases, the laminar flow
number becomes unstable and the transition takes place. This is a common
route to turbulence. Because of this, the turbulent flows have large
Reynolds numbers.
iv. Three-dimensional – Turbulence is rotational and three-dimensional. Therefore,

fluctuations of turbulence is characterised by a high degree of fluctuating
vorticity vorticity motion. Random waves in the oceans are not turbulent,
because they are essentially irrotational.
v. Dissipation – Turbulence energy is dissipated by the viscous motion at small

scale. Therefore, if energy is not supplied to the turbulent flow or
produced within the flow the turbulence will decay. The random
waves are not turbulent because they are not dissipative.
vi. Continuous – The smallest scale of turbulence is much greater than the size of
molecules. Therefore turbulence is a continuum phenomenon,
which is described by the equations of fluid motion.
The turbulent motion is an irregular motion. The irregularity associated with turbulence is such that
it can be described by the laws of probability and turbulent fluid motion can be considered as irregular
condition of flow in which various quantities (such as velocity components and pressure) show a
random variation with time and space in such a way that the statistical average of those quantities can
be quantitatively expressed.
Irregular motion is associated with random fluctuations. It is postulated that the fluctuations
inherently come from disturbances (such as roughness of a solid surface) and they may be either
dampened out due to viscous damping or may grow by drawing energy from the free stream. At a
Reynolds number less than the critical (𝑹𝒆 < 𝟓 × 𝟏𝟎𝟓 ), the kinetic energy of flow is not enough to
sustain the random fluctuation against the viscous damping and in such case laminar flow continues
to exist. At somewhat higher Reynolds number than the critical Reynolds number (𝑹𝒆 > 𝟓 × 𝟏𝟎𝟓 ),
the kinetic energy of flow supports growth of fluctuations and transition to turbulence take place.
The presence of turbulence in the

boundary layer considerably
modifies the velocity distribution
and the boundary layer grows at
faster rate in the direction of flow.
The turbulent boundary layer is
distinguished from laminar
boundary layer by its larger
thickness. The process of
transverse mixing tends to make
the boundary layer flow more
𝑢⁄𝑈∞ 𝑢⁄𝑈∞
uniform throughout the cross
Figure 1.8 Turbulent boundary layer profile compared to the laminar profile.
section (resulting in logarithmic
velocity distribution) except in the immediate vicinity of the boundary layer. This means a rapid
change in velocity (from zero to 𝑈∞ ) take place in the small distance near the wall, thereby giving a
high value of velocity gradient. This in turn gives a larger wall shear stress (𝜏 ∝ 𝜕𝑢⁄𝜕𝑦) at a section
as compared to laminar layer.
Typical velocity profiles for fully developed laminar and turbulent flows are given in Figure 1.8.
Note that the velocity profile is parabolic in laminar flow but is much fuller in turbulent flow, with
a sharp drop near the pipe wall. Turbulent flow along a wall can be considered to consist of four
regions, characterized by the distance from the wall. The very thin layer next to the wall where
viscous effects are dominant is the viscous sub-layer (or laminar or linear or wall). The velocity
profile in this layer is very nearly linear, and the flow is streamlined. Next to the viscous sub-layer is
the buffer layer, in which turbulent effects are becoming significant, but the flow is still dominated
by viscous effects. Above the buffer layer is the overlap layer (log-law region), also called the
inertial sub-layer, in which the turbulent effects are much more significant, but still not dominant.
Above that is the outer (or turbulent) layer in the remaining part of the flow in which turbulent effects
dominate over molecular diffusion (viscous) effects.
Flow characteristics are quite different in different regions, and thus it is difficult to come up with an
analytic relation for the velocity profile for the entire flow as we did for laminar flow. Thus, the
experiment had been conducted by Nikuradse. It showed that Eq. (1.51) is in good agreement with
experiment result and actually approximates nearly the entire velocity profile, except the outer flow.
The inner-wall law typically extends over less than 2% of the profile and can be neglected. Thus we
can use Eq. (1.51) as an excellent approximation to solve nearly every turbulent-flow problem.
𝑢+ = 2.5 ln 𝑦 + + 5 (1.51)
Experimental Data
Figure 1.9 Comparison of the law of the wall and the logarithmic-law velocity profiles
with experimental data for fully developed turbulent flow in a pipe.
We begin with the momentum integral equation for flat plate boundary layer which is valid for both
laminar and turbulent flow. Invoking the definition of average skin friction as in Eq. (1.33),
Eq. (1.45b) can be rewritten as
𝑑𝛿 ∗∗
𝐶𝑑 = 2 (1.52)
𝑑𝑥
It is beyond the scope of this text to derive or

attempt to solve the turbulent flow boundary
layer equations. Expressions for the boundary
layer profile shape and other properties of the
turbulent boundary layer are obtained
empirically (or at best semi-empirically), since
we cannot solve the boundary layer equations
for turbulent flow. Note also that turbulent
flows are inherently unsteady, and the
instantaneous velocity profile shape varies
with time (Fig. 1.10). Thus, all turbulent
Figure 1.10 Illustration of the unsteadiness of a turbulent
expressions discussed here represent time- boundary layer; the thin, wavy black lines are instantaneous
averaged values. One common empirical profiles, and the thick blue line is a long time-averaged profile.
approximation for the time-averaged velocity
profile of a turbulent flat plate boundary layer is the one-seventh-power law,
1
𝑢 𝑦 7
=( ) (1.53)
𝑈∞ 𝛿
where this velocity profile is valid for turbulent boundary layer with 𝑹𝒆 < ~𝟓 × 𝟏𝟎𝟕 . Substitute
Eq. (1.53) into Eq. (1.24) and integrate to obtain
∞ 1 1
∗∗
𝑦 7 𝑦 7 7
𝛿 = ∫ ( ) [1 − ( ) ] d𝑦 = 𝛿 (1.54)
0 𝛿 𝛿 72
Also, substitute Eq. (1.54) into Eq. (1.45b) to obtain

𝑑 7
𝜏𝑤 = 𝜌𝑈∞2 ( 𝛿) (1.55)
𝑑𝑥 72
Consequently, the law of shear stress (derivation is beyond the scope) for the flat plate is found out
by making use of the pipe flow expression which is given by
1
𝜈 4
𝜏𝑤 = 0.0225𝜌𝑈∞2 ( ) (1.56)
𝑈∞ 𝛿
Equating Eq. (1.55) and Eq. (1.56), we obtain

1
𝑑 7 𝜈 4
𝜌𝑈∞2 ( 𝛿) = 0.0225𝜌𝑈∞2 ( )
𝑑𝑥 72 𝑈∞ 𝛿
1
1 𝑑𝛿 𝜈 4
𝛿4 = 0.2314 ( )
𝑑𝑥 𝑈∞
1
5 𝜈 4 (1.57)
𝛿4 = 0.2892 ( ) 𝑥 + 𝐶
𝑈∞
For simplicity, if we assume that the turbulent boundary layer grows from the leading edge of the
plate we shall be able to apply the boundary condition 𝑥 = 0, 𝛿 = 0 which will yield 𝐶 = 0, and
Eq. (1.57) will become
1
5 𝜈 4 5 1
𝛿4 = 0.2892 ( ) 𝑥 4−4
𝑈∞
5 1
𝛿 4 𝜈 4
( ) = 0.2892 ( )
𝑥 𝑈∞ 𝑥
0.371𝑥
𝛿0.99 = 5 (1.58)
√𝑅𝑒𝑥
From Eqs (1.52), (1.54) and (1.58), it is possible to calculate the average skin friction coefficient
(total skin friction) on a flat plate as following
𝑑 7
𝐶𝐷 = 2 ( 𝛿)
𝑑𝑥 72
7 𝑑𝛿 7 0.371 0.072 (1.59)
𝐶𝐷 = = =
36 𝑑𝑥 36 5√𝑅𝑒𝑥 5√𝑅𝑒𝑥
The Eq. (1.60) is found to be in good agreement with the experimental results in the range of Reynolds
number between 5 × 105 → 5 × 107 . This equation is a widely accepted correlation for the average
value of turbulent skin friction coefficient on a flat plate. Table 1.3 gives a summary of different
power-law velocity profile for various range of Reynolds number.
Table 1.3 Results of Momentum Integral Method Applied to Turbulent Boundary-Layer Flow over a Flat Plate
Skin Friction Coefficient over a Flat

Flow type Reynolds Number Range
Plate
1.328
Laminar Flow 𝐶𝐷 = 𝑅𝑒 < 5 × 105
√𝑅𝑒𝐿
Turbulent Flow 0.074
𝐶𝐷 = 5 5 × 105 < 𝑅𝑒 < 107
(Empirical correlation) √𝑅𝑒𝑥
0.455
Turbulent Flow 𝐶𝐷 = 2.58 5 × 105 < 𝑅𝑒 < 109
√log 𝑅𝑒𝐿
1.7 Boundary Layer Separation & Wake Formation

Consider a Couette flow between two parallel plates where one plate is at rest and the other is moving
with a velocity 𝑈∞ . Let us assume the plates are infinitely large in 𝑧-direction, so the 𝑧 dependence
d𝑃 d2 𝑢
is not there and the governing equation is =𝜇 subjected to the boundary conditions at 𝑦 = 0,
d𝑥 d𝑦 2
𝑢 = 0 and 𝑦 = 𝛿, 𝑢 = 𝑈∞ . We get
1 d𝑃 2
𝑢= 𝑦 + 𝐶1 𝑦 + 𝐶2
2𝜇 d𝑥
Invoking the condition (at 𝑦 = 0, 𝑢 = 0), 𝐶2 becomes equal to zero. Invoking the other condition
(at 𝑦 = 𝛿, 𝑢 = 𝑈∞ ),
𝑈∞ 1 d𝑃
𝐶1 = − 𝛿
𝛿 2𝜇 d𝑥
So,
1 d𝑃 2 𝑈∞ 1 d𝑃
𝑢= 𝑦 +( − 𝛿) 𝑦
2𝜇 d𝑥 𝛿 2𝜇 d𝑥
𝑦 1 d𝑃 𝑦 𝛿 2 d𝑃 𝑦 𝑦 2
𝑢 = 𝑈∞ − (𝛿𝑦 − 𝑦 2 ) = 𝑈∞ − ( − ) (1.60)
𝛿 2𝜇 d𝑥 𝛿 2𝜇 d𝑥 𝛿 𝛿 2
𝑢 𝑦 𝛿 2 d𝑃 𝑦 𝑦 2
= − ( − )
𝑈∞ 𝛿 2𝜇𝑈∞ d𝑥 𝛿 𝛿 2
𝑢 𝑦 𝑦 𝑦2
= + 𝑃 ( − 2) (1.61)
𝑈∞ 𝛿 𝛿 𝛿
𝛿 2 d𝑃
where 𝑃=−
2𝜇𝑈∞ d𝑥
Equation (1.61) describes the velocity distribution in non-dimensional form across the channel with
𝑃 as a parameter known as the non-dimensional pressure gradient. When 𝑃 = 0 , the velocity
𝑢 𝑦
distribution across the channel is reduced to = .
𝑈∞ 𝛿
This particular case is known as simple

Couette flow. When 𝑃 > 0, for example,
for a negative or favourable pressure
gradient (− d𝑃 ⁄d𝑥 ) in the direction of
motion, the velocity is positive (flow is
accelerating) over the whole gap between
the channel walls. For negative value of
(𝑃 < 0) , there is a positive or adverse
pressure gradient in the direction of motion
and the velocity over a portion of channel
width can be negative (flow is
decelerating) and back flow may occur Figure 1.11 Velocity profile for various values of pressure gradient.
near the wall which is at rest. Figure 1.11
shows the effect of dragging action of the upper plate exerted on the fluid particles in the channel for
different values of pressure gradient.
Figure 1.12 Real flow around a cylinder
Unlike an ideal fluid that can slide freely around a cylinder, a real fluid has viscosity, and as a result,
the fluid will tend to form a boundary layer and cling to the cylinder’s surface as it flows around it.
This phenomenon was studied in the early 1900s by the Romanian engineer Henri Coanda and is
called the Coanda effect. To understand this viscous behaviour, we will consider the long cylinder
shown in Fig. 1.12a. The flow starts from the stagnation point at A, and thereafter it forms a laminar
boundary layer on the cylinder as the fluid begins to travel around the surface. The favourable
pressure gradient (pressure decrease) within this initial region increases the velocity, Fig. 1.12b.
Because the flow must overcome the drag effect of viscous friction within the boundary layer, the
minimum pressure maximum velocity will occur at point B’. This is sooner than in the case of ideal
fluid flow.
Although the boundary layer continues to grow in thickness farther downstream of point B’, the
velocity decreases here because the adverse pressure gradient (increasing pressure) acting within this
region. Point C’ marks a separation of flow from the cylinder since the velocity of the slower-moving
particles near the surface is finally reduced to zero at this point. Beyond C’, within the boundary layer,
the flow will begin to back up and move in the opposite direction to the free-stream flow. This
ultimately will form a vortex, which will shed from the cylinder, as shown in Fig. 1.12b. A series of
these vortices or eddies produce a wake. The eddies in the wake are kept in motion by the shear
stresses between the wake and the separated current (mainstream flow). They consume considerable
mechanical energy and may lead to a large pressure loss (energy will eventually dissipate as heat) in
the fluid. The resultant of the entire pressure distribution around the cylinder produces the pressure
drag (also known as form drag), 𝐹DP . Notice that the magnitude of this force depends to some degree
on the location or point C’ where the flow separates from the cylinder.
If the flow within the boundary layer is completely turbulent then separation will occur later than if
the boundary layer flow is laminar (Refer back to BTME2213 Fluid Mechanics, Chapter 7). This
happens because within a turbulent boundary layer, the fluid has more kinetic energy than in the
laminar case. As a result, the adverse pressure gradient will take longer to arrest the flow, and so the
point of separation is farther back on the surface. Consequently, the resultant of the pressure
distribution will create a smaller pressure drag than in the case of laminar flow. Since surface
roughness has the effect of producing turbulent boundary layers, it may seem counterintuitive, but
one way to reduce pressure drag is to roughen the cylinder’s front surface.
Unfortunately, the actual point C’ of flow separation for either laminar or turbulent boundary layers
cannot be determined analytically, except through approximate methods. However, experiments have
shown that, as expected, the point of transition from laminar to turbulent flow is a function of the
Reynolds number, and therefore the pressure drag, like viscous or friction drag, will be a function of
this parameter. For a cylinder, the “characteristic length” for finding the Reynolds number is its
diameter D, so
𝜌𝑈∞ 𝐷
Re = (1.62)
𝜇
Methods for Reducing Drag: Besides moving the point of flow separation farther back on the
object, there are some popular methods for this purpose which are stated as follows:
a. By giving the profile of the body a streamlined shape as shown in Figure 1.14. This has an
elongated shape in the rear part to reduce the magnitude of the pressure gradient. The optimum
contour for a streamlined body is the one for which the wake zone is very narrow and the form
drag is minimum.
Figure 1.14 Reduction of drag coefficient by giving the profile a streamlined shape
b. The injection of fluid through porous wall can also control the boundary layer separation. This
is generally accomplished by blowing high energy fluid particles tangentially from the
location where separation would have taken place otherwise. This is shown in Figure 1.15.
The injection of fluid promotes turbulence and thereby increases skin friction. But the form
drag is reduced considerably due to suppression of flow separation and this reduction can be
of significant magnitude so as to ignore the enhanced skin friction drag.
Figure 1.15 Boundary layer control by blowing
Chapter Review:
 When pressure, gravity and viscous forces are all taken into account, the equations of motion
are expressed as the Navier-Stokes equations. Along with the continuity equation, only a
limited number of solutions have actually been obtained, and these are for laminar flow.
 The boundary layer is a very thin layer of fluid located in a region just above the surface of a
body. Within it, the velocity changes from zero at the surface to the free-stream velocity of
the fluid.
 The fluid within the boundary layer formed over the surface of a flat plate will be laminar up
to the critical distance 𝑥cr . In this text, this distance is determined from
𝜌𝑈∞ 𝑥cr
Recr = = 𝟓 × 𝟏𝟎𝟓
𝜇
 The velocity profile for a laminar flow boundary layer has been solved by Blasius method.
Knowing this velocity profile (𝑢/𝑈∞ is unknown or not given), one can find the thickness of
the boundary layer and the friction drag that the flow exerts over a flat plate.
2
5𝑥 0.664𝜌𝑈∞ 𝐴
𝛿0.99 = 𝐹𝐷 =
√Re √𝑅𝑒𝐿
 The thickness and shear-stress distribution for both laminar and turbulent boundary layers can
be determined by an approximate method, using the momentum integral equation.
o First step: Determine momentum thickness using Eq. (1.24) and leave it in term of 𝛿.
𝛿
𝑢 𝑢
𝛿 ∗∗ = ∫ (1 − ) d𝑦
0 𝑈∞ 𝑈∞
o Second step: Determine shear stress using Eq. (1.30) and leave it in term of 𝛿.
𝜕𝑢
𝜏𝑤 = 𝜇 |
𝜕𝑦 𝑦=0
o Third step: Determine boundary layer thickness using Eq. (1.45b). Refer to the
discussed examples.
𝑑𝛿 ∗∗ 𝜏𝑤
=
𝑑𝑥 𝜌𝑈∞2
o Apply general equation to determine the flow properties.
𝜏𝑤
𝐶𝑑 =
Skin coefficient (local) 1 2
2 𝜌𝑈∞
𝐹𝐷
𝐶𝐷 =
Drag coefficient (total/overall) 1 2
2 𝜌𝑈∞ 𝐴
𝐿
1 2
Total/Drag force 𝐹𝐷 = 𝜌𝑈∞ 𝐴𝐶𝐷 or 𝐹𝐷 = 𝑏 ∫ 𝜏𝑤 d𝑥
2 0
 Since turbulent flow creates a larger shear stress on a surface, compared to laminar flow,
turbulent boundary layers create a larger friction drag on the surface.
 Boundary separation is due to adverse pressure gradient (increasing pressure). Beyond the
separation point within the boundary layer, the flow will begin to back up and move in the
opposite direction to the free-stream flow and ultimately formation of vortices is taking place.
 A series of these vortices or eddies produce a wake. They consume considerable mechanical
energy and may lead to a large pressure loss. Hence, it produces form drag.

Chapter 1 - Viscous Incompressible Flow (Rev200) (ANSWER)

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chapter 1 - Viscous Incompressible Flow (Rev200) (ANSWER)

Uploaded by

Copyright:

Available Formats

CHAPTER ONE [BTME2223 APPLIED FLUID MECHANICS]

Chapter One: Viscous Incompressible Flow

1.2 Navier-Stokes Equations

Figure 1.1c Net force in y-direction

Figure 1.1d Net force in z-direction

Net surface force in 𝑥-direction:

𝜕𝜎𝑥 𝜕𝜏𝑦𝑥 𝜕𝜏𝑧𝑥

Net surface force in 𝑦-direction:

𝜕𝜏𝑥𝑦 𝜕𝜎𝑦 𝜕𝜏𝑧𝑦

Net surface force in 𝑧-direction:

𝜕𝜏𝑥𝑧 𝜕𝜏𝑦𝑧 𝜕𝜎𝑧

𝜏𝑥𝑦 = 𝜏𝑦𝑥 ; 𝜏𝑥𝑧 = 𝜏𝑧𝑥 ; 𝜏𝑦𝑧 = 𝜏𝑧𝑦

Total force in 𝑥-direction:

Total force in 𝑦-direction:

Total force in 𝑧-direction:

The well-known Newton’s viscosity law is

Invoking the incompressible continuity equation into the Navier-Stokes equations

Boundary Condition: Since boundary conditions are so critical to

A degenerate form of the interface boundary condition occurs at the free

The plate moves in x-direction and hence, Eq. (1.12) is applied.

The equation can be simplified further based on the flow

Integrate the function,

The boundary conditions:

For zero net flow, 𝑄 = 0.

The effect of the viscous flow over the solid

Thus, both terms are of order 1, and either can be neglected.

The x-component of the Navier–Stokes equation is, neglecting gravity,

𝜕(𝑢∗ 𝑈∞ ) 𝜕(𝑢∗ 𝑈∞ ) 1 𝜕(𝑃∗ 𝜌𝑈∞

The y-component of the Navier–Stokes equations is

where 𝛿 ≪ 𝐿. Therefore, 𝑂(𝜀) ≪ 1⁄𝑂(𝜀) and the term 𝜕 2 𝑣 ∗ ⁄𝜕𝑥 ∗ 2 is neglected.

Compare the term

c. Momentum thickness, 𝛿 ∗∗ . It is defined as the distance by which the boundary should be

1.4 Blasius Flow over a Flat Plate

a. The velocity field solely within the boundary layer,

b. The boundary layer thickness, 𝛿0.99 ,

c. The shear stress distribution on the plate, and

d. The drag force on the plate.

The boundary conditions are

at 𝑦 = 0, 𝑢=𝑣=0 (Condition 1 & 2)

2𝑓 ′′′ (𝜂) + 𝑓(𝜂)𝑓 ′′ (𝜂) = 0 (1.29)

The boundary conditions are

η F F' 0.5(ηF' - F) F" η F F' 0.5(ηF' - F) F"

Blasius Boundary Layer

With the profile known, wall shear stress can be evaluated as

and the local skin friction coefficient is

Eqs (1.32) to (1.36) will be applied to solve the problem.

𝜌𝑈∞ 𝑥 5.0𝑥 1.7206𝑥 0.644𝑥 0.664 1.328

Distance, m Reynolds number 𝛿, mm 𝛿 ∗ , mm 𝛿 ∗∗ , mm 𝐶𝑑 𝐶𝐷

1.5 Momentum Integral Equation for Boundary Layer

We begin with the control volume sketched in Figure 1.6.

and the outgoing mass through the right face as

We now apply conservation of x-momentum for the chosen control volume.

⃗⃗⃗ Figure 1.7 Mass flow balance

The Bernoulli’s equation applied to the outer inviscid flow is given by

We previously defined displacement thickness 𝛿 ∗ in Eq. (1.23) and momentum thickness 𝛿 ∗∗ in

𝑦 = 𝛿, 𝑢 = 𝑈∞ --- 2nd Boundary Condition

These requirements will yield

1.02 × 10−3 × 𝜋 × 0.15

1.6 Turbulent Boundary Layer

i. Irregularity – The turbulence is irregular and random. Because of this, a

ii. Diffusivity – The turbulence is diffusive. In fact, this is a key feature of

iv. Three-dimensional – Turbulence is rotational and three-dimensional. Therefore,