POTENTIAL FLOWS

• An IDEAL FLOW has a non-zero tangential velocity at a solid surface

• A REAL FLUID with a non-zero viscosity must satisfy a no-slip boundary condition

• The viscous effects in a real flow are confined to thin layers close to solid surfaces if
the fluid viscosity is small

• The viscous layers are thin not just when the viscosity is small, but when Re >> 1.
The thickness of such boundary layers within which the viscous diffusion of
voriticity is important approaches zero as Re  

Ideal flow with  = 0 Flow at high Re

Comparison of a completely irrotational flow and a high Reynolds number flow
• Fluid elements starting from rest, or from any other irrotational region, remain
irrotational unless they move into these boundary layers.

• The flow field can be divided into regions namely

• Outer region – Flow is inviscid and irrotational
• Inner region – viscous diffusion of vorticity is important

• The outer flow can be approximately predicted by ignoring the existence of thin
boundary layer and applying the irrotational flow theory around the solid object.

• Once the outer problem is solved, viscous flow equations within the boundary
layer can be solved and matched to the outer solution.
 HIGHLY SEPARATING FLOWS
• An important exception to this classification of inviscid and viscid region is where
the solid object has sucha shape that the boundary layer spearates from the
surface, giving rise to eddies in the wake.

• In this case, viscous effects are not confined to thin layers around solid surfaces,
and the real flow in the limit Re   is quite different from the ideal flow ( = 0)

• Ahead of the point of separation, however, irrotational flow theory is still a good
approximation of the real flow

Upstream of the point of separation, irrotational flow theory is a good approximation of the
real flow
Circulation Theorem
Consider a closed curve C placed in a moving fluid.
The circulation around the contour C is defined as the summation of the velocity
components tangential to C times the elemental length of the contour dr at every
point on the contour.
Let the velocity field be q at any arbitrary point and dr be the directed element of
the arc at the same point, then the circulation  is given by

   q .dr   q cos  dr 
C C
  udx  vdy  wdz 
C

The integral symbol denotes the line integral

around a closed curve in a counterclockwise
direction.
The concept of circulation is similar to the work
done in a vector force field discussed in solid
mechanics. In a vector force field, the line
integral
 F .dr
C

Represents the work done in moving around the

Stokes Theorem
Let A be a surface bounded by the contour C and n be the unit vector normal to this
surface at point P.
Stokes theorem may then be expressed as

   q .dr    .n dA
C A

The importance of the equation above is that it relates

the circulation along a closed circuit C to the vorticity
within the surface bounded by that line.

In order to prove the above equation, let us divide the

space A into a very large number of infinitesimally
small triangles A. If the lines are drawn from the
vertices of these triangles parallel to the x,y and z axes,
a series of infinitesimally small tetrahedrons is formed.
Elementary tetrahedrons, PABC with edges PA, PB and PC equal to dx, dy and dz respectively
The velocity elementary tetrahedrons, PABC with edges PA, PB and PC equal to dx, dy and dz
respectively

dx dy dy dz dz dx
AB   ; BC   ; CA  
cos  sin  cos  sin  cos  sin 

The velocity components at the midpoints of AB, BC and CA which contribute to the
circulation have the following values (they are calculated from the velocity components at the
vertices P, A, B and C)
1 u 1 u 1 u 1 u 1 v 1 v
u xy  u  dx  dy u xz  u  dx  dz v yx  v  dy  dx
2 x 2 y 2 x 2 z 2 y 2 x
1 v 1 v 1 w 1 w 1 w 1 w
v yz  v  dy  dz w zx  w  dz  dx w zy  w  dz  dy
2 y 2 z 2 z 2 x 2 z 2 y

Accordingly, the circulation along the sides of the triangle ABCA is given by (circulation is
considered positive if it rotates according to the right handed screw rule with normal outward)
   u xy cos   AB   v yx sin   AB   v yz cos   BC   w zy sin   BC   w zx cos   CA  u xz sin   CA
d   u xy cos   AB   v yx sin   AB   v yz cos   BC   w zy sin   BC   w zx cos   CA  u xz sin   CA

dx dy dy dz dz dx
AB   ; BC   ; CA  
cos  sin  cos  sin  cos  sin 

 1 u 1 u  dx  1 v 1 v  dy  1 v 1 v  dy
d   u  dx  dy  cos    v  dy  dx  sin    v  dy  dz  cos  
 2  x 2  y  cos   2  y 2  x  sin   2  y 2  z  cos 
 1 w 1 w  dz  1 w 1 w  dz  1 u 1 u  dx
 w  dz  dy  sin  w dz  dx  cos  u dx  dz  sin 
 2 z 2 y  sin   2 z 2 x  cos   2 x 2 z  sin 

 1 u 1 u   1 v 1 v   1 v 1 v 
d    u  dx  dy dx   v  dy  dx  dy   v  dy  dz  dy 
 2 x 2 y   2 y 2 x   2 y 2 z 
 1 w 1 w   1 w 1 w   1 u 1 u 
 w  dz  dy dz   w  dz  dx dz   u  dx  dz dx
 2  z 2  y   2  z 2  x   2  x 2  z 

1 u 1 v 1 v 1 w 1 w 1 u
d   dxdy  dx dy  dydz  dydz  dx dz  dz dx
2 y 2 x 2 z 2 y 2 x 2 z

 w  v   u w   v u 
2 d    dydz    dz dx    
 y z   z x    x y 
  
 dydz
2 .n dA  2 i x  j y  k z .  i j
dzdx
k
dxdy 

 2 2 2 
i j k
1  w v  1   v u 

1
  V   1   
 i 
1  w u 
   j     k   
2 2 x y z 2   y z  2  x z  2  x y 
u v w

1w v 1u w 1 v u

 x      y      z    
2  y z 2 z x 2 x  y

w v u w  v u

2 .n dA  
  
 
dydz    
 
dzdx     dxdy
 y z z x  x  y
  w v   u w    v u 
2d    dydz    dz dx    
 y z    z x   x y 
 .n dA  d
In the same manner, the circulation around all other elementary
triangles, dA, can be formed for the entire surface A, and since the    q .dr    .n dA
circulation along the elementary sides common to two triangles cancels,
C A
the remaining circulation will be that around the closed contour C.
Hence, the Stokes theorem is verified.
Kelvin’s Theorem (Constancy of circulation)
In an ideal homogenous fluid, the circulation is independent of time along a closed
fluid line when the external forces are conservative and are derived from a single
valued potential.
   q .dr
C

D D D D
  q .dr    q .dr 
Dt Dt C Dt C Dt

D
 q .dr   dr . Dq  q .dq
Dt Dt
q Dq P
Equation of motion is  q .q    f 
t Dt 

Dq P dp
dr .   dr . f  dr .   d f 
Dt  
D
 q .dr    d f  dp  q .dq
Dt 
 D D
  q .dr 
Dt C Dt

D dp
 q .dr   d f   q .dq
Dt 

D D dp
  q .dr     d f   q .dq
Dt C Dt C 
D  p q2 
   f   
Dt   2 c

It, as assumed, f, p and q are single-valued functions of r, then the right hand side of the
above equation is equal to zero. Above equation is the rate of change of circulation along a
closed curve and Kelvin’s theorem has been proved
This theorem is valid for any kind of flow of an inviscid fluid as long as forces are conservative.

If for every value of x in the interval, y = f(x) has only one value, y is a single valued function of
x in the interval; if y has more than one value for each value of x, it is a multiple valued
function of x
Circulation  of a circulatory flow (vr = v = 0) with constant vorticity (z = constant)
from both the circulation concept and Stokes theorem
The circulatory flow with constant vorticity is shown in fig below

v v
  cons tan t
r r
From continuity equation,
1 v
0
r r  1   v r  1   v  v r

t r r

r 

r
0

v
 cons tan t  v  f  r 
r
   q .dr    iv r  jrv   kv z  . idr  jd  kdz 
C C
2
   q .dr   rv d  2 r v
C 0
     .n dA
A

2 r dv v
     z r dr d z  
0 0 dr r

 dv v 
2 r
     r dr d
0 0  dr r 

 dv
 v r   dr d
2 r
  
0 0  dr 
2
   v r d  2 rv
0

Hence, the Stokes theorem is satisfied


v 
2 r
The motion in concentric circles is shown in figure. In this case, we consider a flow for which
an integral portion of the fluid bounded by a circle (the core) is undergoing rigid body motion
with constant angular velocity z while a concentric outer portion of the fluid represents a
vortex free circulatory flow. Calculate the tangential velocity in both the regions.

dv v
z  
dr r

v  f  r 

In the outer region

dv v dv dr v
z    0   0  ln v  ln r  ln C  C
dr r v r r
C
v 
r
This represents a vortex-free circulatory flow.
In the inner region
dv v
z    A1 A1  Cons tan t
dr r

1 d  d A1 A1 2
  
v r   A1   
v r   v r  r C
r  dr  dr r 2

v  0 r  0  C  0

A1 2
v r  r
2

v  Ar
This represents either a rotating solid body or a fluid rotating like a solid body in the nucleus.
This concept is employed in the two dimensional airfoil theory where in vortex-free flow exists
outside the boundary of a solid body such as an airfoil.
Velocity potential – Irrotational flow
The movement of ideal fluids is, in general, of two types, each differing from the other
physically and mathematically.
• Rotational flows
• Irrotational flows

   q .dr    .n dA
C A

  q  0    0
Let us consider a closed contour C with circuit ABDEA in the fluid. The circuit ABDEA in this
case must lie in a simply connected region. This means that the closed contour must include
only fluid and can be contracted to a point of the region without ever passing out of the
region.

ABDEA
 q .dr   q .dr   q .dr  0
ABD DEA

 q .dr    q .dr   q .dr   q .dr

ABD DEA ABD AED
D x , y ,z  D x , y ,z 

 q .dr   q .dr   q .dr   d  

ABD AED A A
D  A
 LINEAR MOTION AND DEFORMATION

Element at t0 Element at t0+δt

= + + +

General Translation Linear Angular

motion Rotation
deformation deformation
 TRANSLATION

O’
v vδt
u
O

uδt

If all points in the element have the same velocity which is

only true if there are no velocity gradients, then the
element will simply TRANSLATE from one position to
another.
 LINEAR DEFORMATION
u
B u C u x
x B C C’

δy δy

u
u x
u x

O δx A O A A’
δx
 u 
 x  x   t
 

Because of the presence of velocity gradients, the element will

generally be deformed and rotated as it moves. For example,
u
consider the effect of a single velocity gradient
x
On a small cube having sides  x ,  y and  z
x component of velocity of O and B = u
u
x component of velocity of A and C = u  x
x
This difference in the velocity causes a “STRETCHING” of the volume
element by a volume
 u 
  x    y  z    t 
x 

Rate at which the volume V is changing per unit volume due

u
the gradient
x
1 d   V   u
 
x
 t
 u
 Lim 
V dt  t 0 t  x
 
 v w
If the velocity gradients & are also present
y z

1 d  V  u v w
  
V dt x y z

This rate of change of volume per unit volume is called the

VOLUMETRIC DILATION RATE
Volume of the fluid may change as the element moves from one
location to another in the flow field
Incompressible fluid – volumetric dilation rate = zero
Change in volume element = zero; fluid density = constant
(The element mass is conserved)
 Variations in the velocity in the direction of velocity cause
LINEAR DEFORMATION
u v w
, &
x y z
Linear deformation of the element does not change the shape of
the element
Cross derivates cause the element to ROTATE and undergo
ANGULAR DEFORMATION
u v
,
y x
Angular deformation of the element changes the shape of the
element
 ANGULAR MOTION AND DEFORMATION
 u 
  y  t
 y 
u
u y
y
B C B B’ C


δy δy A’
v  v 
v
v x  x  x   t
  
u x
O δx A O δx A

Consider x-y plane. In a short time interval t line segment OA and

OB will rotate through angles  and  to the new positions OA’
and OB’
Angular velocity of OA, OA

 oA  Lim
 t 0  t

For small angles

 v  x  t   v  t  oA  Lim 
 
 v
x
t 
 v

Tan     x 
 t  0  t  x
x x 

 v  x  - positive  oA - counterclockwise

 oB  Lim
 t 0  t
 u   y  t   u  
 y    y   t 
  u  oB  Lim      u
Tan      t
 y y  t  0  t  y

 

 u 
 y 
  - negative  oB - clockwise
Rotation z of the element about the z-axis is defined as the average
of the angular velocities oA and oB of the two mutually
perpendicular lines OA and OB. Thus, if counterclockwise rotation is
considered positive, it follows that
1 v u
 z    
2 x  y

Rotation x of the element about the x-axis

1  w v 
x    
2 y z

Rotation yof the element about the y-axis

1u w
y    
2z x

   x î   y ĵ   z k̂
 1 1
  curl V    V
2 2
Vorticity  is defined as the vector that is twice the rotation vector

  2   V
Fluid element will rotate about the z axis as an undeformed block
u v
(ie., oA = - oB ) only when  y  
x
Otherwise, the
rotation will be associated with an angular deformation
v u
  Rotation around the z axis is zero.
x y
Rotation and vorticity are zero;
 V  0
FLOW FIELD IS IRROTATIONAL
 u v
In addition to rotation associated with derivatives &
y x
These derivatives can cause the fluid element to undergo an angular
deformation which results in change of shape
Change in the original right angle formed by the lines OA and
OB is SHEARING STRAIN 
 =  + 
 is positive if the original right angle is decreasing
Rate of Shearing Strain or Rate of Angular Deformation

  
 v
 x t  
 u  
 y 
 

t
 u v
  Lim    Lim    
t  0  t  t  0  t  y x
 
 u v
  
y x
Rate of angular deformation is related to a corresponding shearing
stress which causes the fluid element to change in shape
u v

y x

Rate of angular deformation is zero;

Element is simply rotating as an
Rotation undeformed block
 IRROTATIONAL FLOWS
 V  0
i j k
     w v    w u   v u 

 i 
   j    k   
x y z  y  z    x  z   x y 
u v w

1 w v  w v
 x     
2 y z y z
1u w u w
y     
2z x z x
1 v u v u
z     
2 x  y x y
A general flow field would not satisfy these three equations. However, a uniform flow is an
example of irrotational flow
Uniform flow in the x-direction

w v
 0
y z
u w
 0
z x
v u
 0
x y

A uniform flow field ( in which there are no velocity gradients) is

certainly an example of irrotational flow
VARIOUS REGIONS OF FLOW
 Velocity potential: Laplace equation
For an irrotational flow, the velocity gradients are related through the following
equations
w v u w v u
 0  0  0
 y z z x x y
The velocity components can be expressed in terms of a scalar function  (x, y, z) as

  
u v w or V  
x y z
The scalar function  (x, y, z) is called the velocity potential
w v  2  2 This verifies that a velocity field defined is
  
 y z z y y z indeed irrotational.
u w  2  2
   For an irrotational flow, the velocity is
z x z x x z expressible as the gradient of a scalar function 
v u  2  2
  
x y  x  y  y x
For an incompressible fluid, conservation of mass states that

 .V  0
For an incompressible and irrotational flow

V  

 0
2

2    . V Laplacian operator

In cartesian co-ordinates
 2  2  2
 2  2 0
x 2
y z
Inviscid, Incompressible and Irrotational flows are governed by Laplace equation. This type of
flow is called a potential flow. To complete the mathematical formulation of a given problem,
boundary conditions have to be specified. These are usually velocities specified on the
boundaires of the flow field of interest. It follows that if the potential function be determined,
then the velocity at all points in the flow field can be determined and the pressure at all points
can be determined from the Bernoulli’s equation.
 BASIC PLANE POTENTIAL FLOWS
A major advantage of Laplace’s equation is that it is a linear partial differential equation.

Since, it is linear, various solutions can be added to obtain other solutions –

If 1 (x, y, z) and 2 (x, y, z) are two solutions to Laplace’s equation, then 3 = 1 + 2

The practical implication of this result is that if we have certain basic solutions we can
combine them to obtain more complicated and interesting soutions.
Plane flows are considered Two dimensional flows
Cartesian co-ordinates
 
u v
x y
Cylindrical co-ordinates
 1 
vr  v 
r r 
Stream function for Plane flows
 
u v
y x
1  
vr  v  
r  r
Stream function is defined such that the conservation of mass is satisfied
For an irrotational flow
u v

 y x
In terms of stream function
       
    
y  y  x  x 

 2  2
 2 0
x 2
y

Thus, for a plane irrotational flow we can use either the velocity potential or the stream
function – both must satisfy Laplace’s equation in two dimensions.
 Relation between stream function and potential function
The change in  as we move from one point (x, y) to a nearby point (x + dx, y + dy) is given by
 
d  dx  dy
x y

Along a constant line of  (equipotential line), d = 0

0  u dx  v dy
dy u

dx v
For a streamline
V  dr  0

i j k
dy v
u v w  udy  vdx  0  
dx dy dz dx u
Lines of constant  (equipotential line) are orthogonal to lines of constant  (streamlines) at
all points where they intersect. Two lines are orthogonal if the product of their slopes is -1.
A flow net consists of a family of streamlines and equipotential lines.
A flownet is useful in visualising flow patterns and can be used to obtain graphical solutions by
sketching in streamlines and equipotential lines and adjusting the lines until the lines are
approximately orthogonal all points where they intersect.

Velocities can be estimated from the flow

net, since the velocity is inversely
proportional to the streamline spacing.

Velocity near the inside corner will be

higher than the velocity along the outer
part of the bend.

Flownet through a 90 bend

UNIFORM FLOW - streamlines are straight and parallel, and the magnitude of the
velocity is constant
uU v0

u    ux  C
x
C is an arbitrary constant which can be set equal to zero
  Ux
 
U 0    Uy
y x

Uniform flow can be simply described by either a stream function or a velocity potential.
SOURCE AND SINK – represents a purely radial flow
Consider a fluid flowing radially outward from a line through the origin perpendicular to the x-
y plane.
m - Volume rate of flow emanating from the line (per unit length)
By conservation of mass
 2 r  v r  m
m
vr 
2 r
Since the flow is purely radial flow, v = 0, the
corresponding velocity potential can be obtained by
integrating the equations
 m 1 
vr   v  0
r 2 r r 
m
 ln r
2

If m is positive, the flow is radially outward, and the flow is considered to be a source flow.
If m is negative, the flow is toward the origin, and the flow is considered to be a sink flow.
The flowrate, m, is the strength of the source or sink.
We note that at the origin where the velocity becomes infinite, which is of course
physically impossible.

Thus, sources and sinks do not really exist in real flow fields, and the line representing
the source or sink is a mathematical singularity in the flow field.

However, some real flows can be approximated at points away from the origin by using
sources or sinks.

Also, the velocity potential representing this hypothetical flow can be combined with
other basic velocity potentials to describe approximately some real flow fields.

The streamfunction for the source can be obtained by integrating the relationships
1  m 
vr   ; v   0
r  2  r r

m
 
2
It is apparent that the streamlines (lines of  = constant) are radial lines, and the equipotential
lines (lines of  = constant) are concentric circles centered at the origin.
Cylindrical co-ordinates

 ds  2   dz  2   rd  2   dr  2
 ds  2   h1dx1  2   h2dx2  2   h3dx3  2

h1  1 , h2  r , h3  1
dx1  dz , dx 2  d , dx 3  dr
Spherical co-ordinates

 ds  2   d  2    d  2    sin  d  2
 ds  2   h1dx1  2   h2dx2  2   h3dx3  2

h1  1 , h2   , h3   sin 
dx1  d , dx 2  d , dx 3  d
 i1 ri 2 i3
1   
 V 
r z  r
vz rv vr

1  v r   rv   ri 2  v r v z     rv  v z 
  V  i1        i3   
r   r  r  z r   z  
 vz  0
v  v  r 
vr  0

1  v r   rv   ri 2  v r v z     rv  v z 
  V  i1        i3    0
r   r  r  z r   z  

1 d
 rv   0
r dr
r dv v
 0
r dr r

dv v
 0
dr r
 dv v
 0
dr r
dv dr
 0
v r
Integrating

ln v  ln r  ln C1

C
v 
r