CEE 262A

HYDRODYNAMICS

Lecture 3
Autumn 2007-2008
Kinematics of Fluid Motion
A. (a) Definitions, Conventions & Concepts
• Motion of fluid is typically described by velocity V

V  V ( x, y , z , t )
 
Dimensionality Steady or Unsteady
• Given above there are two frames of reference for describing
this motion
Lagrangian Eulerian
“moving reference system” “stationary reference system”
• • Focus on behavior of group of
Pathline particles at a particular point

• Focus on behavior of particular

particles over space of interest
Plan

River y
Flow
x

Measuring
Station
X
V
5
unsteady
3 V3
steady

t3 t t3 t
Lagrangian position X(t) Eulerian Velocity V(t)
 v
v Streamlines
Steady
Flow v
y

• Individual particles must travel on paths whose

tangent is always in direction of the fluid velocity at
the point.

 In steady flows (Lagrangian) path lines are the same

as (Eulerian) streamlines.
(b) Pathlines, Streaklines & Streamlines
nozzle
t0
a b c d e
nozzle

t1 a b c d

(a) Pathline: Line joining positions of particle “a” at successive times

Streaklines: Line joining all particles (a, b, c, d, e) at a particular
instant of time
Sreamlines: Trajectories that at an instant of time are tangent to the
direction of flow at each and every point in the flowfield
(c) Streamtubes

• No flow can pass through a streamline because velocity

is always tangent to the line.
• Concept of streamlines being “solid” surfaces forming
“tubes” of flow and isolating “tubes of flow” from one
another.
B. Some Formalism
(a) Lagrangian vs. Eulerian
X2
t1
t0

x x particle path
~0 ~

X1
* Following individual particle as it moves along path…
At t = t0 position vector x
~0

Any flow variable can be expressed as F ( x , t )

~0
Particle position can be expressed as x( x , t )
~ ~0

Lagrangian
* Concentrating on what happens at spatial point x
~

Any flow variable can be expressed as F ( x, t )

~

Eulerian
Lagrangian description: xi
ui 
t
U i  2 xi
ai   2
t t
This only describes local change at point xin Eulerian description!
~

Need material derivative “translates” Lagrangian

concept to Eulerian Language.
(b) Material Derivative (Substantial or Particle)
Consider F ( x, y, z , t ) ; x~  ( x, y, z )
• As particle moves distance d x in time dt
~
F F
dF  dt  dxi -- (1)
t xi
• If increments are associated with following a specific particle
whose velocity components are such that
dxi  ui dt -- (2)
Substitute (2) (1) and dt
dF F F
  ui -- (3)
dt t xi

DF F F
  ui
Dt t xi

Local rate of change Advective change

at a point x past x
~ ~
(i) Vector Notation:
DF F
Dt

t
 u  F
~
 
(ii) ESN: e.g. if F  a j
Da j a j a j
  ui
Dt t xi

(iii) Along ‘Streamlines’:

DF F F
n  q
Dt t s
s

Magnitude of u
~
(c) Streamlines, Pathline & Streakline
(i) Streamline
No flux By definition: (i) U  ds  0
U  (u , v, w) ~
~ 1 2
ds  (dx, dy, dz )  ( wdy  vdz ) a  (udz  wdx) a
~ ~
~ ds 3 1 2 3
 (vdx  udy ) a  0 a  0 a  0 a
s ~ ~ ~ ~
c
 wdy  vdz; udz  wdx; vdx  udy
dx dy dz
 
u v w
(ii) Pathline dx dy dz
u ; v ; w
dt dt dt
t1 x1 dx t1 y1 dy t1 dzz1
  dt   ;  dt   ; t0 dt  z0 u
t0 x0 u t0 y0 v
 Example 1: Stagnation point flow
2

1.5

0.5

x3 0
x2

-0.5

-1

-1.5

-2
-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
x1

u1  x1 u3   x3
 u1  a  t  x1
First, let’s examine a simple flow:
u3  a  t  x3
(a) Calculate streamlines
dx1 dx3 dx3 u3 a  t  x3 x3
    
u1 u3 dx1 u1 a  t  x1 x1
dx3 x
  3 (an ode in x3 and x1 )
dx1 x1
or...
dx1 dx3 dx1 dx3 dx1 dx3
     
u1 u3 ax1  ax3 ax1 ax3
a ln  x1    a ln  x3   a ln  C  Cleverly chosen
integration constant
C
x1   x1 x3  C  streamlines are hyperbolae
x3
 Stagnation pt flow with a=1
10

0
x3

-2

-4

-6

-8

-10
-10 -5 0 5 10
x1
(b) Calculate pathlines

dx1 dx1
 u1  ax1   adt which can be integrated
dt x1
x1 ( t )
dx1
t
 x1 (t ) 
   adt  ln    at or x1 (t )  x1 (0) exp  at 
x
x1 (0) 1 0  x1 (0) 
dx3
By a similar argument using  u3 , we find that
dt
x3 (t )  x3 (0) exp   at 
But, despite all, x1 x3  x1 (0) x3 (0)  C '
 pathlines are (also) hyperbolae
Example 2: a (more complicated) velocity field: in a surface gravity wave:
Velocity vectors Streamlines
1 1
0 0
-1 -1
-2 -2
0 2 4 6 8 10 0 2 4 6 8 10

Pathline Streakline
1 1
0 0
-1 -1
-2 -2
0 2 4 6 8 10 0 2 4 6 8 10

Stream/streak/path lines are completely different.

 C. Relative Motion near a Point
(1) Basic Motions
(a) Translation
X2 X2

X1 X1
t0 t1

(b) Rotation
X2
• No change in dimensions
of control volume

X1
t1
(c) Straining (need for stress):
Linear (Dilatation) – Volumetric Expansion/Contraction
X2 X2

t0 X1 t1 X1
(d) Angular Straining – No volume change
X2

t1 X1
Note: All motion except pure translation involves
relative motion of points in a fluid
General motion of two points:

u du P’
x2 P ~ ~ x2 '
x
x  x x
~
~ ~ ~
u O’
O ~
x
~

t0 x1 t1 x1

Consider two such points in a flow, O with velocity u ( x, t0 )

~ ~

And P with velocity {u~  d u~ }( x~  ~x, t0 ) moving to O’ and P’

respectively in time t
Therefore, after time t O() means “order of” =
O’ x~  u~ ( x~ , t0 )t “proportional to”
: ( x  x)  u ( x  x, t0 )t
~ ~ ~ ~ ~
P’:
 ( x  x)  {u ( x, t0 )  x  u  O(|| x ||2 )}t
~ ~
 ~ ~    ~  ~   ~  
Taylor series expansion of u~ ( x~  ~x, t0 )

x  ( x  x)  u ( x, t0 )t  x  u t  O(|| x ||2 )t  x  u ( x, t0 )t

'
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~

x  x  x  u t
'
~ ~ ~ ~
to first order
ui
 Change in separation  x  x  x  u t  x j
'
t -- (A)
~ ~ ~ ~ x j

 Relative motion of two points depends on the velocity

ui
gradient, , a 2nd-order tensor.
x j
(2) Decomposition of Motion
“…Any tensor can be represented as the sum of a symmetric
part and an anti-symmetric part…”
ui 1  ui u j  1  ui u j 
   
x j 2  x j xi  2  x j xi 
  

 eij  rij
={ rate of strain tensor} + {rate of rotation tensor}
 u1 u1 u2 u1 u3 
 2   
x1 x2 x1 x3 x1  Note:

1  u 2 u1 u2 u2 u3  (i) Symmetry about
eij   2 
2  x1 x2 x2 x3 x2  diagonal
 u u u3 u2 u3 
 3 1  2 (ii) 6 unique terms
 x1 x3 x2 x3 x3 

Linear & angular straining

  u1 u 2 u1 u3  Note:
 0   
 x2 x1 x3 x1  (i) Anti-symmetry about
1  u 2 u1 u 2 u3  diagonal
rij   0 

2 x1 x2 x3 x2 
 u u u3 u 2  (ii) 3 unique terms (r12, r13, r23)
 3 1  0 
 x1 x3 x2 x3 

Rotation
Terms in rij   12 ijk (  u~ ) k   12 ijk k

 0 3 2  e.g . r12   12 1233   12 3

1
rij   3 0 1  r32   12 3211   12 (1 )  12 1
2
 2 1 0 
Let’s check this assertion about rij

um
ijk k  ijk  klm
xl
um
  jik  klm
xl
The recipe:
um
 im  jl  il  jm  (a) m = i and l =j
xl
(b) l = i and m = j
 ui u j 
   gives
 x j xi 
 2rij
 ui
x  x   x j t
'
Interpretation ~ ~ x j
'
x  x ui
ui  ~ ~
 x j
t x j

ui  x j (eij  rij )

1
 eij x j  ijk k x j
2
1
 eij x j  ikj k x j
2
1
ui  eij x j  ( x)i
2 ~ 

Relative velocity due Relative velocity due

to deformation of to rotation of
fluid element 
element at rate 1/2

 x3 Consider solid body rotation about x2
axis with angular velocity 
 vt
  vt   a 2  
   
x1

General result:
 = 2 x {Local rotation rate of fluid elements)


x3
(a) Simple examples: u3(x1) u1(x3)

u   a  t  x3 , 0, a  t  x1 

u1 u3 x1
2    2a
x3 x1
e0

 (b) Conversely, the flow

u   a  t  x1 , 0, a  t  x3 
x3

r 0

a 0 0
e 0 0 0 x1

0 0 a

What happens to t0
the box?
t1
t2

It is flattened and stretched

(3) Types of motion and deformation .
(i) Pure Translation
u1t
X2
t1 x2
x1 t1  t0  t
u2t t0  x 2
x1
X1
(ii) Linear Deformation - Dilatation
X2 x1' t1  t0  t
a u2
t1 x ’ a x2 t
2
x2
x2 t0
u1
x1 b
b x1t
X1
x1
 In 2D - Original area at t0  x1x2
 x1 x2  (x1  a )(x2  b)
' '
- New area at t1
' ' u1 u
x1 x2  (x1x2  x1tx2  2 x2tx1  O(t ) 2  ...)
x1 x2
u1 u2
 x1x2 [(1  (  )t )  O(t 2 )]
x1 x2
x1'x2 '  x1x2 Area Strain
 Area Strain = and Strain Rate =
t
x1x2
1 dA 1 A 1 A  u1 u2 
 lim and      O(t )
A0 dt t 0 A0 t A0 t  x1 x2 

1 dA u1 u2
  
A0 dt x1 x2
 In 3D
1 dV u1 u2 u3 ui
     u 
V0 dt x1 x2 x3 ~ xi

* Diagonal terms of eij are responsible for dilatation

 
In incompressible flow,   U  0 ( U is the velocity)

Thus (for incompressible flows),

(a) in 2D, areas are preserved
(b) in 3D, volumes are preserved
 (iii) Shear Strain – Angular Deformation

x2
t1  t0  t
u1 u1
u1  x2 x2t
x2 A x2
A
X2
 x 2 t0 d t1
u 2
d B x1t
u1 B x1
O x1 x1 O X1
u 2
u2 u2  x1
x1
 Shear Strain Rate  Rate of decrease of the angle formed by 2
mutually perpendicular lines on an element
d  d 1  1 u1 1 u2 
   ( x2t )  ( x1t )
Iff d , d small t t  x2 x2 x1 x1 
u1 u2
 
x2 x1

1  u1 u2 
 Average Strain Rate      eij
2  x2 x1 
i j

The off diagonal terms of eij are responsible for angular

strain.
 (iv) Rotation

x2
t1  t0  t
u1
 x2t
A x2
u1
 (u1  x 2 )
x2
A B
x 2 t0 d t1
u 2
x1t
 u1
d x1
O B x1
x1 O
u2 u 2
u2  x1
x1
 u2 u2  1  u2 
Rotation due to : d  t    x1t  
x1 x1  x1  x1 
u1 u1  1  u1 
due to x : d  t    x2t  
2 x2  x2  x2 

 Average Rotation Rate (due to superposition of 2 notions)

1  d  d  1  u2 u1 
        rij i j
2  t  2  x1 x2 
Summary

1. Relative motion near a point is caused by

ui
x j
 u
~
 
2. This tensor can be decomposed into a symmetric and
an anti-symmetric part.
(a) Symmetric eij
 
* i  j : Dilation of a fluid volume   u
~

* i  j : Angular straining or shear straining

(b) Anti-symmetric rij
* i j: 0
* i  j : Rotation of an element