Chapter 3.

Fluid kinematics
* For now we are not concerned with the
actual forces producing the fluid motion.
* We shall discuss velocity and acceleration
fields, as well as flow field description and
visualization.
The velocity field [See Video 4.1]
• Recall that the velocity vector and density
are related to molecular mass and velocity
by
P
miVi
V(r, t) = lim P ,
δV →ǫ mi
P
mi
ρ(r, t) = lim .
δV →ǫ δV
* Note that 1/n ≪ ǫ ≪ L3, where n and L
are the number density of the molecules
and the smallest significant (macroscopic)
length scale of the flow, respectively.

Fluid Mechanics, Prof. T.-S. Yang 3-1

 • Fluids in the news: Follow those particles
* Superimpose two photographs of a
bouncing ball taken a short time apart
and draw an arrow between the two images
of the ball.
* This arrow represents an approximation of
the velocity (displacement/time) of the ball.
* The particle image velocimeter (PIV) uses
this technique to provide the instantaneous
velocity field for a given cross section of a
flow.
* The flow being studied is seeded with
numerous micron-sized particles which are
small enough to follow the flow yet big
enough to reflect enough light to be
captured by the camera.
* The flow is illuminated with a light sheet
from a double-pulsed laser.
* A digital camera captures both light pulses
on the same image frame, allowing the
movement of the particles to be tracked.

Fluid Mechanics, Prof. T.-S. Yang 3-2

 * By using appropriate computer software to
carry out a pixel-by-pixel interrogation of
the double image, it is possible to track
the motion of the particles and determine
the two components of velocity in the given
cross section of the flow.
* By using two cameras in a stereoscopic
arrangement it is possible to determine all
three components of velocity.

† See Problem 4.62.

Fluid Mechanics, Prof. T.-S. Yang 3-3

Eulerian and Lagrangian descriptions
1) Eulerian viewpoint:
• The independent variables are the spatial
coordinates (x, y, z) and time t, i.e.,
attention is focused on the fluid passing
through a control volume (CV) fixed in
space.
• The CV generally consists of different “fluid
particles” at different instants.
2) Lagrangian approach:
• Attention is focused on a material volume
(MV), i.e., a CV always consisting of the
same fluid particles.
• The independent variables therefore are t
and the coordinates (x0, y0, z0) that specify
a fluid particle at some reference instant t0.
§ The two descriptions are related, of course,
and the choice of which one to use is largely
a matter of taste.
§ However, it perhaps is more natural to
apply conservation laws to a (Lagrangian)
MV, while the Eulerian system is preferred
for solving most problems.

Fluid Mechanics, Prof. T.-S. Yang 3-4

Dimensionality
• Three dimensional (3D) in general (or, in
reality): V = V(r, t).
* In Cartesian (rectangular) coordinates, we
usually write

V = u(x, y, z, t) i+v(x, y, z, t) j+w(x, y, z, t) k.

* e.g., flow past an airfoil of finite span.

Fluid Mechanics, Prof. T.-S. Yang 3-5

 • Oftentimes the flow is approximately 2D:

V = u(x, y, t) i + v(x, y, t) j.
* e.g., flow past an airfoil of ‘infinite span’.

• Sometimes the flow is further simplified to

1D: V = u(x, t) i.

Fluid Mechanics, Prof. T.-S. Yang 3-6

Steady and unsteady flows
* Eulerian description
1) Steady: ∂ V/∂t = 0
* Must be laminar flow
2) Unsteady: ∂ V/∂t 6= 0
* Clearly much more difficult to analyze
* The flow may be periodic or aperiodic
(laminar) or random (turbulent).

† Video 4.2: Flow past a wing

† Video 4.3: Flow types

• Fluids in the news:

New pulsed liquid scalpel
* High-speed liquid-jet cutters are used for
cutting a wide variety of materials such
as leather goods, jigsaw puzzles, plastic,
ceramic, and metal.
* Typically, compressed air is used to produce
a continuous steam of water that is ejected
from a tiny nozzle.

Fluid Mechanics, Prof. T.-S. Yang 3-7

 * As this stream impacts the material to
be cut, a high pressure (the stagnation
pressure) is produced on the surface of the
material, thereby cutting the material.
* Such liquid-jet cutters work well in air, but
are difficult to control if the jet must pass
through a liquid as often happens in surgery.
* Researchers have developed a new pulsed
jet cutting tool that may allow surgeons to
perform microsurgery on tissues that are
immersed in water.
* Rather than using a steady water jet, the
system uses unsteady flow.
* A high-energy electrical discharge inside the
nozzle momentarily raises the temperature
of the microjet to approximately 10,000◦ C.
* This creates a rapidly expanding vapor
bubble in the nozzle and expels a tiny fluid
jet from the nozzle.
* Each electrical discharge creates a single,
brief jet, which makes a small cut in the
material.

Fluid Mechanics, Prof. T.-S. Yang 3-8

Streamlines, streaklines, and pathlines
1) Streamlines:
• Lines whose tangents are everywhere
parallel to the fluid velocity vector

V = (u, v, w).
• Accordingly,
dx dy dz
= = ≡ ds,
u v w
where s is zero at some reference point in
space, and increases along the streamline
(a parameter used to describe the curve).
* Here s has the dimension of time.
* In a steady flow (or a snapshot of an
unsteady flow), ds = dS/V , where dS is
differential arc-length along the streamline
and V = |V|.
* One may then think of ds as the time taken
for a fluid particle (or a fictitious fluid
particle in an unsteady flow) to travel a
distance dS along the streamline at the
local speed V .

Fluid Mechanics, Prof. T.-S. Yang 3-9

 • Equivalently,
dxi
= ui(xi; t) (i = 1, 2, 3),
ds
with t fixed.
* In the case of an unsteady flow (where the
velocity components are time dependent)
the instantaneous streamlines for a fixed
value of t are considered.

• Initial condition:

x = x0 , y = y0 , z = z0 (s = 0),
which picks out the streamline passing
through the point (x0, y0, z0).

• Symbolically, the solution is

xi = xi(s; x0, y0, z0; t).

* As s takes on all real values, this traces out
the required streamline.

Fluid Mechanics, Prof. T.-S. Yang 3-10

2. Pathlines:
• A pathline is a line traced out in time by a
given fluid particle as it flows.
• So,
dxi
= ui(xi, t) (i = 1, 2, 3).
dt
* Note that t now is the independent variable.

• Initial condition:

x = x0 , y = y0 , z = z0 (t = 0),
which specifies the initial position of the
particle.

• The solution is

xi = xi(t; x0, y0, z0).

* As t takes on all positive values, this traces
out the required pathline.

Fluid Mechanics, Prof. T.-S. Yang 3-11

3. Streaklines:
• A streakline is a line traced out by a
neutrally buoyant marker fluid continuously
injected into a flow field at a fixed point in
space.
• So, as for a pathline,
dxi
= ui(xi, t) (i = 1, 2, 3).
dt

• Important difference in the initial condition:

x = x0 , y = y0 , z = z0 (t = τ ),
where −∞ < τ ≤ t (t being the time of
observation).

• The solution is

xi = xi(t; x0, y0, z0, τ ).

* As τ takes on all the values τ ≤ t, these
equations will define the instantaneous
location (at time t) of the streakline
issuing from the point (x0, y0, z0).

Fluid Mechanics, Prof. T.-S. Yang 3-12

 • Note that for steady flows, the streamline,
pathline, and streakline are all the same.

Fluid Mechanics, Prof. T.-S. Yang 3-13

 • Read Examples 4.2 & 4.3, Munson et al.
• Example: (2-D, unsteady flow)

u = x(1 + 2t), v = y, w=0

1) Streamline:
dx dy
= x(1 + 2t), = y.
ds ds
* So, the general solution is

x = C1e(1+2t)s, y = C2es
(t considered constant).
* For the streamline passing through (1, 1),
we require x = y = 1 (s = 0).
* Hence

x = e(1+2t)s, y = es ,
or, eliminating s,

x = y 1+2t.
* The instantaneous streamline at t = 0 that
passes through (1, 1) therefore is the
straight line x = y.

Fluid Mechanics, Prof. T.-S. Yang 3-14

2. Pathline:
dx dy
= x(1 + 2t), = y,
dt dt
yielding the general solution

x = C1et(1+t), y = C2et.
* For the fluid particle at (1, 1) when t = 0,
require x = y = 1 (t = 0).
* Hence
x = et(1+t), y = et ,
or, eliminating t,

x = y 1+ln y .

3. Streakline: As for pathline,

x = C1et(1+t), y = C2et.
* But, for the streakline issuing from (1, 1),
require that x = y = 1 (t = τ ).
* This yields

x = et(1+t)−τ (1+τ ), y = et−τ .

Fluid Mechanics, Prof. T.-S. Yang 3-15

 • The instantaneous streakline at t = 0 can
be described by
x = e−τ (1+τ ), y = e−τ ,
or
x = y 1−ln y .

§ Recall also that for steady flows, the stream-

line, pathline, and streakline are all the same.
† Video 4.4: Jupiter red spot
† Video 4.5: Streamlines
† Video 4.6: Pathlines

Fluid Mechanics, Prof. T.-S. Yang 3-16

 • Fluids in the news:
Winds on Earth and Mars
* The wind has considerable temporal and
spatial variation.
* For example, the lowest monthly average
wind speed in Chicago (dubbed “the Windy
City” though there are several windier
cities) occurs in August (8.2 mph); the
highest occurs in April (11.9 mph).
* The wind speed variation throughout the
day has a shorter time scale.
* The normally calm morning wind increases
during the day because solar heating
produces a buoyancy-driven vertical motion
that mixes the faster moving air aloft with
the slower moving air near the ground..
* When the sun sets, the vertical motion
subsides and the winds become calm.
* The quickest variations in the wind speed
are those irregular, turbulent gusts that are
nearly always present.
§ i.e., “The wind is 20 mph with gusts to
35”.

Fluid Mechanics, Prof. T.-S. Yang 3-17

 * Spacecraft data have shown that the wind
on Mars has similar characteristics.
* These Martian winds are usually fairly light,
rarely more than 15 mph, although during
dust storms they can reach 300 mph.
* Recent photos from spacecraft orbiting
Mars show swirling, corkscrew paths in the
dust on the Martian surface (a visualization
of surface pathlines) indicating ‘dust devil’
type flows occur on Mars just as they do
on Earth.

Fluid Mechanics, Prof. T.-S. Yang 3-18

The material (time) derivative
• Let
φ(x, y, z, t)
be any field variable. (Eulerian description)
* e.g., ρ, T , V = (u, v, w), etc.
• Observe a specified fluid element for a short
period of time δt as it flows (Lagrangian).
• During that time, the position of the fluid
element will change by amounts δx, δy, δz,
while its value of φ will change by
∂φ ∂φ ∂φ ∂φ
δφ = δt + δx + δy + δz
∂t ∂x ∂y ∂z
(calculus).
• As we are following the fluid,

(δx, δy, δz) = (u, v, w) · δt.

• So,
Dφ δφ ∂φ ∂φ ∂φ ∂φ
≡ lim = +u +v +w ,
Dt δt→0 δt ∂t ∂x ∂y ∂z
which is called the material, total, or
substantial derivative.

Fluid Mechanics, Prof. T.-S. Yang 3-19

 • Use the notation D/Dt to emphasize that
the material derivative is the rate of change
seen by an observer “following the fluid.”

• The material derivative expresses a

Lagrangian derivative in terms of Eulerian
derivatives.

• In vector form,
Dφ ∂φ
= + (| V ·{z∇)φ} .
Dt ∂t
|{z} (2)
(1)
1) Local derivative: unsteady effects
2) Convective derivative: convective effects

Fluid Mechanics, Prof. T.-S. Yang 3-20

Mass conservation of a fluid particle
• Let

V(x, y, z, t) = u(x, y, z, t) i + v(x, y, z, t) j

+ w(x, y, z, t) k
be the velocity filed of a flow, and

ρ(x, y, z, t)
the fluid density field.
† This is the Eulerian description that takes
x, y, z, and t as independent variables, i.e.,
observe at a point that may be occupied
by different fluid particles at different time
instants.

Fluid Mechanics, Prof. T.-S. Yang 3-21

 • Mass flowrate on each side of the control
surface (CS):
" #
∂(ρu) δx
ṁx− = ρu − δy δz,
∂x 2
" #
∂(ρu) δx
ṁx+ = ρu + δy δz,
∂x 2
" #
∂(ρv) δy
ṁy− = ρv − δz δx,
∂y 2
" #
∂(ρv) δy
ṁy+ = ρv + δz δx,
∂y 2
" #
∂(ρw) δz
ṁz− = ρw − δx δy,
∂z 2
" #
∂(ρw) δz
ṁz+ = ρw + δx δy.
∂z 2
• Net mass flowrate into the control volume
(CV):
( )
X ∂(ρu) ∂(ρv) ∂(ρw)
ṁ = − + + δx
|
δy
{z
δz}
∂x ∂y ∂z
| {z } δV
∇·(ρV)
∂(ρδV ) ∂ρ
= −∇ · (ρV) δV = = δV.
∂t ∂t

Fluid Mechanics, Prof. T.-S. Yang 3-22

 • So we obtain the continuity equation for
fluid flow:
∂ρ
+ ∇ · (ρv) = 0.
∂t

• Alternative form:
∂ρ
+ v · ∇ρ +ρ∇ · v = 0
|∂t {z }
Dρ/Dt
Dρ
⇒ + ρ∇ · v = 0.
Dt

• The two vector forms of the continuity eqn

given above are independent of coordinate
system.
* But they can be expanded explicitly once a
particular coordinate system is specified.
* One way to do that will be explained later.

Fluid Mechanics, Prof. T.-S. Yang 3-23

Physical significance of ∇ · v
• From the continuity equation we obtain
1 Dρ 1 D(1/ρ)
∇·v = − =
ρ Dt 1/ρ Dt
1 D(δm/ρ)
=
δm/ρ Dt
1 D(δV )
,=
δV Dt
where δm is the constant mass of a fluid
particle (control mass), and δV = δm/ρ is
its volume.
• So, ∇ · v is the rate of change of volume,
per unit volume, of a fluid particle, as
seen by an observer moving with the fluid
particle.
* Also called the volume dilation rate.
* Having a precise physical meaning, ∇ · v
could not depend on choice of coordinates.

• Special case: For an incompressible fluid,

Dρ/Dt = 0 (ρ not necessarily constant),
and so ∇ · v = 0.

Fluid Mechanics, Prof. T.-S. Yang 3-24

 • An alternative derivation: (see §6.1.2)

* Displacement of lower left corner: (u, v)δt

* Upper right corner:
!
∂u ∂v
u+ δx, v + δy δt
∂x ∂y
* Change in volume: [δV = δx · δy · δz]
(  !
∂u ∂v
∆(δV ) = 1+ δx · 1 + δy
∂x ∂y
  
∂w
· 1+ δz − δx · δy · δz δt
∂z
!
∂u ∂v ∂w
= δV + + δt + H.O.T.
∂x ∂y ∂z
| {z }
∇·V
1 ∆(δV )
⇒ = ∇ · V.
δV δt

Fluid Mechanics, Prof. T.-S. Yang 3-25

 • As we shall show shortly, if the flow is
‘irrotational’, the fluid velocity can be
written as the gradient of a scalar function
φ(x, y, z; t), called the ‘velocity potential’.
* If the fluid is incompressible in addition,
then
∇ · V = ∇ · ∇φ
! !
∂ ∂ ∂ ∂φ ∂φ ∂φ
= i+ j+ k · i+ j+ k
∂x ∂y ∂z ∂x ∂y ∂z
∂ 2φ ∂ 2φ ∂ 2φ 2
= + + ≡ ∇ φ = 0.
∂x2 ∂y 2 ∂z 2
* The ‘Laplacian’ operator ∇2 has different
expressions in different coordinate systems,
but the result ∇2φ is the same in all such
systems.
† Later, we shall briefly discuss the potential
flow theory.
† The Laplace equation

2 ∂ 2φ ∂ 2φ ∂ 2φ
∇ φ= 2
+ 2
+ 2
=0
∂x ∂y ∂z
is one of the most important PDE in physics
and engineering.

Fluid Mechanics, Prof. T.-S. Yang 3-26

Vorticity
• Mathematical definition: ω = ∇ × V
* In Cartesian coordinates,
i j k
∂ ∂ ∂
ω= = ωxi + ωy j + ωz k;
∂x ∂y ∂z
u v w

∂w ∂v
− , ωx =
∂y ∂z
∂u ∂w
ωy = − ,
∂z ∂x
∂v ∂u
ωz = − .
∂x ∂y
• Physical significance:

Fluid Mechanics, Prof. T.-S. Yang 3-27

 * Average rate of CCW rotation about the
fluid element’s centroid:
AV
dθxy
dt
( )
1 (∂v/∂x)δxδt/δx (∂u/∂y)δyδt/δy
= −
2 δt δt
!
1 ∂v ∂u ωz
= − = .
2 ∂x ∂y 2
* Similarly,
AV AV
dθyz
ωx dθzx ωy
= , = .
dt 2 dt 2
* So, the vorticity (vector)

ω = 2Ω,
Ω being the angular velocity (vector) of a
fluid particle w.r.t. its own centroid.
• A flow with ω = 0 everywhere is said to
be irrotational. Fluid particles in such a
flow may still follow curvy (e.g., circular)
trajectories, however.
• The flow is rotational otherwise; fluid
particles may follow straight trajectories.

Fluid Mechanics, Prof. T.-S. Yang 3-28

 † We emphasize again that the length and
direction of ω = ∇ × V are independent of
the choice of coordinate systems in space.
• Note also that
∇ × (∇f ) = 0,
with f being any twice differentiable scalar
function.
* This can be readily verified by substitution:
i j k
∂ ∂ ∂ !
∂ 2f ∂ 2f
∂x ∂y ∂z = − i + ···
∂f ∂f ∂f ∂y ∂z ∂z ∂y
∂x ∂y ∂z
= 0.
* So, as pointed out before, the velocity of
an irrotational flow can be written in terms
of the velocity potential: V = ∇φ.

Fluid Mechanics, Prof. T.-S. Yang 3-29

 † Another useful vector identity:
∇ · (∇ × V) = 0,
with V being any twice differentiable vector
function (field), including the fluid velocity.
† Verify by substitution:
! !
∂ ∂w ∂v ∂ ∂u ∂w
− + −
∂x ∂y ∂z ∂y ∂z ∂x
!
∂ ∂v ∂u
+ − = 0.
∂z ∂x ∂y
Shear deformation rate [Video 6.1; read §6.1.3]

• (x, y) plane: (∂v/∂x) + (∂u/∂y)

• (y, z) plane: (∂w/∂y) + (∂v/∂z)
• (z, x) plane: (∂u/∂z) + (∂w/∂x)

Fluid Mechanics, Prof. T.-S. Yang 3-30

Acceleration of a fluid particle
Dv ∂V
a= = + (| V ·{z∇)V}
Dt ∂t
|{z} (2)
(1)

1) Local acceleration
2) Convective acceleration

• In Cartesian coordinates,

a = axi + ay j + az k,
with
∂u ∂u ∂u ∂u
ax = +u +v +w ,
∂t ∂x ∂y ∂z
∂v ∂v ∂v ∂v
ay = +u +v +w ,
∂t ∂x ∂y ∂z
∂w ∂w ∂w ∂w
ax = +u +v +w .
∂t ∂x ∂y ∂z

Fluid Mechanics, Prof. T.-S. Yang 3-31

∇ · V and other stuffs in cylindrical coordinates

• Unit vectors in cylindrical coordinates:

er = i cos θ + j sin θ
eθ = −i sin θ + j cos θ
ez = k.
* Note that
∂ er ∂ eθ
= eθ , = −er .
∂θ ∂θ
* Other derivatives of unit vectors are zero.
• A function may be expressed in terms of
both (x, y, z) and (r, θ, z).
• Also, using calculus, we have
∂f ∂f ∂f
df = dx + dy + dz
∂x ∂y ∂z
∂f ∂f ∂f
= dr + dθ + dz ≡ ∇f · dr.
∂r ∂θ ∂z

Fluid Mechanics, Prof. T.-S. Yang 3-32

 * This can be taken as the definition (or
meaning) of ∇f (gradient of a scalar field),
which is invariant through coordinate
transformation.
† Suppress possible dependence on t for now.

• We may write

dr = idx + jdy + kdz

= er dr + eθ (rdθ) + ez dz.
* So, in cylindrical coordinates
∂f 1 ∂f ∂f
∇f = er + eθ + ez .
∂r r ∂θ ∂z
* Or, symbolically,
∂ 1∂ ∂
∇ = er + eθ + ez .
∂r r ∂θ ∂z

• Now, let us express the velocity vector V

in cylindrical coordinates:

V = Vr er + Vθ eθ + Vz ez .

Fluid Mechanics, Prof. T.-S. Yang 3-33

1) Expanding, and differentiating, one obtains:
∇
 ·V 
∂ 1∂ ∂
= er + eθ + ez
∂r r ∂θ ∂z
· (Vr er + Vθ eθ + Vz ez )
∂Vr 1 ∂Vθ ∂Vz Vr
= + + +
∂r r ∂θ ∂z r
1∂ 1 ∂Vθ ∂Vz
= (rVr ) + + .
r ∂r r ∂θ ∂z
2) Similarly,
∂ Vθ ∂ ∂
V · ∇ = Vr + + Vz .
∂r r ∂θ ∂z
3) Hence the convective acceleration
(V · ∇) V
 
∂ Vθ ∂ ∂
= Vr + + Vz
∂r r ∂θ ∂z
(Vr er + Vθ eθ + Vz ez )
( )
∂Vr Vθ ∂Vr ∂Vr 1 2
= er Vr + + Vz − V
∂r r ∂θ ∂z r θ
( )
∂Vθ V ∂V ∂V 1
+ eθ Vr + θ θ + Vz θ + Vr Vθ
∂r r ∂θ ∂z r
 
∂Vz Vθ ∂Vz ∂Vz
+ ez Vr + + Vz .
∂r r ∂θ ∂z

Fluid Mechanics, Prof. T.-S. Yang 3-34

 * Centrifugal and Coriolis accelerations
4) Laplacian:

∇2T = ∇ · (∇T )
 
∂ 1∂ ∂
= er + eθ + ez
∂r r ∂θ ∂z
 
∂T 1 ∂T ∂T
· er + eθ + ez
∂r r ∂θ ∂z
∂ 2T 1 ∂ 2T ∂ 2T 1 ∂T
= + + +
∂r2 r2 ∂θ 2 ∂z 2 r ∂r
 
1∂ ∂T 1 ∂ 2T ∂ 2T
= r + 2 2 + .
r ∂r ∂r r ∂θ ∂z 2
5) Vorticity:

ω =∇×V
 
∂ 1∂ ∂
= er + eθ + ez
∂r r ∂θ ∂z
×(Vr er + Vθ eθ + Vz ez )
   
1 ∂Vz ∂Vθ ∂Vr ∂Vz
= er − +eθ −
| r ∂θ {z ∂z } | ∂z {z ∂r }
≡ωr ≡ωθ
 
∂Vθ 1 ∂Vr V
+ ez − + θ .
| ∂r r {z∂θ r }
=[∂(rVθ )/∂r−∂Vr /∂θ]/r≡ωz

Fluid Mechanics, Prof. T.-S. Yang 3-35

Streamline coordinates
(Read §4.2.4, Munson et al.)
• Streamline: a line everywhere parallel to
the local fluid velocity
• So, if we denote the local unit tangent of
a streamline by es, then the fluid velocity
V = V es,
V being the magnitude of |V| (speed).
* s: along streamline, in direction of fluid
flow
* n: normal to streamline, away from center
of curvature
* l: ‘binormal’ direction; el = es × en

Fluid Mechanics, Prof. T.-S. Yang 3-36

 § Shown above is an example of a streamline
coordinate system for 2-D flow.
• Now, note that in streamline coordinates,
∂ ∂ ∂ ∂
∇ = es + en + el ⇒ V · ∇ = V .
∂s ∂n ∂l ∂s
• Also, ∂ es/∂s = −en/R, R being the local
radius of curvature.

Fluid Mechanics, Prof. T.-S. Yang 3-37

 • So, for a steady flow (with ∂ V/∂t = 0), we
have the fluid particle acceleration
DV
a = = (V · ∇)V
Dt
 
∂ ∂V V2
= V (V es) = V es − en .
∂s | ∂s
{z } | R{z }
(1) (2)

1) Convective acceleration in the streamline

direction
2) Centripetal acceleration normal to stream-
line

Fluid Mechanics, Prof. T.-S. Yang 3-38

Stream function
• For incompressible flow, the continuity eqn
reads ∇ · V = 0.
• Suppose that the flow is 2-D in addition;
then in Cartesian coordinates we have
∂u ∂v
+ = 0.
∂x ∂y
• May introduce stream function ψ such that
∂ψ ∂ψ
u= , v=− .
∂y ∂x
* The continuity eqn then is automatically
satisfied.

§ In polar coordinates, the continuity eqn

reads
1∂ 1 ∂Vθ
(rVr ) + = 0,
r ∂r r ∂θ
and we may write
1 ∂ψ ∂ψ
Vr = , Vθ = − .
r ∂θ ∂r

Fluid Mechanics, Prof. T.-S. Yang 3-39

Lines of constant ψ

• Along a constant-ψ line,

∂ψ ∂ψ
dψ = dx + dy = −v dx + u dy = 0
∂x ∂y
∂ψ ∂ψ
= dr + dθ = −Vθ dr + Vr r dθ,
∂r ∂θ
i.e.,
   
dy v 1 dr Vr
= , and = .
dx ψ u r dθ ψ Vθ
• So, constant-ψ lines are streamlines (hence
the name).

Fluid Mechanics, Prof. T.-S. Yang 3-40

Physical significance of ψ

• The volumetric flowrate (per unit breadth)

g
through AB:
Z B Z B
Q′12 = (u dy − v dx) = (Vr r dθ − Vθ dr)
A A
Z B !
∂ψ ∂ψ
= dy + dx
A ∂y ∂x
Z B 
∂ψ ∂ψ
= dθ + dr
A ∂θ ∂r
Z B
= dψ = ψ2 − ψ1.
A
* Integration of an exact differential is path-
independent.

Fluid Mechanics, Prof. T.-S. Yang 3-41

Boundaries conditions for ψ

1) Solid boundaries that allow no permeation

of fluid must have a constant value of ψ,
and hence coincide with streamlines.
2) Suppose the flow is uniform at the inlet,
with constant velocity V1.
† i.e., V = V1i at x = x1.
* Then we have ψ = ψ0 + V1y at x = x1.
* The constant ψ0 can be arbitrarily chosen,
say zero.
3) Suppose the fluid is incompressible; then
by mass (or volume) conservation the mean
R h2
velocity V2m = { 0 u dy2}/h2 = V1h1/h2.
* But usually other BC must be given to
uniquely determine the flow; e.g. specified
exit pressure.
• Governing equations for ψ will be discussed
elsewhere.

Fluid Mechanics, Prof. T.-S. Yang 3-42