Water Science and Engineering

Free Surface Hydrodynamics

y
y
A part of Module 2: Hydraulics and Hydrology
Water Science and Engineering

Dr. Shreedhar Maskey

Senior Lecturer
UNESCO-IHE Institute for Water Education

S. Maskey (2011)

About Module 2 Hydraulics and Hydrology

Free surface hydrodynamics (35%)

Engineering hydrology (35%)

GIS and remote sensing (30%)

S. Maskey (2011)

Part-1: Fundamental principles and equations

Continuity and Momentum principles

Euler and Navier-Stokes equations
Bernoulli equation and applications

Continuity principle

Expresses the conservation of mass

in a control volume occupied by a
fluid.
Obtained byy equating
q
g the change
g in
fluid mass to the difference between
the rate of mass IN and OUT in a
given time.
Change in mass in time t is

( (

uAx Ax

u
x
x

uAx
X

Rate of fluid mass inflow and

outflow for the continuity equation.
Y

t ))xyz txyz
t
t

Change in mass due to the change in

and u (in x-direction)

( uAx Ax

u
u
u
x uAx )t Ax x
t
txyz
x
x
x
4

S. Maskey (2011)

Continuity principle

uAx Ax

u
x
x

uA x

General form of continuity equation:

u v w

u
0

v
w
t
x
y
z
x y z

For incompressible fluid with constant :

v w

0
y z

For a steady and incompressible flow in

a pipe:

Q V1 A1 V 2 A2
Also valid for a steady continuous flow
in an open channel.
5

Continuity principle

For unsteady flow in an open channel:

h
0
t

Q A

0
x t

S. Maskey (2011)

Momentum principle

Inertia force: To change an existing motion of a body of

mass M, it is necessary to apply a force F to this mass,
which causes an acceleration a = dV/dt, such that

dV
dV
dt

dV
(unit volume)
dt

This is the Newton's Equation of Motion. The product

M(dV/dt) is the inertia force.

Two types of inertia forces are considered

Due to local accelation change in velocity in time

Due to convective acceleration change in velocity

over a distance

dV
dt
V

dV
dx
7

Momentum principle

Inertia force components in 3 dimensions (x, y, z axes):

u

u
u
u
v
w
z
x
y

v
v
v
v w
x
y
z

w
w
w
v
w
x
y
z

Local
acceleration
term

Convective
Convective
acceleration
terms

S. Maskey (2011)

Momentum principle

Pressure forces acting on two

opposite faces on the yz-plane.

Applied forces:
Gravity force (Fg = Mg)
Pressure force

p p x yz
x

p(yz)

p
p

p (yz ) p x yz xyz
x
x

X
Temperature
-t(oC)

Dynamic
Viscosity
-(N s/m2) x
10-3

Kinematic
Viscosity
-(m2/s) x 10-6

1.787

1.787

1.519

1.519

10

1.307

1.307

20

1.002

1.004

30

0.798

0.801

40

0.653

0.658

50

0.547

0.553

60

0.467

0.475

70

0.404

0.413

80

0.355

0.365

90

0.315

0.326

100

0.282

0.294

Viscous force (recall Newtons Law

of Viscosity)

ddu

dy
Shear stress
Coefficient of viscosity

Momentum principle

zy

zx

In 3 dimensions
yz

yx
xx

x(yz )
y (xz ) zx z (xy )
x
y
z
zy
yy
xy
x(yz )
y (xz )
z (xy )
z
y
x
yz
xz

x(yz )
y (xz ) zz z (xy )
y
z
x

xz

xy

yx
X

Convention used to represent

shear stresses on various
planes. The first subscript
represents the axis normal to the
plane and the second subscript
represents the axis along which
the force is acting.

10

S. Maskey (2011)

Momentum equation

zx

Obtained from balancing inertia

and applied forces

yz

Euler equation neglects viscous

forces
1 p
u
u
u
u
u
v
w

x
t
x
y
z
1 p
v
v
v
v
u v w
t
x
y
z
y
w
1 p
w
w
w
u
v
w

g
t
x
y
z
z

zy

xz

xy

yx
X

Convention used to represent

shear stresses on various
planes. The first subscript
represents the axis normal to the
plane and the second subscript
represents the axis along which
the force is acting.

11

Momentum equation

zx

Navier-Stokes equation also includes

viscous forces
du 1 p xx yx zx

dt x
x
y
z
dv 1 p xy yy zy

z
x
y
dt y
dw 1 p xz yz zz
g

x
y
z
dt
z

zy

yz

xz

xy

yx
X

Convention used to represent

shear stresses on various
planes. The first subscript
represents the axis normal to the
plane and the second subscript
represents the axis along which
the force is acting.

12

S. Maskey (2011)

Bernoulli equation

Can be derived from Euler equation

Steady flow assumption is used
Has wide application in hydraulics
Represents the total energy and signifies that total the
energy remains constant

V2
p

z Constant
2 g g

TOTALENERGYLINE
V12 2 g

Sf

Energyline
V22 2 g

V1

h2 V2

h1
S0

Freesurface

Channelbottom
13

Part-2: 1D Channel Flow

Steady-uniform flow
Friction coefficient and velocity distribution
Specific discharge,
discharge critical depth
Non-uniform gradually varied flow (backwater
curve)
Unsteady flow

St-Venant equation
Kinematic wave

14

S. Maskey (2011)

Steady-uniform flow

Velocity is constant (i.e. V/t = 0) and is uniform in

space (i.e. in one dimension V/x = 0).
The water surface line remains parallel to the channel
bed profile.
p
Gravity force to be balanced by the shear force due to
the bed resistance.

gxA sin x b dP

b g sin

A
gR sin
P

Remember:
P is wetted perimeter
Hydraulic radius, R = A / P

15

Steady-uniform flow

For small slope

sin tan

z b
S0
x

b gRS 0

For turbulent flows, the shear stress can be assumed to

be related to the average velocity V and friction
coefficient Cf, as
2

b C f V

Finally

g
RS 0
Cf
16

S. Maskey (2011)

Steady-uniform flow

Chezys formula

V C RS 0
Q CA RS0 C

3
2

S 02

P2
Remember:

Mannings formula
2

Discharge, Q = A x V
2

1
V R 3 S 02
n

5
3

Hydraulic radius, R = A / P

1
1 A 2
Q AR 3 S 02
S0
n
n 23
P

17

Steady-uniform flow

Normal depth depth based on uniform flow

From Mannings formula
5
1

1 A3 2
Q
S0
n 23
P
Rectangular channel

( Bhn ) 5 / 3
nQ

2/3
( B 2 hn )
S0

Remember:
For a wide
F
id rectangular
t
l
channel (i.e. B >> h)

A5 / 3
nQ

2/3
P
S0

PB
Rh

Wide rectangular channel

nQ
hn
B S
0

Can you work out similarly the relationship for the normal depth using the
Chezys formula?
18

S. Maskey (2011)

Uniform flow computation in a compound channel

The flow velocity and the carrying capacity of the channel are
normally different for the main channel and the over bank flow
or the floodplain.
The roughness coefficient may also vary in the main channel
and the floodplain.
2

A2 h 2

b22

A1

b21

The usual approach is

Qtotal Q1 Q2 Q3

b32

b31

h1
b11

b12

3
h 3 A3

b13

That is, the total discharge is the sum of the

discharges through each segment of the cross section.

1
2/3
Ai Ri S 1/ 2
ni

19

Specific energy

The energy (in water head)

due to the depth of water (h)
plus its velocity (the velocity
head) is the specific energy.

V2
Q2
E h
h
2g
2 gA 2

Differentiating w.r.t. h
dE
Q 2 dA
1 3
dh
gA dh

Supecriticalflowrange
hc
Emin

dE
V2
1
dh
gD

For a given discharge there

exists a min. specific energy
Emin (at dE/dh = 0):

V
gD

With dA/dh = B and D = A/BT.

V
gh

For a rectangular channel

20

S. Maskey (2011)

10

Critical depth, Froude number

Water depth at Emin is called a

critical depth, hc.
V
gD

Q2
A D
g

1
3

Q
q
hc 2
gB
g
2

1
3

Supecriticalflowrange
hc
Emin

Froude
F
d number
b (Fr)
(F ) represents
t
the ratio of inertia forces to
gravity forces, given by
V
Fr
At critical depth (hc), Fr = 1.
gD
21

Subcritical and supercritical flows

Subcritical flow:

Fr < 1 V < (gD)

Critical flow:

Fr = 1 V = (gD)

Supercritical flow: Fr > 1 V > (gD)

Subcritical flow
hn

hc

Supercritical flow
hc

hn

22

S. Maskey (2011)

11

Celerity, subcritical and supercritical flow

The critical velocity (gD) is also the speed, or more precisely,

"celerity" of small gravity waves that occur in shallow water in
open channels as a result of a disturbance to the free surface.

gD

gh

For a rectangular channel

A gravity wave can be propagated upstream in water of subcritical

flow (c > V), but not in water of supercritical flow (c < V).

Wave patterns created by disturbances: (a) Still water, V = 0; (b) subcritical flow, c > V;
(c) critical flow, c = V; and (d) supercritical flow, c < V.
23

Non-uniform, steady flow

Gradually varied flow- backwater curve computation

(non-uniform, steady flow)

24

S. Maskey (2011)

12

Gradually varied (non-uniform) steady flow

Gradually varied (non-uniform) steady flow equations

can be derived either by
(1) Taking / t terms equal to zero (because steady flow) in the
St. Venant equation, or
(2) From the consideration of total energy (Bernoulli equation)
between two sections in a uniform channel.

25

Gradually varied (non-uniform) steady flow

Equation in terms of Froude Number (Fr):

dh S 0 S f

dx 1 Fr 2
OR

Sf

4/3
n 2Q 2
Mannings
2 2 P
n
Q

2 43
10 / 3 formula
AR
A

Fr 2

BT Q 2
gA3

dh S 0 S f

dx dE dh
Sf = friction slope, S0 = channel bed slope, P = wetted perimeter, A = cross-section area, h = water depth,
x = length along the channel, n = Mannings coefficient, BT = top width of channel cross section, Fr =
Froude number, g = acceleration due to gravity, Q = discharge, E = specific energy.
Exercise:
Rewrite the formula for Sf for a wide rectangular channel. Can you workout Sf using Chezys formula?
26

S. Maskey (2011)

13

Gradually varied (non-uniform) steady flow

Equation in terms of normal and critical depths for

a wide-rectangular channel:

dh
h 3 hn3
S0 3
dx
h hc3

Q2
h 2 2
C B S0
3
n

Q2
h
gB 2
3
c

C = Chezys coefficient, B = constant bed width, hn = normal depth, hc = critical depth.

27

Gradually varied flow flow profiles

M1 (backwater curve)

P4/3
S f n Q 10 / 3
A
2

M2
(drawdown curve)

Region 1
yn

M3

Region 2
yc

Region 3
28

S. Maskey (2011)

14

Solution of gradually varied steady flow equation

Numerical solution (direct step solution)

dh S 0 S f

dx 1 Fr 2

1 Fr 2
x h
S0 S f
OR

h x

S0 S f
1 Fr 2

The problem here is Sf and Fr

are dependent
p
on h, i.e. the
equation is implicit on h.

Starting at the downstream boundary,

compute the horizontal distance x
that corresponds to a given change in
depth h.

Starting at the downstream boundary,

compute change in depth h for a
chosen horizontal distance x.
29

Solution of gradually varied steady flow equation

Numerical solution (improved Eulers method)

dh S 0 S f

dx 1
1 Fr 2
h j 1/ 2 h j

Remember!
When starting from downstream towards
upstream, make sure that x is negative.

x S 0 S f Compute Sf and Fr with h = hj.

2 1 Fr 2 h
j

S 0 S f Compute
C
t Sf and
dF
Fr with
ith h = hj+1/2
h j 1 h j x
2
1 Fr h j1 / 2
hj = water depth at section j, hj+1 = water depth the section next to section j,
hj+1/2 = water depth between sections j and j+1.
30

S. Maskey (2011)

15

Solution of gradually varied steady flow equation

Numerical solution (Predictor-corrector method)

Remember!
dh S 0 S f

When starting from downstream towards

dx (1 Fr 2 )
upstream,
t
make
k sure th
thatt x
is
i negative.
ti

Predictor :

Corrector :

S 0 S f Compute Sf and Fr with h = hj.

h jpre
h
x
1
j

2
1 Fr h j

cor
j 1

S 0 S f
x S 0 S f
hj

2 1 Fr 2 h 1 Fr 2 h pre
j
j 1

Repeat Corrector step with a new value of predictor h until the predictor h is close enough to
corrector h. The new value of predictor is normally taken as the current value of the corrector h.
31

Hydraulic jump

Occurs when supercritical flow changes

into a subcritical flow.

h2 1
1 8Fr12 1

h1 2

Hydraulic jump

Supercritical
h1

V1

Subcritical

h2

V2

Practical application: to dissipate energy

downstream of dams,
dams weirs
weirs, etc.
etc

Enery loss, E E1 E2

h2 h1

4h1h 2

32

S. Maskey (2011)

16

Unsteady, non-uniform flow

Unsteady Flow, St. Venant Equation

Flood Wave Propagation

33

Unsteady non-uniform flow: moving flood wave

H, Q

Longitudinal view of water surface in an open channel

H, Q

H, Q

at time T1

at time T2

at time T3
X
34

S. Maskey (2011)

17

Unsteady non-uniform flow

Longitudinal Profile: Water surface profile

along a river channel at a given time.
Represents variation in space
T

Hydrograph: Discharge at various times

through a given section of a river.
Represents variation in time

Computation of unsteady
non-uniform flow considers
variation in both space and
time together.
35

Saint Venant equations for 1D channel flow

Assumptions of Saint-Venant equation

One-dimensional flow
Gradually varying flow in space and time
Hydrostatic pressure prevails,
prevails vertical acceleration is
negligible
Stable and relatively small bed slope
Longitudinal axis of the channel is approximated as a
straight line.
Incompressible, homogeneous and constant viscosity fluid

Consists of Continuity and Momentum equations

Momentum equation can be derived from Euler
equation with an added bed shear force.
36

S. Maskey (2011)

18

Saint Venant equations

Momentum force:

Gravity force component:

Pressure force:

V
V
V
Ax
x
t

gAxS 0

h
xA
x

gAxS f

Shear force due to friction:

V12 2 g

Energyline

Sf

V22 2 g

V1
h1

Freesurface

h2 V2

S0

Channelbottom

37

Saint Venant equations

Equating momentum and applied forces:

V
V
h
V
g
g S0 S f 0
t
x
x

V = Q/A

Q Q 2

t x A
Local
acceleration term

h
gA gAS 0 S f 0
x

Convective
acceleration term

Pressure
force term

Gravity
force term

Friction
force term

38

S. Maskey (2011)

19

Saint Venant equations

Various flood wave approximations from the SaintVenant equation:

Continuity equation
Q A
0

x t
Momentum equation
Q Q 2
h
gA gA( S 0 S f ) 0

t x A
x
Kinematic wave
Diffusion wave
Full dynamic wave
39

Kinematic wave approximation

From continuity equation:

Q A

0
x t

e at c wave
a e app
approximation
o at o neglects
eg ects
Kinematic
acceleration and pressure terms of the
momentum equation, i.e.

Note: If there is lateral inflow q, the

continuity equation becomes:
Q
A
q

x
t

S0 S f
This assumption allows to use Manning or
Chezy equation:

1
AR 2 / 3 S 1/ 2 OR Q CAR1/ 2 S 1/ 2
n

40

S. Maskey (2011)

20

Kinematic wave approximation

In both Manning and Chezy equation, A can be
expressed in terms of Q as

A Q
Using Manning equation, the coefficients alpha
and beta are given by
Exercise:
0.6

nP 2 / 3
1/ 2 and 0.6
S

Verify that the given and are for the

Manning equation and derive them for the
Chezy equation.

41

Kinematic wave approximation

In continuity equation, replacing A with
Q and simplifying.

Q
Q

Q 1
0
x
t

Q 1 Q

0
x ck t

ck is the kinematic wave celerity.

Where
ck

1
1 Q
5Q 5

V
1

Q
Q
3A 3

42

S. Maskey (2011)

21

Unsteady flow and flood wave propagation

A note on numerical methods for the solution

1D unsteady flow

43

Unsteady flow and flood wave propagation

In practice almost always solved using numerical

methods.

Finite Difference Method is one of the widely used

numerical methods.

Discretization is used in space and time to approximate

the partial
partial difference equation
equation to a finite
finite difference
(algebraic) equation.

44

S. Maskey (2011)

22

Finite difference numerical solution

Understanding space-time discretisation of

numerical schemes
t

Q nj11

Q nj 1

n+1
n
n-1
n=0
j=0

known points
unknown point

Qnj

Q nj1

j-1

x
j

j+1

initial condition
boundary condition

45

Finite difference numerical solution

Explicit and Implicit schemes:

In an explicit scheme the value of the variable at time level

n+1 can be computed directly (or explicitly) from the values
at time level n.
In an implicit scheme the computation of the value of a
variable at n+1 involves one or more of the values from the
same time level (i.e. n+1).

46

S. Maskey (2011)

23

Numerical solution of kinematic wave (an example)

An example of an explicit scheme

t

Q 1 Q

0
x ck t

Partial differential equation

TO
Finite difference equation

Q nj1 Q nj
x

Q nj1

Q nj11

Q nj

Q nj1

n+1

n 1
n
1 Q j 1 Q j 1
0
ck
t

n
n-1
j-1

Q nj11 Cr Q nj (1 Cr )Q nj1

j+1

known points
unknown point

Where
t
x

C r ck

Cr is the Courant Number.

47

Numerical solution of kinematic wave (an example)

An example of an implicit scheme

t

Q 1 Q

0
x ck t

Partial differential equation

TO
Finite difference equation

n 1
j 1

n 1
j

Q nj11

n 1
j 1

1 Q Q
t
ck

n
j 1

Q nj1

Q nj11

Q nj

Q nj1

n+1
n

0
Multiply both
sides by x

1
Cr
Q nj1
Q nj1
1 Cr
1 Cr

n-1
j-1

j+1

known points
unknown point

Where
C r ck

t
x

Cr is the Courant Number.

48

S. Maskey (2011)

24

Finite difference numerical solution

Stability of an scheme

In numerical methods, the choice of the space step (x)

and time step (t) is important!
An explicit scheme becomes unstable when the Courant
number (Cr = ct/x) is greater than 1. Where c is the
celerity.
An implicit solution is always stable. However, Cr >> 1 is
nott recommended.
d d

49

Finite difference numerical solution

Initial and boundary conditions:

Numerical schemes are based on

the assumption
p
that variable
values are known at all points on
space at time level zero also
called an initial condition.
A numerical scheme also requires
that values at all time levels at the
starting point on space (called
upstream boundary condition)
and/or at the end point on space
(called the downstream boundary
condition).

50

S. Maskey (2011)

25

Commonly used initial and boundary conditions

Initial condition:

Upstream boundary condition:

Global or local water depth and steady state computation

Discharge hydrograph

Downstream boundary condition:

Q-h relationship
p

End of Presentation
51

S. Maskey (2011)

26