Chapter1 3
Chapter1 3
1. Introduction
1
Hydraulic gradient is coincident with the free surface of water and
sectional area of flow expands or contracts in accordance the discharge
Atmospheric pressure (ignored in many cases) but pressure of the fluid
itself on the surrounding is considered
Flow driven by gravity or potential energy (unlike pressure in pipes)
2
balance is characterized by no change of flow cross-section, depth
or velocity. This is uniform flow.
Flow
Unsteady Steady
3
The flow is steady when at a particular section, the depth of liquid and other
parameters such as velocity, area of cross-section, discharge do not change with
time.
Steady v / t = 0
Uniform v / x = 0 or y / x = 0
4
The opposite is non-uniform flow, where the depth varies along the length of
the channel (e.g. between converging or diverging boundaries). Non-uniform
flow is also called varied flow.
Varied v / x # 0 or y / x # 0
Non-uniform flow can be further divided into two types: Gradually varied flow
and Rapidly varied flow.
In gradually varied flow, the variation of the depth of liquid along the length is
gradual. In this type of flow, we find water surface profiles and their lengths.
Flow through hydraulic structures is a typical example.
y1
y2
Dam
In rapidly varied flow, the change in depth is sudden (e.g. water flow over an
overflow dam, there is a sudden rise of water at the toe of the dam and a
hydraulic jump (low stage to high stage) forms. Another common phenomena is
a hydraulic drop (from high stage to low stage). We are generally interested in
the location of the rapidly varied phenomenon and its effect on the flow.
These are basic relationships that we use to analyze open channel flow. They
account of the mass, energy and momentum in a system:
5
Conservation of matter (or mass) – The continuity equation
Mass (matter) can not be destroyed nor created. In fluid mechanics
& in open channel hydraulics, conversion by chemical activity is not
considered
Conservation of energy – The energy equation
Energy budget
Conservation of momentum – The momentum equation
Momentum budget
1.9 Definitions
Area of cross-section (or wetted area), A – is the area of the liquid
surface cut when a cross-section is taken normal to the direction of flow.
Depth of flow, y – is the vertical distance between the lowest point of the
channel section and the free surface.
Depth of flow section, d – is the depth of liquid at the section measured
normal to the direction of flow. A relation between the depth of flow, y and
the depth of flow section, d is: d = y cosθ, where θ is the angle, which the
bed line makes with horizontal. For a channel with small slope, the value
of cosθ will be approximately unity. Therefore, d ≈ y.
Wetted perimeter, P – is the length of the line (or curve) of intersection of
the wetted surface with a plane normal to the direction of flow. In other
words, it is the length of the perimeter of the section, which is wetted. The
free surface is not included in the wetted perimeter as there is no
boundary surface to cause friction.
Hydraulic radius, R – is the ratio of the area of cross-section to the
wetted perimeter (R = A/P).
Hydraulic depth (Hydraulic mean depth), D – is the ratio of the wetted
area, A to the top width, T (D = A/T).
Stage – is the elevation or vertical distance of the free surface above a
datum. If the lowest point of the channel section is taken as the datum, the
stage is the same as the depth of flow, y.
6
Hydraulic Gradient Line (HGL) – is the line indicating the pressure at
various sections along the channel. In case of open channels, the HGL
coincides with the liquid surface. If a piezometer is inserted in an open
channel, the liquid will rise in the tube to the level of the liquid surface. The
liquid surface slope, Sw is the slope of the liquid surface or HGL.
Energy Gradient Line (EGL) – is the line indicating the total energy of the
liquid with respect to a selected datum. The EGL is above the HGL. The
vertical distance between the two lines is αV 2/2g, where α is the velocity
distribution factor, which takes into account the non-uniform distribution of
velocity across the section. The value of α varies from 1.1 to 1.2 for the
turbulent flow in open channels. However, for simplicity it is usually taken
as unity.
Hydraulic slope, S – is the slope of the total energy line: S = h f / L. Where
L is the length of the channel in which the drop of the total energy line is
hf. For uniform flow, S = S w = So. For small So, So = tangent or sine of
angle.
<End of 17/2/2004 Introductory Lecture>
2. Velocity distribution
There is free surface and friction along the channel wall of open channel. The
velocity at a solid boundary must be zero and in open channel flow it generally
increases with distance from the boundaries (non-uniformly distributed in the
channel section). The maximum velocity does not occur at the free surface but is
usually below the free surface a distance of 0.05 to 0.25 of the depth.
7
Elemental layer
x
y
x
vmax
dcos d
v
The component of the gravitational force parallel to the slope is g(d – y)sin and
this is resisted by the shear force, . Note that y is measured from the channel
bed to the water surface.
= g(d – y)sin
By the Newton’s law of viscosity, = (dv / dy); is dynamic viscosity [N.s/m2]
(dv / dy) = g(d – y)S; S = sin, the longitudinal slope
dv = g(d – y)S dy
dv = (gS/) (d – y)dy
= / is kinematic viscosity [m2/s]
dv = (gS/) (d – y)dy
dv = (gS/) (d – y)dy
v = (gSy/) (d – y/2) + C; C is a constant of integration
For v = 0 when y = 0, C = 0
v = (gSy/) (d – y/2)
v = gSyd/ – gSy2/2, which is an equation of a parabola.
Therefore, the velocity distribution is parabolic under uniform flow. The
roughness of the channel will cause the curvature of the vertical-velocity-
distribution curve to increase (deviates from parabolic).
8
2.1.1 Maximum velocity determination
dv / dy = gSd/ – gSy/
For maximum velocity, dv / dy = 0:
gSd/ = gSy/
y=d
Therefore, maximum velocity occurs at the surface.
Substituting y =d in the velocity distribution equation above,
vmax = gSd2/ – gSd2/2
vmax = gSd2/2
V = gSd2/3 = (2/3)vmax
Therefore, the mean velocity over the depth d is (2/3)v max.
9
Water surface
80 100
40
Cite the general patterns for velocity distribution in other several channel
sections of channel shapes:
Rectangular
Triangular
Trapezoidal
Shallow ditch
Narrow rectangular section
Pipe
10
2.3 Approximation methods for mean velocity
(2) For a closer approximation, is the average of the velocities at 0.2 and 0.8 x
bed depth, which accords with a parabolic distribution, although there is no
theoretical justification for this.
- The mean velocity varies between 0.8 and 0.95 times the surface
velocity, the usual value being about 0.85.
- In the transverse direction the distribution is much more irregular,
except that in general the velocity adjacent to the sides is nearly
always less than that nearer to the center. Because of the
inconsistent nature of the distribution, it is not possible to define
the location of any singe point of measurement that will yield the
mean velocity over the entire cross-section.
11
From which,
Mean velocity,
And the height ym above the bed at which the filament (path line) velocity has the
mean value is obtained from:
or ym = 0.37d
Which compares very favorably with the conventional (theoretical) 0.4d derived
above.
And at 0.8d:
12
For very wide-open channels, observations have shown that the velocity
distribution in the central region of the section is essentially the same as it
would be in a rectangular channel of infinite width.
The sides of the channel have practically no influence on the velocity
distribution in the central region in the 2D flow
Wide open channel can safely be defined as a rectangular channel
whose width is greater than 10 times the depth of flow.
The true velocity head may be expressed as V2/2g, where is known as the
energy coefficient or Coriolis coefficient. And the true Kinetic Energy (K.E.)
may therefore be computed as AV3/2 if and V are known.
Experimental data indicate that the value of varies from about 1.03 to
1.36 for fairly straight prismatic channels. The value is generally higher for
small channels and lower for large streams of considerable depth.
= 1 for uniform velocity distribution across the channel section. Also,
assumed for regular cross-sections.
= 1.6 for complex cross-sections
is very large in steep channels than in flat channels.
In most practical problems dealing with regular channels it is not
necessary to consider the variation of velocity throughout the cross-
section, since use of the average velocity will give the accuracy required.
The non-uniform distribution of velocities also affects the calculation of
momentum in open channel flow. Therefore, a momentum coefficient, β is
introduced in the momentum equation.
13
3. Pressure distribution
Water surface
A
h
HGL
h = hs h
The pressure at any point on the section is directly proportional to the depth
of the point below the free surface and equal to the hydrostatic pressure
corresponding to this depth. In other words, the distribution of pressure over
the cross-section of the channel is the same as the distribution of hydrostatic
pressure, i.e., the distribution is linear and can be represented by a straight
line AB.
For parallel flow, the streamlines have neither substantial curvature nor
divergence. As a result, there are no appreciable acceleration components
normal to the direction of flow that would disturb the hydrostatic pressure
distribution in the x-section of a parallel flow. Gradually varied flow may also be
regarded as parallel flow, since the change in depth of flow is so mild that the
streamlines have neither appreciable curvature nor divergence. For practical
purposes therefore, the hydrostatic law of pressure distribution is applicable to
gradually varied flow as well as to uniform flow.
14
3.2 Curvilinear flow
hs
B B’
Fig. 8: Concave flow
15
Let the deviation from an otherwise hydrostatic pressure hs in a curvilinear flow
be designated by c. Then the true pressure or piezometric height h = hs + c.
The c is positive for concave flow, negative for convex flow and zero for parallel
flow.
For simplicity, the pressure head of a curvilinear flow may be represented by
’y, where ’ is a correction coefficient for the curvature effect. The correction
coefficient is referred to as a pressure distribution coefficient. Since this
coefficient is applied to a pressure head, it may be specifically called a pressure
coefficient.
Q is total discharge, y is depth of flow, ’ is greater than 1.0 for concave flow, ’
is less than 1.0 for convex flow and ’ is equal to 1.0 for parallel flow.
In rapidly varied flow the change in depth of flow is so rapid and abrupt that the
streamlines posses substantial curvature and divergence. Consequently, the
hydrostatic law of pressure distribution does not hold strictly for rapidly varied
flow.
16
Fig. 9a: Pressure distribution in large-slope channels
The above equation states that the pressure at any vertical depth y is equal to
the hydrostatic pressure ( ) multiplied by . If the angle is small
will tend to unity. This lead to a definition of channels of small and large
slopes. If the angle does not exceed 60 (i.e. channel slope not steeper than 1
in 10), the pressure distribution may safely be taken as hydrostatic. Such
channels are known as having small bottom slope.
On the other hand, if the channels are steeper than 1 in 10 ( > 60), the
hydrostatic pressure needs to be corrected for the effect of channel steepness
(i.e. by a correction factor ). These channels are known as large sloped
channels.
17
A’ h=dcos
A
h=ycos2 d=ycos
y
y
B B’ C dL
dcos
Pressure distribution,
on vertical section A’C C
B
Fig. 9b: A straight sloping channel of unit width and slope angle
Where d = ycos, is the depth measured perpendicular from the water surface in
other words the depth of flow section.
Horizontal line parallel to datum
dH = hL1-2
V2/2g Energy line slope, Sf
y
dcos d
z
dx
Datum
18
1 2
Fig. 10: Geometry for derivation of the gradually varied flow equation
From the geometry in the above figure, h = ycos2 does not apply strictly to
varied flow, particularly when is very large, h becomes bigger whereas h =
dcos still applies.
The equation h = ycos2 states that the pressure head at any vertical depth is
equal to this depth (y) multiplied by a correction factor cos2. Apparently, if the
angle is small, this factor will not differ appreciably from unity.
The correction tends to decrease the pressure head by an amount less
than 1% until is nearly 6o, a slope of about 1 in 10. Since the slope of
ordinary channels is far less than 1 in 10, the correction for slope
effect can usually be safely ignored. However, when the channel slope
is large and its effect becomes appreciable, the correction should be
made if accurate computation is desired. A channel of this type, say,
with a slope greater than 1 in 10, is hereafter called a channel of large
slope.
For a channel of large slope, where a longitudinal vertical profile of
appreciable curvature, the pressure head should be corrected for the
effect of the curvature of streamlines. Simply, the pressure head may be
expressed as ’ ycos2, where ’ is the pressure coefficient.
The need for flow equations arises following development from the industrial
revolution and hence great constructional works. A number of empirical formulae
19
that fit field and experimental data were developed and published. Only the three
most important ones are considered here: Chezy, Darcy-Weisbach and
Manning formula. Of these the Manning formula is the one that is now regarded
as the most appropriate (e.g. it is robust even to non-uniform flow conditions).
(4.1.1)
y1 = y2 P1 = P2 V1 = V2, which affects Fs as it is equal to q(V2 – V1)
For uniform flow, the gravity (ALS) and resistance (oPL) forces are in
equilibrium so that oPL = ALS
20
y1
V1 Wsin
y2
V2
W Rf
The shear stresses are assumed proportional to the square of the mean
velocity.
L
P
Where o is shear stress at the boundary (assumed constant over the surface or
a mean value), P is wetted perimeter and L is length of the channel.
Rf = V2PL (4.1.2)
Substitute Eq. 4.1.1 in Eq. 4.1.2:
21
ALS = V2PL
; R = Hydraulic radius
as
(1) Ganguillet & Kutter (swiss Engineers) formula (1869) in Metric units:
22
(2) Bazin formula (1897) (French Hydraulic Engineer) in English units:
or in Metric units,
The equation is less satisfactory than the Ganguillet & Kutter form because of
calibration from small experiment channels.
(3) The Powel formula (1950) which is a logarithmic formula for the roughness of
artificial channels (Implicit function of C) in English units:
(4) From the Manning formula (1889) (Irish Engineer) in Metric units:
The equation is obtained by comparing Manning and Chezy’s formula:
23
Where n is coefficient of surface roughness almost identical numerically with
Kutter’s n.
later.
L P
Where o is shear stress at the boundary (assumed constant over the surface or
a mean value), P is wetted perimeter and L is length of the channel.
For equilibrium:
ALS = oPL
24
o = RS (4.1.3)
Where, is the friction factor, h f is the friction loss, V is the mean velocity and D
is the pipe diameter.
25
So
The formula is widely used today (because of its simplicity and therefore straight
forward application, compares well within the normal limits of accuracy of
discharge measurement) and is also applicable to pipes.
SI units (Metric):
English units:
Where n is dimensionless.
26
Developed empirically from 7 different formulas, based on Bazin’s
experimental data and further verified by 170 observations.
Most widely used of all uniform-flow formula for open-channel flow (simple
and practical)
For practical purposes Manning’s n and Kutter’s n can be considered
identical for S 0.0001 and R between 1.0 and 30.0 ft.
Determination of Manning’s n:
27
A channel section is said to be the most efficient if it gives the maximum
discharge for the given shape, slope, area and roughness. The most efficient
cross-section offers least resistance to flow.
The most efficient section is usually the most economical section. The
cost of construction depends to a large extent on the quantity of
excavation. The quantity of excavation is proportional to the area. The
cost would be a minimum when the excavation is minimum for a given
discharge. In other words, for a given excavation, when the discharge
is a maximum. This condition is the same as that for the most efficient
section.
Moreover, for the channel of the most efficient cross-section, because the
wetted perimeter is a minimum, the cost of the channel grading and
lining, which depends on the wetted perimeter, is also a minimum.
5.1 Formulation:
28
From the Manning’s formula:
Common shapes:
29
Rectangular
Trapezoidal
Triangular
A = bd; P = b + 2d
For a given A, P = A / d + 2d
For minimum P, dP / dd = -A /d2 + 2 = 0
bd / d2 = 2 or b = 2d
Therefore, the best proportions for a rectangular channel are such that the
inscribed semi-circle is tangential to the bed and sides (best rectangular
section).
d 1
30
b
If the side slope is s horizontally to 1 vertically, then
With given A,
Top width,
The most efficient hydraulic section is when the top width is twice the length
of a sloping side.
Example:
A trapezoidal channel carries a discharge of 2.5 cumecs. Design the section if
the slope is 1 in 1200 and the side slopes are 1 in 1. Use Chezy’s formula, C =
55.
Solution:
For the trapezoidal section to be the most efficient, R = D / 2 and from above
, Substituting in the equation for area A, gives:
Q = AV
31
From
32