SCMD 09 CH07 Flow in Open Channels

STORMWATER CONVEYANCE
MODELING AND DESIGN

Authors
Haestad Methods
S. Rocky Durrans

Managing Editor
Kristen Dietrich

Contributing Authors
Muneef Ahmad, Thomas E. Barnard,
Peder Hjorth, and Robert Pitt

Peer Review Board
Roger T. Kilgore (Kilgore Consulting)
G. V. Loganathan (Virginia Tech)
Michael Meadows (University of South Carolina)
Shane Parson (Anderson & Associates)
David Wall (University of New Haven)

Editors
David Klotz, Adam Strafaci, and Colleen Totz
HAESTAD PRESS
Waterbury, CT USA

Click here to visit the Bentley Institute
Press Web page for more information

C H A P T E R
7
Flow in Open Channels
Open channels are ubiquitous in both natural and constructed environments. They
consist of natural stream channels as well as natural swales and depressions along
which water runs following rainfall events. In the constructed or urbanized environ-
ment, an open channel may take the form of a street gutter, drainage ditch, pipe (if
there is a free water surface at atmospheric pressure), lined or unlined stormwater col-
lection channel, or natural or modified stream channel.
In open-channel flow, unlike closed-conduit flow, the cross-sectional area of flow is
not a constant that is governed by the shape and size of the conveyance. Rather, the
cross-sectional area and associated flow descriptors, such as the average flow veloc-
ity, are unknowns and must be determined as part of the solution to a problem. Further
complicating analyses of open-channel flow is the fact that there is not, in general, a
unique or one-to-one relationship between discharge Q and depth of flow y in an open
channel. In most instances of open-channel flow, the depth of flow for any specified
channel discharge depends on channel shape and roughness characteristics, as well as
on specified boundary conditions at flow control sections (that is, cross sections where
a known relationship between depth and discharge exists). Thus, open-channel flow is
analytically much more difficult to treat than is pressure flow, but the challenge
imposed by the additional difficulty also makes it more interesting.
This chapter presents methods for analysis of the hydraulics of steady flows in open
channels. Unsteady flows are viewed as an advanced topic and are introduced in
Appendix C. The reader may consult Chow (1959); Abbott (1979); Cunge, Holly, and
Verwey (1980); French (1985); Fread (1993); and Chaudhry (1993) for more detailed
coverage of unsteady flows in open channels.
Section 7.1 of this chapter describes the classification of the types of open channels
and gives an overview of their geometric properties. Energy and constitutive relation-
ships for the estimation of energy losses in open-channel flow are presented in Sec-
tion 7.2.
Sections 7.3 and 7.4, respectively, deal with the fundamentally important concepts of
uniform flow and critical flow in open channelsthe only types of flow for which
there is a one-to-one relationship between channel discharge and flow depth. Uniform
218 Flow in Open Channels Chapter 7
flow is the basis for many types of channel design; the subject of design is taken up in
detail in Chapter 8. The concept of critical flow is fundamental to the analysis of var-
ied flows in open channels and to understanding many types of flow controls. Estab-
lishing critical flow in an open channel can also provide a basis for flow
measurement.
The momentum principle, as applied in the analysis of open-channel flow, is taken up
inSection 7.5. As in pipe flow, the momentum principle is applied in situations where
the engineer desires an estimate of the hydraulic forces acting on a flow obstruction
such as a bridge pier. The momentum principle is also fundamental to the analysis of
the rapidly varied flow phenomenon known as the hydraulic jump.
Sections 7.6 and 7.7 address the topics of weirs and orifices, respectively, which are
commonly used structures for flow control and measurement. Weirs and orifices may
be used as outlets for stormwater detention facilities. The weir and orifice equations
also play a role in describe the hydraulic performance of stormwater inlets and culvert
entrances.
Gradually varied flow is discussed in Section 7.8. Gradually varied flows are those in
which the depth of flow changes along the length of a channel as a consequence of
convective accelerations of the flow. These accelerations occur where flow controls
are such that the flow depth at the control is different from the depth of uniform flow
and forces on the water are not balanced. Such situations are common upstream of
culvert entrances, at channel entrances and exits, and at locations where channel geo-
metric properties change.
Spatially varied flowsflows for which the discharge is not a constant along the
length of a channelare discussed briefly in Section 7.9. These flows may have
either an increasing or a decreasing discharge in the direction of flow, and different
methods exist for the analysis of each flow type. Spatially varied flows with a
decreasing discharge occur, for example, in gutters at stormwater inlet locations. Spa-
tially varied flows with an increasing discharge may occur in side overflow spillway
channels. Flow in a natural stream channel is usually spatially varied because of the
increase in contributing drainage basin area as one proceeds downstream.
7.1 CLASSIFICATION AND PROPERTIES OF OPEN
CHANNELS
Open channels may be natural or man-made and come in a wide variety of cross-sec-
tional shapes. For analytical purposes, channels may be classified as prismatic or non-
prismatic. In a prismatic open channel, the descriptive geometric variables such as its
width, side slopes, and longitudinal slope, do not change along the length of the chan-
nel. Practically speaking, one would typically not expect to encounter a perfectly pris-
matic channel outside of the laboratory, but many constructed channels can, for the
purposes of analysis, be considered prismatic. A concrete curb and gutter along the
edge of a paved roadway with a constant slope is one example of a prismatic channel.
Another example is shown in Figure 7.1.
Section 7.1 Classification and Properties of Open Channels 219
Figure 7.1
A riprap-lined
prismatic channel
In a nonprismatic channel the descriptive geometric variables (shape, roughness, etc.)
change with location along the channel length. Natural stream channels, which typi-
cally have variations in width and other properties, are examples of nonprismatic open
channels (see Figure 7.2). Coverage of open-channel flow in this text deals primarily
with prismatic channels. Nevertheless, certain principles associated with the hydrau-
lics of flow in nonprismatic channels will be addressed as necessary.
Open channels can also be classified on the basis of whether they are lined or unlined.
In a lined channel, the channel bottom and sides have been paved with an erosion-
resistant material, such as concrete, asphalt, metal, riprap, or vegetation. An unlined
channel is one in which the channel bottom and sides consist of natural geologic
materials such as earth, gravel, or rock. As shown in Chapter 8, the lining material (or
lack thereof) has practical significance in the design of the open channel.
Geometric Properties of Open Channels
The longitudinal slope of an open channel, denoted by S
o
, is a critically important
geometric variable. An alternative way of describing the channel slope is by use of its
angle of inclination above the horizontal, denoted by u (seeFigure 7.3). Note that S
o
=tanu. Note also that if u is small (less than about 5 degrees), then S
o
~ sinu. A hori-
zontal open channel has a longitudinal slope of zero, and an adverse open channel has
a negative slope (that is, the channel bottom elevation increases in the downstream
direction).
Figure 7.2
A nonprismatic
natural channel
Figure 7.3
Depth of flow
measurements
Most geometric properties of open channels are those related to a channels cross sec-
tion. Strictly speaking, channel cross sections should be taken perpendicular to the
channel bottom (that is, normal to the direction of flow). The depth of flow measured
Section 7.1 Classification and Properties of Open Channels 221
perpendicular to the channel bottom is denoted by d in Figure 7.3. However, in most
open channels, the slope is small, and vertical channel cross sections can be employed
with little error. The depth of flow measured vertically is denoted by y (see Figure
7.3). The relationship between these depths is d =ycosu.
The most common cross-sectional shapes of prismatic channels are rectangular, tri-
angular, trapezoidal, and circular. Other shapes are those associated with curbs and
gutters, which are discussed in Section 10.1. Figure 7.4 illustrates the dimensions
used to describe common channel shapes. B denotes the channel bottom width, y
denotes the depth of flow, D denotes the diameter of the channel (if a circular pipe), T
denotes the top width of flow, and z represents a measure of steepness of the channel
sides. For circular channels, | represents the angle subtended by the intersections of
the water surface with the channel or conduit surface, as shown in the figure. Note
that rectangular and triangular channels are merely special cases of a trapezoidal
channel, the former in the case of z =0, and the latter in the case of B =0.
The cross-sectional area, wetted perimeter, hydraulic radius, top width T, and hydrau-
lic depth can be determined on the basis of the depth of flow y and the properties
shown in Figure 7.4. The hydraulic radius is defined as
R = A/P (7.1)
where R =hydraulic radius (ft, m)
A =cross-sectional area (ft
2
, m
2
)
P =wetted perimeter (ft, m)
The definition of hydraulic depth is
D
h
= A/T (7.2)
where D
h
=hydraulic depth (ft, m)
T =top width (ft, m)
Table7.1 summarizes the relationships for each of the channel cross-sectional shapes
shown in Figure 7.4. Note that the formulas from triangular and trapezoidal channels
are only applicable to cross sections with equal side slopes, and the depth y in an open
channel is always taken as the greatest depth at any point in a cross section. The
hydraulic depth can be thought of as the average depth of flow in a cross section. Use
of the hydraulic depth can also be viewed as a type of shorthand, as the ratio D
h
=A/T
occurs frequently in equations describing open-channel flow.
The cross-sectional properties of irregularly shaped channels can be determined by
dividing a cross section into a set of rectangles and/or triangles and applying basic
geometric formulae related to areas, perimeters, and slopes. Clearly, this process is
quite tedious and time-consuming; for that reason, computer-based methods for solv-
ing open-channel flow problems are especially useful for irregularly shaped channels.
Figure 7.4
Dimensions of
common channel
shapes
Table 7.1 Cross-sectional properties of prismatic open channels
Channel Shape Area (A) Wetted Perimeter (P) Hydraulic Radius (R) Top Width (T)
Hydraulic Depth (D
h
)
Rectangular By B +2y B y
Trapezoidal B +2zy
Triangular
zy
2
2zy
Circular
a
a. Angle | is measured in radians. 1 radian ~ 57.3 degrees.
2
By
B y +
2
By zy +
2
2 1 B y z + +
2
2 1 B y z + +
2
2
By zy
B zy
+
+
2
2 1 y z +
2
2 1
zy
z + 2
y
2
( sin )
8
D | |
2
D| sin
1
4
D |
|
| |
|
\ .
sin
2
D
| | |
|
\ .
sin
8 sin( / 2)
D | |
|
| |
|
\ .
Section 7.2 Energy and Constitutive Relationships 223
Example 7.1 Geometric Properties for a Trapezoidal Channel. An open channel
has a cross-sectional shape as shown in Figure E7.1.1. If the depth of flow is 4 ft, determine the cross-
sectional area, wetted perimeter, hydraulic radius, top width, and hydraulic depth.
Figure E7.1.1 Open channel for Example 7.1
Solution: Because the cross-sectional shape of the channel does not coincide with any of those shown
inFigure 7.4, the properties of the channel section must be derived on the basis of its geometric prop-
erties.
The cross-sectional area can be constructed as the sumof the areas of a rectangle and a triangle, and is
A =12(4) +0.5(4)[2(4)] =64 ft
2
The wetted perimeter can be computed as the sumof the water depth against the right-hand side of the
channel, the channel bottomwidth, and the length along the sloping side of the channel:
The hydraulic radius is
R = A/P =64/24.94 =2.57 ft
The top width is equal to
T =12 +2(4) =20 ft
and the hydraulic depth is
D
h
= A/T =64/20 =3.20 ft
7.2 ENERGY AND CONSTITUTIVE RELATIONSHIPS
In open-channel flow problems, the energy equation takes a somewhat different form
than that presented in Equations 3.9 and 6.1. The difference arises in the pressure
head term, which, instead of being written as p/, is written as dcosu, where d is the
depth of flow normal to the channel bottom and u is the angle of inclination of the
channel. Thus, the energy equation for flow in an open channel is
(7.3)
2 2
4 12 4 8 24.94 ft P = + + + =
2 2
1 1 2 2
1 1 2 2
cos cos
2 2
L
V V
Z d Z d h
g g
o o
u u + + = + + +
where Z =channel bottom elevation (ft, m)
d =depth of flow normal to the channel bottom (ft, m)
u =channel slope angle
o =a velocity distribution coefficient defined by Equation 6.3
V =average flow velocity (ft/s, m/s)
g =gravitational acceleration constant (ft/s
2
, m/s
2
)
h
L
=energy loss between section 1 and section 2 (ft, m)
Figure 7.5 depicts the terms in Equation 7.3, as well as the next two equations.
Table7.2 gives values of o and | (for an explanation of |, see Section 7.5) for open-
channel flows. (Estimation of o for compound channel sections is presented in Sec-
tion 7.3.)
Becaused =ycosu (see Figure 7.3), an alternative expression for the energy equation
for open-channel flow is
(7.4)
wherey
1
and y
2
are the flow depths (measured vertically) at sections 1 and 2, respec-
tively (ft, m).
When the angle of inclination, u, is small, this equation simplifies to the version most
commonly applied in practice:
(7.5)
Figure 7.5
Definition figure for
energy equation for an
open channel
2 2
2 2 1 1 2 2
1 1 2 2
cos cos
2 2
L
V V
Z y Z y h
g g
o o
u u + + = + + +
2 2
1 1 2 2
1 1 2 2
2 2
L
V V
Z y Z y h
g g
o o
+ + = + + +
Pressure Variations in Open-Channel Flow
Replacement of the pressure head term, p/, with dcosu in Equation 7.3 means that the
fluid pressure at depth d in open-channel flow is equal to
(7.6)
where p =fluid pressure at depth d (lb/ft
2
, N/m
2
)
d =depth measured normal to the flow streamlines (ft, m)
=fluid specific weight (lb/ft
3
, N/m
3
)
If one desires instead to use the vertical depth, y, the pressure is found to be
(7.7)
Because cosu is less than unity in all nonhorizontal channels, this expression indicates
that the pressure at depth y in an open-channel flow is less than the pressure at the
same depth in a static fluid.
Example 7.2 Pressure Variation over a Spillway. Flow occurs on a spillway with a
depth d =1.00 m. The slope of the spillway is 15 percent. Determine the fluid pressure on the face of
the spillway and compare that pressure to the pressure existing at a depth of 1.00 min standing water.
Solution: At a depth of 1.00 min standing water, the pressure is found to be
p = d =1,000(1.00) =1,000 kg/m
2
The pressure at the same depth on the face of the spillway can be evaluated using Equation 7.6, where
u =arctanS
o
=arctan(0.15) =8.53.
p =1,000(1.0)(cos 8.53) =988.9 kg/m
2
The pressure reduction caused by the slope of the channel is slightly more than 1 percent in this case.
Constitutive Relationships for Open-Channel Flow
Energy losses represented by the h
L
term in the energy equation can be viewed as con-
sisting of localized and distributed losses. Local losses, often referred to as minor
losses, occur at locations where the flow velocity changes in either magnitude or
direction. They are described later in this section on page 230. Distributed losses are
Table 7.2 Typical values of o and | in open-channel flow (Chow, 1959)
Channel Description
o |
Min Avg Max Min Avg Max
Regular channels, flumes, spillways 1.10 1.15 1.20 1.03 1.05 1.07
Natural streams and torrents 1.15 1.30 1.50 1.05 1.10 1.17
Rivers under ice cover 1.20 1.50 2.00 1.07 1.17 1.33
River valleys, over flooded 1.50 1.75 2.00 1.17 1.25 1.33
cos p d u =
2
cos p y u =
due to friction exerted on the flow by the channel bottom and sides. Distributed losses
may be described by expressing the head loss for a channel reach as the product of the
energy gradient, S
f
, and the length, L, of the reach:
h
L
= S
f
L (7.8)
where S
f
=energy gradient (ft/ft, m/m)
L =reach length (ft, m)
A number of constitutive relationships are available to express the energy gradient as
a function of flow and channel characteristics. The most commonly used relationships
are described in the subsections that follow.
The Manning Equation. TheManning equation (Manning, 1891) is an empiri-
cally derived expression obtained from observations of flows in laboratory channels.
It was developed as a formula for uniform flow but has also been successfully applied
for analyses of gradually and spatially varied flows. It is the most widely used of all
constitutive relationships for open-channelflow analysis.
The Manning equation can be written as
(7.9)
where V =cross-sectionally averaged flow velocity (ft/s, m/s)
C
f
=unit conversion factor (1.49 for US customary units, 1.00 for SI)
n =Mannings roughness coefficient
R =hydraulic radius of the flow (ft, m)
Note that S
f
is equal to the channel slope S
o
in the case of uniform flow, and is differ-
ent from S
o
in the case of varied flow. The numerical value of Mannings n is indepen-
dent of the set of units employed.
Table7.3 provides a listing of typical values of Mannings n for various channel types
and surfaces. Note that n is not a constant for a particular channel; it depends on a
number of factors, including flow depth, channel alignment and irregularity, the pres-
ence of vegetation and flow obstructions, and the amount of sediment being trans-
ported by the stream. Chow (1959) provides a detailed discussion of the effects of
these factors.
Because computations of flow velocities and energy gradients via the Manning equa-
tion are fairly sensitive to the assumed value of n, an understanding of influential fac-
tors should enable one to make more informed choices regarding this important
parameter. For example, Chow (1959), in a theoretical discussion of channel rough-
ness and uniform flow formulas, showed that the value of Mannings n is related to
the channel roughness height as:
(7.10)
2/ 3 1/ 2
f
f
C
V R S
n
=
1/ 6
R
n k
k
| |
=
|
\ .
where (R/k) =a value that is a function of the dimensionless ratio of the hydraulic
radius R to the roughness height k
k =channel roughness height (ft, m)
Table 7.3 Typical values of Mannings n (adapted from Chow, 1959)
Channel Description Minimum n Average n Maximum n
Closed conduits flowing partly full
Brickwork 0.011 0.014 0.017
Cast iron, coated 0.010 0.013 0.014
Cast iron, uncoated 0.011 0.014 0.016
Clay, vitrified 0.011 0.014 0.017
Concrete, finished 0.011 0.012 0.014
Concrete, unfinished 0.012 0.014 0.020
Corrugated metal 0.021 0.024 0.030
Rubble masonry 0.018 0.025 0.030
Steel, welded 0.010 0.012 0.014
Steel, riveted and spiral 0.013 0.016 0.017
Wood stave 0.010 0.012 0.014
Lined or built-up channels
Asphalt 0.013 0.016
Brickwork 0.011 0.014 0.018
Concrete, troweled 0.011 0.013 0.015
Concrete, unfinished 0.014 0.017 0.020
Corrugated metal 0.021 0.025 0.030
Rubble masonry 0.017 0.029 0.035
Steel, smooth 0.011 0.013 0.017
Wood, planed and creosoted 0.011 0.012 0.015
Wood, planed and untreated 0.010 0.012 0.014
Excavated or dredged channels
Earth, straight and uniform 0.016 0.022 0.033
Earth, winding and sluggish 0.023 0.035 0.050
Overgrown with vegetation 0.040 0.080 0.140
Rock cuts, smooth and uniform 0.025 0.035 0.040
Rock cuts, jagged and irregular 0.035 0.040 0.050
Natural streams
Flood plains, pasture 0.025 0.033 0.050
Flood plains, cultivated 0.020 0.035 0.160
Flood plains, trees 0.050 0.080 0.200
Major streams (width >100 ft) 0.025 0.100
Minor streams on plain areas 0.025 0.045 0.150
Mountain streams 0.030 0.045 0.070
The value of the function (R/k) does not vary much over a wide range of R/k. Strick-
ler (1923) showed that, on average, (R/k) =0.0342 when the roughness height (in ft)
is taken as the median grain size of the channel bed material. (When the median grain
size is expressed in units of meters, (R/k) =0.0281.) This relationship indicates that
Mannings n is proportional to the one-sixth power of the median sediment size in a
stream.
Example 7.3 Determining the Energy Gradient with the Manning Equation.
A discharge of 300 cfs occurs at a depth of 2.50 ft in a concrete-lined rectangular open channel with a
width of 10 ft and n =0.013. What is the energy gradient S
f
for this flow? If the channel bottomslope
isS
o
=0.001, is this flow uniformor varied?
Solution: The cross-sectional area, wetted perimeter, and hydraulic radius of the flow are
A =10(2.50) =25 ft
2
P =10 +2(2.50) =15 ft
R =A/P =25/15 =1.67 ft
V =Q/A =300/25 =12 ft/s
Substituting these values into the Manning equation and rearranging yields the energy gradient:
Because the energy gradient is not equal to the channel slope, the flow must be varied.
The Chzy Equation. The Chzy equation, developed around 1769 by the
French engineer Antoine Chzy, is probably the oldest constitutive relationship for
open-channel flow (Herschel, 1897). Unlike the Manning equation, the Chzy equa-
tion can be derived theoretically by equating the force resisting a flow to the gravita-
tional force driving the flow (a condition that must be true for uniform flow). Because
the resisting force is proportional to the square of the flow velocity, the net result is
the expression
(7.11)
where C =a roughness coefficient known as Chzys C (ft
0.5
/s, m
0.5
/s)
Like Mannings n, Chzys C depends on a number of factors and thus is not a con-
stant. In 1869, the Swiss engineers Ganguillet and Kutter (1869) published a formula
relating C to the roughness, slope, and hydraulic radius of a channel as
(7.12)
where S =channel slope (ft/ft, m/m)
n =Kutters n
k =a constant equal to 1 for U.S. Customary units or 1.811 for SI units
2 2
2 4/ 3 2 4/ 3
( ) [0.013(12)]
0.00555
1.49 1.49 (1.67)
f
nV
S
R
= = =
f
V C RS =
0.00281 1.811
41.65
0.00281
41.65
S n
C
n
k
S R
+ +
=
| |
+ +
|
\ .
For practical purposes, Kutters n is the same as Mannings n when the channel slope
is larger than 0.0001 and the hydraulic radius is between 1 and 30 feet (0.3 and 9 m).
Bazin (1897) published a formula for C in which slope is not considered. His equation
is
(7.13)
where m =a roughness coefficient (see Table 7.4)
Although considerably simpler than the Ganguillet and Kutter (1869) formula,
Bazins is considered to be less reliable.
Powell (1950) suggested that Chzys C can be expressed as
(7.14)
where C =Chzys C (ft
0.5
/s)
Re =Reynolds number
c =measure of roughness having the tentative values shown in Table 7.5
R =hydraulic radius (ft)
For open-channel flow, the Reynolds number can be computed as
(7.15)
where v =fluid kinematic viscosity (ft
2
/s, m
2
/s)
=fluid density (slugs/ft
3
, kg/m
3
)
=fluid dynamic viscosity (lbf-s/ft
2
, N-s/m
2
)
The reader may want to compare Equation 7.15 to Equation 6.5, which is used to
compute the Reynolds number in the case of a circular pipe.
Table 7.4 Proposed values of Bazins m (Chow, 1959)
Channel Description m
Very smooth cement or planed wood 0.11
Unplaned wood, concrete or brick 0.21
Rubble masonry or poor brickwork 0.83
Earth channels in good condition 1.54
Earth channels in ordinary condition 2.36
Earth channels in rough condition 3.17
157.6
C
m
k
R
=
+
10
42log
4Re
C
C
R
c | |
= +
|
\ .
Re
VR VR
v
= =
.
The Powell formula, being an implicit function of C, requires a trial-and-error solu-
tion and therefore is less convenient than other approaches. It should be used with
caution, as further research is required for determination of proper values of the
roughness parameter c.
Example 7.4 Determining the Energy Gradient with Chzys Equation. Repeat
Example 7.3 using the Chzy equation to compute the energy gradient. Use the Ganguillet and Kutter
formula to determine the value of Chzys C. In Example 7.3, a discharge of 300 cfs occurs at a depth
of 2.50 ft in a concrete-lined rectangular open channel with a width of 10 ft and n =0.013.
Solution: Using n =0.013 and R =1.67 ft fromExample 7.3, the Ganguillet and Kutter formula may
be stated as
Substitution of this expression, V =12 ft/s, and R =1.67 ft into Equation 7.11 leads to an expression
that is implicit in S
f
. Solving by trial and error,
S
f
=0.00534
Again, because the energy gradient is not equal to the channel slope, the flow must be varied.
Minor Losses in Open-Channel Flow
As indicated in Section 6.3, the classical method of predicting minor energy losses
caused by flow expansions and contractions is to express the energy loss as being pro-
portional to the absolute difference between the upstream and downstream velocity
heads:
(7.16)
where h
L
=energy loss (ft, m)
K =minor loss coefficient
Table 7.5 Tentative values of Powells c (Chow, 1959)
Channel Description c
Neat cement 0.00020.0004
Unplaned plank flumes 0.00100.0017
Concrete-lined channels 0.0040.006
Earth, straight and uniform 0.04
Dredged earth channels 0.10
0.00281
180.96
0.0000283
1.42
f
f
S
C
S
+
=
+
2 2
2 1
2
L
V V
h K
g
=
Section 7.3 Uniform Flow and Normal Depth 231
The minor loss coefficient, K, depends on the geometry of the flow contraction or
expansion. For channel transitions such as entrances and exits to and from flumes,
tunnels, or siphons, Chow (1959) presents the recommended design values shown in
Table7.6. Note that the loss coefficient for a flow expansion is always greater than the
coefficient for a flow contraction.
In non-prismatic channels, where the cross-sectional area and hence the flow velocity
may change from one cross section to another, eddy losses occur and can be predicted
using Equation 7.16. For gradually converging and diverging reaches, respectively,
recommended values of K are about 0.1 and 0.3 [Chow, 1959; U.S. Army Corps of
Engineers (USACE), 1991].
7.3 UNIFORM FLOW AND NORMAL DEPTH
Uniform flow in an open channel occurs when the depth, velocity, cross-sectional
area, and discharge are constant along the channel length. As a practical matter, this
condition can exist only in a prismatic channel, and only if the flow is steady. When
uniform flow does exist, the constancy of the flow variables (such as velocity and
depth) can be used to recognize that the slopes of the channel bottom (S
o
), water sur-
face (S
w
), and energy grade line (S
f
) are all equal.
Uniform flow occurs in a channel reach between two cross sections when all forces
acting on the water in the reach are in balance and there is no flow acceleration. With
reference to Figure 7.6, these forces are the hydrostatic pressure forces, P
1
and P
2
,
acting at the upstream and downstream ends of the reach, the frictional force, F
f
,
opposing the flow, and the gravitational force, Wsinu (where W is the weight of the
water), that is driving the flow. Because the flow is uniform, P
1
and P
2
must be equal
and the frictional force must be equal to the driving force, or F
f
=Wsinu.
If the frictional force is expressed as the product of a shear stress and the area of the
channel bottom and sides to which that stress is applied, one can write F
f
=tPL,
wheret is the shear stress, P is the wetted perimeter of the flow, and L is the length of
the channel reach (Figure 7.6). Similarly, if the weight of the water is expressed as
W = AL, where A is the cross-sectional flow area, and if the slope S
o
is approximated
by sinu, then Wsinu =ALS
o
. Equating the frictional and driving forces leads to Equa-
tion 7.17:
t = RS
o
(7.17)
Table 7.6 Minor loss coefficients for open-channel flow (Chow, 1959)
Type of Channel Transition Contraction Expansion
Cylinder quadrant 0.15 0.25
Square-ended 0.30 or more 0.75
Straight line 0.30 0.50
Warped 0.10 0.20
where t =average shear stress (lb/ft
2
, N/m
2
)
=fluid specific weight (lb/ft
3
, N/m
3
)
R =hydraulic radius (ft, m)
S
o
=channel slope (ft/ft, m/m)
Thus, when flow is uniform, Equation 7.17 gives the average shear stress acting on
the channel bed and sides. This parameter is important in the design of unlined chan-
nels, as it has a direct bearing on whether the flow will mobilize soil particles on the
channel boundaries.
Other than its use for the derivation of Equation 7.17, the concept of the force balance
that must exist in uniform flow illustrates that uniform flow represents a condition of
equilibrium. The depth at which this equilibrium condition is reached is called the
normal depth and is denoted by y
n
. The normal depth depends on the discharge, chan-
nel cross-sectional geometry, channel roughness, and channel slope.
If a flow control exists that causes the actual depth of flow at a cross section to be dif-
ferent from the normal depth, then a varied flow profile will exist as a transition
between the depth at the flow control and the normal depth in the channel. In other
words, despite the existence of the flow control, there will be a region of the channel
in which the flow adjusts itself to seek out its equilibrium condition. Depending on
the length of the channel, the flow may or may not have the opportunity to actually
attain its normal depth condition.
Figure 7.6
Forces on a fluid
element for a uniform
flow condition
Uniform Flow in Simple Channels
A simple open channel is one with constant roughness and without overbank (or
floodplain) areas. For the purpose of this text, a simple channel will also be taken as
one without a gradually closing top, such as occurs in a pipe. In a simple open chan-
nel, the relationships between flow variables under conditions of uniform flow are
given by a constitutive relationship for which it is assumed that the energy gradient S
f
is equal to the channel bottom slope S
o
. As discussed in Section 7.2, the most com-
mon constitutive relationships are the Manning and Chzy equations.
If either the Manning or the Chzy equation is multiplied by the cross-sectional area A
of flow, it can be expressed in terms of the discharge rather than the velocity of the
flow. That is,
(7.18)
where Q =discharge (cfs, m
3
/s)
C
f
=is a conversion factor equal to 1.49 for U.S. Customary units or 1.00 for
SI units
A = cross-sectional area of flow (ft
2
, m
2
)
n =Mannings roughness coefficient
or
(7.19)
where C =Chzys C (ft
0.5
/s, m
0.5
/s)
It can be seen from these expressions that, for a simple channel of fixed slope and dis-
charge, there is a unique value of flow depth (recall that R and A depend on depth)
that will balance each equation. That unique depth is the normal depth y
n
. Similarly, if
the slope and normal depth are fixed, there is a unique discharge that results. Finally,
if the discharge and normal depth are fixed, there is a unique channel slope that results
(recognizing that the slope represented by the negative square root can be disre-
garded). That slope is called the normal slope of the channel for the given discharge
and depth, and is denoted by S
n
.
Computation of uniform flow involves determining the depth, discharge, or slope that
satisfies Equation 7.18 or 7.19 when the other variables are specified. As illustrated in
the following examples, determination of the normal depth generally requires a trial-
and-error solution. Determination of either the discharge or the normal slope can be
accomplished without iterating.
Example 7.5 Determining Discharge with the Manning Equation. A trapezoidal
open channel has a bottomwidth B =10 ft, side slopes with z =3 (see Figure E7.5.1), a slope S
o
=
0.002, and n =0.035. Determine the discharge in the channel if the flow is uniformand its depth is
3.20 ft. Use the Manning equation.
2/ 3 1/ 2
f
o
C
Q AR S
n
=
o
Q CA RS =
Figure E7.5.1 Trapezoidal channel for Example 7.5
Solution: Fromthe expressions given in Table7.1, the cross-sectional area and wetted perimeter of
the channel can be computed as
A =10(3.20) +3(3.20)
2
=62.72 ft
2
P =10 +2(3.20)(1 +3
2
)
1/2
=30.24 ft
The hydraulic radius is therefore R =62.72/30.24 =2.07 ft. Substitution into Equation 7.18 yields
Example 7.6 Determining Normal Slope with the Manning Equation. Deter-
mine the normal slope of the channel in Example 7.5 if the discharge is Q = 350 cfs.
Solution: By algebraic rearrangement, Equation 7.18 can be expressed in the form
Substituting the given values, the normal slope is
S
n
=0.00651
Example 7.7 Determining Normal Depth with the Manning Equation. A rect-
angular open channel has a width of B = 3.6 m, n =0.013, and S
o
=0.015. Determine the normal
depth of flow if the discharge is Q =7.0 m
3
/s.
Solution: The cross-sectional area of flow can be written as
A = 3.6y
The wetted perimeter can be written similarly as
P = 3.6 +2y
Substitution of these expressions into Equation 7.18 yields
or
By trial and error, the value of y that balances this equation is y
n
=0.42 m.
2/ 3 1/ 2
1.49
(62.72)(2.07) (0.002) 194 cfs
0.035
Q = =
2
2/ 3
1.49
n
nQ
S
AR
| |
=
|
\ .
Q 7.0
1.00
0.013
------------- 3.6y ( )
5 3
3.6 2y + ( )
2 3
0.015
1 2
= =
3.6y ( )
5 3
3.6 2y + ( )
2 3
0.74 =
Uniform Flow in a Channel with a Gradually Closing
Top
A pipe, whether it is circular, elliptical, or another shape, is an example of a channel
with a gradually closing top. In a channel with a gradually closing top, there may be
more than one depth of flow that satisfies the normal depth criterion of equilibrium.
Furthermore, the maximum discharge that can be conveyed by the channel under uni-
form flow conditions occurs at a depth of flow smaller than that corresponding to full
flow in the conduit.
Figure 7.4on page 222 illustrates a circular pipe of diameter D in which the depth of
flow is denoted by y. The angle | defined by the edges of the free water surface is also
shown. Note that | =0 when y =0, and that | =2t radians when y = D. The relation-
ship between y, D, and | (in radians) is given by
(7.20)
From Table7.1, the cross-sectional area and wetted perimeter are given by
(7.21)
(7.22)
Substitution of these expressions into the Manning equation yields
(7.23)
3
/s)
C
f
=unit conversion factor (1.49 for US customary units, 1.00 for SI)
n = Mannings n
D =diameter of pipe (ft, m)
| =angle defined by edges of free water surface of circular channel (see
Figure 7.4)
For chosen values of D, n, and S
o
, Equation 7.23 may be applied to develop a graph of
Q as a function of | (or as a function of y by rearranging Equation 7.20 to solve for |
and substituting this expression into Equation 7.23). Thus, | (or y) could be obtained
from the graph if Q were specified. As shown in Figure 7.7, which illustrates such a
graph, two possible depths of flow (y-values) exist for any possible discharge greater
than or equal to the full-flow discharge (that is, the discharge when y = D). It can also
be seen from the graph that the maximum discharge occurs when y ~ 0.94D.
1 cos
2 2
D
y
| ( | |
=
( |
\ .

( )
2
sin
8
D
A | | =
2
D
P
|
=
( )
5/ 3
2/ 3
2
1/ 2
sin
8 2
f
o
C
D D
Q S
n
|
| |
(
| |
= (
|
\ . (

Although Figure 7.7 is useful for demonstrating that there may be more than one pos-
sible depth of uniform flow in a channel with a gradually closing top, it was devel-
oped for the case of a specific pipe diameter, roughness, and slope. Figure 7.8 depicts
a more useful dimensionless graph that can be used for any combination of pipe geo-
metric variables. In this figure, the relative depth of flow, y/D, is shown as a function
of the ratio of actual flow to full flow and the ratio of actual velocity to full velocity.
Example 7.8 illustrates the use of Figure 7.8.
Figure 7.7
Graph of discharge
versus normal depth
for a representative 1-
ft diameter pipe
Figure 7.8
Dimensionless graph
of discharge and
velocity fractions
versus depth fraction
for circular pipe
Example 7.8 Depth and Velocity in a Circular Pipe. A 36-in. diameter pipe with
n =0.013 and S
o
=0.01 conveys a discharge of 45 cfs. Determine the depth and velocity of flow in the
pipe. If the pipe flows at its maximumcapacity (y ~ 0.94D), what is the discharge?
Solution: For full-flow conditions, the pipes cross-sectional area and wetted perimeter are, respec-
tively, 7.07 ft
2
and 9.42 ft. Applying the Manning equation, the full-flow discharge is Q
f
=67 cfs. The
full-flow velocity is
V
f
=67/7.07 =9.46 ft/s
The actual discharge in the pipe is Q =45 cfs, so
Q/Q
f
=45/67 =0.67
FromFigure 7.8, the relative depth of flow is y/D =0.60. Thus, the actual depth of flow is
D(y/D) =3(0.60) =1.80 ft.
Usingy/D =0.60, Figure 7.8 shows that V/V
f
=1.09. Thus, the actual flow velocity is
V
f
(V/V
f
) =9.46(1.09) =10.3 ft/s
At maximumcapacity, the actual discharge is about 1.09 times the full-flow discharge (see Figure
7.8). Therefore, the maximumdischarge is
1.09(67) =73 cfs
Dimensionless graphs of geometric elements similar to the one shown in Figure 7.8
can be developed for pipe shapes other than circular. The Concrete Pipe Design Man-
ual (American Concrete Pipe Association, 2000) and similar publications developed
by associations of pipe manufacturers may be consulted to obtain graphs for other
common pipe shapes.
Uniform Flow in a Channel with Composite Roughness
In a simple channel with composite roughness, the roughness characteristics vary
from one part of the channel boundary to another. For example, a constructed pris-
matic channel might have an earthen bottom and riprap side slopes. A laboratory
channel might have a steel bottom and glass sides. In such channels, a uniform flow
formula such as the Manning equation can be applied if the n-value used is an equiva-
lent n-value computed to represent an average measure of the channel roughness.
There are several ways in which an equivalent n value may be determined. As shown
inFigure 7.9, a channel cross section may be subdivided into a number of subareas N,
each of which has its own roughness n
i
, area A
i
, wetted perimeter P
i
, hydraulic radius
R
i
, and average depth y
i
, where i =1, 2 .N. The wetted perimeter of each subarea is
computed using only that part of the perimeter that coincides with the channel bottom
or sides.
Figure 7.9
Division of channel
for computation of the
equivalent roughness
coefficient
If one assumes that the velocity in each subarea is equal to the mean velocity for the
cross section as a whole, the equivalent n value may be determined as (Horton, 1933;
Einstein, 1934)
(7.24)
where n
e
=equivalent Mannings n
P
i
=wetted perimeter for the i-th channel subarea (ft, m)
n
i
=Mannings n for the i-th channel subarea
If one assumes that the total force resisting the flow is equal to the sum of the resisting
forces in each of the subareas, the equivalent n value is (Pavlovskii, 1931; Muhlhofer,
1933; Einstein and Banks, 1951)
(7.25)
2/ 3
3/ 2
i i
e
i
Pn
n
P

=
`

)
_
_
1/ 2
2
i i
e
i
Pn
n
P

=
`

)
_
_
Lotter (1933) showed that the equivalent n value could be expressed as the following,
assuming that the total discharge is equal to the sum of the discharges in the individ-
ual subareas:
(7.26)
Finally, if the velocity distribution in the cross section is assumed to be logarithmic,
the equivalent n value is (Krishnamurthy and Christensen, 1972)
(7.27)
where y
i
=average depth of the i-th subarea
In an evaluation of these alternative expressions based on data from 36 actual channel
cross sections, Motayed and Krishnamurthy (1980) concluded that Equation 7.24 per-
formed the best.
Example 7.9 Composite Roughness in a Rehabilitated Channel. Figure E7.9.1
illustrates the cross section of an old masonry tunnel whose hydraulic performance is to be improved
by paving the invert with finished concrete. If the depth of flow in the tunnel is 4 ft both before and
after the rehabilitation, determine the amount by which the tunnel discharge capacity increases fol-
lowing the change. The tunnel slope is S
o
=0.002, and the n values for rough masonry and finished
concrete are 0.030 and 0.013, respectively. Use Equation 7.24 to determine the equivalent n value fol-
lowing the rehabilitation.
Figure E7.9.1 Tunnel cross section for Example 7.9
5/ 3
5/ 3
e
i i
i
PR
n
PR
n
=
| |
|
|
\ .
_
3/ 2
3/ 2
ln
ln
i i i
e
i i
P y n
n
P y
=
_
_
Solution: In both cases, the cross-sectional area of flow is A =16 ft
2
and the wetted perimeter is P =
12 ft. The tunnel capacity prior to the change is
The equivalent n value following the rehabilitation is
Using this equivalent value, the capacity following the rehabilitation is
In this instance, the capacity of the tunnel is increased by over 20 percent.
Uniform Flow in Compound Cross-Sections
Figure 7.10, which is a simplified representation of a flooded river valley, illustrates a
compound channel cross-section. It consists of a main channel and overbank (or
floodplain) areas on either side of the main channel that are inundated by high flows.
Roughness is typically higher in the overbank areas than in the main channel, with the
effect that flow velocities are typically lower in the overbank areas than in the main
channel. This condition causes either of two possible problems to arise in the analysis
of the flow.
Figure 7.10
A compound channel
cross-section
To illustrate the first type of problem, assume that at a given cross section the water
surface elevations are the same in both the main channel and the overbank areas.
Under this assumption, there must be three different energy grade line elevations
within the cross section: one for the main channel and one for each overbank area.
The multiple energy grade lines occur because the velocities, and hence the velocity
heads, are different in each area of the cross section.
Alternatively, one might assume that the energy grade line elevation at a cross section
is the same for all three parts of the flow area. In this case, again because of velocity
differences, one comes to realize that the water surface elevations must be different in
the main channel and overbank areas.
To resolve these issues and arrive at both a single water surface elevation for a cross
section and a single energy grade line elevation, the velocity distribution in the cross
5/ 3 2/ 3 1/ 2
1.49
(16) (12) (0.002) 43 cfs
0.030
Q

= =
{ }
2/ 3
3/ 2 3/ 2 3/ 2
2/ 3
4(0.030) 4(0.013) 4(0.030)
0.025
12
e
n
+ +
= =
5/ 3 2/ 3 1/ 2
1.49
(16) (12) (0.002) 52 cfs
0.025
Q

= =
section must be accounted for. The velocity distribution coefficient o is used for this
purpose.
If a channel cross section is subdivided into N subareas, each of which has an area A
i
and average velocity V
i
, wherei =1, 2 .N, then o can be computed using an approx-
imation of Equation 6.3:
(7.28)
where o =velocity distribution coefficient
A
i
=cross-sectional area of the i-th subarea (ft
2
, m
2
)
V
i
=average velocity of the i-th subarea (ft/s, m/s)
A =total cross-sectional area (ft
2
, m
2
)
V
m
=the mean velocity in the entire cross section (ft/s, m/s)
V
m
is computed as
(7.29)
Using these expressions, the velocity head at a cross section may be expressed as
oV
m
2
/2g.
The total discharge at a cross section may be computed as
(7.30)
where Q =total discharge at a channel cross section (cfs, m
3
/s)
Q
i
=discharge at a cross-sectional subarea (cfs, m
3
/s)
If one assumes that the energy gradient S
f
is equal for each subarea, it may be shown
that
(7.31)
where k
i
= theconveyance of the i-th subarea (cfs, m
3
/s)
S
f
and
(7.32)
where
(7.33)
3
3
i i
m
AV
AV
o =
_
1
m i i
V AV
A
=
_
i i i
Q Q AV = =
_ _
1/ 2
i i f
Q S k =
1/ 2
f
Q S k =
i
k k =
_
In these expressions, k
i
denotes the conveyance of the i-th subarea of the cross sec-
tion. If the Manning equation (Equation 7.9) is employed as the constitutive relation-
ship for the flow, then the conveyance can be written as
(7.34)
where C
f
= unit conversion factor (1.49 for U.S. customary units, 1.00 for SI)
Dividing Equation 7.31 by 7.32, one obtains
(7.35)
This expression shows that the discharge in the i-th subarea is a fraction of the total
discharge, where the fraction is computed as the ratio of the conveyance for the sub-
area in question to the total conveyance for the cross section.
Use of these equations is typically made in either of two ways. If one desires to find
the depth of uniform flow for a specified set of discharge and channel characteristics,
the solution involves iterative calculations using assumed water surface elevations.
For an assumed water surface elevation, the velocity and area of each subarea can be
determined, and a total discharge Q can be computed using Equations 7.30 and 7.31.
If the computed flow rate is different from the known discharge, then a new water sur-
face elevation is assumed, and the procedure is repeated until agreement is attained.
After the water surface elevation is found, the mean velocity of flow can be deter-
mined as V
m
=Q/A, and i can be determined by using Equation 7.28. Finally, the
velocity head is computed as oV
m
2
/2g and added to the water surface elevation to
obtain the energy grade line elevation.
A second way in which these equations can be applied occurs when the discharge,
channel characteristics, and energy grade line elevation are known at a cross section.
This situation occurs in gradually varied flow profile analysis via the standard-step
method (see Section 7.8). As in the previous case, iterations are made on assumed
water surface elevations. For an assumed water surface elevation, one can determine
the areas and conveyances of each of the subareas of the channel section. The dis-
charge in each subarea can then be determined using Equation 7.35, and the velocity
in each subarea can be found as the ratio of the discharge to the area. The mean veloc-
ity can be found by using Equation 7.29, and o can be found by using Equation 7.28.
Finally, the velocity head can be computed as oV
m
2
/2g and added to the assumed
water surface elevation to obtain a computed energy grade line elevation. If the com-
puted energy grade line elevation is not the same as the known elevation, a new water
surface elevation is assumed, and the procedure is repeated.
2/ 3
f
i i i
i
C
A R
n
k =
i
i
i
Q Q
k
k
=
_
Example 7.10 Uniform Flow in a Compound Cross Section. Figure E7.10.1 illus-
trates a compound channel section consisting of two rectangular elements. Determine the depth of
uniformflow in this channel section if the total discharge is Q =450 cfs and the channel slope is S
o
=
0.005. Also, determine the velocity distribution coefficient o for this uniformflow. Mannings n val-
ues for the main channel and overbank area are 0.020 and 0.030, respectively.
Figure E7.10.1 Compound channel cross section for Example 7.10
Solution: For any assumed depth of flow y
1
greater than 3 ft, the areas and wetted perimeters of the
two subareas can be found as
A
1
=10y
1
A
2
=12y
2
=12(y
1
3)
P
1
=13 +y
1
P
2
=12 +y
2
=12 +(y
1
3)
where the subscripts 1 and 2 refer to the main channel and overbank, respectively.
These expressions can be used with Equations 7.31 and 7.34 to find the discharge in each subarea, and
then a total discharge as the sumof the two subarea discharges. Because the flow is uniform, S
o
is
used in place of S
f
in Equation 7.31. By trial and error, the correct depth is found when the computed
discharge is equal to the known discharge.
If the initial assumption for depth is y
1
=4 ft, then A
1
=40 ft
2
, A
2
=12 ft
2
, P
1
=17 ft
2
, and P
2
=13 ft
2
.
Therefore, fromEquations 7.31 and 7.34,
So, according to Equation 7.30,
( )
2 3
2 3
1 1 1
1
1 49 40
40 5 272 cfs
0 020 17
/
/
.
,
.
f
C
A R
n
k
| |
= = =
|
\ .
( )( )
1 2
1 2
1 1
5272 0 005 373 cfs
/
/
.
o
Q S k = = =
( )
2 3
2 3
2 2 2
2
1 49 12
12 565 cfs
0 030 13
/
/
.
.
f
C
A R
n
k
| |
= = =
|
\ .
( ) ( )
1 2
1 2
2 2
565 0 005 40 cfs
/
/
.
o
Q S k = = =
Because 413 cfs is less than the given flow of 450 cfs, a slightly higher water surface elevation must
be assumed and the flow computed again until the correct elevation is found. For this example, trial
and error leads to y
n
=4.18 ft.
For the depth of 4.18 ft, the individual subareas and velocities are
A
1
= 41.8 ft
2
A
2
=14.16 ft
2
V
1
=9.52 ft/s
V
2
=3.67 ft/s
Therefore, fromEquations 7.28 and 7.29:
As shown, the velocity distribution coefficient is 1.26.
7.4 SPECIFIC ENERGY AND CRITICAL FLOW
Critical flow, like uniform flow, is a case in open-channel flow for which a unique
relationship between depth and discharge exists. The occurrence of critical flow cor-
responds to the point of minimum specific energy for a given discharge and channel
geometry; thus, the study of critical flow is facilitated by an understanding of specific
energy concepts.
Specific Energy
Specific energy is the energy head relative to the channel bottom elevation. For a
channel of small slope, the specific energy E is
(7.36)
where E =specific energy (ft, m)
y =depth of flow (ft, m)
o =velocity distribution coefficient
V =cross-sectionally averaged flow velocity (ft/s, m/s)
g = gravitational acceleration (32.2 ft
2
/s, 9.81 m
2
/s)
Q =discharge (cfs, m
3
/s)
A =cross-sectional area (ft
2
, m
2
)
373 40 413 cfs
i
Q Q = = + =
_
( )( ) ( ) ( )
1 1
41.8 9.52 14.16 3.67 8.04 ft/s
41.8 14.16
m i i
V AV
A
( = = + =

+
_
( )( ) ( )( )
( )( )
3 3
3
3 3
41.8 9.52 14.16 3.67
1.26
41.8 14.16 8.04
i i
m
AV
AV
o
+
= = =
+
_
2 2
2
2
2
V Q
E y y
g
gA
o o
= + = +
Section 7.4 Specific Energy and Critical Flow 245
Clearly, the specific energy at a channel cross section consists of the sum of the pres-
sure head (the depth of flow) and velocity head terms from the energy equation
(Equation 7.25).
For a given discharge Q and channel cross-sectional geometry, inspection of Equation
7.36 reveals that E depends only on the depth of flow. A graph showing the depen-
dence of E on y for a given discharge and channel geometry is called a specific energy
curve. Qualitatively, the characteristics of a specific energy curve can be deduced
from Equation 7.36. For small values of y, the velocity will be large and the second
term on the right side of the equation will be much larger than the first. Thus, the spe-
cific energy curve will approach the asymptote E = oV
2
/2g. Conversely, for large val-
ues of y, the velocity will be small and the first term will be much larger than the
second. In that case, the curve will asymptotically approach E = y. Figure 7.11 illus-
trates the characteristics of a typical specific energy curve.
Because a specific energy curve shows the dependence of E on y for a fixed discharge
and channel cross-sectional geometry, a different curve must arise if either the dis-
charge or geometry changes. (Note, however, that changing only the slope simply
means moving another point on the same curve.) It is difficult (if not impossible) to
state any generally applicable effects of geometric changes, but Equation 7.36 shows
that E increases when the discharge increases. That is, if two specific energy curves
for a given channel geometry were plotted on the same graph, with one curve corre-
sponding to a discharge Q
1
and the second corresponding to a discharge Q
2
>Q
1
, the
result would be a shift of the curve for Q
2
to higher values of E. This shift is shown in
Figure 7.11.
Figure 7.11
Example of specific
energy curves for two
flows in the same
channel
Inspection of a specific energy curve reveals that there are two possible depths of flow
corresponding to any specific energy greater than the minimum specific energy level
denoted by E
c
. These two depths are called alternate depths. The smaller of the two
depths corresponds to a shallow and rapid flow known as supercritical flow. The
larger of the two depths corresponds to a deep and tranquil flow known as subcritical
flow. The two limbs of a specific energy curve join at a point where the specific
energy is E
c
and there is a single corresponding value of depth denoted by y
c
(see Fig-
ure 7.11). The depth y
c
is called the critical depth, and a flow at this depth is corre-
spondingly called a critical flow.
Example 7.11 Specific Energy in a Rectangular Channel. Compute and graph a
specific energy curve for a rectangular open channel with B =10 ft and Q =200 cfs.
Solution: To graph the curve, one may assume depths of flow and compute the corresponding E val-
ues. For an assumed depth y, the cross-sectional area is A =10y and the velocity is
V = Q/A =20/y
The specific energy (assuming o =1) is therefore
Figure E7.11.1 shows this specific energy curve for this example.
Figure E7.11.1 The specific energy curve for Example 7.11
Critical Depth and Critical Flow
As shown in the preceding section, critical depth and the corresponding critical flow
occur when the specific energy is a minimum for a fixed discharge and channel cross-
sectional geometry. Computation of the critical depth y
c
can be accomplished in a
trial-and-error manner to minimize E. An alternative (and more enlightening) way of
2
(20/ )
2
y
E y
g
= +
computing critical depth is to recognize that it occurs when dE/dy is equal to zero (see
Figure 7.11). Differentiating Equation 7.36 and assuming o = 1, one obtains
(7.37)
where T =dA/dy =top width of the flow
Equating this expression to zero, critical depth occurs when
(7.38)
TheFroude number of an open-channel flow is a dimensionless quantity defined as
(7.39)
where Fr =the Froude number
D
h
=hydraulic depth (see Equation 7.2) (ft, m)
Comparing Equations 7.38 and 7.39, it can be seen that critical depth and critical flow
occur when the Froude number is equal to unity. Subcritical flow occurs when Fr <1,
and supercritical flow occurs when Fr >1.
An additional fact that can be deduced from Equations 7.38 and 7.39 is that when a
flow is at critical depth, the velocity head is equal to one-half of the hydraulic depth:
(7.40)
This relationship can be quite useful for practical problem solving, particularly when
a channel is rectangular in shape.
The reader should note that the hydraulic depth D
h
of a channel is generally not the
same as the depth of flow y. For common prismatic channel shapes, the hydraulic
depth is smaller than the depth of flow, and is equal to the depth of flow only in rect-
angular channels.
Example 7.12 Finding Critical Depth and Minimum Specific Energy in a Rect-
angular Channel. Determine the critical depth and minimumspecific energy for the channel in
Example 7.11.
Solution: The criterion for critical depth is that Equation 7.38 must be satisfied. For the channel of
interest, B = T =10 ft, Q =200 cfs, and A =10y. Substituting and solving for y, one obtains the critical
depth as
2 2
3 3
1 1
dE Q dA Q T
dy dy
gA gA
= =
2
3
1
Q T
gA
=
1/ 2
2
3
Fr
h
Q T V
gD gA
| |
= = |
|
\ .
2
2 2
h
D V
g
=
FromEquation 7.40, the velocity head for critical flow is one-half the hydraulic depth, or 1.16 ft in
this example. Thus, the minimumspecific energy is
E
c
=2.32 +1.16 =3.48 ft
Example 7.13 Finding the Critical Depth in a Trapezoidal Channel. A trapezoi-
dal open channel has a bottomwidth of B = 3 mand side slopes with z =2. Determine the critical
depth if Q =7 m
3
/s.
Solution: The cross-sectional area of flow, in terms of the depth, is
A =3y +2y
2
The top width is
T = 3 + 4y
Substituting into Equation 7.38, one obtains
By trial and error, the depth that satisfies this expression is y
c
=0.70 m.
Compound Channel Sections
In a compound channel section, consisting of a main channel and overbank areas,
there may be more than one critical depth for a given discharge. Further, determina-
tion of the number of critical depths and their values is more complex than simply
locating all local minima on a specific energy curve, because a critical flow condition
may exist even in situations for which the specific energy is not at a local minimum
(Chaudhry and Bhallamudi, 1988).
The procedure described by Chaudhry and Bhallamudi (1988) for determining the
critical depths in a compound channel is presented here. The procedure first deter-
mines the number of critical depths that occur within a channel section. It then deter-
mines each of the critical depths. It assumes that the channel cross section is
symmetrical about its centerline, with dimensional terminology as shown in Figure
7.12.
Based on these dimensions, one can define a number of dimensionless parameters:
(7.41)
where y
r
=ratio of the depth y to the vertical distance from the bottom of the over-
bank to the bottom of the main channel y
f
(7.42)
where b
r
=ratio of the width of an overbank B
f
to the width of the main channel B
m
2
3
2
200
2.32 ft
32.2(10)
c
y = =
Q
2
T
gA
3
----------
7
2
3 4y + ( )
9.81 3y 2y
2
+ ( )
3
--------------------------------------- 1 = =
r
f
y
y
y
=
f
r
m
B
b
B
=
Figure 7.12
determining critical
depths in a compound
channel
Credit: Chaudry, 1993. Reproducedby permissionof PearsonEducation, Inc., Upper SaddleRiver, N.J .
(7.43)
where b
f
=ratio of the width of an overbank B
f
to the vertical distance from the bot-
tom of the overbank to the bottom of the main channel y
f
(7.44)
where n
r
=ratio of Mannings n for the main channel n
m
to Mannings n for an
overbank n
f
As described by Chaudhry and Bhallamudi (1988), one can use these dimensions and
dimensionless ratios to write an expression of the form
(7.45)
where
(7.46)
and
(7.47)
f
f
f
B
b
y
=
m
r
f
n
n
n
=
2 3
2
m f
gB y
C
Q
=
1 2
C c c = +
2 2
1
1 1 1
2 ( 1) 1 2
r r r r r r
m m
c
y b y y y b

| | | |
= +
| | `
+
\ . \ .
)
(7.48)
(7.49)
where A
1
=flow area for the main channel
A
2
=flow area for one of the overbanks
P
1
=wetted perimeter for the main channel (2y
f
+B
m
)
P
2
=wetted perimeter for one of the overbanks (y y
f
+B
f
)
For specified channel dimensions, roughness values, and a discharge, one can deter-
mine a target value of C using Equation 7.45. Various flow depths (and hence values
of y
r
) can be assumed so that Equations 7.46 through 7.49 can be applied and C
graphed as a function of y
r
. The values of y
r
for which C is equal to the target value
(the right side of Equation 7.45) are the critical depths for flows over the floodplain.
An additional critical depth, depending on the discharge, may also occur at a depth
smaller than y =y
f
. Example 7.14, adapted from Chaudhry (1993), illustrates the pro-
cedure for computing critical depths in a compound channel.
Example 7.14 Determining the Critical Depths in a Compound Channel.
Determine the critical depths for Q =1.7 and 2.5 m
3
/s in a channel having the following dimensions:
B
m
=1.0 m, B
f
=3.0 m, y
f
=1.0 m, and n
r
=0.9 (see Figure 7.12).
Solution: Consider first the possibility that a critical depth might occur within the main channel (i.e.,
at a depth smaller than y
f
). Applying Equation 7.38 with T =B
m
andQ =1.7 m
3
/s, one finds a critical
depth of y
c
=0.67 m. Applying the same equation, but with Q =2.5 m
3
/s, one finds a critical depth of
y
c
=0.86 m. Thus, for both of these discharges, a critical depth occurs within the main channel. It
must now be determined if additional critical depths occur once the flow rises to a depth greater than
y
f
.
Figure E7.14.1 is a graph of C versus y
r
obtained by applying Equations 7.46 through 7.49. Note that
the minimumvalue of y
r
employed is 1.00, which corresponds to an actual depth of y = y
f
. Note also
that C approaches 1 when y
r
approaches 1 m, and that C attains a maximumvalue of C
max
= 2.88
wheny
r
=1.033 m.
UsingEquation 7.45 with Q =1.7 m
3
/s yields C = 3.39. This is shown as the upper of the two hori-
zontal lines in Figure E7.14.1. Because it does not intersect the C- y
r
curve in that figure, there is no
critical depth larger than y
f
when the discharge in the channel is 1.7 m
3
/s.
UsingEquation 7.45 with Q =2.5 m
3
/s yields C = 1.57. This is shown as the lower of the two hori-
zontal lines in Figure E7.14.1. Because it intersects the C- y
r
curve in the figure twice, two critical
depths larger than y
f
exist when the discharge in the channel is 2.5 m
3
/s. Those critical depths are
1.002 m and 1.12 m.
Figure E7.14.2 shows specific energy curves computed for each of the two discharges in the channel.
Note that when the discharge is Q =1.7 m
3
/s, only one critical depth exists: 0.67 m. However, when
the discharge is Q =2.5 m
3
/s, there are three critical depths, at values of 0.86, 1.002, and 1.12 m. As
shown in the figure, not all of the critical depths correspond to local minima on the specific energy
curve.
2
2 (1 ) 5 2 1 1
3[ 2 ( 1)] ( 1) 1 2 1
r r r r r f r r r r
m m m m
c
y b y y y b y y b y

| |
=
` |`
+ +
\ . ) )
5/ 3 2/ 3
2 1 1 2
1
1 2 ( / ) ( / )
r
m
n A A P P
=
+
Figure E7.14.1 C versus y
r
for the compound channel in Example 7.14
Figure E7.14.2Specific energy curves for two discharges in a compound channel for Example 7.14
Flow Control
A control section in open-channel flow is said to occur at a channel location for which
there is a known or unique value of the depth of flow for a specified discharge. The
depth of flow at a control section is essentially a boundary condition that provides a
unique solution to the differential equation describing a varied flow in a channel. This
concept is discussed further in subsequent sections of this chapter.
Because a unique relationship exists between discharge and flow depth for a critical
flow, a channel cross section at which critical flow occurs is necessarily a flow control
section. Common occurrences of such critical flow control sections occur at free over-
falls and at locations where a mildly sloping channel ends and becomes a steeply
sloping channel. Rouse (1936) has shown that, at a free overfall, critical depth actu-
ally occurs at a distance of about 3y
c
to 4y
c
upstream of the brink of the overfall.
In open-channel flow, flow control sections can exist at locations other than critical
flow sections. For example, the flow depth created at the upstream end of a culvert is
often a flow control section, as is the depth of a supercritical flow issuing from
beneath a sluice gate. Many other examples of flow control sections can be found.
Flow Measurement
Critical flow is an instance in which there is a unique definitive relationship between
the depth of flow and discharge in an open channel. Thus, if one can cause a flow to
pass through its critical depth, measurement of that depth can be used to deduce the
discharge.
There are a number of ways that one can force a flow to pass through critical depth.
One way is to install a weir in a channel (see Section 7.6). Weirs, however, tend to
cause large head losses in the flow, which may not be acceptable in some circum-
stances. Critical depth can be caused by a flow contraction, and it occurs naturally at a
free overfall or at a location where a mild and steeply sloping channel join. A Parshall
flume is another widely used device for measurement of open-channel flow, though
critical depth may or may not occur in such a flume.
Analysis of Flow in Channel Transitions
A transition occurs in an open channel at a location where the cross-sectional geome-
try of the channel changes. For example, a transition might correspond to a change in
the width of a rectangular channel, or in the diameter of a pipe (such as at a manhole).
A transition may also be as simple as a step or drop in the bottom elevation of the
channel, with no corresponding change in cross-sectional geometry.
When there is a reduction in cross-sectional area and a corresponding increase in
velocity at a transition, the flow is contracted. A flow expansion occurs when the
cross-sectional area increases and the velocity decreases.
Consider a transition consisting of a simple step in the bottom elevation, as shown in
Figure 7.13. The discharge, Q, is constant through the transition, and the cross-
sectional geometry of the channel is the same both upstream and downstream of the
step. If the depth of flow at a cross section downstream of the step is known and des-
ignated by y
2
, the corresponding specific energy E
2
can easily be found. If one
assumes that there is no energy loss in the transition (that is, that the energy grade line
has a negligible slope in the transition), the specific energy E
1
at a cross section
upstream of the step can be determined as E
1
=E
2
+AZ, where AZ is the height of the
step. A specific energy relationship can then be applied to determine the two alternate
depths of flow at the upstream section.
Figure 7.13
A simple step channel
transition and the
resulting specific
energy curve
Determining which of the two alternate depths is the correct upstream depth requires
one to consider the concept of attainability. Because neither the channel geometry nor
the discharge changes in the transition, the single specific energy curve shown in Fig-
ure 7.13 applies to both the upstream and downstream portions of the transition. To
move from one location on the curve (represented by a particular combination of
depth and specific energy) to another location on the curve requires that one traverse
along the path of the curve. Point A on the curve represents the known depth and spe-
cific energy downstream of the transition, and points B and C represent the alternate
depths upstream of the transition. Note that movement along the curve from point A
to point B requires an increase in specific energy, which occurs in this situation since
E
1
is larger than E
2
by the amount AZ. On the other hand, movement along the curve
from point A to point C requires that the specific energy first decrease to E
c
, and then
increase to the amount E
1
.
Noting that specific energy may be viewed graphically as the distance from the
energy grade line to the channel bottom, it is seen in Figure 7.13 that in moving from
the downstream to the upstream cross section, the specific energy only increases. It
does not decrease and then increase, and thus point C on the specific energy curve is
not attainable from point A in this instance. Thus, the correct upstream depth is the
larger of the two alternate depths. Had there been a hump or sill in the channel bottom
between the two cross sections such that the specific energy at the hump were equal to
E
c
, then the smaller of the two alternate depths would have been the correct solution.
This concept of attainability is applicable to any type of flow in a channel transition,
whether it is subcritical or supercritical, provided that neither the discharge nor the
channel geometry changes in the transition. When the geometry changes, the concept
of attainability remains valid, but its analysis becomes more complicated. The com-
plication arises because the specific energy relationships upstream and downstream of
the transition are different from one another, and there is generally a continuum of
specific energy relationships due to the gradually changing geometry of the transition
along its length.
In the preceding discussion, it was assumed that energy losses in a transition were
negligible. In truth, the energy losses occurring in a channel transition consist of fric-
tional losses along the length of the transition and minor losses caused by velocity
changes. Because channel transitions are often short in length, the frictional losses are
usually small compared to the minor losses and can be safely ignored. Minor loss
coefficients applicable to channel transitions were presented in Section 7.2 (page
230).
Section 7.5 Momentum in Open-Channel Flow 255
Incorporation of energy losses into an analysis of flow through a transition requires a
trial-and-error approach. To illustrate, if the downstream depth and specific energy are
known, a first estimate of the upstream depth and specific energy can be made using
the procedures discussed previously. With an initial estimate of the upstream depth,
the velocities at the upstream and downstream cross sections can be determined, and
the minor energy loss can be computed using Equation 7.16. A refined estimate of the
upstream specific energy is then determined by recognizing that the minor loss causes
the energy grade line elevation to increase by the amount h
L
. This refined estimate of
the upstream specific energy can be used to obtain a refined estimate of the upstream
depth, and the procedure is repeated until convergence is obtained.
Example 7.15 Bottom Transition in a Rectangular Channel. A sudden drop of
0.50 ft occurs in the channel described in Examples 7.11 and 7.12. Recall that the channel is rectangu-
lar with B =10 ft and Q =200 cfs. The depth of flow downstream of the drop is y
2
=4.80 ft. Deter-
mine the depth upstreamof the drop assuming that energy losses are negligible.
Solution: The specific energy downstreamof the drop is
Based on the assumption of zero energy loss, the specific energy upstreamof the 0.5-ft drop must
therefore be
E
1
=5.07 0.5 =4.57 ft
The alternate depths upstreamof the drop are the two positive solutions of the following expression:
The two alternate depths are found, by trial and error, to be y
1
=4.22 ft and y
1
=1.40 ft.
FromExample 7.12, the critical depth in this channel is 2.32 ft, and the corresponding minimumspe-
cific energy is 3.48 ft. Because the specific energy in the channel decreases from5.07 to 4.57 ft, but
never drops to a value as small as 3.48 ft, the larger of the two alternate depths is the only attainable
one. Therefore, the upstreamdepth of flow is y
1
=4.22 ft.
7.5 MOMENTUM IN OPEN-CHANNEL FLOW
Equation 3.12, the momentum principle, as applied to a coordinate direction x for a
steady flow, states that the sum of the external forces acting on the water within a con-
trol volume is equal to the net rate of momentum outflow from the control volume.
This principle is applied when determining the upstream and downstream depths for a
hydraulic jump and when evaluating forces on submerged structures.
Figure 7.14(a) illustrates a reach of channel (a control volume) of length L between an
upstream cross section 1 and a downstream cross section 2. The coordinate direction x
is taken as positive in the downstream direction and parallel to the channel bottom.
External forces acting on the water in the control volume consist of the pressure force
P
1
acting at cross section 1, the pressure force P
2
acting at cross section 2, the fric-
tional force F
f
opposing the flow, and the gravitational force Wsinu driving the flow.
( )( )
{ }
( )
2
2
2
2 2
200 10 4 8
4 8 5 07 ft
2 2 32 2
/ .
. .
.
V
E y
g
(

= + = + =
2
1
1 1
(200/10 )
4.57 ft
2(32.2)
y
E y = + =
Unlike the case in uniform flow, the pressure forces P
1
and P
2
shown in Figure 7.14
are not necessarily equal. The sum of the forces is equal to the change in momentum
across the control volume.
Figure 7.14
forces acting on a
fluid element with
non-uniformflow
Substituting the sum of the external forces into Equation 3.12 and equating the sum to
the net momentum outflow rate, one obtains
(7.50)
where _F
x
=sum of external forces in the x direction
P
1
and P
2
=forces exerted on cross section due to hydrostatic pressure (see
Equation 7.53) (lb, N)
W =weight of fluid element (lb, N)
u =angle that channel bed makes with horizontal
F
f
=frictional force (lb, N)
=fluid density (slugs/ft
3
, kg/m
3
)
3
/s)
|
1
and |
2
=velocity distribution coefficients as defined by Equation 6.30
V
1
and V
2
=average flow velocities (ft/s, m/s)
Typical values of | for open-channel flow are presented in Table7.2. In compound
channels, where the velocity may vary greatly from one part of the cross section to
another, an estimate of | may be obtained as
(7.51)
where | =velocity distribution coefficient
A
i
=cross-sectional area of the i-th subarea (ft
2
, m
2
)
V
i
=average velocity of the i-th subarea (ft/s, m/s)
n = number of subareas
A =total cross-sectional area (ft
2
, m
2
)
V =average velocity for the entire cross section (ft/s, m/s)
Thepressure force P exerted at a cross section in an open-channel flow (see Equation
7.50) can be expressed as
P = zA (7.52)
where =fluid specific weight (lb/ft
3
, N/m
3
)
z =distance from the free water surface to centroid of cross-sectional area
[see Figure 7.14(b)] (ft, m)
A =cross-sectional area of flow (ft
2
, m
2
)
Table7.7 summarizes expressions for z for rectangular, trapezoidal, triangular, and
circular cross-sectional shapes (for dimension definitions, see Figure 7.4). (Note that
both side slopes must be equal for triangular and trapezoidal channels if these equa-
tions are used.) For other channel shapes, z can be computed as
1 2 2 2 1 1
sin ( )
x f
F P P W F Q V V u | | = + =
_
2
1
2
n
i i
i
AV
AV
|
=
=
_
(7.53)
where z
i
=vertical distance from free water surface to centroid of cross section of
i-th subarea (ft, m)
If other forces are exerted on the water in the control volume, such as by a step in the
channel bottom or by a bridge pier, then those forces should be included in Equation
7.50. This point is illustrated in the Forces on Submerged Structures discussion
later in this section.
Specific Force
If Equations 7.50 and 7.52 are applied to a short reach of a horizontal channel, the
frictional force F
f
will be small and Wsinu will be zero. Thus, if |
1
and |
2
are
assumed to be 1, the momentum equation simplifies to
(7.54)
or
F
s1
= F
s2
(7.55)
where
(7.56)
and F
s
=specific force (lb, N)
Table 7.7 Centroidal distance z for prismatic channels
a
a. Angle | is measured in radians. 1 radian ~ 57.3 degrees.
Channel shape Area z
Rectangular By
Triangular
zy
2
Trapezoidal
By + zy
2
Circular
a
1
n
i i
i
A z
z
A
=
=
_
y
2
---
y
3
---
y
3
3B 2zy ( )
6 By zy
2
+ ( )
--------------------------------
D
2
| | sin ( )
8
--------------------------------
2D
|
2
---
\ .
| |
sin D |
|
2
---
\ .
| |
cos sin
3 | | sin ( )
-----------------------------------------------------------------
D
|
2
---
\ .
| |
cos
2
----------------------
2 2
1 1 2 2
1 2
Q Q
z A z A
gA gA

| | | |
+ = + | |
| |
\ . \ .
2
s
Q
F zA
gA
| |
= + |
|
\ .
The quantity F
s
expressed by Equation 7.56 is known as the specific force in an open-
channel flow. It consists of two terms, one representing the hydrostatic pressure force,
and the second representing the additional dynamic force caused by the flow momen-
tum.
Specific force, like specific energy, depends on the flow depth, channel geometry, and
discharge. For a specified discharge and channel shape, a graph of specific force can
be developed as a function of flow depth. This graph is known as a specific force
curve.
Like a specific energy curve, a specific force curve contains two limbs: the first corre-
sponds to subcritical flow, and the second corresponds to supercritical flow. The spe-
cific force curve also has a minimum force given by F
sc
, which occurs at the critical
depth y
c
. Note that for any value of the specific force greater than F
sc
, there are two
possible depths of flow. Again, this is exactly analogous to the case of a specific
energy curve. However, in the case of a specific force curve, the two depths are called
sequent depths (some texts refer to them as conjugate depths).
Example 7.16 Specific Force in a Trapezoidal Channel. Develop a specific force
curve for a trapezoidal open channel with B =8 ft, z =2, and Q =250 cfs.
Solution: Using expressions for A and z fromTable7.7, one can substitute into Equation 7.56 to
obtain an expression for F
s
as a function of flow depth y:
A graph of the resulting relationship is shown in Figure E7.16.1, which is the specific force curve.
Note that the specific force has a minimumat the critical depth of 2.51 ft.
Figure E7.16.1 Graph of specific force versus depth for channel in Example 7.16
2 3 2
2
2 3
( )
s
By zy Q
F
g By zy
| |
| = + +
|
+
\ .
The Hydraulic Jump
The hydraulic jump is a phenomenon that occurs when flow in an open channel
increases from a small supercritical depth to a larger subcritical depth with a corre-
sponding decrease in energy caused by turbulent eddies. This loss of energy can be
considerable. For this reason, hydraulic jumps are often employed to dissipate other-
wise harmful energy in a channel.
To illustrate the characteristics of a hydraulic jump, consider a jump in a horizontal
trapezoidal channel with dimensions and discharge as specified in Example 7.16 on
page 259. Assume that the depth of flow upstream of the jump is y
1
=1.00 ft (0.3 m),
as shown in Figure 7.15, and that the length of the jump is short. Because the channel
is horizontal and the jump is short, conditions assumed at the beginning of the preced-
ing subsection (that is, F
f
is small and Wsinu is 0) are applicable, and one can con-
clude that F
s1
=F
s2
. Using the information in Figure E7.16.1 with y
1
=1.00 ft (0.3 m)
yields Fs
1
=Fs
2
=12,403 lb (5,625 kg) and y
2
=4.93 ft (1.5 m). Thus, for a hydraulic
jump in a horizontal channel, the relationship between the depths upstream and down-
stream of the jump is given by the sequent depths on a specific force curve.
The energy loss in the hydraulic jump can be assessed by computing the specific
energy upstream and downstream of the jump. Upstream of the jump where the depth
isy
1
=1.00 ft (0.3 m), the corresponding specific energy (from Equation 7.36) is E
1
=
10.70 ft (3.3 m). Downstream of the jump, where the depth is y
2
=4.93 ft (1.5 m), the
specific energy is E
2
=5.06 ft (1.54 m). The difference between the two specific ener-
gies is the head loss in the jump, which is h
L
=5.64 ft or 1.72 m (see Figure 7.15).
Note that the alternate depths do not play a role in the analysis of a hydraulic jump,
but the specific energy curve does provide information relevant to the head loss that
occurs in the jump.
Figure 7.15
Definition figure for a
hydraulic jump
As described in the previous paragraphs, the characteristics of a hydraulic jump in a
horizontal channel of any cross-sectional shape can be determined by use of specific
force and specific energy curves. However, in the special case of a horizontal, rectan-
gular channel, one can show that the sequent depths are related as
(7.57)
wherey
1
, y
2
=flow depths at sections 1 and 2, respectively (ft, m)
Fr =Froude number
and
(7.58)
where g =gravitational acceleration constant (32.2 ft/s
2
, 9.81 m/s
2
)
In such a channel, one can also show that the energy loss is
(7.59)
where h
L
=head loss in hydraulic jump (ft, m)
E
1
, E
2
=total energy (ft, m)
If a hydraulic jump occurs in a sloping channel such that the term Wsinu in the
momentum equation cannot be neglected, the specific force approach to estimating
the relationship between the depths upstream and downstream of the jump cannot be
applied. Instead, one must rely on an experimental relationship, such as that illus-
trated in Figure 7.16. The figure illustrates the ratio of the downstream depth to the
upstream depth (y
2
/y
1
measured vertically or d
2
/d
1
measured normal to the channel
bottom) as a function of the channel slope (S
o
) and the Froude number of the
upstream flow (Fr
1
). For this figure, a rectangular cross-section shape is assumed.
The length of a hydraulic jump is a characteristic that must be evaluated on the basis
of experimental data. Figure 7.17, which is based on data collected by the U.S.
Bureau of Reclamation (1955), illustrates the length L of a jump as a function of the
downstream depth, the channel slope, and the Froude number of the upstream flow.
This figure was developed for rectangular channels; however, in the absence of other
data, it can be used to approximate jump length in nonrectangular channels.
2 2
1
1
1
1 8Fr 1
2
y
y
| |
= +
|
\ .
2
2 1
1
1
Fr
V
gy
=
3
2 1
1 2
1 2
( )
4
L
y y
h E E
y y
= =
Figure 7.16
Ratio of downstream
to upstreamdepths for
a hydraulic jump in a
rectangular channel as
a function of the
channel slope and
upstreamFroude
number
Credit: Chow, 1959.
Figure 7.17
Graph for determining
the length of a
hydraulic jump
Credit: Chow, 1959.
Example 7.17 Hydraulic Jump in a Rectangular Channel. A hydraulic jump forms
in a horizontal rectangular open channel with B =3 mand Q =6 m
3
/s. The depth of the approaching
flow is y
1
=0.40 m. Determine the depth downstream of the jump, the energy loss in the jump, and the
length of the jump.
Solution: Because the channel is horizontal and rectangular, Equation 7.57 may be used to determine
the downstreamdepth. The Froude number of the approaching flow is
Therefore, applying Equation 7.57, the downstreamdepth is
The energy loss in the jump may be determined using Equation 7.59, and is
The length of the jump can be estimated using Figure 7.17. Fromthat figure, the dimensionless ratio
L/y
2
is estimated to be about 4.8. Therefore,
L =4.8y
2
=4.8(1.24) =5.95 m
Forces on Submerged Structures
When an open-channel flow occurs past a submerged structure, such as a bridge pier,
the flow exerts a hydraulic force on that structure. The structure in turn exerts an equal
and opposite force on the flow. The force exerted on the flow by the structure must be
considered as an external force on the flow when applying the momentum equation.
Consider flow in an open channel of small slope and short length where there is a sub-
merged structure (such as a bridge pier) in the flow path. In this case, the weight of
the water and frictional forces may be neglected, and the momentum equation
becomes
(7.60)
where F
p
=force exerted on the flow by the pier (lb, N)
Note that it is assumed in this expression that the force exerted by the pier opposes the
flow direction (that is, it is in the negative x-direction).
A comparison of Equation 7.60 to Equations 7.54 and 7.55 indicates that
F
s1
= F
s2
+ F
p
(7.61)
Thus, the pier force F
p
can be determined graphically from a specific force curve if
the depths of flow y
1
and y
2
upstream and downstream of the pier are known. The
depth change due to the pier may be found through an application of the energy equa-
tion based on an estimate of the energy loss due to the pier.
Fr
V
1
gy
1
------------
6 3 0.04 ( )
9.81 0.4
---------------------------- 2.52 = = =
y
2
y
1
2
----- 1 8Fr
2
+ 1 ( ) 1.24 m = =
h
L
y
2
y
1
( )
3
4y
1
y
2
-----------------------
1.24 0.40 ( )
3
4 1.24 0.40
------------------------------------ 0.30 m = = =
1 1 2 2 2 2 1 1
( )
p
z A z A F Q V V | | =
Example 7.18 Force in a Channel Transition. Figure E7.18.1 illustrates a channel tran-
sition consisting of a step of height AZ. Determine the force exerted on the step by the water if the
channel is rectangular with B =6 m, y
1
=2 m, Q =12 m
3
/s, and AZ =0.25 m. Assume that the energy
loss in the transition is negligible.
Figure E7.18.1 Channel transistion for Example 7.18
Solution: First, the downstreamdepth y
2
must be determined. The specific energy upstreamof the
step is
and the specific energy downstreamof the step is
E
2
=E
1
AZ = 2.05 0.25 =1.80 m
The equation for E
2
can then be written as
By trial and error, the downstreamdepth is y
2
=1.74 m.
Assuming that the step exerts a force on the flow in the upstreamdirection (the negative x-direction),
that force may be determined fromEquation 7.61 as
F
p
= F
s1
F
s2
or, substituting fromEquation 7.56,
Because the sign of this result is positive, the assumed direction of the force exerted by the step is cor-
rect. Therefore, the force exerted on the step by the water is in the downstreamdirection and has a
magnitude of 26,800 N.
( )( )
( )
2
2
1
1 1
12 2 6
2 2 05 ft
2 2 9 81
/
.
.
V
E y
g
(

= + = + =
( )
( )
2
2
2
12 6
1 8
2 9 81
/
.
.
y
y
(

= +
2 2
1 1 2 2
1 2
p
Q Q
F z A z A
gA gA

| | | |
| | = + +
| |
\ . \ .
2 2
12 1.74 12
9, 810 1(12) (1.74)(6) 26, 800 N
9.81(12) 2 9.81(1.74)(6)
p
F

( (

= + + = ( (
`
( (

)
Section 7.6 Weirs 265
7.6 WEIRS
A weir in an open channel is a structure that may be employed for flow measurement
or as a flow control structure. Weirs may be used as outlet structures for water storage
facilities such as detention ponds or reservoirs. Weir flows over roadway embank-
ments may also occur during high flow events at culvert crossings. Finally, some
hydraulic structures, notably stormwater inlets, behave as weirs during some types of
flow conditions.
Weir Characteristics
Weirs are classified as being either sharp-crested or broad-crested. A sharp-crested
weir is one in which a relatively thin metal, concrete, or plastic plate forms the crest
of the weir. For a broad-crested weir, the weir crest thickness is relatively large, and
flow over the weir crest is critical.
As illustrated in Figure 7.18, additional classifications for weirs concern the attributes
of the crest cross-sectional shape (such as rectangular, triangular, or trapezoidal), and
whether there are contractions of the flow at the ends of the weir. A contracted weir
[Figure 7.18(b)] is a weir type that causes the flow to contract. Flow over a suppressed
or uncontracted weir [Figure 7.18(a)], however, has no end contractions.
Figure 7.18
Definition figures for
rectangular and
triangular weirs
Credit: Chin, 2000
The jet of water that issues in a downstream direction from the crest of a sharp-crested
weir is called the nappe. Ideally, the nappe should spring from the crest and should
have atmospheric pressures prevailing at both its upper and lower surfaces. If the
lower face of the nappe is at a pressure below atmospheric, the nappe may cling to the
downstream face of the weir and change its hydraulic performance. If the possibility
of a clinging nappe presents itself in a particular application (usually it is only a prob-
lem with suppressed weirs), vents should be installed so as to enable atmospheric
pressure to exist on the underside of the nappe [Figure 7.18(a)].
Sharp-Crested Weirs
Theoretically, the discharge over a sharp-crested weir may be determined via the
energy equation. The procedure amounts to equating the total energy head upstream
of the weir to the total energy head on the weir crest. The resulting equation for dis-
charge in terms of the upstream energy level depends on the cross-sectional shape of
the weir, and inevitably also involves some simplifying assumptions. An empirical
weir discharge coefficient, C
d
, is introduced to account for the simplifying assump-
tions and to correct the theoretical discharge. (Note that a weir coefficient is some-
times used instead of the discharge coefficient C
d
described here. The weir coefficient
is equivalent to C
d
(2g)
1/2
. Whereas C
d
is unitless, the weir coefficient value depends
on the units being used.)
Two types of sharp-crested weirs, rectangular and triangular, are discussed in this sec-
tion.
Rectangular Weirs. The discharge Q over a suppressed rectangular sharp-
crested weir [Figures 7.18(a) and 7.19] may be expressed in terms of the head H on
the weir as
(7.62)
3
/s)
C
d
=discharge coefficient determined using Equation 7.63
g =gravitational acceleration constant (32.2 ft/s
2
, 9.81 m/s
2
)
L =length of the weir crest (ft, m)
H =height of the energy grade above weir crest (ft, m)
Figure 7.19
Rectangular sharp-
crested weir
3/ 2
2
2
3
d
Q C g LH =
If the velocity upstream of the weir is low, as in the case of a pond spillway, the veloc-
ity head is often considered negligible, and H is taken as the difference in elevation
between the upstream water surface and the weir crest. H should be measured far
enough upstream of the weir that drawdown effects are avoided.
Fundamentally, the discharge coefficient in Equation 7.62 depends on the Reynolds
and Weber numbers of the flow, and on the dimensionless head H/H
w
, where H
w
is the
height of the weir crest above the bottom of the channel (Figure 7.18). However, for
practical purposes, experiments have shown that the discharge coefficient can be rep-
resented as (Rouse, 1946; Blevins, 1984)
(7.63)
whereH
w
=height of weir crest above bottom of channel (ft, m)
Equation 7.63 is valid for H/H
w
less than 5 and is approximate up to H/H
w
=10. For
H/H
w
greater than 15, the discharge can be computed from the critical flow equation
(seeSection 7.4) by assuming that y
c
= H (Chaudhry, 1993).
When flow occurs over a contracted rectangular weir [see Figure 7.18(b)], the end
contractions cause the effective crest length of the weir to be reduced from the actual
crest length. In this instance, the head-discharge relationship for the weir may be
expressed as
(7.64)
where N =number of weir end contractions (N =1 or 2; usually, N =2)
This equation assumes that L is greater than 3H. Note that the effective length of the
weir is L 0.1NH, which is 0.1NH shorter than the actual length of the weir crest.
A Cipolletti weir is a contracted weir that, rather than a rectangular cross-section, has
a trapezoidal cross section with 1:4 (horizontal:vertical) side slopes to compensate for
the flow contraction. Denoting the base width of the trapezoid (that is, the weir crest
length) as L, Equations 7.62 and 7.63 can be used to determine the head-discharge
relationship. (Note that Equation 7.64 is not used in this case because the sloped sides
of the trapezoid offset the effective crest length reduction that otherwise occurs with a
contracted weir.)
Example 7.19 Discharge over a Contracted Rectangular Weir. Arectangular
weir with contractions on both sides (N =2) has a crest length L =3 m, H
w
=2.5 m, and H =0.30 m.
Determine the discharge over the weir.
Solution: Using Equation 7.63, the discharge coefficient is
Applying Equation 7.64, the discharge is
0.611 0.075
d
w
H
C
H
= +
3/ 2
2
2 ( 0.1 )
3
d
Q C g L NH H =
0.30
0.611 0.075 0.62
2.5
d
C = + =
{ }
3/ 2 3
2
(0.62) 2(9.81) 3 0.1(2)(0.30) (0.30) 0.88 m s
3
Q = =
Triangular Weirs. A triangular, or v-notch, weir has a v-shaped cross section
[see Figures 7.18(c) and 7.20]. This type of weir can control flow more accurately
than a rectangular weir when the discharge and head on the weir are small. Because
the discharge over a triangular weir for any given weir head H is much smaller than
that for a rectangular weir with the same head, triangular weirs are usually used only
in instances where the discharge is small.
Figure 7.20
Triangular sharp-
crested weir
The discharge over a sharp-crested triangular weir can be shown to be
(7.65)
where u =angle formed by vertex of v-shaped weir cross section (radians)
Values of the discharge coefficient for use in this expression are shown in Figure 7.21.
The solid curves on the graph should be used for nearly all design purposes, espe-
cially when the head on the weir is 2 in. (5 cm) or more. For smaller heads, both vis-
cous and surface-tension effects are important determinants of weir behavior, and the
dashed lines reflect this combination of effects. Although the dashed lines extend
above H =0.2 ft in the graph, it is recommended that the solid lines be used for larger
values of H. As a practical matter, the difference in using the dashed or solid curves
will be small, as changes in C
d
are typically less than 0.01.
Broad-Crested Weirs
Broad-crested weirs, as the name implies, have weir crests that are broad as opposed
to sharp-edged. For example, a roadway embankment containing a culvert behaves as
a broad-crested weir if the water is high enough to flow over the roadway surface.
Theoretically, the flow profile over a broad-crested weir can be evaluated through the
concepts of specific energy and critical depth. Figure 7.22(a) illustrates a broad-
crested weir of height H
w
=AZ with an approaching flow of discharge Q and depth y
1
.
The thickness of the weir crest in the direction of flow is Ax, and the cross section is
rectangular with a weir crest length of L normal to the plane of the page.
5/ 2
8
2 tan
15 2
d
Q C g H
u | |
=
|
\ .
Figure 7.21
Discharge coefficient
values for triangular
weirs (Franzini and
Finnemore, 1997)
Credit: FinnemoreandFranzini, 2002. Reproducedwithpermissionof theMcGraw-Hill Companies.
Figure 7.22
Definition figure and
specific energy curve
for a broad-crested
weir
The specific energy upstream of the weir is E
1
. Assuming no energy loss, the specific
energy on the crest of the weir is (E
1
AZ) =E
2
, from which the depth y
2
may be
determined using the specific energy curve [Figure 7.22(b)]. Thus, if the weir crest
height H
w
=AZ is not excessively large and the assumption of zero energy loss is
made, then the flow profile over the weir may be evaluated using a specific energy
curve. Further, the specific energy E
3
and depth y
3
downstream of the weir will be
equal to E
1
and y
1
, respectively, if the channel bottom elevation is the same upstream
and downstream of the weir. If desired, a minor loss can be introduced to account for
the energy loss that actually takes place.
If, assuming fixed values for E
1
, y
1
, and Hw =AZ, E
2
is found to be less than E
c
, the
assumptions described in the preceding paragraph are not valid. This situation is
impossible because E
c
represents the lowest energy on the specific energy curve [see
Figure 7.22(b)]. What actually occurs is that the flow is choked, meaning that a back-
water effect is created by the weir. In such a case, one must set y
2
=y
c
and E
2
=E
c
.
The specific energy upstream of the weir is then found as E
1
=E
c
+H
w
, and the depth
y
1
is determined from the specific energy curve.
When a flow is choked (that is, when a structure behaves as a broad-crested weir), the
discharge may be computed as a function of the head H =E
1
H
w
as
(7.66)
where values of the discharge coefficient can be estimated from
(7.67)
These expressions are valid provided that 0.08 <(y
1
H
w
)/Ax <0.50. Head losses
across the crest of the weir cannot be neglected if (y
1
H
w
)/Ax <0.08. If (y
1
H
w
)/Ax
>0.50, streamlines over the weir crest are not horizontal. In these cases, the flow pro-
file should be evaluated using gradually varied flow profile analysis methods (see
Section 7.8).
7.7 ORIFICES
Like a weir, an orifice functions as a hydraulic control. For example, in stormwater
systems, an orifice may be part of a detention pond outlet structure configuration in
the form of a restrictor plate or an opening in a riser pipe or inlet box (see Chapter
12). Culvert entrances and stormwater inlets may also function hydraulically as ori-
fices.
The discharge through a single orifice opening depends on the area of the opening, the
effective head on the orifice, and the type of edge around the opening. The relation-
ship between these variables, which are shown in Figure 7.23, is expressed as:
Q = C
d
A
0
(2gh
0
)
1/2
(7.68)
3/ 2
2
3
d
H
Q C gL
| |
=
|
\ .
( )
1/ 2
0.65
1 /
d
w
C
H H
=
+
Section 7.8 Gradually Varied Flow 271
3
/s)
C
d
= discharge coefficient depending on the orifice edges
A
0
= area of the orifice opening (ft
2
, m
2
)
g =acceleration of gravity (ft/s
2
, m/s
2
)
h
0
=effective head on the orifice (ft, m)
Brown et al. (1996) indicate that C
d
can be taken as 0.60 if the edges of the orifice
opening are uniform and square-edged. For ragged edges, such as those resulting
from use of an acetylene torch to cut an opening, a value of 0.40 should be used.
The effective head on an orifice depends on whether it has a free or submerged dis-
charge. If the discharge is a free outfall, the effective head is the difference between
the upstream water surface elevation and the elevation at the centroid of the orifice
opening. If tailwater submerges the outlet of an orifice, the effective head is equal to
the difference between the water surface elevations immediately upstream and down-
stream of the orifice.
Figure 7.23
Orifice flow diagram
7.8 GRADUALLY VARIED FLOW
For gradually varied flow in an open channel, the depth and velocity of a steady flow
change along the length of the channel. As the name implies, the rates of change of
depth and velocity are gradual and there are no abrupt changes such as hydraulic
jumps. Because the curvatures of streamlines are small, hydrostatic pressures can be
assumed to exist in the flow (that is, the pressure head can be taken as the depth of
flow), and the constitutive relationships developed for uniform flow, such as the Man-
ning equation, can be assumed to be valid for analyses of gradually varied flow.
Gradually varied flow occurs in an open channel when a flow control causes the depth
of flow at a cross section to be different from normal depth in the channel. This situa-
tion can occur, for instance, upstream of culverts, at locations where the slope of a
channel changes, and at channel entrances. Depending on the depth at the control sec-
tion relative to the normal depth, the flow may be accelerated or decelerated in the
channel.
The evaluation of gradually varied flow requires computations to determine flow pro-
file type and flow depths at multiple locations along the length of the channel. This
section presents an example in which a gradually varied flow profile is evaluated.
However, because of the number of calculations involved, engineers typically use
spreadsheets or other computer programs (such as StormCAD or HEC-RAS) when
performing this type of analysis.
Water Surface Profile Equation
For a channel of small slope, the total energy head H at a cross section consists of the
sum of the channel bottom elevation Z, the depth of flow y, and the velocity head
oV
2
/2g. Taking the x-direction to be horizontal and positive in the downstream direc-
tion, the derivative dH/dx is
(7.69)
where H =total energy head at a cross section (ft, m)
x =distance along channel (ft, m)
Z = channel bottom elevation (ft, m)
y = vertical flow depth (ft, m)
o = velocity distribution coefficient
Q = discharge (cfs, m
3
/s)
g =gravitational constant (32.2 ft
2
/s, 9.81 m
2
/s)
A = area of cross section (ft
2
, m
2
)
Noting that the cross-sectional area A depends on the depth of flow y, the chain rule
can be applied to rewrite this expression as
(7.70)
where Fr =Froude number
BecausedH/dx =S
f
is the slope of the energy grade line, and because dZ/dx =S
o
is
the channel bottom slope, this equation can be further simplified to read
(7.71)
where S
o
= channel bottom slope (ft/ft, m/m)
S
f
This equation is the governing equation of gradually varied flow in an open channel.
It describes how the depth changes with distance along the channel as a function of
channel and flow characteristics. A unique solution to this differential equation that
describes the flow profile for a particular case of interest requires knowledge of a
boundary condition (that is, a known depth of flow at a particular channel cross
section).
2
2
2
dH dZ dy d Q
dx dx dx dx
gA
o
| |
= + + |
|
\ .
( )
2
2
3
1 1 Fr
dH dZ dy Q T dZ dy
dx dx dx dx dx
gA
o o
| |
= + = + |
|
\ .
2
1 Fr
o f
S S
dy
dx
o

Classification of Flow Profiles
For a specified discharge Q in a prismatic open channel of known dimensions, one
can apply the procedures described in Section 7.3 and Section 7.4 to determine the
normal and critical depths y
n
and y
c
. Knowledge of these two depths, as well as the
depth at a flow control section, enables one not only to classify the channel and flow
type, but also to determine the direction in which numerical calculations of the flow
profile must proceed.
When one compares the normal and critical depths for a particular channel geometry
and discharge, several possibilities arise. If one finds that y
n
is greater than y
c
, then the
channel is said to have a hydraulically mild (M) slope. If y
n
is equal to y
c
, the channel
has a critical (C) slope. If y
n
is less than y
c
, the channel is said to have a hydraulically
steep (S) slope. If the channel has a slope of zero, then normal depth does not exist,
and the channel is said to be horizontal (H). Finally, if the channel bottom elevation
increases as one moves in a downstream direction, the channel is said to be adverse
(A), and the normal depth is imaginary (that is, it involves the square root of a nega-
tive quantity).
Generally, the range of possible flow depths in an open channel can be subdivided
into three subranges, called zones. Zone 1 corresponds to all flow depths that are
larger than both the normal and critical depths. Zone 2 corresponds to flow depths
between normal and critical depths, and Zone 3 corresponds to flow depths smaller
than both normal and critical depths. For example, in a hydraulically mild (M) chan-
nel, where y
n
is greater than y
c
, Zone 1 corresponds to y >y
n
, Zone 2 corresponds to
y
c
<y <y
n
, and Zone 3 corresponds to y <y
c
. In a critical (C) channel, where y
n
equals y
c
, Zone 2 does not exist. Similarly, in horizontal (H) and adverse (A) chan-
nels, where normal depth is either non-existent or imaginary, Zone 1 does not exist.
By comparing a known depth of flow at a cross section with the normal and critical
depths for that discharge, one can conclude the flow profile type. For example, if the
known depth at a particular cross section in a mild (M) channel lies in Zone 1, the
flow profile is of the M1 type. If the known depth of flow at a cross section in a steep
(S) channel lies in Zone 3, the flow profile is of the S3 type. Figure 7.24 illustrates the
various flow profile types that may exist in prismatic channels and indicates the
shapes of the water surface profiles in terms of their curvatures.
Classification of the flow profile type in a channel for a particular discharge is an
important first step that must take place before the actual computation of the water
surface elevations along the length of the channel. The classification aids in the under-
standing of what the profile should look like, and hence provides a check on subse-
quent calculations. Understanding the flow profile type also helps in terms of
classifying the flow as being subcritical or supercritical. Regardless of the channel
slope classification (M, C, S, H, or A), a flow is subcritical when y is greater than y
c
and supercritical when y is less than y
c
.
Knowledge of whether a flow profile is subcritical or supercritical has implications
for profile computations. Computation of a subcritical profile should always proceed
in an upstream direction from a known depth at a downstream control section. Con-
versely, computation of a supercritical profile should always proceed in a downstream
direction from a known depth at an upstream control section. This distinction arises
because of fundamental differences in the ranges of influences in open-channel flow,
particularly with reference to the celerities (velocities) of small gravity waves caused
by flow disturbances.
Figure 7.24
Flow profile types for
prismatic channels
Credit: Chow, 1959.
Profile Calculations
Equation 7.71 is the fundamental equation describing a gradually varied flow profile.
Given a known flow depth at a control section (that is, a boundary condition), a solu-
tion of that equation to define depths at additional locations (cross sections) of interest
can be accomplished in a number of ways. Chow (1959) provides a discussion of
many of the historically used methods, as well as of methods that continue to be used
today. More modern methods involving numerical integration of the governing equa-
tion can also be developed.
There are two basic classifications of methods for solving Equation 7.71. In the first,
one solves for values of xthe cross section locationscorresponding to assumed
flow depths y. This type of method is generally useful only for prismatic channels
because the cross-sectional geometry is independent of location. The direct-step
method is an example of this type of procedure.
The second solution method classification is useful for non-prismatic channels. In this
instance, methods for solution of Equation 7.71 are designed to determine the flow
depth y at each specified channel location x. The most widely used of these
approaches is called the standard-step method. Further discussion of this method can
be found in texts on the subject of open-channel flow such as those by Chow (1959)
and Chaudhry (1993).
Because the focus of this text is prismatic channels, the remainder of the discussion is
limited to the direct-step method. To develop the direct-step method, the form of the
energy equation given in Equation 7.5 is written for a reach of channel as
(7.72)
whereZ
1
and Z
2
=channel bed elevation at sections 1 and 2, respectively (ft, m)
E
1
and E
2
=total energy at sections 1 and 2, respectively (ft, m)
h
L
=head loss due to friction (ft, m)
If the length of the channel reach between the two cross-sections is denoted by Ax,
then
h
L
= S
f
Ax (7.73)
where S
f
=friction slope
Ax =distance between cross sections 1 and 2
and
Z
1
= Z
2
+ S
o
Ax (7.74)
where S
o
=channel bed slope
Making these substitutions, Equation 7.72 can be rewritten in terms of Ax as
(7.75)
1 1 2 2 L
Z E Z E h + = + +
2 1
o f
E E
x
S S
A =

This last expression, while not resembling the differential Equation 7.70, provides the
basis for the flow profile computation. For a given discharge Q and channel cross-
sectional geometry, and for known or assumed flow depths y
1
and y
2
, one can deter-
mine the specific energies E
1
and E
2
. Using a constitutive relationship such as the
Manning equation, one can also determine the friction slopes S
f1
and S
f2
at each of the
two cross sections. If the average of those two slopes is used in Equation 7.74, one
can finally find the distance Ax between the two cross sections. The procedure is best
explained through an example.
Example 7.20 Gradually Varied Flow Analysis of a Channel Slope Transi-
tion. A slope change occurs in a trapezoidal open channel with B =5 mand z =2. Upstreamof the
break point, the channel slope is S
o
=0.001, and downstream of the break point the slope is
S
o
=0.020. Figure E7.20.1 illustrates the channel profile. If n = 0.025 and Q =15 m
3
/s, compute the
flow profile in the vicinity of the slope change. Assume that the channel lengths upstreamand down-
streamof the slope change are long.
Figure E7.20.1 Channel profile for Example 7.20
Solution: The critical depth in the channel is found, by trial and error, to be y
c
=0.86 m. Upstreamof
the slope break point, the normal depth is y
n1
=1.48 m, and downstreamof the slope break point the
normal depth is y
n2
=0.65 m. Thus, upstreamof the break point the channel is mild (M), and down-
streamof the break point it is steep (S).
In this situation, the control section occurs at the break point and the depth at the control section is the
critical depth of 0.86 m. Moving in an upstreamdirection fromthe break point, the flow profile depth
should converge toward the normal depth, and the flow profile is of the M2 type. Downstreamof the
break point, the flow profile is of the S2 type. Because an M2 profile is subcritical and an S2 profile is
supercritical, the entire flow profile must be computed in two parts. The M2 profile is computed in an
upstreamdirection fromthe control section, and the S2 profile is computed in a downstreamdirection
fromthe control section.
Calculations for the M2 profile are illustrated in Table7.8. The first column contains assumed flow
depths, beginning with the known depth at the flow control section and proceeding in an upstream
direction. [Note that as one moves upstream, the depth increases until it becomes the normal depth
(seeFigure 7.21). Assumed flow depths were developed by subtracting from the normal depth in 0.1-
m increments.] The second through fourth columns are computed on the basis of the depth and chan-
Section 7.9 Spatially Varied Flow 277
nel geometry. The fifth column, velocity, is computed as the ratio of the discharge to the cross-sec-
tional area, and the specific energy in the sixth column is computed using Equation 7.36. Friction
slopes, in the seventh column, have been computed for this example using the Manning equation.
The average friction slope for successive rows in the table is shown in column eight, and the corre-
sponding Ax, determined fromEquation 7.75, is shown in column nine. A running summation of Ax
values is shown in the last column of the table. Those values represent the distances upstreamof the
slope break at which the corresponding flow depths occur.
Table7.9 is identical to Table7.8, except that it pertains to the steep reach of the channel downstream
of the slope break point. (Assumed flow depths were developed by adding to the normal depth in 0.1-
mincrements.) Again, the summed Ax values shown in the last column of the table represent locations
at which the corresponding depths occur. In this case, however, the distances are measured down-
streamrather than upstreamfromthe slope break point. The computed water surface profile is illus-
trated in Figure E7.20.1.
7.9 SPATIALLY VARIED FLOW
With spatially varied flow in an open channel, the discharge varies along the channel
length, even in otherwise steady flow conditions. This situation occurs when flow is
either added to or diverted from the channel along its length. Flow may be added, for
example, in a side-channel spillway. A stormwater inlet on a continuous grade (dis-
cussed in Section 10.2) is an example where flow is diverted from the channel as it
passes the inlet opening.
Table 7.8 Direct-step flow profile calculations
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
y (m)
A (m
2
)
P (m) R (m) V (m/s) E (m) S
f
Avg S
f
Ax (m) EAx (m)
0.86 5.78 8.85 0.65 2.60 1.203 0.007427 0
0.98 6.82 9.38 0.73 2.20 1.226 0.004624 0.006026 4.6 4.6
1.08 7.73 9.83 0.79 1.94 1.272 0.003238 0.003931 15.4 20.1
1.18 8.68 10.28 0.85 1.73 1.332 0.002334 0.002786 33.7 53.8
1.28 9.68 10.72 0.90 1.55 1.402 0.001722 0.002028 68.5 122.3
1.38 10.71 11.17 0.96 1.40 1.480 0.001297 0.001510 152.1 274.4
1.48 11.78 11.62 1.01 1.27 1.563 0.000995 0.001146 565.7 840.1
Table 7.9 Direct-step flow profile calculations
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
y (m)
A (m
2
)
P (m) R (m) V (m/s) E (m) S
f
Avg S
f
Ax (m) EAx (m)
0.86 5.78 8.85 0.65 2.60 1.203 0.007427 0
0.80 5.28 8.58 0.62 2.84 1.211 0.009633 0.008530 0.7 0.7
0.75 4.88 8.35 0.58 3.08 1.233 0.012134 0.010884 2.3 3.0
0.70 4.48 8.13 0.55 3.35 1.271 0.015510 0.013822 6.3 9.3
0.65 4.10 7.91 0.52 3.66 1.334 0.020163 0.017837 28.9 38.2
Methods for dealing with spatially varied flow are distinguished in terms of whether
flow is added to or diverted from the channel. When flow is added to a channel, as in
the case of a side channel spillway, turbulence generated in the flow is such that
energy losses are greater than those due to friction alone, and the energy equation is
not easily applied. Therefore, in a channel with an increasing discharge, momentum
rather than energy should be used in computing the flow profile. In the case of a chan-
nel with a decreasing discharge, experience has shown that the energy equation can be
applied for flow profile computations because of the general lack of turbulence gener-
ated in this sort of flow.
Spatially Varied Flow with an Increasing Discharge
A derivation of the fundamental differential equation describing a spatially varied
flow with an increasing discharge is accomplished by applying the momentum princi-
ple to the flow in a channel reach. Assuming that the coefficients o and | are unity,
the resulting equation is (Chow, 1959)
(7.76)
where y =vertical depth of flow (ft, m)
x =distance along channel bottom (ft, m)
S
o
= channel bottom slope (ft/ft, m/m)
S
f
3
/s)
q =dQ/dx (cfs/ft, m
3
/s/m)
g = gravitational acceleration (32.2 ft/s
2
, 9.81 m/s
2
)
A = cross-sectional area of channel (ft
2
, m
2
)
Fr =Froude number
Note that this equation is identical to Equation 7.71 with the exception of the third
term in the numerator. In the third term, Q, is the discharge, which depends on the
location x along the length of the channel, and q =dQ/dx is the rate at which flow is
added to the channel along its length. Note that q has units of cfs per foot of channel
length, or of m
3
/s per meter of channel length. The friction slope S
f
appearing in the
equation, which accounts for friction along the channel bottom and sides, may be
evaluated using a constitutive relationship such as the Manning equation.
If the depth of flow is known at a control section, a numerical solution of Equation
7.76 may be used to determine the water surface profile in a channel. French (1985)
has shown that the location x
c
of the control section in a channel (measured from
upstream end of the reach) with Q =0 when x =0 is
(7.77)
2
2
2
1 Fr
o f
Qq
S S
dy gA
dx

=
2
3
2 4/ 3
2
1/ 3
8
c
c
c o
f c c
q
x
gn P
gT S
C T A
=
| |
|
|
\ .
Section 7.10 Chapter Summary 279
where x
c
=location of control section (ft, m)
T
c
=top width of flow at control section (ft, m)
n = Mannings roughness coefficient
P
c
=wetted perimeter at control section (ft, m)
C
f
=unit conversion factor (2.21 U.S. Customary units, 1.0 SI units)
A
c
=cross-sectional area of flow at control section (ft, m)
Critical depth occurs at the control section location determined by Equation 7.77.
Solution of Equation 7.77 must be accomplished in an iterative, trial-and-error man-
ner. One begins by assuming a value x
c
, from which Q = qx
c
. For this discharge, one
determines the corresponding critical depth y
c
, and then the values of A
c
, P
c
, and T
c
.
Finally, Equation 7.76 is applied to compute a new value of x
c
. If the computed value
of x
c
agrees with the assumed one, convergence is attained. Otherwise, the procedure
is repeated with a new trial value.
If the location x
c
of the control section lies within the length L of the channel that is
subjected to lateral inflow, then flow upstream of that point will be subcritical, and
flow downstream of that point will be supercritical. If x
c
is greater than L, then flow in
the channel will be subcritical throughout and will be controlled by conditions exist-
ing at the location x = L. Even if x
c
is less than or equal to L, the critical section may
be drowned by downstream conditions.
Spatially Varied Flow with a Decreasing Discharge
The equation describing spatially varied flow with a decreasing discharge is derived
using the energy principle, as shown by Chow (1959). Assuming that the coefficients
o and | are unity, the equation is
(7.78)
Note that this expression is identical to Equation 7.76, except for the coefficient on
the last term in the numerator. Note also that in the case of spatially varied flow with a
decreasing discharge, the quantity q = dQ/dx is negative.
As in the case of Equation 7.75, a numerical method may be employed to solve Equa-
tion 7.78 to yield a flow profile.
7.10 CHAPTER SUMMARY
Open channels are extremely common in both natural and constructed environments,
and include streams, swales, ditches, gutters, and gravity-flow pipes. Unlike closed-
conduit flow analysis, descriptors such as cross-sectional flow area and average
velocity are unknown variables when evaluating flow in open channels. Their values
depend on channel shape, slope, and roughness characteristics, and on channel bound-
ary conditions.
2
2
1 Fr
o f
Qq
S S
dy gA
dx

=

Open channels may be classified as either prismatic or non-prismatic. Common cross-
sectional shapes for prismatic channels are rectangular, trapezoidal, triangular, and
circular. Channels may also be classified on the basis of whether they are lined or
unlined.
When evaluating open-channel flow, the energy equation takes the form
(7.79)
Uniform flow in an open channel means that the flow depth, discharge, and other flow
properties are constant. This condition can occur in a sufficiently long prismatic chan-
nel. The depth at which uniform flow occurs is called normal depth. Constitutive rela-
tionships such as the Manning and Chzy equations can, in addition to being used to
compute friction losses, be used to predict the normal depth for a given channel and
discharge.
Critical flow is another condition for which a unique relationship exists between
depth and discharge. It also corresponds to the point of minimum specific energy (the
energy relative to the channel bottom) for a given channel geometry and discharge.
Critical flow often occurs at channel control sections, such as at free overfall loca-
tions. The concept of specific energy of flow can also be applied to the evaluation of
flow properties at channel transitions such as expansions, contractions, and steps.
Momentum and specific force concepts are used in evaluating the depths upstream
and downstream of hydraulic jumps, and in computing the forces on hydraulic struc-
tures such as channel transitions and bridge piers. The momentum equation for open-
channel flow is written as
(7.80)
2 2
1 1 2 2
1 1 2 2
cos cos
2 2
L
V V
Z d Z d h
g g
o o
u u + + = + + +
1 2 2 2 1 1
sin ( )
x f
F P P W F Q V V u | | = + =
_
References 281
For a short, horizontal channel reach, specific force can be computed as
(7.81)
Weirs are used to control or measure flow in open channels. Weirs may be classified
as either sharp- or broad-crested. They may also be classified by shape; common
shapes are rectangular and triangular. Finally, weirs can be classified as to whether the
weir flow is contracted or suppressed. Equations and coefficients used in computing
the flow or head for some types of weirs were presented in the chapter.
When depth and velocity gradually change along the length of a channel, the condi-
tion is termed gradually varied flow. Evaluation of gradually varied flow involves
classifying the type of profile based on the relative values of normal depth and critical
depth for the channel, and on the location of the water surface relative to these depths.
Once the profile type is known, corresponding flow depths and locations along the
channel reach are determined.
If discharge varies along a channel reach, spatially varied flow exists. If the discharge
increases along the channel reach, momentum concepts (rather than energy) should be
used to evaluate the flow with a trial-and-error procedure. If the discharge is decreas-
ing, energy concepts can be used.
REFERENCES
Abbott, M. B. 1979. Computational Hydraulics: Elements of the Theory of Free Surface Flows. London:
Pitman.
American Concrete Pipe Association (ACPA). 2000. Concrete Pipe Design Manual. 13th Printing
(revised). Irving, Texas: ACPA.
Bazin, H. 1897. A New Formula for the Calculation of Discharge in Open Channels. Mmoire No. 41,
Annales des Ponts et Chauses, Paris 14, Ser. 7, 4
th
trimester: 2070.
Blevins, R. D. 1984. Applied Fluid Dynamics Handbook. New York: Van Nostrand Reinhold.
Chow, V. T. 1959. Open-Channel Hydraulics. New York: McGraw Hill.
Chaudhry, M. H. 1993. Open-Channel Flow. Englewood Cliffs, New J ersey: Prentice Hall.
Chaudhry, M. H., and S. M. Bhallamudi. 1988. Computation of Critical Depth in Compound Channels.
Journal of Hydraulic Research 26, no. 4: 377395.
Cunge, J . A., F. M. Holly, and A. Verwey. 1980. Practical Aspects of Computational River Hydraulics.
London: Pitman.
Einstein, H. A. 1934. The Hydraulic or Cross Section Radius. Schweizerische Bauzeitung 103, no. 8: 89
91.
Einstein, H. A., and R. B. Banks. 1951. Fluid Resistance of Composite Roughness. Transactions of the
American Geophysical Union 31, no. 4: 603610.
Finnemore, E. J . and J . Franzini. 2000. Fluid Mechanics with Engineering Applications. New York:
McGraw-Hill.
Franzini, J . and E. J . Finnemore. 1997. Fluid Mechanics, 9th Ed. New York: McGraw-Hill.
2
s
Q
F zA
gA
| |
= + |
|
\ .
Fread, D. L 1993. Flow Routing in Maidment, D. R., ed. Handbook of Hydrology. New York: McGraw-
Hill.
French, R. H. 1985. Open-Channel Hydraulics. New York: McGraw-Hill.
Ganguillet, E., and W. R. Kutter. 1869. An Investigation to Establish a New General Formula for Uniform
Flow of Water in Canals and Rivers (in German), Zeitschrift des Oesterreichischen Ingenieur und
Architekten Vereines vol. 21, no. 1: 625; no. 2: 4659. Translated into English as a book by R. Hering
and J .C. Trautwine, J r., A General Formula for the Uniform Flow of Water in Rivers and Other Chan-
nels, J ohn Wiley and Sons, New York, 1888.
Herschel, C. 1897. On the Origin of the Chzy Formula. Journal of the Association of Engineering Soci-
eties 18: 363368.
Horton, R. 1933. Separate Roughness Coefficients for Channel Bottomand Sides. Engineering News-
Record 111, no. 22: 652653.
Krishnamurthy, M., and B. A. Christensen. 1972. Equivalent Roughness for Shallow Channels. Journal
of the Hydraulics Division, ASCE 98, no. 12: 22572263.
Lotter, G. K. 1933. Considerations of Hydraulic Design of Channels with Different Roughness of Walls.
Transactions All Union Scientific Research, Institute of Hydraulic Engineering 9: 238241.
Manning, R. 1891. On the Flow of Water in Open Channels and Pipes. Transactions, Institution of Civil
Engineers of Ireland 20: 161207.
Motayed, A. K., and M. Krishnamurthy. 1980. Composite Roughness of Natural Channels. Journal of the
Hydraulics Division, ASCE 106, no. 6: 11111116.
Muhlhofer, L. 1933. Rauhigkeitsunteruchungen in einemstollen mit betonierter sohle und unverkleideten
wanden. Wasserkraft und Wasserwirtschaft 28(8): 8588.
Pavlovskii, N. N. 1931. On a Design Formula for UniformMovement in Channels with Nonhomogeneous
Walls (in Russian), Izvestiia Vsesoiuznogo Nauchno-Issledovatelskogo Insituta Gidrotekhniki 3: 157
164.
Powell, R. W. 1950. Resistance to Flow in Rough Channels. Transactions, American Geophysical Union
31, no. 4: 575582.
Rouse, H. 1936. Discharge Characteristics of the Free Overfall. Civil Engineering 6(7): 257.
Rouse, H. 1946. Elementary Fluid Mechanics. New York: Wiley.
Strickler, A. 1923. Some Contributions to the Problemof the Velocity Formula and Roughness Factors for
Rivers, Canals and Closed Conduits (in German). Mitteilungen des Eidgenossischen Amtes fur Wasser-
wirtschaft 16.
U.S. Army Corps of Engineers (USACE). 1991. HEC-2, Water Surface Profiles, User's Manual. Davis,
California: USACE Hydrologic Engineering Center.
U.S. Bureau of Reclamation (USBR). 1955. Research Studies on Stilling Basins, Energy Dissipators, and
Associated Appurtenances, Hydraulic Laboratory Report No. Hyd-399. Lakewood, Colorado: USBR.
Problems 283
PROBLEMS
7.1 Fromthe natural channel cross-section data given in the table, determine the area, wetted perimeter,
hydraulic radius, top width, and hydraulic depth if the depth of flow is 4 ft.
7.2 Determine the area of flow, wetted perimeter, hydraulic radius, top width, and hydraulic depth for
each channel described. The flow depth for all channels is 1.22 m.
a) rectangular channel with a bottomwidth of 1.5 m
b) triangular channel with side slopes of 2 horizontal to 1 vertical
c) semi-circular channel with a radius of 2.1 m
d) trapezoidal channel with a bottomwidth of 3.0 mand side slopes of 3 horizontal to 1 vertical
7.3 Demonstrate that the hydraulic depth D
h
in a rectangular channel is equal to the actual depth of flow.
Also demonstrate that the hydraulic depth is smaller than the actual depth of flow in trapezoidal and
triangular channels.
7.4 Determine the discharge in a 3-ft diameter sewer pipe if the depth of flow is 1 ft and the slope of the
pipe is 0.0019. Use a value of 0.012 for the Manning roughness coefficient.
7.5 An asphalt-lined trapezoidal channel has the dimensions indicated in the figure. The channel slope is
0.001. Use the Chzy equation to find the discharge if B =4.9 mand y =1.4 m.
7.6 By comparing the Manning and Chzy equations, show that the relationship between n and C is C =
1.49R
1/6
/n if R is in feet.
7.7 Explain why uniformflow cannot occur in frictionless and horizontal channels.
7.8 Determine the normal and critical depths in a triangular channel if Q =5 cfs, n =0.045, and
S
o
=0.008. Is uniformflow in this channel subcritical or supercritical?
7.9 A 36-in. diameter pipe with n =0.013 and S
o
=0.01 conveys a discharge of 25 cfs. Determine the
depth and velocity of flow in the pipe.
Station (ft) Elevation (ft)
0 10.1
5 7.8
15 5.5
18 6.0
22 8.2
26 9.8
30 10.7
7.10 Determine the critical depth in a trapezoidal channel if the discharge is 14.2 m
3
/s. The width of the
channel bottomis 6.1 m, and the side slopes are 1:1.
7.11 In the triangular section shown, the flow is 350 cfs. Using n =0.012, calculate the critical depth and
the critical velocity.
7.12 The discharge is 7.1 m
3
/s in a 3-m-wide rectangular section. Plot the specific energy as a function of
depth of flow.
7.13 Calculate the critical depth of the channel in Problem7.12 using Equation 7.38.
7.14 Determine the sequent depth and length of a hydraulic jump in a rectangular channel with B =5 ft, Q
=150 cfs, n =0.013, and S
o
=0.05. Flow upstreamof the jump is uniform. What is the energy loss in
the jump?
7.15 Water flows through a rectangular section at a rate of 14.2 m
3
/s. The channel is 6.1 mwide, and the
upstreamdepth is 0.94 m. The flow undergoes a hydraulic jump as indicated in the diagram. Com-
pute the:
a) downstreamdepth
b) energy loss in the hydraulic jump
Problems 285
7.16 A sharp-crested, contracted, rectangular weir (N =2) is 20 ft long. The height of the weir crest above
the channel is 8 ft. Develop and plot a rating curve (that is, a relationship between Q and H) for this
weir. Use H values ranging from 0 to 3 ft to develop the rating curve.
7.17 Water flows in a 6.1-m-wide rectangular concrete channel at a rate of 14.2 m
3
/s. The slope changes
from0.0125 to 0.05. Plot the water surface profile through the transition.
Stormwater is
conveyed through
this rectangular,
concrete channel,
which is fenced for
safety reasons.

SCMD 09 CH07 Flow in Open Channels

Uploaded by

Copyright:

Available Formats

SCMD 09 CH07 Flow in Open Channels

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SCMD 09 CH07 Flow in Open Channels

Uploaded by

Copyright:

Available Formats

STORMWATER CONVEYANCE

MODELING AND DESIGN

Section 7.8 Gradually Varied Flow 273

276 Flow in Open Channels Chapter 7

280 Flow in Open Channels Chapter 7

You might also like