2/10/23

UNIFORM FLOW

What do we need to consider in OCF design?

• Largely based on the concept of uniform flow

• Two main quantities require our attention:
1. Design discharge
• Acceptable risk and frequency analysis
• Channel slope – topography dependent, soil conditions etc.
• Cross-sectional area – topography dependent
2. Resistance coefficient
• Vital in the design – shear stress
• Unique value of depth of flow – normal depth

1
 2/10/23

Example 1. Sewer Design

• A typical problem encountered in practice may take the below
form:
“Design a concrete sewer that has a maximum design discharge of
1.0 m3/s and a minimum discharge of 0.2 m3/s if its slope is 0.0018.”

Q: How is this determined?

A: By assuming a uniform flow condition and carrying out the
required calculations

We will return to this idea once we have explored the concept

of uniform flow in greater detail.

Review: What do we know about uniform flow?

• Gravity drives the flow
• The weight of water in the flow
direction causes acceleration
• Shear stresses at the bottom and y
Sw
sides of the channel resist the flow x

• For long prismatic channels with So

constant So, roughness, and A, a
balance of forces is eventually
attained
• From that point the depth y and
velocity V will stay constant, resulting
in a condition of uniform flow (UF)
• Thus, the specific energy E also
remains constant
4

2
 2/10/23

Review: What do we know about uniform flow?

• The flow depth in UF is called the normal
depth yn
• Since y = yn at all x in uniform flow, So (the
bed slope) is parallel to the free surface slope
Sw (dy/dx) y
Sw
x
• What are the objectives of uniform flow
analysis? So

1. Develop an equation relating the bottom

shear stress to different flow variables.
2. Understand the use of various empirical
resistance formulae that are used in UF.
3. Establish a procedure for computing the
normal depth for a specified discharge in
channel with known properties.

A QUICK REVIEW : CONSERVATION OF MASS AND ENERGY EQUATIONS

Continuity equation Q = A1V1 = A2V2

PA VA2 r ga V 2 V2
H A = zA + + = z + ( y - a) + + = z+ y+
r g 2g r g 2g 2g

V2
H = z+ y+
2g
V12 V2
Energy equation: z1 + y1 + = z2 + y2 + 2 + hL
2g 2g
L V2 L V2
hL = f = f
Dh 2 g Rh 8 g
z1 - z2 z1 - z2 The slope of the
S0 = tan a = @ bottom of the channel
x2 - x1 L

If the channel bottom is straight (the bottom slope is constant):

V12 V2
Energy equation: y1 + + S0 L = y2 + 2 + hL
2g 2g

3
 2/10/23

Friction Slope
• In OCF design, So is selected such that it provides adequate elevation drop (Δz) to
overcome the frictional head loss (hL) and thus maintain the desired Q
• Therefore, there is a close connection between hL and So , and it makes sense to
express hL as a slope (or the tangent of an angle). This is done by defining a friction
slope Sf as
ℎ"
𝑆! =
𝐿
• The energy equation then takes the form
𝑉#$ 𝑉$$
𝑦# + = 𝑦$ + + 𝑆! − 𝑆% 𝐿
2𝑔 2𝑔
• From the above, Sf is equal to So when hL = z1 - z2
• Recall from specific energy considerations the EGL is located along z + y + V2/2g
above some horizontal reference datum.
• The EGL is typically sloped down like the channel itself due to frictional losses, the
vertical drop being equal to the head loss hL and thus the slope being the same as Sf

Analysis of Uniform Flow

• Consider the forces on a CV comprising a section of uniform flow

• We have
• Fp1 and Fp2 (the pressure forces on the US and DS faces of the CV)
• The weight of water W, equal to the product of the specific weight of water
γ, the cross-sectional area of flow A and the length of the CVΔL
• The resisting shear force τoPΔL, where P is the wetted perimeter and τo is
the mean boundary shear stress
8

4
 2/10/23

Analysis of Uniform Flow

• In this case, the free surface is
assumed to be non-deforming since the
flow is uniform and y is unchanging
• Recall the momentum equation for
steady UF with one inlet and one outlet

̇ out − 𝑉in ) = ∑𝐹⃗&'

𝑚(𝑉
• In the above expression, the momentum flow rates in and out of the CV cancel out,
as well as the pressure forces at the outlet and inlet, which leaves

𝜏𝑜𝑃Δ𝐿 = 𝛾𝐴Δ𝐿 𝑠𝑖𝑛𝜃

• Dividing both sides by PΔL and understanding that the hydraulic radius R = A/P
𝜏𝑜 = 𝛾𝑅 𝑠𝑖𝑛𝜃

Analysis of Uniform Flow

• For small values of θ, sinθ ≈ tanθ =
the bed slope So
• Recall that for steady UF,

𝑉"# 𝑉##
𝑦" + + 𝑆$𝐿 = 𝑦# + + ℎ%
2𝑔 2𝑔
ℎ%
𝑆$ =
𝐿 hL

• From above, it can be seen that So

is equal to the slope of the EGL
since the head loss is equal to the
elevation drop

𝜏𝑜 ℎ % 𝜏𝑜 = 𝛾𝑅𝑆𝑜
=
𝛾𝑅 𝐿
10

5
 2/10/23

Bed Shear Stress

• This allows us to return to our first objective:

Develop an equation relating the bottom shear stress to different flow variables.
• First, we need to consider the variables which affect the mean bed shear stress τo
• From dimensional analysis, we can get
𝜏𝑜 = 𝑓 𝜌, 𝜇, 𝑔, 𝑉, 𝑅, 𝑘𝑠, 𝐶& , 𝑁, 𝑈
• In the above expression, Cs, N, and U are non-dimensional and represent the effect of
cross-sectional shape, degree of non-uniformity and unsteady effects, respectively
• Using the Buckingham Pi theorem to form dimensionless groups gives

𝜏𝑜 𝜌𝑉 4𝑅 𝑘𝑠
= 𝑓(𝑅𝑒 = , 𝑅𝑟 = , 𝐹 , 𝐶 , 𝑁, 𝑈)
𝜌𝑉2 𝜇 4𝑅 ' &

11

The Chezy Coefficient and Chezy’s Equation

• In order to now understand the use of various empirical resistance formulae that are
used in UF (i.e. objective 2), we consider
𝜏𝑜 𝜌𝑉 4𝑅 𝑘𝑠
= 𝑓(𝑅𝑒 = , 𝑅𝑟 = , 𝐹 , 𝐶 , 𝑁, 𝑈)
𝜌𝑉2 𝜇 4𝑅 ! "
• From the above expression, we can see that τo is proportional to ρV2
τo 𝛼 𝜌𝑉2 or alternatively τo = 𝑘𝜌𝑉2
where k is a dimensionless constant
• Rearranging and substituting 𝜏𝑜 = 𝛾𝑅𝑆$ gives

𝜏% 𝑔
𝑉= = 𝑅𝑆 = 𝐶 𝑅𝑆$
𝑘𝜌 𝑘 $

• In the above expression, C

is the Chezy coefficient (sometimes called Chezy’s
constant), which has the dimensions length1/2/time

12

6
 2/10/23

The Chezy Coefficient and Chezy’s Equation

• This can also be derived from our previous analysis of UF, which gave
𝜏𝑜 ℎ&
= = 𝑆%
𝛾𝑅 𝐿
• Recall the Darcy-Weisbach friction factor, f
8𝜏𝑜
𝑓=
𝜌𝑉2
,! ! ''
• Substituting gives 𝑆+ = = , from which
-. /. $0

10
𝑉= 𝑅𝑆% = 𝐶 𝑅𝑆% where 𝐶 = 8𝑔/𝑓
!
()
where 𝑓 = 2.0 log 14.8𝑅 /𝜀

• The Chezy coefficient ranges from about 30 m1/2/s for small channels with rough surfaces to 90
m1/2/s for large channels with smooth surfaces

13

The Modified Moody Diagram

• Values of the Chezy coefficient for various channel roughness types have been evaluated
experimentally
• The modified Moody diagram can also be utilized to
estimate C
• The shear velocity u* has been defined as
𝜏%
u∗ = = 𝑔𝑅𝑆%
𝜌

• The shear velocity is used as the velocity scale in

the Reynolds number, giving a value known as the
roughness Reynolds number, Re* which can be used
to define surfaces as smooth, transitional or fully
rough
𝑅𝑒∗ = 𝑘𝑠𝑢∗ /𝜇
• Here, ks is the roughness height Re = 4VR/n
Smooth: Re* < 4
Transitional: 4 < Re* < 100
Fully rough: Re* > 100

14

7
 2/10/23

Moody
chart

15

The Chezy Coefficient

• There are expressions for calculation of C; however, as they are based on
pipe data, they are valid only for small channels with smooth surfaces
• For flows over smooth surfaces (sometimes called smooth flows):

𝐶 = 28.6𝑅𝑒 "/0 𝑖𝑓 𝑅𝑒 < 101

𝑅𝑒 8𝑔
𝐶 = 4 2𝑔 log"$ 𝑖𝑓 𝑅𝑒 > 101
2.51𝐶
• For flows over rough surface (sometimes called rough flows):

𝑘 2.5
𝐶 = −2 8𝑔 log"$ +
12𝑅 𝑅𝑒 𝑓

16

8
 2/10/23

The Manning Equation

• The Chezy equation was first developed in 1968 but estimation thereof can be complicated
since, as noted in the previous slide, there are many parameters which are required for
estimation of C
• Accordingly, Manning developed a new formula for estimation of V or Q, known as the
Manning equation:
𝒌𝒏 𝟐/𝟑 𝒌𝒏 𝟐/𝟑
𝑽= 𝑹𝒉 𝑺𝟏/𝟐 𝑸= 𝑨 𝑹𝒉 𝑺𝟏/𝟐
𝒏 𝒏

• The value of kn is dependent on whether or not V is being calculated using SI units or imperial
units
𝑘3 = 1 m4/5 /s = 3.2808 ft 4/5 /s = 1.486 ft4/5 /s
• Here, n is the Manning coefficient and has the dimensions of (length)1/3/time
• Like the Chezy coefficient, n has been determined experimentally for various surfaces
• Note: different sources may provide slightly different values of n for the same surface type, but
this should not significantly affect calculation of V

17

18

9
 2/10/23

Mean values of the Manning coefficient n

for water flow in open channels*
From Chow (1959).

Wall Material n
A. Artificially lined channels
Glass 0.010
Brass 0.011
Steel, smooth 0.012
Steel, painted 0.014
Steel, riveted 0.015
Cast iron 0.013
Concrete, finished 0.012
Concrete, unfinished 0.014
Wood, planed 0.012
Wood, unplaned 0.013
Clay tile 0.014
Brickwork 0.015
Asphalt 0.016
Corrugated metal 0.022
Rubble masonry 0.025
B. Excavated earth channels
Clean 0.022
Gravelly 0.025
Weedy 0.030
Stony, cobbles 0.035
C. Natural channels
Clean and straight 0.030
Sluggish with deep pools 0.040
Major rivers 0.035
Mountain streams 0.050
D. Floodplains
Pasture, farmland 0.035
Light brush 0.050
Heavy brush 0.075
Trees 0.150
* The uncertainty in n can be ±20 percent or more.

19

20

10
 2/10/23

Example 1. Determining Q using Manning’s equation

Water is flowing in a weedy excavated earth channel of trapezoidal cross section with a bottom
width of 0.8 m, trapezoid angle of 60° and a bottom slope angle of 0.3°. If the flow depth is
measured to be 0.52 m, determine the flow rate of water through the channel. Repeat for a
bottom slope of 1°.
6& )/5
• Recall that 𝑄 = 𝐴 𝑅7 𝑆 4/) y = 0.52 m
3
θ = 60°

b = 0.8 m

21

Example 1. Determining Q using Manning’s equation

Water is flowing in a weedy excavated earth channel of trapezoidal cross section with a bottom
width of 0.8 m, trapezoid angle of 60° and a bottom slope angle of 0.3°. If the flow depth is
measured to be 0.52 m, determine the flow rate of water through the channel. Repeat for a
bottom slope of 1°.
6& )/5
• Recall that 𝑄 = 𝐴 𝑅7 𝑆 4/) y = 0.52 m
3
θ = 60°
• Refer to geometric properties
b = 0.8 m
8 $.=)
• 𝐴 =𝑦 𝑏+ = 0.52 m 0.52 m + = 0.5721 m2
9:3; 9:3>$°
)8 )($.=) C)
• 𝑝=𝑏+ = 0.8 m + = 2.001 m
"A3; "A3>$°
E 0.5721 m2
• 𝑅7 = = = 0.2859 m
F ).$$4 C

• 𝑆% = 𝑡𝑎𝑛0.3° = 0.005236
4\1/3/s ' )
• 𝑄= (0.5721 m2) 0.2859 m ( (0.005236 )' = 0.60m3/s
$.$5

22

11
 2/10/23

Example 1. Determining Q using Manning’s equation

Water is flowing in a weedy excavated earth channel of trapezoidal cross section with a bottom
width of 0.8m, trapezoid angle of 60° and a bottom slope angle of 0.3°. If the flow depth is
measured to be 0.52 m, determine the flow rate of water through the channel. Repeat for a
bottom slope of 1°.
6& )/5
• Recall that 𝑄 = 𝐴 𝑅7 𝑆 4/) y = 0.52 m
3
θ = 60°
• Refer to Table 1-1 or similar
b = 0.8 m
8 $.=)
• 𝐴 =𝑦 𝑏+ = 0.52 m 0.52 m + = 0.5721 m2
9:3; 9:3>$°
)8 )($.=) C)
• 𝑝=𝑏+ = 0.8 m + = 2.001 m
"A3; "A3>$°
E 0.5721 m2
• 𝑅7 = = = 0.2859 m
F ).$$4 C

23

Normal Depth, yn

• You may have noted that both the Chezy and Manning equation are dependent on the
dimensions of flow, including the flow depth, y
• As previously described, typical channel design is carried out for a desired flowrate Q at a
uniform flow conditions; a specified flow depth is rarely prescribed (although there may be
limits on a maximum value depending on the local topography)
• Our role is then to establish a procedure for computing the normal depth for a specified
discharge in channel with known properties (i.e. objective 3)
• From the Manning equation, we can decipher that for a given flowrate, channel material and
bed slope, we are left with a function of yn
𝑘3 )/5
𝑄= 𝐴 𝑅7 𝑆 4/)
𝑛
• Determination of yn is then an iterative exercise using trial-and-error or an equation solver or
select design charts

• In this course, we will predominantly use trial-and-error

24

12
 2/10/23

s = side slope

Co = kn = 1 in SI
units and 1.49 in
FPS

25

Example 2. Determining yn for Given Q in a Rectangular Channel

Water is to be transported in an unfinished concrete rectangular channel with a bottom width of

4 ft at a rate of 51 ft3/s. The terrain is such that the channel bottom drops 2 ft per 1000 ft length.
Determine the minimum height of the channel under uniform-flow conditions. What would
your answer be if the bottom drop is just 1 ft per 1000 ft length?
6& )/5 Q = 51 ft3/s
• Recall that 𝑄 = 𝐴 𝑅7 𝑆 4/)
3

b = 4 ft

26

13
 2/10/23

Example 2. Determining yn for Given Q in a Rectangular Channel

Water is to be transported in an unfinished concrete rectangular channel with a bottom width of

4 ft at a rate of 51 ft3/s. The terrain is such that the channel bottom drops 2 ft per 1000 ft length.
Determine the minimum height of the channel under uniform-flow conditions. What would your
answer be if the bottom drop is just 1 ft per 1000 ft length?
6& )/5 Q = 51 ft3/s
• Recall that 𝑄 = 𝐴 𝑅7 𝑆 4/)
3
4.IJ KL)/( /M 8& ) KL
• 51 ft3/s = 4 ft (𝑦3 )( ))/5 ( )4/)
$.$4I IN8& 4$$$ KL y
• Expanding and simplifying gives a nonlinear equation in yn
• We can either use an equation solver or trial-and-error to solve for yn
• This gives yn = 2.5 ft b = 4 ft

27

Example 2. Determining yn for Given Q in a Rectangular Channel

Water is to be transported in an unfinished concrete rectangular channel with a bottom width of

4 ft at a rate of 51 ft3/s. The terrain is such that the channel bottom drops 2 ft per 1000 ft length.
Determine the minimum height of the channel under uniform-flow conditions. What would your
answer be if the bottom drop is just 1 ft per 1000 ft length?
6& )/5 Q = 51 ft3/s
• Recall that 𝑄 = 𝐴 𝑅7 𝑆 4/)
3

b = 4 ft

28

14
 2/10/23

Example 3. Determining yn for Given Q in a Trapezoidal Channel

Compute the normal depth in a trapezoidal channel having a bottom-width of 10 m, side slopes
of 2H to 1V and carrying a flow of 30 m3/s. The slope of the channel bottom is 0.001 and n =
0.013.
6& )/5
• Recall that 𝑄 = 𝐴 𝑅7 𝑆 4/) yn
3 1
2

b = 10 m

29

Example 3. Determining yn for Given Q in a Trapezoidal Channel

Compute the normal depth in a trapezoidal channel having a bottom-width of 10 m, side slopes
of 2H to 1V and carrying a flow of 30 m3/s. The slope of the channel bottom is 0.001 and n =
0.013.
6& )/5
• Recall that 𝑄 = 𝐴 𝑅7 𝑆 4/) yn
3 1
2
• 𝐴 = 𝑦3 𝑏 + 2𝑦3
b = 10 m
• 𝑝 = 𝑏 + 2 5𝑦3
E+
• 𝑅7 =
F
[𝑦3(20 + 2𝑦3)]=/5 − 12.33(10 + 4.47𝑦3))/5 = 0

30

15
 2/10/23

𝒌𝒏
Slope classification Mild (yo > yc) 𝑸=
𝟐/𝟑
𝑨 𝑹𝒉 𝑺𝟏/𝟐
𝒏
Critical (yo = yc)
Steep(yo < yc)

Mild (So < Sc)

Critical (So = Sc)
Steep(So > Sc)

31

For a value of So = 0.004,

The figure on the right is drawn for a specific value

of So = 0.004 and n = 0.015

32

16
 2/10/23

Rectangular channel

We now know: For a rectangular channel: and q2 = g yc3

Using expressions for area and hydraulic radius:

yc/b = 0.1667

In a non-dimensional form:

33

34

17
 2/10/23

Why is this important?

35

Example

We can solve for yc. Using trial and error

we can get two roots, one for yc/b > 1/6
and another for yc/b < 1/6. We get: yc/b =
Using the limit slope equation: 1.098 and 0.0115, from which yc = 3.29 m
and 0.0345 m.

and qc = 0.02 m2/s (0.22 ft2/s)

Qc = qcb = 56.1 m3/s and 0.06 m3/s

0.007 in:

36

18
 2/10/23

Using at F = 1

If the slope could be constructed to be less

than the limit slope of 0.00409, then it would
be mild for all discharges

37

Compound Channels

• Thus far, we have dealt only with channels that have a uniform surface roughness
• However, it should be noted that surface roughness and thus n for most natural and some
human-made channels vary along the wetted perimeter and sometimes even along the
channel

• For example, a natural river may have a stony bottom for its regular bed but a surface covered
with bushes for its extended floodplain
The McPartlan canal
connecting Lake Pedler to
Lake Gordon, Tasmania, a
trapezoidal concrete-lined
channel (Chanson 2004)

Experimental flume
experiment with a central
main channel and two
adjacent floodplains lined
with artificial grass (Dupuis
2016)

38

19
 2/10/23

Compound Channels

• For a low flood stage, we may have only to be concerned with flow in the main channel
• During a flood event of some magnitude, when water levels rise above the main channel and
enter the surrounding floodplain, we have a compound channel with different sections of
varying n

The McPartlan canal

connecting Lake Pedler to
Lake Gordon, Tasmania, a
trapezoidal concrete-lined
channel (Chanson 2004)

Experimental flume
experiment with a central
main channel and two
adjacent floodplains lined
with artificial grass (Dupuis
2016)

39

Manning’s n for Compound Channels

• There are several methods for solving such problems:

• We can determine an effective Manning coefficient n for the entire channel cross section
• There are several empirical formulae for calculating an effective/equivalent n
• We can also consider the channel in subsections and apply the superposition principle
• The channel cross section is divided into N subsections, each with its own n and Q
• When determining the perimeter of a section, only the wetted portion of the
boundary for that section is considered
• The imaginary boundaries are ignored
• Q is the sum of the flow rates through all sections
• We are assuming that there are no losses due to shear stress between the
subsections

40

20
 2/10/23

Expressions for Calculation of Equivalent n, also called ne

• If we consider a channel section that is divided into N subsections, each with a wetted
perimeter Pi and Manning coefficient ni where i = 1,2,…, N.
• From this, several expressions for determining ne have been derived:
• Horton’s formula: if we assume that the mean flow velocity in each of the subsections is
equal to the mean flow velocity:
)/5
∑ 𝑃A 𝑛A 5/)
𝑛O =
∑ 𝑃A

• An alternate expression for ne is derived by assuming that the total force resisting the flow
is equal to sum of the forces resisting the flow in each subarea (Einstein and Banks
1934)
4/)
∑ 𝑃A 𝑛A )
𝑛O =
∑ 𝑃A 4/)

41

Expressions for Calculation of Equivalent n, also called ne

• Assuming that the total discharge is equal to sum of the discharge in each subsection
(Lotter 1933):
𝑃𝑅=/5
𝑛O =
∑ 𝑃A 𝑅A =/5
𝑛A

• Assuming a logarithmic velocity distribution (Krishnamurthy and Christensen 1972):

∑ 𝑃A 𝑦A 5/) ln(𝑛A )
𝑛O =
∑ 𝑃A 𝑦A 5/)

42

21
 2/10/23

Example 4. Comparison of Expressions for ne

A trapezoidal channel has vegetated banks (n = 0.040) and a stable bottom (n = 0.025). The
channel bottom width is 10 ft and the side slopes are 4:1. Find the equivalent or composite
value of n using the four empirical methods described, if the flow depth is given as 3 ft.

)/5
∑ 𝑃A 𝑛A 5/)
Horton 𝑛O = yn = 3 ft
∑ 𝑃A 1
4
4/)
∑ 𝑃A 𝑛A )
Einstein and Banks 𝑛O = b = 10 ft
∑ 𝑃A 4/)
𝑃𝑅 =/5
𝑛O =
Lotter ∑ 𝑃A 𝑅A =/5
𝑛A
Krishnamurthy and ∑ 𝑃A 𝑦A 5/) ln𝑛A
Christensen 𝑛O =
∑ 𝑃A 𝑦A 5/)

43

Example 4. Comparison of Expressions for ne

A trapezoidal channel has vegetated banks (n = 0.040) and a stable bottom (n = 0.025). The
channel bottom width is 10 ft and the side slopes are 4:1. Find the equivalent or composite
value of n using the four empirical methods described, if the flow depth is given as 3 ft.
)/5
∑ 𝑃A 𝑛A 5/)
Horton 𝑛O =
∑ 𝑃A

1 3
2 yn = 3 ft
Subsection 5/) 1
Pi ni 𝑃A 𝑛A
No. 4
1 12.37 0.040 0.099
b = 10 ft
2 10 0.025 0.040
3 12.37 0.040 0.099 )/5
0.238
𝑛O = = 0.036
∑ 34.74 0.238 34.74

44

22
 2/10/23

Example 4. Comparison of Expressions for ne

A trapezoidal channel has vegetated banks (n = 0.040) and a stable bottom (n = 0.025). The
channel bottom width is 10 ft and the side slopes are 4:1. Find the equivalent or composite
value of n using the four empirical methods described, if the flow depth is given as 3 ft.
Repeat for three remaining empirical methods.

1 3
2 yn = 3 ft
1
4

b = 10 ft

45

Example 5. Using Manning’s Equation for Compound Channels

The banks of a rigid boundary channel have flooded, causes water to flow over one lined and
one natural floodplain on either side, as shown in the figure below. Estimate the flowrate
through the channel if the elevation drop is 1 ft over a channel reach that is 500 ft in length.
6& )/5
• Recall that 𝑄 = 𝐴 𝑅7 𝑆 4/)
3

• So = 1 ft/500 ft = 0.002
• Note: n is not symmetrical in this channel

A p Rh Q
Subsection No. n
(ft2) (ft) (ft) (ft3/s)

1 1.8 0.6 + 3 = 3.6 ft 0.5 0.020

0.8 + 2 + 0.8 =
2 2.8 0.778 0.015
3.6

3 1.8 0.6 + 3 = 3.6 ft 0.5 0.03

∑ 16.8

46

23
 2/10/23

Example 5. Using Manning’s Equation for Compound Channels

Estimate the flowrate through an asymmetric channel, as shown in the figure below. The bed
slope of the channel is given as 0.003. Let n1 = 0.030 and n2 = 0.050.
6&
6m 8m
)/5
• Recall that 𝑄 = 𝐴 𝑅7 𝑆 4/)
3

2m 2
1

3m

Subsectio Ac p Rh Q
n
n No. (ft2) (ft) (ft) (ft3/s)
1

47

Return to Example 1. Sewer Design

• A typical problem encountered in practice

may take the below form:
“Design a concrete sewer that has a maximum
design discharge of 1.0 m3/s and a minimum
discharge of 0.2 m3/s if its slope is 0.0018.”
Q: With our new understanding of uniform
flow, how is this best determined?
A: Manning’s equation can be used to
determined yn at a uniform flow condition,
which provides the minimum required
conduit dimensions. Design curves can
also be used to check calculations.

48

24
 2/10/23

Example 1. Sewer Design

It is good practice to design the pipe for a filling ratio of 0.8 or less at maximum design flow. At y/d = 0.8, from
the figure below, nQ/knS1/2d8/3 = 0.305. The initial diameter is calculated assuming full rough turbulent flow
using:

If Manning’s equation is not applicable, use Chezy’s equation with Colebrook-White expression for the friction
factor.
The initial diameter is rounded up to the next commercial pipe size and the actual flow depth is calculated for the
commercial diameter

If yo/d is any other

value, use
appropriate value
of AR2/3/d8/3 to
evaluate diameter
of the culvert

49

Sewer Design: Stage 2

• In addition to determining the channel dimensions, we may be asked to check for additional
design considerations such as a phenomenon known as self-cleansing

50

25
 2/10/23

Design Stage 2 - Alternate

It has been suggested by a researcher (Hager,

1999) that a critical shear stress tc of 2.0 Pa (0.042
lbs/ft2) is required for self-cleansing.
From an earlier slide:
𝜏𝑜 = 𝛾𝑅𝑆$
The corresponding critical velocity and it’s
variation with filling ratio can be obtained by
setting to = tc so that slope So = tc/gR in Manning’s
equation. The critical velocity in dimensionless
form is:
𝑉P 𝑛 𝑔 𝑅 "/S
𝑉P∗ = =
𝐾R 𝑢∗P 𝑑"/S 𝑑
Here 𝑢∗P = critical value of shear velocity = 𝜏P /𝜌.
The dimensionless cleansing velocity is a unique
function of the filling ratio, y/d, as shown in the
figure.
51

Back to our design problem

To obtain a first estimate of diameter for the maximum design discharge of 1.0 m3/s occurring at y/d
= 0.8:

Round up the result to a commercial diameter of 1.067 m (42 in. – this is from US). You can find the
eManual from Lafarge Concrete and others
Next calculate the normal depth for the selected commercial size from Manning's equation:

from which θ = 4.341 rad and y0 = (1.067/2)[1 – cos(4.341/2)] = 0.835 m.

The corresponding cross-sectional area, A = [4.341 – sin(4.341)]× 1.0672/8 = 0.7504 m2
V = Q/A = 1.0/0.7504 = 1.33 m/s.
From figure, Vc* = 0.82 for y0/d = 0.835/1.067 =0.78.
Using the definition of Vc* from equation provided earlier, and an assumed critical shear stress of 2.0 Pa, we have

52

26
 2/10/23

so that the design velocity at maximum discharge is greater than the critical velocity.

Repeating these calculations for the minimum discharge of 0.2 m3/s, we obtain
θ = 2.301 rad, which gives y0 = 0.316 m and A = 0.2214 m2.
The velocity V = Q/A = 0.2/0.2214 = 0.90 m/s.

For the minimum discharge, y0/d = 0.316/1.067 = 0.30 and Vc* = 0.74 from the figure below
Then the critical velocity, Vc = 0.71 m/s, and once again, V > Vc.

53

f = 8gA2RS/Q2

As in pipe flow problems, let us assume f = 0.015

Calculate friction factor with 4R as length scale in Re (smooth pipe)

Recalculate f using this Re

54

27
 2/10/23

above

Equation on right

55

Channel Cross Section Shapes

• So far, we have examined various cross-sectional shapes for open-channel flows

• When designing a channel, if we are asked to select a channel shape (sewers and similar
conduits notwithstanding), how is this decision made?

• As practicing engineers, there are economical constraints to consider

• Open channel systems are typically designed to transport a liquid to a location at a lower
elevation at a specified flowrate under the influence of gravity at the lowest possible cost
• Since no energy input is required, the cost of an open-channel system consists mainly of the
construction costs, which are proportional to the size of the system
• This translates to the necessity of minimizing the cross-sectional area of the channel

• Although the governing design concerns such as channel stability or Froude number
maximization will take precedence, once these limits have been established, we are then free
to assume that our main area of concern is cost optimization

56

28
 2/10/23

Channel Cross Section Shapes

• So, let us safely assume that our main area of concern is A

• Returning to Manning’s equation
𝑘3 )/5
𝑄= 𝐴 𝑅7 𝑆 4/)
𝑛
• If Ac is to be minimized for a fixed Q, then flow velocity V is maximized since flow is assumed
to be uniform across the channel cross-section
𝑘3 )/5 4/)
𝑉= 𝑅 𝑆
𝑛 7
• If So and n have been pre-determined, Rh needs to be maximized in order for V to be
maximized
• Now our problem is reframed with the singular focus of Rh
• Since Rh = A/p, we essentially want p to be minimized for a fixed value of A
• This would give us the “best” cross-section for channel design

57

Best Hydraulic Cross-Section

• The shape that we know has the minimal perimeter per unit area is a circle
• Therefore, the best cross-section for an open channel will be a b
semicircular one

• However, sometimes other cross-sectional shapes are actually sometimes y

less expensive to construct, or are specified due to topographical concerns
• How then do we optimize design if we can’t use a semi-circular conduit?
• Consider both rectangular and trapezoidal channels

y
θ
y
b

58

29
 2/10/23

Best Hydraulic Cross-Section: Rectangular Channel

• For a rectangular cross-section

𝑃 = 𝑏 + 2𝑦
𝐴
𝑏 = y
𝑦
𝐴
𝑃 = + 2𝑦
𝑦
PQ P E (E b
• Now = + 2𝑦 = +2
P8 P8 8 8'

PQ
• Setting = 0 gives 𝐴 = 2𝑦 ) = 2𝑦 𝑦
P8

• Since 𝐴 = 𝑏𝑦, 𝑏 = 2𝑦 ideally b/y is sometimes called the “Aspect ratio” of the channel
• Essentially, the best rectangular cross-section is one in which a semi-circle can be inscribed

59

Best Hydraulic Cross-Section: Trapezoidal Channel

• For a trapezoidal cross section

2𝑦
𝑃=𝑏+
𝑠𝑖𝑛𝜃
𝑦
𝐴= 𝑏+ 𝑦 y
𝑡𝑎𝑛𝜃 θ
E 8
• From which 𝑏 = − b
8 9:3;

E 8 )8
• So then 𝑃 = − +
8 9:3; "A3;

PQ (E 4 )
• Now, = − +
P8 8' 9:3; "A3;

PQ R"A3;
• Setting = 0 and substituting the above expression for A gives 𝑦 =
P8 4(4(S%";)

• It can be observed that the best hydraulic cross-section for a trapezoid is one with θ = 60°
• The shape is that of a half-hexagon, inside which a semi-circle can be inscribed

60

30
 2/10/23

Example 6. Best Cross-Section of an Open Channel

Water is to be transported at a rate of 2 m3/s in uniform flow in an open channel whose surfaces
are asphalt lined. The bottom slope is 0.001. Determined the dimensions of the best cross
section if the shape of the channel is (a) rectangular and (b) trapezoidal.
6& )/5
• Recall 𝑄 = 𝐴 𝑅7 𝑆 4/)
3

• Recall that 𝑃 = 𝑏 + 2𝑦
y

61

Example 6. Best Cross-Section of an Open Channel

Water is to be transported at a rate of 2 m3/s in uniform flow in an open channel whose surfaces
are asphalt lined. The bottom slope is 0.001. Determined the dimensions of the best cross
section if the shape of the channel is (a) rectangular and (b) trapezoidal.
6& )/5
• Recall 𝑄 = 𝐴 𝑅7 𝑆 4/)
3

• Recall that 𝑃 = 𝑏 + 2𝑦
• n = 0.026 for an asphalt-lined channel
• We know that the best hydraulic cross-section for a rectangular channel occurs when 𝑦 =
𝑏/2
• 𝐴 = 𝑏𝑦 = 𝑏2/2 and 𝑅ℎ = 𝐴/𝑝 = 𝑏/4
4 R& R )/5
• 2 m5 /s = ( ) (0.001)4/) y
3 ) I

• From the above expression, b = 1.84 m and y= 0.92 m

62

31
 2/10/23

Example 6. Best Cross-Section of an Open Channel

Water is to be transported at a rate of 2 m3/s in uniform flow in an open channel whose surfaces
are asphalt lined. The bottom slope is 0.001. Determined the dimensions of the best cross
section if the shape of the channel is (a) rectangular and (b) trapezoidal.
6& )/5
• Recall 𝑄 = 𝐴 𝑅7 𝑆 4/)
3
)8 8 y
• Recall that 𝑃 = 𝑏 + and 𝐴 = 𝑏 + 𝑦 θ = 60°
"A3; 9:3;

R"A3; b
• For the best hydraulic cross-section, 𝑦 =
4(4(S%";)

63

Concrete-lined channels

Note, y in the figure is the distance from the bed and

H is the depth of flow, AR = channel width/depth
64

32
 2/10/23

Example with practical limitations

tables

65

SL = 0.0022

So/SL = 0.23

from earlier analysis.

Froude number will be less that 0.48 for all
discharges below 150 m3/s
66

33