Chapter 4

UNIFORM FLOW

4.1 INTRODUCTION
When uniform flow occurs in a channel, (i) the discharge, the velocity and the depth remain
constant along the length of the channel, (ii) the energy line (Sf), the water surface (Sw) and the channel
bottom (S0) are parallel, i.e. Sf = Sw = S0, and (iii) water is neither accelerated nor retarded and the net
external force on water is zero.
The flow in a long straight prismatic channel under normal condition, i.e. when there is no inflow
or out flow or no transition or control like sluice gates, weirs, dams etc., tends to be uniform.
Uniform flow is considered to be steady only, since unsteady uniform flow is not practically
possible. True uniform flow does not normally occur in natural channels, because changes in the cross-
section along the length of the channel induce non-uniform flow conditions still, the concept hydraulics. In
fact, the resistance relations developed for uniform flow are used to predict channel resistance in gradually
varied flows, both steady and unsteady.

4.2 ESTABLISHMENT OF UNIFORM FLOW

The condition for the establishment of uniform flow in an open channel can be determined
considering the momentum equation, Eq. (2.16). For uniform flow, U1 = U2, Fp1 = Fp1 and 1 = 2 Hence,
Eq. (2.16) reduces to

W sin  Ff (4.1)
which indicates that when uniform flow occurs in a channel, the active component of the gravity force
causing the flow is equal to the total force of friction or resistance.
The above condition implies that (i) flow cannot be uniform in a horizontal channel for which =0
and for uniform flow to occur, the channel must have a slope in the downstream direction, (ii) flow connot
be uniform in an adverse slope channel in which both W sin  and Ff act in the same direction, which is
opposite to the direction of flow, (iii) flow cannot be uniform in frictionless channel for which Ff = 0, and
uniform flow of an ideal fluid is impossible, since an ideal fluid has no friction.
The condition for the establishment of uniform flow W sin  Ff , can be used to explain why
the flow in a long straight prismatic channel under normal condition tends to be uniform. Suppose that the
flow condition at some location of a channel is such that W sin  Ff and flow is non-uniform. As the
flow proceeds downstream, the flow is accelerated and the flow velocity increases. Since Ff U2, the
friction or resistance force also increases and a balance between W sin  and Ff tends to reach and the
flow tends to be uniform. On the other hand, if W sin  < Ff at some location of a channel, the flow is
retarded and the flow velocity decreases. Hence, the friction or resistance force also decreases and a
balance between W sin  and Ff tends to reach and the flow tends to be uniform as the flow proceeds
downstream. Thus, uniform flow seems to be self-adjusting and any departure form the condition
W sin  Ff tends to reestablish this condition.

1
4.3 VELOCITY DISTRIBUTION

Bed Shear Stress and Friction Velocity

When water flows in a channel, the pull of water produces a force that acts on the channel bed
in the direction of flow. This force is known as the shear or tractive or drags force and is equal to the
friction or resistance force Ff. If the average value of this force per unit wetted area, which is known
as the shear stress, is denoted by  0 , then
F f   0 PL (4.2)
Where P is the wetted perimeter and L is the length of the channel. Therefore, from Eq. (4.1)

W sin   Ff   0 PL (4.3)
When the angle of bottom slope θ is small, sinθ≈tanθ. Also, tanθ=S0 and when the flow is uniform,
S0=Sf. Therefore, the active component of the gravity force=Wsinθ=γALsinθ≈ γALtanθ =γALS0,
where γ is the specific weight of wat0er, A is the cross-sectional area and S0 is the channel bottom
slope. Therefore,

 ALS0   0 PL (4.4)
or
A
0   S 0  RS 0  gRS0 (4.5)
P

For a wide channel, R  h, hence Eq. (4.5) becomes

 0  hS0  ghS0 (4.6)

The quantity  0 /  has the dimensions of velocity and the bed shear stress  0 is expressed as
 0   u *2 (4.7)
or
u*   0 /  (4.8)
where u* is known as the shear or friction or drag velocity, since it varies with the boundary friction
 0. It does not represent a velocity which is physically real. However, it is used as the velocity scale
in the study of velocity distribution in open channels.
Using Eqs. (4.5) and (4.8) it can be shown that

u*  gRS0 (4.9)
and when the channel is wide
u*  gRS0 (4.10)

2
Boundary Layer
The velocity distribution across a channel section is not uniform
owing to the presence of boundary or wall roughness. The boundary layer is
the flow layer above the boundary in which the flow velocities are retarded
by the effect of boundary roughness. Generally, the thickness of the
boundary layer is defined as the distance from the boundary surface to the
point where u = 0.99u0, when u0 is the velocity in the outer layer (Fig. 4.1a).

Fig. 4.1a Velocity distribution over a

smooth channel surface

Laminar or Viscous Sublayer

Even in a turbulent boundary layer, there is a very thin layer near the boundary in which the
flow is laminar and is known as the laminar or viscous sublayer. The thickness of this layer is given
by

11.6v
v  (4.11)
u*

Smooth and Rough Boundaries

As stated earlier, velocity distribution across a channel section is not uniform owing to the
presence of boundary (wall) roughness. The effect of boundary roughness on the velocity distribution
in turbulent flow was first investigated for pipe flow by Nikuradse who introduced the concept of
equivalent sand grain roughness (ks) as standard for all other types of roughness elements. The ratio
ks/R of the roughness height to the hydraulic radius is known as the relative roughness. The roughness
element mainly influence the velocity distribution close to surface generating eddies which affect the
turbulence structure close to the boundary. Schlichting (1968), in connection with flat surfaces and
pipes, experimentally determined the following criteria, based on the ratio of Nikuradse roughness (ks)
and a length scale of the viscous sublayer (v/u*), for classifying boundary surfaces:
1. Hydraulically smooth boundary

u * ks
5 (4.12)
v

The roughness elements are well-covered by the viscous sublayer (ks<δv) and do not affect the
velocity distribution outside the sublayer. The velocity distribution depends on the viscosity of the
fluid.

3
 2. Hydraulically rough boundary

u * ks
 70 (4.13)
v
The roughness elements project through the viscous sublayer (ks>δv) and the velocity distribution
outside the sublayer is affected by the surface roughness. The viscosity of the fluid has no effect on
the velocity distribution.
3. Transition boundary

u * ks
5  70 (4.14)
v
The velocity distribution is affected both by viscosity of the fluid and the bottom roughness.

Velocity Distribution in Turbulent Boundary Layer

Along a Vertical
The velocity distribution along a vertical in a wide channel in turbulent flow is given by

uz 1 z z z
 ln  2.50 ln  5.75log (4.15)
u *  z0 z0 z0

where z0 is the zero velocity level, i.e. u = 0 at z = z0 and depends on whether the boundary is
hydraulically smooth or rough. Equation (4.15) indicates that the velocity distribution in turbulent
flow is logarithmic and is commonly known as the Prandtl von Karman universal velocity
distribution law.
Experimental evidence suggests that the logarithmic velocity profile is a good approximation
for the full depth of the flow. The following values of z0 are obtained.
1. Hydraulically smooth surface (u*ks/v5)

1 v v
z 0  0.11 (4.16)
9 u* u*
2. Hydraulically rough surface (u*ks/v70)

1
z0  ks  0.033k s (4.17)
30
3. Transition regime (5<u*ks/v70)

v
z0  0.11  0.033k s (4.18)
u*

Depth-Averaged Velocity
As stated earlier in Art. 1.6, the generally accepted procedure to determine the depth averaged
velocity in a stream using the area-velocity method is to average the velocity measurements at 0.2h

4
and 0.8h depths from the free surface, or alternatively, to measure velocity at 0.6h depth from the free
surface, when h is the total depth of flow. For logarithmic velocity distribution, Eq. (4.15), Vanoni
(1941) showed that the flow velocity, measured at 0.632h from the free surface, is equal to the depth-
averaged velocity in the vertical.

Cross-Sectional Mean Velocity

The mean velocity for turbulent flow in open channels and derived the following equations:
1. Hydraulically smooth surface (u*ks/v5)
U u*R  3.64u * R 
 5.75 log  3.25  5.75 log  (4.19)
u* v  v 
2. Hydraulically rough surface (u*ks/v70)
U R  12.2 R 
 5.75 log  6.25  5.75 log  (4.20)
u* ks  ks 
3. Transition regime (5<u*ks/v<70)
U  12.2 R 
 5.75 log  (4.21)
u*  s
k  3.35v / u * 

Example 4.1
A rectangular channel is 6 m wide and laid on a slope of 0.25%. The channel is made of
concrete (ks = 0.002m) and carries water at a depth of 0.50m. Compute the mean velocity of flow.

Solution
S0 = 0.25/100=0.0025
R = A/P = (6  0.50)/(6 + 2  0.50) = 0.4286 m
u* = gRS0  9.81 0.4286 0.0025  0.1025m / s
k su * 0.002  0.1025
  205  70
v 10 6
Hence, the boundary is hydraulically rough and the mean velocity of flow is obtained by Eq. (4.55),
i.e.
U 12.2 R 12.2  0.4286
 5.75 log  5.75 log  19.65
u* ks 0.002
U  19.65  0.1025  2.014m / s

4.4 UNIFORM FLOW FORMULAS

Chezy Formula
The Chezy formula can be found mathematically from two assumptions. The first assumption
states that, in steady uniform flow the active component of the gravity force causing the flow is equal
to the total force of friction or resistance, as indicated by Eq. (4.1). When the angle of bottom slope 

5
is small, sin  tan . Also, tan  = S0 and when the flow is uniform, S0 = Sf. Therefore, the active
component of the gravity force = W sin  AL sin  AL tan   ALS 0  ALS f , where  is the
unit weight of water, A is the cross-sectional area, L is the length of the channel and Sf is the slope of
the energy line.
The second assumption states that, in turbulent flow the resistance force per unit wetted are
varies as the square of the mean velocity. The total wetted area is the product of the wetted perimeter
P and the length of the channel L. Hence, the total force of resistance is given by
Ff  kPLU 2 (4.22)
where k is a constant of proportionality. Hence, Eq.(4.2) gives
ALS f  kPLU 2 (4.23)

 A 
U2  . S f  .R.S f
k  k

U 
1 1
.R 2 .S f 2
k
 U  C.R 2 .S f 2
1 1

 Q  CA.R 2 .S f 2
1 1

Equation (4.66) can be rearranged to yield

U  C.R 2 .S f 2
1 1
(4.24)

where  / k is written as one constant C.

Equation (4.24) is probably the first steady uniform flow formula developed by the French
engineer Antoine Chezy in 1769. The resistance factor C is referred to as the Chezy’s C.
The Chezy formula can be used in any systems of units. The dimensions of Chezy’s C in Sl
units are m /s and C / g is dimensionless. The numerical value of Chezy’s C varies with the
1/2

systems of units. For the rivers of Bangladesh, the numerical value of C varies from 30 m1/2/s to 80
m1/2/s and the mean value of C may be taken as 50 m1/2/s.
The Chezy formula is applicable for steady uniform and nearly uniform flows. It is widely
used in Europe.

Darcy-Weisbach formula
The Darcy-Weisbach formula, firs presented by julius Weisbach in 1845 and primarily
developed for pipe flow, is given by
L U2
hf  f (for pipe flow), for circular section (4.25)
d0 2g
where hf is the frictional loss, f is the friction factor, L is the length of the pipe, d 0 is the diameter of
the pipe, U is the mean velocity of flow and g is the acceleration due to gravity. Since d 0 = 4R and the
energy gradient Sf = hf/L, the above formula may be written as

6
 8 g 12 12
U R .S f (4.26)
f
This formula is same in all the systems of units and may be applied to uniform and nearly uniform
flows in open channels. The friction factor f is dimensionless and its numerical value remains same in
all the systems of units.
Manning Formula
In 1889 the Irish engineer Robert Manning presented a formula for steady uniform flow in
open channels. This formula is completely empirical in nature. In SI and English units this formula is
given by
1 2 / 3 1/ 2
U  R Sf (4.27)
n
and
1.486 2 / 3 1 / 2
U R Sf (4.28)
n
respectively, where U is the mean velocity, n (s/m1/3 or sec/ft1/3) is the Manning’s roughness
coefficient, specifically known as Manning’s n, R is the hydraulic radius and Sf is the slope of the
energy line.
The numerical value of n is the same in all the systems of units, but the coefficient in the
Manning formula is different in different systems of units. For example, when the Manning formula in
SI units, Eq. (4.27), is converted to English units. Eq. (4.28), it has to be multiplied by the factor
(3.28)1/3 = 1.486, as shown below.

n in s / m1 / 3 
n
1/ 3
in sec/ ft1 / 3 1m  3.28 ft 
3.28
1 2 / 3 1/ 2
U  R S f in SI units
n
1

1/ 2
1/ 3
R 2 / 3 S f in English units
n / 3.28
3.281 / 3 2 / 3 1 / 2
 R S f in English units
n
1.486 2 / 3 1 / 2
 R S f in English units
n
The Manning formula has been verified by many laboratory and field measurements and
found to give satisfactory results. Therefore, it has been the most widely used of all the uniform flow
formulas for open channel flow computation. However, in applying the Manning formula the main
difficulty lies in the determination of the roughness coefficient n, for there is no exact method of
selecting the value of n. Even at the present stage of knowledge, in addition to the information and the
methods available for selecting n, the veteran engineers have to exercise sound engineering judgement
and experience in selecting the proper value of n.
Typical values of Manning’s roughness coefficient n for some open channel surfaces are
presented in Table 4.1.

7
 Table 4.1 Typical values of Manning’s roughness coefficient n for some open channel surfaces
Surface Value of n Surface Value of n
Glass 0.010 Plastic 0.010
Cement 0.011 Concrete 0.013
Wood 0.015 Earth canals 0.025
Rivers 0.025 Flood plains 0.040

Chow (1959) presented an extensive table of n-values for various types of channels. A set of
photographs for channels with a wide range of resistance coefficients are presented by Chow (1959),
French (1986) and Chaudhry (1993).
The Manning formula is valid for fully rough turbulent flow and the point at which fully
rough flow begins can be defined in terms of Manning’s n in SI units as
n6 RS f  2.5331013 (4.72)
Tests on concrete pipes up to 6 m in diameter as well as a reexamination of Nikuradse date
have shown that Eq. (4.70) is valid for values of R/ks up to about 1500. As such, one can use Manning
formula with confidence for predicting the resistance of rough open channels.

Relationship Between Chezy’s C, Darcy-Weichbach friction Factor f and Manning’s n

Using Eqs. (4.24), (4.26) and (4.27), we obtain the following relationships between C and n,
C and f and n and f in SI units.
1 16
i) C  R (4.29)
n
C 8
ii)  (4.30)
g f

1 f
iii) n  R 6 (4.31)
8g
Also, combination of Eqs.(4.9), (4.26) and (4.30) yiclds with S0 = Sf
U C 8
  (4.32)
u* g f

which indicates that C / g is dimensionless.

4.5 ESTIMATION OF RESISTANCE COEFFICIENTS

The most common parameters which describe the resistance to flow in open channels are (i)
the Chezy’s resistance factor C, (ii) the Darcy-Weisbach friction factor f, and (iii) the Manning’s
roughness coefficient n

Estimation of Chezy’s C (Based on White-Colebrook Formula)

For hydraulically smooth flow (u*ks/v  5)

8
  3.64u * R   11.4 Re 
C  18 log   18 log  (4.33)
 v   C 
in which Re (= UR/v) is the Reynolds number. Application of Eq. (4.78) requires an iterative
calculation procedure.
For hydraulically rough flow (u*ks/v  70)
 12.2 R 
C  18 log  (4.34)
 ks 
And for Transition boundary (5<u*ks/v<70), C is obtained from
U C  12.2 R 
  5.75 log  (4.35)
u* g  s
k  3.35v / u * 

Estimation of Manning’s n
Stickler formula: The most simple and the best-known of the methods used for estimating
Manning’s n is the empirical formula presented by the Swiss engineer Strickler in 1923. This formula
is based on data from (i) streams with beds consisting of coarse material and free from bed
undulations, and (ii) fixed bed channels with grains pasted to the bottom and sides. The formula
originally proposed by strickler is
1/ 6
d50
n  0.047d 20
1/ 6
(4.36)
21.1
where d50 is the median diameter or the diameter of the bed material in meters such that 50 percent of
the material by weight is smaller.
The Strickler formula has two major advantage: (i) it relates n to the size of the grains which
can be measured easily, and (ii) since d50 is raised to 1/6th power, an error in estimating its value has a
corresponding less effect on the computed value of n.

Factors Affecting Manning’s n

The value of n is highly variable and depends on a number of factors, which are to some
extent interdependent. The factors that exert the greatest influence upon Manning’s n in both natural
and artificial channels are briefly described below.
i) Roughness of the surface: The value of Manning’s n depends on the roughness of the
surface which in turn depends on the size and shape of the grains of the material forming the channel
perimeter. In general, fine-grained soils (e.g. clay, silt and sand) result in a low value of n and coarse-
grained soils (e.g. gravels and boulders) result in a high value of n.
ii) Vegetation: The presence of vegetation in a channel retards the flow and increases n
depending on the height, density, distribution and type of vegetation. Owing to the seasonal growth of
aquatic plants, the value of n may increase in the growing season and diminish in the dormant
seaseon.
iii) Channel irregularity: Channel irregularities include sand bars, depressions, holes,
humps, etc. and increase the value of n.

9
 iv) Channel alignment: The value of n is low for straight channels and high for curved
channels and increases with the curvature of the channel.
v) Silting and scouring: In general, silting converts an irregular channels into a regular one
and decreases n, whereas scouring does the reverse and increases n.
vi) Obstruction : The presence of obstructions like logs, bridge piers etc. tends to increase n
depending on the size, shape, number and distribution of the obstructions.
vii) Stage and discharge: The value of n generally decreases with increase in stage and
discharge. However, the value of n may be high when the flood plains in a river are submerged at high
stages.
viii) Suspended material and bed load: The suspended material and bed load cause an
increase in Manning’s n because additional energy is required to move the sediment.

Example 4.2
An open channel lined with concrete (d50 = 1.5mm) is laid on a slope of 0.1 %. The channel is
trapezoidal with b = 6 m and s = 2. Compute the uniform flow discharge in the channel if the depth of
flow is 2 m. Also compute the numerical values of Chezy’s C and friction factor f and check if the
Manning’s formula is applicable.

Solution Sf = S0 = 0.001, d50 = 1.5 mm = 1.5/1000 m

n = 0.047d501/6 = 0.047 x (1.5/1000)1/6 = 0.016
A = (b + sh)h = (6 + 2 x 2) x 2 = 20 m2

P  b  2 1  s 2 h  6  2 1  22  2  14.94m
R = A/P = 1.34 m
Q = (1/n) AR2/3S01/2 = (1/0.016) x 20 x 1.34 2/3 x 0.0011/2
= 48.29 m3/s
C = (1/n) R1/6 = (1/0.0159) x 1.3381/6 = 66.01 m1/2/s
f = 8g/C2 = 8 x 9.81/66.012 = 0.018
Now,
n6 RS f  0.0166  1.34  0.001  5.9101013  2.5331013

Hence, the Manning’s formula is applicable.

10
4.6 NORMAL DEPTH, SETION FACTOR, CONVEYANCE AND
HYDRAULIC EXPONENT

Normal Depth
The depth of uniform flow is known as normal depth and is designated by h n. Similarly, the
discharge of uniform flow is known as the normal discharge designated by Qn, the velocity of uniform
flow is known as the normal velocity designated by U n and so on. The Chezy and the Manning
formulas for the discharge can then be written as
1 1
Qn  CAn Rn S 2 2
n (4.37)
and,
1
1 2
Qn  An Rn 3 Sn2 (4.38)
n
respectively.

Section Factor
The product of flow area and two-thirds power of the hydraulic, i.e. AR2/3, is known as the
section factor in connection with the Manning formula. It can be computed if the flow the channel
section and depth h are given.
When the flow is uniform, the product of flow area and two-thirds power of the hydraulic
radius. i.e. AnRn2/3, is known as the section factor for uniform flow computation in connection with
the Manning formula. From Eq. (4.38), we obtain
2 nQn
An Rn 3  (4.39)
Sn
Thus, AnRn2/3 can be computed either if the channel section and the normal depth h n are specified, or
alternatively, using Eq. (4.39) from given values of the Manning’s n, the normal discharge Qn and the
normal slope Sn.
When the Chezy formula is used, then AR1/2 is the section factor and AnRn1/2 is the section
factor for uniform flow computation. Using Eq. (4.37), it can be shown that
2 Q
An Rn 3  (4.40)
C Sn
The section factor is and important element in the computation of uniform flow.

Conveyance
The conveyance for a channel section in terms of the Manning formula is given by
1
K AR1 / 3 (4.41)
n
which can be computed when the section, the roughness coefficient n and the depth h are given.
When the flow is uniform, then using the Manning formula (Eq. 4.86), it can be shown that
Qn  K n S n (4.42)
where
1
Kn  A2 Rn2 / 3 (4.43)
n
is the conveyance for uniform flow. It can be computed using Eq. (4.42) when Qn and Sn are specified
or using Eq. (4.43) from the given section, n and the normal depth hn.

11
 In terms of the Chezy formula, the conveyance is given by
1
K  CAR 2 (4.44)
and using Eq. (4.37), we obtain
Qm  K n S n (4.45)
when
K n  CAn Rn1 / 2 (4.46)
is the conveyance for uniform flow.
For unit longitudinal slope of the channel, i.e. When Sn = 1, Eq.(4.90) shows that Qn = Kn, i.e.
the conveyance is numerically equal to the discharge. Also, for a given longitudinal slope, Qn  Kn.
Thus, the conveyance is a measure of the carrying capacity of the channel.

Hydraulic, Exponent for Uniform Flow Computation

The conveyance K is a function of the depth of flow for a given channel section and
roughness and it is convenient to express K in the form
K2 = C2hN (4.47)
where C2 is a coefficient and N is an exponent which is known hydraulic exponent for uniform flow
computation.
Assume that n and N are independent of h. Taking logarithm of both sides of Eq. (4.47) and
then differentiating with respect to h, one obtains
d (ln K ) N
 (4.48)
dh 2h
Also, using R = A/P in Eq. (4.41), taking logarithm of both sides of Eq. (4.41), then
differentiating with respect to h and using dA/dh = B, we get
d (ln K ) 1  dP 
  5B  2 R  (4.49)
dh 3A  dh 
Equating the right sides of Eqs. (4.48) and (4.49) and solving for N, we obtain
2h  dP 
N  5B  2 R  (4.50)
3A  dh 
which is the general equation for the hydraulic exponent N when the conveyance is expressed in terms
of the Manning’s formula.
When the conveyance is expressed in terms of the Chezy formula, Eq. (4.44), it can be shown
in a similar way that
h dP 
N  3B  2 R  (4.51)
A dh 
Equations (4.50) and (4.51) indicate that the numerical value N depends on the channel shape
and the depth of flow. It also depends on whether the conveyance is expressed in terms of the
Manning or the Chezy formula. The values of N for different channel sections are given in Table 4.2.

12
 Table 4.2 Values of N for different channel sections
Conveyance computed by Conveyance computed by
Channel section
Manning equation Chezy equation
1. Wide (b >> h) 3.33 3.00
2. Triangular 5.33 5.00
2 4( h / b )  2(h / b)
5 3
3  1  2(h / b) 

3. Rectangular
1  2(h / b)

2  1  2 s ( h / b)  1  2 s ( h / b)
5  3
3  1  s ( h / b)  1  s ( h / b)
4. Trapezoidal
1  s 2 ( h / b) 2 1  s 2 ( h / b)
4 
1  2 1  s 2 ( h / b) 1  2 1  s 2 ( h / b)

16h  5 sin  / 2 1  4h  6 sin  / 2 1 

5. Circular      
3d0    sin   sin  / 2  d0    sin   sin  / 2 
6. Parabolic (perimeter 4 3
3 3
equation z = cy2) 3  2ch 3  2ch

Example 4.3
Derive the expression for the hydraulic exponent for uniform flow computation N for a
rectangular channel based on the Manning formula. Then, compute the numerical values of N for (i) a
wide channel, and (ii) a narrow channel.

Solution: For a rectangular channel, A = bh, B = b, P = b + 2h, R = A/P, dP/dh = 2

2h  dP  2h  bh 
N  5B  2 R   5b  2   2
3A  dh  3bh  b  2h 
2 4h  2  4(h / b) 
 5    5 
3 b  2h  3  1  2(h / b) 
This is the expression for N for a rectangular channel based on the Manning formula.

(i) For a wide channel, h/b  0. Hence,

2 4(h / b)  2  4  0  10
N 5   5    3.33
3  1  2(h / b)  3  1  2  0  3
(ii) For a narrow channel, h/b  . Hence,
2 h( h / b)  2  4  2 4 
N   5   5    5  2
3 1  2( h / b)  3  1 / ( h / b)  2  3  0  2

Example 4.4
Compute the hydraulic exponent for uniform flow computation N of a trapezoidal channel
with b = 6.1 m, s = 2 and h = 2 m based on the Manning formula.

13
Solution
h 2
  0.328
b 6.1
2 1  2 s ( h / b) 2 1  s 2 ( h / b)
N   5   4
3 1  s ( h / b) 3 1  2 1  s 2 ( h / b)

2 1  2  2  0.328 2 1  2 2  0.328
  5   4
3 1  2  0.328 3 1  2 1  2 2  0.328
 4.653 0.793  3.860
Alternative solution: A = 20.2 m2, P = 15.044m, B = 14.1m, dP/dh = 2 1  s 3  2 5

N 
2h 
 5B  2 R
3A 
dP 

2 2
dh  3  20.2
 
5  14.1  2  1.342 2 5  3.860

Computation of Normal Depth

Analytical Method
The normal depth is an important parameter in the analysis of open channel flow. It may be
computed using the Manning or the Chezy formula when the channel section, the discharge Q, the
bottom slope S0 and the Manning’s n or the Chezy’s C are given. For wide and triangular channels,
the following analytical (explicit) expressions for the normal depth can be easily obtained.

a) Using the Manning formula

i. Wide channel
3/ 5
 nq 
hn    (4.52)
 S 
 0
ii. Triangular channel
3/8
21/ 4 (4  s 2 )1/ 8  nQ 
hn  (4.53)
s5/8  S 
 0
b) Using the Chezy formula
i. Wide channel
2/3
 q 
hn    (4.54)
C S 
 0 
ii. Triangular channel

 
2/5
21/ 5 1  s 2
1 / 10
 Q 
hn    (4.55)
s3/ 5 C S 
 0 
Example 4.5

14
 A wide channel with S0 = 0.0025 carries a discharge of 3 m2/s. Compute the normal depth
and velocity (a) using the Manning formula when n = 0.020, and (b) using the Chezy formula when C
= 45 m 1/2/s.

Solution
(a) Using the Manning formula
3/ 5
 mq   0.020 3 
3/ 5

hn       1.12m
 S   0.0025 
 0
q 3
Un    2.69m / s
hn 1.12
(b) Using the Chezy formula
2/3
 q   3 
2/3

hn       1.21m
C S   45  0.0025 
 0 
q 3
Un    2.48m / s
hn 1.21

Example 4.6
For a triangular channel with side slopes of 2 : 1, a longitudinal slope of 0.0016 and n =
0.015, determine the normal depth if Q = 10 m3/s.

Solution s = 2, S0 = 0.0016, n = 0.015, Q = 10 m3/s

hn 
21/ 4 (1  s 2 )1/ 8  nQ  21/ 4 1  22

 
1/ 8
 0.01510 
 
3/ 8

 1.55m
s5/8  S  25 / 8  0.0016 
 0
A  sh 2  2  1.552  4.79m
Q 10
Un    2.09m / s
An 4.79
Alternative solution
nQ 0.01510
AR 2 / 3    3.75
S0 0.0016
2/3
 sh 
 sh 2

  3.75
 2 1 s 
2

2/3
 2hn 
or, 2h  2
n

  3.75
 2 1 2 
2

which gives hn8/3 = 3.206

 hn = 1.55 m
Then, A  sh  2 1.55  4.79m
2 2 2

15
 Q 10
U n    2.09m / s
An 4.79

Trial - and - Error Approach

For other simple geometric channel sections, like the rectangular, trapezoidal, circular and
parabolic sections, the computation of normal depth can be conveniently carried out by the trial-and-
error solution of Eq. (4.39).
Example 4.7
For a rectangular channel with b = 6.0 m, n = 0.025 and S0 = 0.0025, compute the normal
depth and velocity if Q = 20 m3/s.
Solution
nQ 0.025 20
AR 2 / 3    10.000
S0 0.0025
Now assume several values of h and compute the section factor AR2/3 unit the computed
value of AR2/3 is close to 10.000.
h(m) A(m2) P(m) R(m) AR2/3 Remarks
1.00 6.000 8.000 0.750 4.952 h too small
2.00 12.000 10.000 1.200 13.551 h too large
1.60 9.600 9.200 1.043 9.876
1.62 9.720 9.240 1.052 10.054
1.61 9.660 9.220 0.048 9.965 h closest
Hence, the normal depth, hn = 1.61 m and the normal velocity
Q 20
Un    2.07m / s
A 9.66
Example 4.8
For a trapezoidal channel with b = 6 m, s = 2, n = 0.025 and S0 = 0.001, compute the normal
depth and velocity if Q = 14 m3/s.
Solution
nQ 0.025  14
AR 2 / 3    11.068
S0 0.001
h(m) A(m2) P(m) R(m) AR2/3
1.00 8.000 10.472 0.764 6.684
2.00 20.000 14.944 1.338 24.288
1.30 11.180 11.814 0.946 10.776
1.31 11.292 11.858 03952 10.929
1.32 11.405 11.903 0.958 11.084

Hence, the normal depth, hn = 1.32 m and the normal velocity

Q 14
Un    1.23m / s
A 11.405

16
Example 4.9
Compute the normal depth and velocity in a parabolic channel with Q = 20 m3/s, n = 0.025
and S0 = 0.0025 when the profile of the channel is given by y2 = 4z.

Solution
nQ 0.025 20
AR 2 / 3    10.000
S0 0.0025
For a parabolic channel, the top width B = 2y and h = z. Hence, assume values of h, compute
y using the equation y2 = 4z taking z = h and then compute B, A, P, R and AR2/3 as shown below.
h (m) y (m) B (m) A (m2) P (m) R (m) AR2/3
1.00 2.000 4.000 2.667 4.591 0.581 1.856
2.00 2.828 5.657 7.542 7.192 1.049 7.785
3.00 3.464 6.928 13.856 9.562 1.449 17.743
2.26 3.007 6.013 9.060 7.822 1.158 9.993
2.27 3.013 6.027 9.120 7.846 1.162 10.008

Hence, the normal depth, hn = 2.26 m and the normal velocity

Q 20
Un    2.21m / s
A 9.060

Example 4.10
A circular channel 2 m in diameter is laid on a slope of 0.001 and carries a discharge of 4
3
m /s. Compute the normal depth and velocity when n = 0.013.

Solution
nQ 0.013 4
AR 2 / 3    1.644
S0 0.001
(rad) A (m2) P(m) R(m) AR2/3
1 0.079 1 0.079 0.015
2 0.545 2 0.273 0.229
3 1.429 3 0.476 0.610
4 2.378 4 0.595 1.682
3.94 2.328 3.94 0.591 1.639
3.95 2.337 3.95 0.592 1.647

Hence, n = 3.95 rad and the normal depth

d0   2 3.95 
hn  1  cos   1  cos   1.39m
2  2  2 2 
and, the normal velocity
Q 4
Un    1.71m / s
An 2.337
17
Numerical Methods
The numerical methods, used for solving nonlinear algebraic equations involving a single
variable, e.g. the method of bisection, thse method of iteration, the method of false position, the secant
method, the Newton-Raphson method etc., as stated in Art. 3.2, can be conveniently used to compute
the normal depth for rectangular trapezoidal, circular and parabolic channel sections. The computation
of normal depth using the bisection and the Newton-Raphson methods is considered here.

Bisection method: Suppose that we want to compute the normal depth in a channel for a given
section, discharge Q, roughness coefficient n and bottom slope S0. Then the function

f (h)  AR 2 / 3  An Rn2 / 3  AR 2 / 3  nQ / S0 (4.56)

must be satisfied by some positive depth greater than say hmin and less than say hmax. The normal
depth is taken equal to (hmin + hmax)/2 and f(h) is determined. If f(h) is positive then the root is less
than (hmin + hmax)/2 and the upper limit is taken as (hmin + hmax)/2. On the other hand, if f(h) is negative,
then the lower limit is taken as (hmin + hmax)/2. The procedure is repeated till the desired accuracy is
attained.

Example 4.11
For a trapezoidal channel with b = 6 m, s = 2, n = 0.025 and S0 = 0.001, compute the normal
depth by the method of bisection if Q = 14 m3/s.

Solution A  (6  2h)h, P  6  4.472h, R  A / P, An Rn2 / 3  11.068

f (h)  AR 2 / 3  An Rn
2/3

6  2hh5 / 3  11.068
6  4.472h2 / 3
Initially the values of hmin and hmax are taken as 0 and 10m, respectively. The computation is carried
out as follows.
hmin hmax h=(hmin + f(h) Root lies between
hmax)/2
0 10 5 148.647 0 and 5
0 5 2.5 26.562 0 and 2.5
0 2.5 1.25 -1.041 1.25 and 2.5
1.25 2.50 1.875 10.381 1.25 and 1.875
1.25 1.875 1.5625 4.105 1.25 and 1.5625
1.25 1.5625 1.4063 1.394 1.25 and 1.4063
1.25 1.4063 1.3281 0.143 1.25 and 1.3281
1.25 1.3281 1.2891 -0.458 1.2891 and 1.3281
1.2891 1.3281 1.3086 -9.357 1.3086 and 1.3281
1.3086 1.3281 1.3184 -0.009 1.3184 and 1.3281

Hence, the normal depth, hn = 1.32 m

18
Newton-Raphson method: Suppose we want to compute the normal depth in a channel for given
section, Q, n and S0. Obviously, when h = hn
nQ nQ 2 / 3
AR 2 / 3  0 or, A5 / 3  P 0 (since R = A/P)
S0 S0
If we now assume
nQ 2 / 3
f (h)  A5 / 3  P (4.57)
S0
5 dA nQ 2 1 / 3 dP 5 2 / 3 2nQ 1 / 3 dP  dA 
then f (h)  A2 / 3   P  A B P   B  (4.58)
3 dh S0 3 dh 3 3 S0 dh  dh 
For a given channel section, f (h) and f’(h) depend on the depth of flow only and hence can be easily
evaluated.

Example 4.12
For a trapezoidal channel with b = 6 m, s = 2, n = 0.025 and S0 = 0.001, compute the normal
depth by the Newton-Raphson method if Q = 14 m3/s.

Solution A  (6  2h)h P  6  2 5h and B  6  4h

nQ 0.025  14
  11.068
S0 0.001

f (h)  6  2hh

5/3

 11.068 6  2 5h 
2/3
 3.1753  hh
5/3

 17.569 4  5h 
2/3

f (h)  5.291(3  2h)3  hh

2/3

 26.191 3  5h 
1 / 3

The computation of normal depth is carried out as follows.

h f(h) f’(h) f ( h) h  h  h
h  
f (h)
1.000 -20.979 51.579 0.407 1.407
1.407 7.493 89.555 -0.083 1.324
1.324 0.406 81.215 -0.005 1.319
1.319 0.003 79.094 -0.000 1.319

Hence, the normal depth, hn = 1.32 m

Normal Depth for a conduit with Gradually Closing Top:

In some channel sections the top width either remains constant or increases with flow depth.
The section factor AR2/3 for these channel sections increases with an increase in the depth of flow and
there is only one normal depth for a given discharge. The rectangular, triangular, trapezoidal and
parabolic sections fall in this category. However, for channel sections having a gradually closing top,
the section factor at first increases with depth and decreases with depth when the full depth is
approached. Consequently, it is possible to have two normal depths for the same value of the section
factor and, therefore, for the same discharge. The circular section falls in this later category.
In a circular section when the depth is greater than about 0.82 d0, it is possible to have two
different normal depths for the same discharge and a small disturbance in the water surface may
cause the water surface to seek alternate normal depths, thus contributing to the instability of the
water surface. Therefore, it is advisable in the design of a circular section to restrict the depth to a
value less than or equal to 0.80 d0.
19
4.7 COMPUTATION OF NORMAL AND CRITICAL SLOPES
The normal slope (Sn) is the longitudinal slope of the channel that is required to maintain
uniform flow in the channel. When the Manning formula is used
n 2U n2 n 2Qn2
Sn   (4.106)
Rn4 / 3 An2 Rn4 / 3
or, when the Chezy formula is used
U n2 Qn2
Sn  2  2 2 (4.107)
C Rn C An Rn
Equations (4.106) and (4.107) indicate that the normal slope depends on the channel section,
the discharge, the depth and the channel roughness. Thus when the channel section, Q, n or C and hn
are given, the normal slope can be obtained using Eq. (4.106) or (4.107).
The critical slope (Sc) is the longitudinal slope of the channel for which the flow in the
channel is both uniform and critical, i.e. uniform flow occurs in a critical state and Sn = Sc, Un = Uc
and hn = hc. When the channel section, n or C and h or Q are given, the critical slope can be
determined using the Manning formula as
n 2U 2 n 2Q 2
Sc   (4.108)
R4 / 3 A2 R 4 / 3
or, using the Chezy formula as
U2 Q2
Sc  2  2 2 (4.109)
C R C AR
When Q is given, the normal depth hn, which is also equal to the critical depth hc, is first computed
using the critical condition and then the critical slope is computed using Eq. (4.108) or (4.109). On the
other hand, when, hn(=hc) is given, the mean velocity U or the discharge Q is first determined using
the critical condition and then the critical slope is computed using Eq. (4.108) or (4.109).

Example 4.14
A rectangular channel has a bottom width of 6 m,  = 1.12 and n = 0.020. (a) For hn = 1 m
and Q = 11 m3/s, determine the normal slope. (b) Determine the critical slope for Q = 11 m3/s. (c)
Determine the critical slope for hn = 1 m.

Solution Rectangular channel, b = 6m,  = 1.12, n = 0.025

(a) hn  1m Q  11m3 / s
A  bh  6  1  6m2 , P  b  2h  8m, R  A / P  0.75m
2
 0.020  11 
2
 nQ 
 Sn   2/3 
 2/3 
 0.0020
 AR   6  0.75 
(b) Q = 11 m3/s
Q 2 1.12  112
hc  3 3  0.73m
gb2 9.81 62
 hn  hc  0.73m
A  bh  6  0.73  4.36m2 , P  b  2h  6  2  0.73  7.45m, R  A / P  0.58m

20
  nQ   0.020  11 
2 2

 Sc   2 / 3. 
 2/3 
 0.0053
 AR   4.36  0.58 
(c) hc = h n = 1 m
A  bh  6  1  6m2 , P  b  2h  6  2  1  8m, R  A / P  0.75m
U n  U c  gDc /   ghc /   9.81 1 / 1.12  2.96m / s

Q  AU c  6  2.96  17.76m3 / s
or, Q  g /  bhc1.5  9.81/ 1.12  6  11.5  17.76m3 / s

 nQ   0.020  17.76 
2 2

 Sc   2/3 
 2/3 
 0.0051
 AR   6  0.75 

4.8 CHANNEL SECTION WITH COMPOSITE ROUGHNESS

The roughness may vary along the perimeter of a channel as shown in Fig. 4.5 (a). Such a
channel section is known as a channel section with composite roughness. A good example of such a
section is provided by a rectangular flume built with a wooden bottom and glass walls having
different n-values for the bottom and the walls.

Fig. 4.5 Channel section with composite roughness

In applying the Manning or the Chezy formula to compute flow in such a section, it is
necessary to compute an equivalent n-value for the entire perimeter. The flow area is divided into 3
parts, as shown in Fig. 4.5 (b), of which the wetted perimeters P1, P2 and P3 and the coefficients of
roughness n1, n2 and n3 are known. Following Horton (1933), it is assumed that each part of the area
has the same mean velocity that is also equal to the mean velocity of the whole section, i.e.
U1 = U2 = U3 = U
On the basis of this assumption, the equivalent n-value for the entire section is obtained as
follows. Since the energy slope is the same for all the sections, Eqs. (4.70) and (4.110) give
1 2/3 1 2/3 1 2/3 1 2/3
R1  R2  R3  R (4.111)
n1 n2 n3 n
or
2/3 2/3 2/3
1  A1  1 A  1A 
2/3
1 A
    2    3     (4.112)
n1  P1  n2  P2  n3  P3  n P
so that

21
 3/ 2
n   P1 
A1   1   A (4.113a)
n P
3/ 2
n  P 
A2   2   2  A (4.113b)
 n P
3/ 2
n   P3 
A3   3   A (4.113c)
n P
Using the above three equations in
A = A1 + A2 + A3 (4.114)
and simplifying and rearranging, we get

3/ 2
 P1n13 / 2  P2 n23 / 2  P3n33 / 2 
n   
 (4.115)
 P 
which can be used to compute the equivalent n-value for the entire section.

Example 4.15
The sides of a laboratory flume are made of glass (n = 0.010) and the bottom is made of
wood (n = 0.015). The flume is rectangular with b = 1 m and is laid on a slope of 0.001. Compute the
discharge in the flume if hn = 0.4 m.

Solution
Designating the perimeter of the bottom as P1 and the combined perimeter of the two sides as
P2, we have P1 = 1m, P2 = 2 x 0.4 = 0.8m, P = P1 + P2 = 1.8 m, n1 = 0.015, n2 = 0.010. Then

A  1 0.4  0.4m2 , R  A / P  0.4 / 1.8  0.222m

2/3
 P n 3 / 2  P2 n23 / 2 
2/3
 1  0.0153 / 2  0.8  0.0103 / 2 
n   1 1 
     0.013
 P   1.8 
1 1
Q AR 2 / 3 S0   0.4  0.2222 / 3  0.0011 / 2  0.36m3 / s
1/ 2

n 0.013

4.9 COMPOUND CROSS-SECTION

The cross-section of a channel may be composed of several distinct subsections as shown in
Fig. 4.6. Such a channel section is known as a compound section. For example, an alluvial river
subjected to seasonal floods generally consists of a main channel and two side channels. The side
channels are usually shallower and rougher than the main channel, So, the mean velocity in the main
channel is greater than the mean velocity in the side channels.
In dealing with a compound cross-section the subsections (the main and the side channels)
are first separated by drawing vertical lines at the banks (shown as dotted lines in Fig. 4.6). It is
assumed that (i) the water surface and energy line are horizontal across the cross-section, and (ii) the

22
vertical separation lines are the lines of zero shear and, therefore, do not contribute to the wetted
perimeters of the subsections.
The discharges in the subsections are computed by applying the Manning or the Chezy
formula separately to each subsection, i.e.

Fig. 4.6 Compound cross-section

1
Q1 
2/3
A1R1 S 01 / 2 (4.116a)
n1
1
Q2 
2/3
A2 R2 S01 / 2 (4.116a)
n2
1
Q3 
2/3
A3 R3 S01 / 2 (4.116a)
n3
The total discharge for the entire section is equal to the sum of these discharges, i.e.

Q = Q1 + Q2 + Q3 (4.117)
The mean velocity for the entire section is equal to the total discharge divided by the total area, i.e.

U = Q/A (4.118)
Where the total area

A = A1 + A2 + A3 (4.119)
The equivalent n-value for the entire section can be computed using the Manning formula for
the entire section as

n  AR 2 / 3S0
1/ 2
/Q (4.120)
or, using the equation

23
 AR 2 / 3 AR 2 / 3
n  (4.121)
K1  K 2  K 3 K
where K1, K2 and K3 are the conveyances of the individual subsections and K (=K 1 + K2 + K3) is the
conveyance for the entire section.
Owing to the differences that exist among the velocities of the subsections, the energy
coefficient  and the momentum coefficient  of the whole section are considerably different from
those of the subsections (Fig. 4.6). These coefficients can be computed as follows. the discharges Q1,
Q1  K1 S f , Q2  K5 S f , Q3  K3 S f (4.122)
Then
Q1 K1
U1   Sf (4.123a)
A1 A1
Q2 K 2
U2   Sf (4.123b)
A2 A2
Q3 K 3
U3   Sf (4.123c)
A3 A3

Q  UA  Q1  Q2  Q3  K1  K 2  K3  S f  K S f (4.124)
so that
Q K
U  Sf (4.125)
A A
Incorporating the above expressions for U1, U2, U3 and U into Eqs. (1.31) and (1.32) and
simplifying
1 K13 / A12   2 K 23 / A22   3 K 33 / A32
 (4.126)
K 3 / A2
1 K12 / A1   2 K 22 / A2   3 K 32 / A3
 (4.127)
K3 / A
where 1, 2 and 3 are the energy coefficients and 1, 2 and 3 are the momentum coefficients for
the individual subsections, respectively.

Example 4.16
A channel consists of a main section and two side sections as shown in the following figure.
Compute the total discharge and the mean velocity of flow for the entire section when n = 0.025 for
the main section, n = 0.035 for the side sections and S0 = 0.0001. Also, compute the numerical values
of n,  and  for the entire section assuming that  = 1.12 and  = 1.04 for the main and side sections.

24
Solution
The main and the two side (left and right) sections are separated by drawing vertical (dotted)
lines as shown. The computed values of A, P, R, Q and K for the three sections and A, P, Q and K for
the entire section are shown below.
Section A P R Q K
Main 425.0 56.18 7.565 655.12 65512
Left 80.0 24.00 3.333 51.00 5100
Right 60.0 19.00 3.158 36.90 3690
 565.0 99.18 743.02 74302

Hence, the total discharge, Q = 743.02 m3/s.

The mean velocity U, roughness coefficient n, energy coefficient  and momentum

coefficient  for the entire section are obtained as follows.
Q 743.02
U   1.315m / s
A 565.0
n  AR 2 / 3S01 / 2 / Q  565.0  565.0 / 99.18  0.00011 / 2 / 743.02  0.024
2/3

1K13 / A12   2 K 23 / A22   3 K33 / A32


K 3 / A2
1.12  655123 / 4252  1.12  51003 / 802  1.12  36903 / 602
  1.236
743023 / 5652
1K12 / A1   2 K 22 / A2  3 K32 / A3

K2 / A
1.04  655122 / 425  1.04  51002 / 80  1.04  36902 / 60
  1.90
743022 / 565
4.10 COMPUTATION OF FLOOD DISCHARGE BY SLOP-AREA METHOD
The slope-area method is an indirect method of estimating the flood discharge in a river from
past records of stages at different sections using the Manning or the Chezy formula. Flood flows in
natural channels are usually unsteady and varied. The use of a steady uniform flow formula for
computing flood discharge is acceptable only when the changes in flood stage and discharge are

25
sufficiently gradual and the data available are not sufficient to justify the use of a more sophisticated
method.
The selection of a suitable channel reach is probably the most important aspect of the slope-
area method. The following points must be considered in selecting the channel reach.
i) Reliable and good quality high water marks must be available in the selected reach.
ii) A straight and uniform reach is preferred and a gradually contracting reach should be
chosen over an expanding reach if there is a choice. The change in conveyance in the reach should
be less than 30%.
iii) The length of the reach L should be at least 75 times the average depth of flow, the fall of
water surface F should be equal to or greater than the velocity head and the fall should at least be
equal to 0.15 m.

Fig. 4.7 Definition sketch

With reference to Fig. 4.7, applying the energy equation between sections 1 and 2, we get
U12 U 22
F  1  2  h f  hc (4.128)
2g 2g
or
 U2 U2 
h f  F   1 1   2 2   hc (4.129)
 2g 2g 
where hc is the eddy loss, given by

U12 U 22
hc  k 1  2 (4.130)
2g 2g
and the coefficient k is assumed to range between 0 and 0.1 for gradual contractions, between 0 and
0.2 for gradual expansions and to have a value of 0.5 for abrupt expansions or contractions.
The following data are required for the slope-area method: (i) the cross-sectional areas A1 and
A2 of the upstream and downstream sections of the selected reach, (ii) the wetted perimeters P1 and
P2, (iii) the Manning roughness coefficients n1 and n2, (iv) the energy coefficients 1 and 2, (v) the

26
length of the reach L, and (vi) the fall of the water surface F between the two sections. In addition, the
eddy-loss coefficient k is to be specified if the eddy loss is to be included.
The computation of flood discharge using this method involves the following steps:
i) Compute the conveyances K1 and K2 for the two end sections.
ii) Compute the geometric mean conveyance for the reach, i.e.

K  K1 K1 (4.131)

iii) Since the discharge Q is not known initially, as a first approximation assume that h f = F
and hence
hf F
Sf   (4.132)
L L
iv) The first approximation of the discharge (which is also the uniform flow discharge) is than
computed using the equation

Q  K Sf (4.133)
v) A more accurate value of the energy slope is now obtained using the equation

hf
Sf  (4.134)
L
where hf is given by Eq. (4.129). The corresponding discharge is then computed by Eq. (4.133) using
the revised slope given by Eq. (4.134).
v) Repeat step (v) until the assumed and the computed discharges agree.

Example 4.17
Compute the flood discharge through a river a reach of 850 m using the following data:
A1 = 10350 m2, P1 = 2035 m, n1 = 0.030, 1 = 1.15
A2 = 9275 m2, P2 = 1965 m, n2 = 0.030, 2 = 1.18
The fall of water surface in the reach is 0.76 m. Neglect eddy loss.

Solution L = 850 m, F = 0.76 m, hc = 0

 10350 10350/ 2035  1020320

1 1
K1  A1 R12 / 3 
2/3

n1 0.030

 9275 9275/ 1965  869950

1 1
K1  A2 R22 / 3 
2/3

n2 0.030

K  K1K 2  1020320 869950  942140

27
Approxi- Assumed F (m) U2 U2 hf (m) Sf (x10-4) Computed
mation Q(m3/s) 1 1 (m)  2 2 ( m) Q(m3/s)
2g 2g
1st - 0.76 - - 0.7600 8.94 28170

2nd 28170  0.4343 0.5548 0.6395 7.52 25840

3rd 25840  0.3653 0.4668 0.6585 7.75 26220

4th 26220  0.3762 0.4806 0.6556 7.73 26170

5th 26170  0.3747 0.7488 0.6559 7.72 26170

 The flood discharge, Q = 26,170 m3/s

PROBLEMS

4.1 Assuming that the velocity distribution along a vertical in an open channel is logarithmic, compute
the position of the mean velocity below the free surface. Also show that (i) the velocity at 0.6 depths
and (ii) the average of the velocities at 0.2 and 0.8 depths are approximately equal to the mean
velocity in a vertical.

4.2 (a) A trapezoidal channel has a bottom width of 6.0m, side slopes of 1.5H:1V, a depth of flow of
2.0 m, n = 0.025 and S0 = 0.0001. Assuming that the flow is uniform, (i) compute Q, (ii) compute C, f,
0 and u*, and (iii) compute ks, determine whether the channel boundary is smooth or rough and
state if the Manning formula is applicable for computing flow in this channel. Assume that the
velocity distribution is logarithmic.

(b) Consider the following data for the Padma (Ganges) river at the Baruria station in Faridpur on
the 2nd July, 1989: A = 33,500 m2, Q = 56,200 m3/s and B = 3820 m. Assuming that the flow is
uniform, (i) compute n, C, f, u* and 0, and (ii) determine whether the channel boundary is smooth or
rough taking the velocity distribution as logarithmic. Assume that the river is wide. Longitudinal
slope of the river is 4 cm/km.

4.3 Show that for a wide rough channel with logarithmic velocity distribution in the vertical, the
Manning roughness coefficient n may be expressed by

n
r  1h1/ 6
5.57r  0.95
where r (= u0.2/u0.8) is the ratio between the measured velocities at two-tenths and eight-tenths of
depth.

4.4 When the Manning formula is used, show the critical slope at a given normal depth hn may be
expressed by
gn 2 Dn gn 2 n 2 g 10/9
Sc  and that this slope for a wide channel is S c   2/9
Rn4/3 hn1/3 q
Where q is the discharge per unit width.

28