1.

4 Computation of Gradually Varied Flow

• The computation of gradually-varied flow profiles involves basically

the solution of dynamic equation of gradually varied flow. The main
objective of computation is to determine the shape of flow profile.
• Broadly classified, there are three methods of computation;
namely:
1. The graphical-integration method,
2. The direct-integration method,
3. Step method.
The graphical-integration method is to integrate the . dynamic
equation of gradually varied flow by a graphical procedure. There
are various graphical integration methods. The best one is the
Ezra Method.
 The direct-integration method: Thje differentia equation of GVF
can not be expressed explicitly in terms of y for all types of flow
cross section; hence a direct and exact integration of the equation
is practically impossible. In this method, the channel length under
consideration is divided into short reaches, and the integration is
carried out by short range steps.
• The step method: In general, for step methods, the channel is
divided into short reaches. The computation is carried step by step
from one end of the reach to the other.
• There is a great variety of step methods. Some methods appear
superior to others in certain respects, but no one method has been
found to be the best in all application. The most commonly ısed
step methods are:
1. Direct-Step Method,
2. Standart-step Method.
 1.4.1 Direct-Integration Methods
We have seen that the flow equation

dy S o  Sf

dx 1  Fr 2

is true for all forms of channel section, provided that the Froude number
Fr is properly defined by the equation:
V 2T Q2 T
Fr 
2

g A g A3
and the velocity coefficient, α = 1, channel slope q is small enough so
that cos θ =1.
We now rewrite certain other elements of this equation with the aim of
examining the possibility of a direct integration. It is convenient to use
here the conveyance K and the section factor Z.
 The Conveyance of a channel section, K:
If a large number of calculations are to be made, it is convenient to
introduce the concept of “conveyance” of a channel in order to calculate
the discharge. The “conveyance” of a channel indicated by the symbol K
and defined by the equation
Q
Q  KS 1/ 2 or K 
S
This equation can be used to compute the conveyance when the
discharge and slope of the channel are given.
When the Chézy formula is used: K  CAR1 / 2
where c is the Chézy’s resistance factor. Similarly when the Manning
formula is used
1
K  AR 2/3

n
• When the geometry of the water area and resistance factor or
roughness coefficient are given,
One of the above formula can be used to calculate K. Since the
Manning formula is used extensively in most of the problems, in
following discussion the second expression will be used. Either K
alone or the product Kn can be tabulated or plotted as a function of
depth for any given channel section: the resulting tables or curves
can then be used as a permanent reference, which will immediately
yields values of depth for a given Q, S and n. This conveyance factor
concept is widely used for uniform flow computation.
Since the conveyance K is a function of the depth of flow y, it may be
assumed that:
where K 2
 C 1 y N

C1 = coefficient, and
N = a parameter called hydraulic exponent
Taking the logarithms of both sides of above eq. And then differentiating
with respect to y:
2nK  Nny  nC1
dnK  N dnK 
  N  2y
dy 2y dy

On the other hand: from Manning’s Eq.

1
K  AR 2 / 3
n
Taking the logarithm of both sides and differentiating with respect to y:

2
nK  nR  nA  nn
3
dnK  2 dR 1 dA
 
dy 3R dy A dy
• The derivative of hydraulic radius with respect to y:
dR 1 dA A dA T R dP
  2  
dy P dy P dy P P dy

dnK  2 1  T R dP  T 2 T 2 1 dP T
       
dy 3 R  P P dy  A 3 A 3 P dy A
dnK  N dnK 
  N  2y
dy 2y dy
dnK  1  dP 
  5T  2R 
dy 3A  dy 

2y  dP 
N  5T  2R 
3A  dy 
• This is the general Eq. for the hydraulic exponent N. If the channel
cross section is known N can be calculated accordingly provided that
the derivative dP/dy can be evaluated. For most channels, except for
channels with abrupt changes in cross-sectional form and for closed
conduits with gradually closing top, a logarithmic plot of K as ordinate
against the depth as abscissa will appear approximately as straight
line. Thus if any two points with coordinates (K1, y1) and (K2, y2) are
taken from the straight line, the approximate value of N may be
computed by the following Eq.
LnK

log K 1 / K 2 
K2
N2
K1 log y1 / y 2 

y1 y2
Lny
• For wide rectangular channels: R  y
• The Chézy Equation gives the value of K as:

K 2  C2 A 2 R  C2 b 2 y 2 y  C2 b 2 y3  N  3
• On the other hand the Manning Equation gives the value of K as:

1 1 1
K 2
AR2 4/3
 b y y
2 2 4/3
 b y
2 10 / 3

n 2
n n 2

 N  10 / 3
 The Section Factor: Z
• The Section Factor: Z is especially used for critical flow computation.
However it becomes useful to transform the GVF equation into a
form which can be integrated directly.
• We now consider the Froude number
2
Q T
F  r
2
3
g A
• Since this term equals unity at critical flow, then
2 3
Q A
 c
g Tc
• Ac, Tc are the values of A and T at critical flow..
• By definition we introduce the concept of section factor as
2 A
A Z A A D and Z  A D
3
Z 
2 or
2
  2

T T
• The section factor for critical flow becomes Z 2  Ac  Q
3 2

c
Tc g
Q
• or Zc  for critical flow only.
g

Since the section factor z is a function of depth, the equation

indicates that there is only one possible critical depth for maintaining
the given discharge in a channel and similarly that, when the depth is
fixed, there can be only one discharge that maintains the critical flow
and makes the depth critical in that given channel section.
• Since the section factor z is a function of the depth of flow y, it may
be assumed that
Z 2  C2y M

Where C2 is a coefficient and M is a parameter called the hydraulic

exponent. Taking the logarithms on both sides of above equation and
then differentiating with respect to y:
d nZ  M d nZ  d nZ 
  M  2y 2
dy 2y dy d ny 
A3
Now taking the logarithms on both sides of Eq. Z 
2

2nZ  3nA  nT

Take the derivative with respect to y:
 d nZ  3 dA 1 dT 3 1 dT
   T 
dy 2A dy 2T dy 2A 2T dy
• Then M becomes::

y A dT 
M   3T  
A T dy 

• This is a general equation for the hydraulic exponent M, which is a

function of the channel section and the depth of flow.
• For a given channel section M can be computed directly from this
expression, provided that the derivative dT/dy can be evaluated.
However, approximate values of M for any channel section may be
obtained from the following equation:
 logZ 1 / Z 2 
M 2
logy 1 / y 2 

where z1 and z2 are section factors for any two depths y1 and y2 of
the given section, i.e.:
A 3
A 3
Z 12  1 and Z2  2
2
T1 T2
For rectangular sections: A=by, and T=b, hence section factor
becomes:
b y
3 3
Z 
2
 b 2y 3  M  3
b
• Now, coming back to the computation of gradually varied flow; the
term, (S0-Sf )may be written as

 Sf 
S 0  Sf  S 0 1  
 S0 
Q Q
• On the other hand K  and for uniform flow K 0 
Sf S0

• Therefore S f Q 2 / K 2 K o2
 2  2
S0 Q / K o K2

 K o2 
• And hence S 0  Sf  S 0 1  2 
 K 

• Also, it is assumed that K  C1y N and K  C y N

0 1 0
• Therefore
N
 Ko  2  y  
S 0  Sf  S 0 1  2   S 0 1   o  
 K    y  
• On the other hand:

Q T Q 1
2
Z 2 2
Fr 
2
  c
g A3 g Z 2 Z 2
• because for critical flow: Z 2  Q 2

c
g
• Also it is assumed that Z 2  C2y M
Therefore Froude number can be written as:
M
C2y M
 yc 
Fr 
2 c
  
C2y M
 y 
Finally the Gradually Varied Flow Eq. takes the form:
N
 yo 
1   
dy S 0 S f  y 
  S0 M
dx 1  Fr 2
 yc 
1   
 y 
 Bresse’s Method:
• Applicable only to wide rectangular channels. Assumptions:
• Wide rectangular channel,
• Chézy’s formula is applicable and Chézy’s C is constant.
• From Chezy’s eq’n: , K  CA R A=by Ry
• Therefore K 2  C 2 A 2 R  C 2 b 2 y 2 y  C 2 b 2 y 3  N  3

3 3 3
A b y
Z 
2
  b 2 y3  M  3  yo 
3
T b 1   
Therefore, the GVF Equation becomes:
dy
 So  y 
3
dx  yc 
1   
 y 
Manipulating this equation we can integrate it.
   y 3 
 y0 
3
1   c  
  y   Adding and subtracting  
S o dx   3
dy  y 
1   y o  
  y  

 y   yo   yc  
3 3 3

1   o         Rearrange it and multiply and

   y   y   dy
S o dx   
y 3
 yo 
3   y 
 1    
divide by  
 y0 
  y  

  y 3  y   y
3
 3

  c    o    
  y   y  .  yo  
S o dx  1  3 3
dy
  yo   y  
1     
  y   yo  

  y 
3
  3
  c   1  yc 
 1    
  yo      
S o dx  1   dy or S o dx  1 
yo
dy
3
3
  y 
  1  1   y  
  y o  y 
     o  

y
Let U   dy  y o du change variables
yo
  yc  
3

 1    
  yo   Integrating
S o dx  1   y o du
• 1 U 3
 
 
 
  y 
3
 du 
So x  y o u  1   c  
  c
   yo   1u 
3
   
 du
Let f be the integral of f  
1u 3

du 1 u u 1 1 2
3
f  n  tan 1

1u 3
6 u  12
3 2u  1
Therefore, the distance between any two
section becomes:

yo    yc  
3


L  x2  x1  u2  u1   1    f2  f1 
So 
 
  yo  
 

 Bakhmeteff Method
• Bakhmeteff improved Bresse’s method as follows:

 sf 
1  
dy s  s  so  Sf
•  o f
 so Let Fr  
2

dx 1  Fr 2
1  Fr2 So
Sf
1
Therefore dy  S o So Then we can write that:
dx Sf
1 
So
N
S Sf  y o 
1  f
1 S0 On the other hand, we can wite that:   
dx 
S
dy S0  y 
S0 1  f
S0

Hence dx becomes:
 N
 yo 
1      y0 
N
dy  y   
dx  Add and subtract
so y 
N
 y 
1   o 
 y 
N
 yo 
1     N N
dy  y   yo   yo 
dx        and rearrange as:
so  yo 
N
 y   y 
1   
 y 
N
  yo  
 1     
  y  
S 0dx  1  dy
  yo  
N
 1    
  y  
 N
Now multiple and divide by  y 
 
 
 y0 
N
 y 
N
 y  
 1    o    

S 0dx  1   y   yo  
N N dy and rearrange as:
  yo   y  
1     
  y   yo  

 
  y
 1     Let U   dy  y o du
S 0dx  1  N dy yo
•
 1   y  
 y  
 o
• Integrating
y 0  1    
 dx   S 0 1  1  u N dy
y0  du 
x  u  1    N 
S0  1u 

Therefore, the distance between any two

section becomes:

u 2  u1   1  Fu 2 , N   Fu, N 

yo
L  x 2  x1 
so
• Here we are assuming that β is constant
2 2
Sf Q T S Q T
Fr  
2
 3
  0
3
S0 gA S f gA
and hence
Q2
Q  CA RS f Sf  2 2
C AR

Therefore  becomes:

S 0C A R Q T S 0C T S 0C T
2 2 2 2 2
  
Q 2
gA 2
gP g P
• If T/b is constant, then β will be constant.
• For example, for a rectangular channel,
T=b, and P=b+2y, hence

T b b 1
  
P b  2y  y  1  2y / b
b1  2 
 b

• And T/b becomes constant only for wide

rectangular channel.
• For a triangular channel: T=2zy
T
P  2 1 z y2

1
z

T 2zy z
   constant
P 2 1 z 2 y 1 z 2
• The integral in the equation is designated
by F(u,N), that is

• Fu, N   
du
 Varied-Flow Function
o 1 u
N

• The values of F(u,N) have been obtained

numerically, and are given in tabular form
for N ranging from 2.2 to 9.8.
 Example on Bakhmeteff Method:
Water is taken from a lake by a triangular channel,with side slope
of 1V:2H. The channel has a bottom slope of 0.01, and a Manning’s
roughness coefficient of 0.014. The lake level is 2.0 m above the
channel entrance, and the channel ends with a free fall. Determine :
1.The discharge in the channel,
2.The water-surface profile and the length of it by using the direct-
integration method.

2m

n=0.014, S0=0.01 X

1
2
a) To determine the discharge, assume that the channel slope is
steep. Then at the head of steep slope, the depth is critical depth.
Hence 2=Emin
A y A y
A  2 y , P  2 5 y, R  
2
, T  4y, D  
P 5 T 2
DC yC
5
2 E C y C   yC   y C ⇒ y C  1.6 m.
2 4 4
Q 2
A 3
5 . 12 3
A c  2x 1.6 2  5.12 m2  c   20.97
g Tc 4 x 1.6
 Q  14.34 m3 / s let' s compute the normal depth :
2/3 2/3
A 0  A0  2y  2y 
2 2
Q   S o  14.34  
0 0
 0.01
n  Po  0.014  2 y 0 5 
y 80 / 3  1.716  y 0  1.225 m  y c  1.6 m  steep slope &
Q  14.34 m3 / s
2. The water-surface profile and the length of it by using the direct-
integration method.
• The water surface profile will be S2 type, and the depth of flow will
change between the limits: 1.01y0≤y≤yc.
• Since T/P is constant for triangular channels, we can use
Bakhmeteff’s method. Let’s compute the value of :

S 0C 2 T
 Q  CA 0 R 0S0 A 0  3.0 m2
g P
R 0  0.5476 m, T  4.9 m, P  5.48 m
S 0C 2 T
C  64.595   
g P
0.01x 64.6 2 x 4.9
  3.80
9.81x 5.48
• Let’s obtain the value of N

2y  dP  y
N   5T  2R , R  , P  2y 5
3A  dy  5
2y  y 
N  2 
5x 4 y  2 x 2 5
3x 2 y  5 
16
N   5.333
3

u 2  u1   1  Fu 2 , N   Fu, N 

yo
L  x 2  x1 
so
• To obtain the values of F(u,N) we have to use tables with
interpolation.
y (m) u=y/y0 F(u,5.333)
1.6 1.31 0.0811
1.237 1.01 0.622
D -0.296 0.5411

• Interpolation:
• For u=1.30 and N=5 →
• For u=1.32 and N=5 →
• Hence by interpolation:
• For u=1.31 and N=5 →
F(1.30, 5)=0.100 and
F(1.32, 5)=0.093
F(1.31, 5)=0.0965
• Similarly:
• For u=1.30 and N=5.4 → F(1.30, 5.4)=0.081 and
• For u=1.32 and N=5.4 → F(1.32, 5.4)=0.093
• Hence by interpolation:
• For u=1.31 and N=5.333 → F(1.31, 5)=0.0811

yo
L  x 2  x1  u 2  u 1   1   F u 2 ,N   F u ,N 
so
1.225
L  0.296  1  3.80.5411  149.34 m
0.01

• Therefore, the length of S2 profile is 150 m.

 Ven Te Chow Method
• Ven Te Chow improved Bakhmeteff’s method for all types of cross
sections as follows:
N
y 
1   o 
dy
 So  y 
M
dx y 
1   c 
 y 
The gradually-varied flow equation can be written as

1 1  y c / y M
dx  N dy multiply by both numerator and
S o 1  y o / y 
N M
 y   y0 
denominator by   and second term in numerator by  
 yo   yo 
1  y / y o N  y / y o N y c / y M y 0 / y o M 
dx   dy
So  y / y o N  1 
 N M M N
 y   yc   yo   y  
          
1  y o   yo   y   y o  dy
dx 
So  1  y / y o N 
 
 
y
Now let u   dy  y o du then above equation becomes:
yo
:
M
 y  N M 
  c  u u N 
y o  y o  
dx  du
so  1u N
 
 
Adding and subtracting 1 to the numerator, dx becomes::
 M
 y  N M 
1  u N  1   c  u 
yo   yo  
dx  du
Rearrange to obtain
So  1u N
 
 

M
yo  1  y c  u N M 
dx  1     du Integrate
So  1  u N
 y o  1  u 
N


yo  u
du  yc 
Mu
u NM 
x u       du   const.
so  1 u N
 yo  o 1  u N

 o
• The first integral on the RHS of above eq. is designated by or
u
du
F u ,N     Varied - Flow Function
o 1u
N

The second integral may also be expressed in the form of the

varied-flow function. Let
N
v u N /J and J
N  M  1
J J NN
u v J /N du  v dv
N
J
u v N M  J N v
u N M
J v N N J dv
o 1  u N du  N o 1 v J dv  
N o 1 v J
• Therefore:
M
yo  u
du  yc  J v
dv 
x  u         const.
So  o 1  u N
y
 o N 1  v J

 o

or
M
yo  J  yc  
x  u  F u ,N     F v ,J   const.
So  N  yo  

yo
L  x 2  x1  u 2  u1   F u 2 ,N   F u1,N  
So
M
J  yc 
  F v 2 ,J   F v 1J  
N  yo 
• Where the subscripts 1 and 2 refer to sections 1 and 2 respectively.
This eq. contains varied-flow functions, and its solution can be
simplified by the use of the varied-flow-function table. This table
gives values of F(u,N) for N ranging from 2.2 to 9.8. Replacing values
of u and N by corresponding values of .F(v,J).
•
• In computing a flow profile, first the flow in the channel is analyzed,
and the channel is divided into a number of reaches. Then the length
of each reach is computed by the above Eq. from known or assumed
depths at the ends of the reach. The procedure of the computation is
as follows:
1. Compute the normal depth yo and critical depth yc from the given
values of Q and so, n
A 2/3 Q2 T
Q R So  y o 3
 1  yc
n g A
2. Determine the hydraulic exponents N and M for an estimated
average depth of flow in the reach under consideration. It is assumed
that the channel section under consideration has approximately
constant hydraulic exponents
• If any two depths in the section are known then
•
log K 1 / K 2  logz1 / z 2 
• N2 M2
log  y / y  logy1 / y 2 
• 1 2

• N and M can be computed from above expressions or use the

general formulas if dP/dy and dT/dy can be evaluated.
3. Compute J by
N
J
N  M 1
5. Compute values u  y y and v  u
N /J
at the two end
sections of the reach
0

6. From the varied-flow-function table, find values of F(u,N) and F(v,J).

7. Compute the length of reach by the given equation obtained above.
• In doing so a table can be prepared as follows

y u v F(u,N) F(v,J)
y1
y2
y3
D
 Example on Ven Te Chow’s Method:
• Water flows at a uniform depth of 3 m. in a trapezoidal channel. The
trapezoidal section has a bottom width of 5 m., and side slope of
1H:1V. The channel has a bottom slope of 0.001, and a Manning’s
roughness coefficient of 0.013. The channel ends with a free fall.
Determine the water-surface profile and the length of it by using the
direct-integration method.

y0=3 m
1
1
S0=0.001, n=0.013
b=5 m
 First, determine the discharge, and then the type of the slope:
A 2/3
Q R S0 A  by  zy 2  5x3  32  24 m2
n
P  b  2 y 1  z 2  5  2 x3 2  13.46 m.
A
R   1.783 m Q 
24
1.7832/3 0.001  86 m 3 / s
P 0.013
To determine critical depth yc

Q 2

A
 c 
5 yc  y
3
c
2 3

86 2
 753.92
g Tc 5  2 yc 9.81
by trial and error y c  2.6 m

y0=3 m > yc=2.6 m Therefore it is mild slope. Hence M2 profile occurs.

The depth changes between the limits yc = 2.6 m ≤ y ≤ 0.99y0 = 2.97 m

 • Let’s compute the hydraulic exponents
K2 Z2
ln K1 ln Z1
A 2/3 A
N=2 K= R M=2 Z=A
y2 n y2 T
ln y1 ln y1


  
M

y
L  x2  x1  0 u2  u1   F u 2 , N   F u1, N    
y c J
F v 2 , J   F v1, J 
S0 
  y0  N 

y
u v  u N/J
y0
 Computation of Hydraulic Exponents N and M, n=0.013

y (m) A (m2) P (m) R (m) T (m) K Z N M J

2,60 19,76 12,35 1,60 10,2 2078,9 27,50 3,71 3,55 3,19
2,97 23,67 13,40 1,77 10,94 2660,7 34,82

Computation of Length of M2 profile by using

Direct-Integration Method in one step

y0=3.0 m , yc=2.6 m, n=0.013, S0=0.001, N=3.7, M=3.6, J=3.2

y (m) u v F(u,3.7) F(v,3.2) L (m)

2,60 0,87 0,85 1,05 1,043 -341,781

2,97 0,99 0,99 1,7875 2,017

L=342 m
 1.4.2 Step Methods for Uniform and Irregular
Channels
The Direct Step Method (DSM)
• In direct step method, distance is calculated from the depth.
• It is only applicable to prismatic channels. The energy equation
between two sections is:
V12 V 22
z 1  y 1  a1  z 2  y 2  a2  Sf Dx
2g 2g
• Assume that a1=a2=1 z1-z2=S0Dx, Then above
equation can be written as:
S 0 Dx  E 1 E 2Sf Dx
E2  E1
• Solving for Dx: Dx 
So  S f
• The distance between two sections can be calculated from:

 v 
2
D y  a 
E2  E1 DE  2g 
Dx   
So  S f So  S f So  S f

• The slope of energy-grade line can be computed from

Manning’s Equation as:

nV 2 2
Sf  4 / 3 1

Sf  Sf1  Sf 2 
R 2
Consider the M2 profile in a channel section. Suppose that we want to
calculate the length of M2-profile for a given discharge Q, and channel
section.
NDL

M2
CDL

yc y01
yc
A
O
• We know that the depth will be changing between the limits yc≤y <y0.
• On the other hand the distance between two sections can be written as:
  v2 
D y  a 
E 2 - E1 DE  2g  and
Dx   
S o - Sf S o - Sf So - Sf

1 n2 V 2
S f = (S f 1 + S f 2 ) Sf =
2 R4/3

• Start from a control section. In this case the control section is the
downstream section where critical flow occurs.
• Nominate a series values of y, between the range of yc≤y <y0
• Calculate the values of R, V, E, and Sf corresponding to these
assumed depths.
• Calculate Δx for each interval between successive values of y
• NOTE:
• If flow is subcritical computation is from downstream towards
upstream.
• If flow is supercritical computation is from upstream towards
downstream.
• If one of the depth is uniform depth for the section under
consideration, then 1% off value of normal depth y0 must be taken.
For example for profiles like M1 and S2 where normal depth is
approached asymptotically from above 1.01 y0, and for profiles such
as M2 and S3, where normal depth is approached asymptotically
from belove 0.99 y0 must be taken.
The best thing is to prepare a table as follows:

y A P R=A/P U E Sf Dx x
(m) (m2) (m) (m) (m/s) (m) Sf (m) (m)

Yc

Y1

Y2

y3
 The Standard-Step Method
In this method, the depth is calculated from distance.
Applicable to both nonprismatic and prismatic channels. It is a trial and
error process
V22
•Assume y2, compute H2: H2  z2  y2 
2g
1
2 = H1 + h f = H1 + (S f 1 + S f 2 )Δx
•Also compute:
H′
1
2
•Where h f = S f Δx
2
•Compare H2 to H2’:
•if H 2  H 2 then assumed y2 is O.K.

•But, if H2 ≠H1′ , then assume another value of y2

Improve your initial estimate by this amount:

He where H e  H 2 - H 2
Dy 2 
3 s e2
1  Fr 
2
2 Dx
2 R2
For natural rivers, instead of depth y, it is preferable to use the height h
of the water level above some fixed datum. This height, h=z+y, is
known as the STAGE. Hence total head at a section can be written as:

V2 V2
H= y+ z+ α = h+ α
2g 2g
• In nonprismatic channels, the hydraulic elements are no langer
independent of the distance along the channel
• In natural channels, it is generally necessary to conduct a field
survey to collect data required at all sections considered in the
computation.
• The computation is carried on by steps from station to station where
the hydraulic characteristics have been determined. In such cases
the distance between stations is given, and the procedure is to
determine the depth of flow at the stations.
 Step Method-Divided Channels
• In this section we consider cases where the channel cross section is divided into distinct
regions having distinct flow characteristics.
• The most common example of this situation is the one shown in Fig. below, the case of
overbank flow, when the flow over the berms has a different depths and the surface possible
a different roughness, from those existing in the main channel.
• If the channel is straight the water-surface level will remain substantially constant over the
whole section of flow, since the hydrostatic pressure must remain constant along any
horizontal line drawn across the section.
• However, the distinct regions of flow shown in Figure will almost certainly have different
velocities and velocity heads; the problem then is to define a total head H applicable to the
entire cross section.
• The solution is to use the energy coefficient, a; the total head line then extends across the
whole water surface, a distance of above it, as shown in Fig. above. This total head line,
and any head losses deduced from it, are assumed to be applicable to the section as a
whole, and also to each of the individual subsections.
• The algebra required is conveniently handled by the use of the conveyance K. The energy
coefficient , a, is defined by the equation

  A   A   Q
3
1 1
2
3

  A
1
a  1

•  A3
  
3 2

…………………………………... Eq.(1)
Q 
m 1 1

  1
 

• Where subscripts 1, 2, 3,… indicate the distinct subsections of the flow

 Now since we are assuming that the same value of friction slope Sf applies to each
subsection, it follows from the definition of the conveyance K , that

Q1 Q 2 Q 3 Q
Q 1

   ... n  ……………………………… Eq.(2)

K1 K 2 K 3 Kn K 1

And therefore that

2
 
  A1 
   K 13  …………… Eq.(3)
 
a 3   A 2 
   1
  K1 
 
 
 Since each element of Eq. (2) is equal to (Sf)1/2 , it follows that
an
  Q1 
2

  Q2
Sf     ………………………. Eq.(4)
 1 
2
K  
    K1 
 
 

Where Q is the total flow. Thus a and Sf, the two factors which are of critical importance in the tabulation,
can be calculated without explicitly evaluating the discharges etc. The values of K which are to be inserted
in Eqs. (3) and 4) are obtained from the Manning equation as
1
K AR2 / 3
n
We now have all the information required to proceed with the computation.
For a typical flood-level computation shown in Fig.(1), the necessary tabulation is as follows
It is seen that at each major section a separate line is devoted to each subsection.
These lines are labeled M.C. (main channel) and L.B. (left berm) in column 2.
Again is assumed that a plan of each section is available, from which, given the water level, A and P
can be determined for each subsection.
The liquid interface between zones 1 and 2, shown as a broken line in Fig. (1), is not
normally included in estimates of the wetted perimeter P because the shear stress along
this interface is so much lower than at the solid boundaries. ,
The calculation proceeds upstream from known conditions at river km 1.
River
Km
Sub
sec
Stage
,h
A
(m2)
P
(m)
R
(m)
R(2/3) n K K3/A2
a Vm H (Sf) Dx hf H’
(m/s)
(m) (m) av (m) (m)
1 M.C

L.B 1.32

Total …….. ……. ……..

2 M.C 1.25

L.B

Total

3
 1.5 The Discharge Problem
(Side-Channel Spillways) A

A side-channel carries water away

from an overflow spillway in a channel x
parallel to the spillway crest as
L
shown in figure.
A
The discharge over the entire width (L) of
an overflow spillway can be determined by
Eq.(1), and the discharge through any
section of the side channel at a distance x
from the upstream end of
the channel is

Q x  xCH 3a / 2 Eq.(1)
The side-channel spillway must provide a slope steep enough to carry away the accumulating
flow in the channel. However, a minimum slope and depth at each point along the channel is
desired in order to minimize construction costs. For this reason, an accurate water surface
profile for the maximum design discharge is important in the side-channel spillway design.

The flow profile in the side channel cannot be analyzed by the energy principle
(i.e., gradually-varied flow profile) because of the highly turbulent flow conditions
that cause excess energy loss in the channel.

However, an analysis based on the momentum between two adjacent sections,

a distance of ∆x apart, in the side channel,

 F  rQ  DQV  DV  rQV Eq.(2)

where r is the density of water, V is the average velocity, and Q is the discharge
at the upstream section. The symbol ∆ signifies the incremental change at
the adjacent downstream section.
 The forces represented on the left-hand side of Eq.(2) usually include the weight component
of the water body between the two sections in the direction of the flow rgADx sin q ,
the unbalanced hydrostatic forces

 
ρgAy cosθ  ρg  A  ΔA y  Δy  cosθ
 

and a friction force, Ff, on the channel bottom. Here A is the water cross-sectional area, y is the distance
between the centroid of the area and the water surface, and q is the angle of the channel slope.
The momentum equation may thus be written as

ρgAΔx sin θ  ρgAy  ρg  A  ΔA y  Δy cosθ  Ff  rQ  DQV  DV  rQV Eq.(3)

Let So  sin θ for a reasonably small angle, and Q  Q1, V  ΔV  V2 , A  Q1  Q2  / V1  V2 ,
and Ff  ASf Dx ; the above equation may be simplified to
Q1 V1  V2   DQ 
Dy    DV  V2   S0 Dx  Sf Dx
gQ1  Q 2  
Eq.(4)
Q1 
Where Dy is the change in water surface elevation between the two sections.
This equation is used to compute the water surface profile in the side channel.
The first term on the right-hand side represents the change in water surface
elevation between the two sections resulting from the impact loss caused by
the water falling into the channel. The middle term represents the change from
the bottom slope, and the last term represents the change caused by friction
in the channel. Relating the water surface profile to a horizontal datum, we may write

Q1 V1  V2   DQ 
Dz  Dy  S0 Dx    DV  V2   Sf Dx
gQ1  Q 2   Q1 
Eq.(5)

Note that when Q1  Q 2 or when DQ  0 , Eq.(5) reduces to

 V22 V12 
Dz      Sf Dx Eq.(6)
 2g 2g 
which is the energy equation for constant discharge in an open channel as derived in
Computation of Gradually-Varied Flow.
 Example

A 6-m overflow spillway discharge water into a side-channel spillway with a horizontal bottom slope. If
the overflow spillway (C=2.043 m1/2/s) is under a head of 1.28 meter, determine the depth change from
the end of the side channel (after it has collected all of the water from the overflow spillway) to a point
1.5 meter upstream. The concrete (n=0.013) side channel is rectangular with a 3-m bottom. The water
passes through critical depth at the end of the side channel.

Solution

The flow at the end of the side channel (Eq.(1)) is

Q  CLH a   2.04361.28  17.75 m3 / sec

3/ 2 3/ 2

The flow at a point 1.5 meter upstream (Eq.(1)) is

Q  xC H a   4.52.0431.28  13.31 m3 / sec

3/ 2 3/ 2

Solving for critical depth at the end of the channel, we have

q  Q / b  17.75 / 3  5.92 m3 / s / m

yc  q / g 2

1/ 3

 5.92 /9.81
2
1/ 3
 1.528 m
 The solution method employs a finite-difference solution scheme (Eq.(4)) and an iterative processs
can be employed to compute the upstream depth. The upstream depth (or depth change, ∆y) is
estimated, and Eq.(4) is solved for a depth change. The two depth changes are compared, and a
new estimate is made if they are not nearly equal. Table... displays the solution. Because side-
channel spillway profile computations involve implicit equations, computer algebra software (e.g.,
Mathcad, Maple, and Mathematic) or spreadsheet programs will prove very helpful.

DX Dy y A=3y Q (m3/s) V (m/s) Q1+Q2 V1+V2 DQ DV R=3y/(3+2y) Sf Dy

0 1.528 4.584 17.75 3.872
1.5 -0.3 1.828 5.484 13.31 2.427 31.06 6.299 4.44 1.445 0.824 0.001289 -0.755

DX Dy y A Q V Q1+Q2 V1+V2 DQ DV R Sf Dy
0 1.528 4.584 17.75 3.872
1.5 -0.755 2.283 6.849 13.31 1.943 31.06 5.816 4.44 1.929 0.905 0.000729 -0.819

DX Dy y A Q V Q1+Q2 V1+V2 DQ DV R Sf Dy
0 1.528 4.584 17.75 3.872
1.5 -0.819 2.347 7.041 13.31 1.890 31.06 5.763 4.44 1.982 0.915 0.00068 -0.825

DX Dy y A Q V Q1+Q2 V1+V2 DQ DV R Sf Dy
0 1.528 4.584 17.75 3.872
1.5 -0.825 2.353 7.059 13.31 1.886 31.06 5.758 4.44 1.987 0.916 0.000675 -0.826

DX Dy y A Q V Q1+Q2 V1+V2 DQ DV R Sf Dy
0 1.528 4.584 17.75 3.872
1.5 -0.826 2.354 7.062 13.31 1.885 31.06 5.757 4.44 1.987 0.916 0.000675 -0.826