CHAPTER 1

1.1 INTRODUCTION
Computer analysis of steady flow in open channels is a valuable tool for
hydraulic engineers. It enables detailed predictions of flow characteristics
(surface profiles most significantly) for a particular water course (in this case
open channels) under certain conditions; without the need to take field
measurements which can be time consuming, expensive and subject to
instrumental and human errors.
The demand for efficient and accurate software that can deal with problems
faced by hydraulic engineers has lead to numerous hydraulic packages in the
software market. In the past some of the more sophisticated software
available required the use of powerful computers and long run times, as a
result of the level of complexity involved. However this situation has improved
with the advances in computer hardware.
1.2 OPEN CHANNEL HYDRAULICS
An open channel is a conduit in which water flows with a free surface usually
under atmospheric pressure. The classification of open channel flow is made
according to the change in flow depth with respect to time and space.
The flow of water in an open channel is a familiar sight, whether in a natural
channel like that of a river, or an artificial channel like that of an irrigation ditch.
Its movement is a difficult problem when everything is considered, especially
with the variability of natural channels, but in many cases the major features
can be expressed in terms of only a few variables, whose behavior can be
described adequately by a simple theory. The principal forces at work are
those of inertia, gravity and viscosity, each of which plays an important role.
Flow in an open channel is said to be "steady" if the depth of flow does not
change or if it can be assumed to be constant during the time interval under
consideration. The flow is "unsteady" if the depth changes with time.
1
Open channel flow is said to be "uniform" if the depth and velocity of flow is
the same at every section of the channel. Hence it follows that uniform flow
can only occur in prismatic channels. A uniform flow may theoretically be
steady or unsteady, depending on whether or not the depth changes with
time. The establishment of unsteady uniform flow requires that the water
surface fluctuate with time while remaining parallel to the channel bottom.
Since it is impossible for this condition to occur within a channel, steady
uniform flow is the fundamental type of flow treated in open channel
hydraulics.
Flow is "varied" if the depth of flow changes along the length of the channel.
Varied flow may be either steady or unsteady. Since unsteady uniform flow is
rare, the term "unsteady flow" is used to designate unsteady varied flow
exclusively.
Varied flow may be further classified as either "rapidly" or "gradually"
varied. The flow is rapidly varied if the depth changes abruptly over a
comparatively short distance; otherwise, it is gradually varied. Rapidly varied
flow is also known as a local phenomenon; an example of which is the
hydraulic jump.
With these varying conditions, open channel hydraulics can be very complex,
encompassing many different flow conditions from steady uniform flow to
unsteady rapidly varied flow. Most of the problems in storm water drainage
involve uniform, gradually varied or rapidly varied flow situations. In this
project however computational procedures for uniform, gradually varied are
presented.
1.2.1 UNIFORM FLOW
2
For a given channel condition of roughness, discharge and slope, there is
only one (1) possible depth for maintaining a uniform flow. This depth is
referred to as normal depth.
The Manning's Equation is used to determine the normal depth for a given
discharge.
And is given by the expression;
Q = A R
2/3
S
1/2
/n,
Where,
Q = Total discharge, m
3
/s
n = Roughness coefficient
A = Cross-sectional area of channel, m
2
R = Hydraulic radius of channel, m (R=A/P)
S = Slope of the frictional gradient, m/m
P = Wetted perimeter, m
1.2.2 GRADUALLY VARIED FLOW
The most common example of gradually varied flow in urban drainage
systems occurs in the backwater of bridge openings, culverts, storm sewer
inlets and channel constrictions.
Under these conditions, gradually varied flow will be created and the flow
depth will be greater than normal depth in the channel (sub critical flow).
1.2.3 RAPIDLY VARIED FLOW
Rapidly varied flow is characterized by abrupt changes in the water surface
elevation for a constant flow. The change in elevation may become so abrupt
that the flow profile is virtually broken (discontinuity), resulting in a state of
high turbulence. Some common causes of rapidly varied flow in urban
drainage systems are side-spill weirs, weirs and spillways of detention basins.
1.2.4 MANNINGS ROUGHNESS COEFFICIENT
3
Because several primary factors affect the roughness coefficient, a procedure
has been developed to estimate this value, n. By this procedure, the value of
n may be computed by:
n = (n0+ n1+ n2+ n3+ n4) m
The grain roughness component, n0 has a lower limit, which accounts for the
smooth boundary condition. Its value is
n0 = 0.015 dm
1/6

Where;
dm = d50 which is the particle-diameter (mm) at which 50% (by mass) of the
material is larger than that particle-diameter.
The d50 value can be used provided that d50 > 0.05 mm. If d50 < 0.05 mm, a
minimum value of 0.05 mm should be used.
n0 is a basic n value for a straight, uniform, smooth channel in the natural
materials involved, n1 is a value added to n0 to correct for the effect of surface
irregularities; n2 is a value for variations in shape and size of the channel
cross section; n3 is a value for obstructions; n4 is a value for vegetation and
flow conditions; and m is a correction factor for meandering of the channel.
The Mannings roughness coefficient is used in conjunction with the
Mannings equation which the most widely used equation for calculating
uniform flow in open channels. It was published in 1889, and later modified to
read (in metric units),
n S AR Q / ) (
2 / 1 3 / 2

Where,
Q
= Total discharge, m
3
/s
n
= Roughness coefficient
A = Cross-sectional area of channel, m
2
R = Hydraulic radius of channel, m (
P
A
R )
S = Slope of the frictional gradient, m/m
P = Wetted perimeter, m
4
The U.S. Bureau of Reclamation (1957) published a good description of
channels, with their suitable n value, based on the work of Scobey. As shown
below, this description gives good information if n remains below about 0.030
n = 0.012.
For surfaced, untreated lumber flumes in excellent condition; for short,
straight, smooth flumes of unpainted metal; for hand-poured concrete of the
highest grade of workmanship with surfaces as smooth as a troweled
sidewalk with masked expansion joints; practically no moss, larvae, or gravel
ravelings; alignment straight, tangents connected with long radius curves; field
conditions seldom make this value applicable.
n = 0.013
Minimum conservative value of n for the design of long flumes of all materials
of quality described under n = 0.012; provides for mild curvature or some
sand; treated wood stave flumes; covered flumes built of surfaced lumber,
with battens included in hydraulic computations and of high-class
workmanship; metal flumes painted and with dead smooth interiors; concrete
flumes with oiled forms, fins rubbed down with troweled bottom; shot concrete
if steel troweled; conduits to be this class should probably attain n = 0.012
initially.
n = 0.014
Excellent value for conservatively designed structures of wood, painted metal,
or concrete under usual conditions; cares for alignment about equal in curve
and tangent length; conforms to surfaces as left by smooth-jointed forms or
well-broomed shot concrete; will care for slight algae growth or slight deposits
of silt or slight deterioration.
n = 0.015
5
Rough, plank flumes of unsurfaced lumber with curves made by short length,
angular shifts; for metal flumes with shallow compression member projecting
into section but otherwise of class n = 0.013; for construction with first-class
sides but roughly troweled bottom or for class n = 0.014 construction with
noticeable silt or gravel deposits; value suitable for use with muddy gravel
deposits; value suitable for use with muddy water for either poured or shot
concrete; smooth concrete that is seasonally roughened by larvae or algae
growths take value of n = 0.01 5 or higher; lowest value for highest class
rubble and concrete combination.
n = 0.016
For lining made with rough board forms conveying clear water with small
amount of debris; class n = 0.014 linings with reasonably heavy algae; or
maximum larvae growth; or large amounts of cobble detritus; or old linings
repaired with thin coat of cement mortar; or heavy lime encrustations; earth
channels in best possible conditions, with slick deposit of silt, free of moss and
nearly straight alignment; true to grade and section; not to be used for design
of earth channels.
n = 0.017
For clear water on first-class bottom and excellent rubble sides or smooth
rock bottom and wooden plank sides; roughly coated, poured lining with
uneven expansion joints; basic value for shot concrete against smoothly
trimmed earth base; such a surface is distinctly rough and will scratch hand;
undulations of the order of 0.025 m.
n = 0.018
About the upper limit for concrete construction in any workable condition; very
rough concrete with sharp curves and deposits of gravel and moss; minimum
design value for uniform rubble; or concrete sides and natural channel bed;
for volcanic ash soils with no vegetation; minimum value for large high-class
canals in very fine silt.
6
n = 0.020
For tuberculated iron; ruined masonry; well-constructed canals in firm earth or
fine packed gravel where velocities are such that the silt may fill the interstices
in the gravel; alignment straight, banks clean; large canals of classes n =
0.0225.
n = 0.0225
For corrugated pipe with hydraulic functions computed from minimum internal
diameter; average; well-constructed canal in material which will eventually
have a medium smooth bottom with graded gravel, grass on the edges, and
average alignment with silt deposits at both sides of the bed and a few
scattered stones in the middle; hardpan in good condition; clay and lava-ash
soil. For the largest of canals of this type a value of n = 0.020 will be originally
applicable.
n = 0.025
For canals where moss, dense grass near edges, or scattered cobbles are
noticeable. Earth channels with neglected maintenance have this value and
up; a good value for small head ditches serving a couple of farms; for canals
wholly in-cut and thus subject to rolling debris; minimum value for rock-cut
smoothed up with shot concrete.
n = 0.0275
Cobble-bottom canals, typically occurring near mouths of canyons; value only
applicable where cobbles are graded and well packed; can reach 0.040 for
large boulders and heavy sand.
n = 0.030
Canals with heavy growth of moss, banks irregular and overhanging with
dense rootlets; bottom covered with large fragments of rock or bed badly
pitted by erosion.
7
n = 0.035
For medium large canals about 50 percent choked with moss growth and in
bad order and regimen; small channels with considerable variation in wetted
cross section an biennial maintenance; for flood channels not continuously
maintained; for untouched rock cuts and tunnels based on paper cross
section.
n = 0.040
For canals badly choked with moss, or heavy growth; large canals in which
large cobbles and boulders collect, approaching a stream bed in character.
n = 0.050 - 0.060
Floodways poorly maintained; canals two-thirds choked with vegetation.
n = 0.060-0.240
Floodways without channels through timber and underbrush, hydraulic
gradient 0.20 to 0.40 m per 1000 m.
Artificial channels are usually taken to have uniform sectional roughness,
while the roughness of natural channels is usually represented by a
composite sectional roughness.
1.3 STEADY FLOW IN OPEN CHANNELS
All computations will be based on the conservation of mass, momentum and
energy (in the form of Bernoulli's theorem), and the Mannings formula for
frictional resistance.
The assumptions, methods and formulas that are present here work for
problems of water flows on the usual engineering scale. The flows are long
compared to their cross-sections, so that a single velocity can describe the
situation adequately at any cross-section.
8
1.4 ARBITRARY SECTIONS
The sectional geometry of an artificial channel is in most cases regular, while
that of natural channels varies under most circumstances; these factors
contribute greatly to the characteristics of steady flow in both types of
channels.
Given S and
Q
, and the shape of the channel, depth y and velocity V can be
found; therefore given S and y, we can find
Q
. If the given channel is one of
arbitrary shape,
Q
(total discharge) must be used instead of
q

(discharge/unit width), and y is usually the depth from the lowest point in the
channel. The area A occupied by fluid and the wetted perimeter P are
functions of y, and may be expressed analytically for channels of any
geometric shape. The value of y corresponding to a given
Q
and S is
called the normal depth for those conditions.
1.5 IMPORTANCE OF SURFACE PROFILE COMPUTATIONS
This program can be put into diverse use, typical instances where it can be
applied include;
i. Studies involving dam break failures
ii. Flood alleviation schemes
iii Elevations for;
a. pumping plants
b. canal head works
c. Energy dissipaters
iv. Sediment transport estimation
v. River rehabilitation.
vi. Tail water ratings for hydroelectric power plants
vii. Grade lines for
a. Highways
b. Bridges and culverts
1.6 DEFINITION OF TERMS
9
In order to understand the meaning of some important terms associated with
steady flow in open channels, as applied in this project, a brief discussion of
these basic terms is in order
1.6.1 FROUDE NUMBER Fr
It is defined as:
Fr= V/(gDm)
1/2
The knowledge of the Froude number of an open channel flow gives the idea
of the conditions of flow.
When Fr>1, the resulting flow condition is as follows:
Supercritical flow
Water velocity > wave velocity
Disturbances travel downstream
Upstream water levels are unaffected by downstream control
When Fr = 1
critical flow results
When Fr< 1, the resulting flow condition is as follows:
Sub-critical flow
Water velocity < wave velocity
Disturbances dont travel downstream and upstream water levels are
affected by downstream control
1.6.2 CONTROL SECTION
The control of flow is the establishment of a definitive relationship between the
stage (depth) and the discharge of the flow. When the control of flow is
achieved at a section of the channel, this section said to be a control section.
The control section controls the flow in such a way that it restricts the
transmission of the effects of changes in flow condition either in the upstream
10
direction or in a downstream direction, depending on the state of flow in the
channel.
1.6.3 PRISMATIC CHANNEL AND NON- PRISMATIC CHANNELS
Prismatic channels are channels that do not change in section longitudinally.
The computation of surface profiles in such channel is a function of the stage
(flow depth).
Fabricated channels are mostly prismatic. The non-prismatic channels on the
other hand, are such channels that vary in section longitudinally. The
computation of backwater in the non prismatic channels is a function of its
sectional dimensions. Natural courses are categorized as non prismatic
channels.
1.7 OBJECTIVE OF STUDY
This project is aimed at producing software that can compute the depths of
flow at selected sections in an open channel and thus display the surface
profile of the channel of concern. This would be carried out with particular
emphasis on arbitrary sectioned and non prismatic open channels.
The program will be written with the aid of MATLAB, my reason for this
selection being that;
i, it is a proficient programming environment, this allows for less voluminous
programs and thus enables faster development of the required program.
ii, it is an excellent means for production of programs where the display of
graphics is required.
11
CHAPTER 2
2.1 LITERATURE REVIEW
This chapter contains an overview of background information and scientific
literature relevant to this research.
2.2 BACKGROUND OF BACKWATER ANALYSIS
This aspect of hydraulics has been researched upon by a lot of scientists in
the past, this is evidently due to the fact that it was discovered that a thorough
insight of the backwater phenomenon enabled a good prediction to be made
of the flow pattern in most open channels. The fact that a good knowledge of
the backwater equations serves as a great tool in the analysis of gradually
varied steady flow in open channels cannot therefore be overemphasized.
Many hydraulicians (mostly in the 18
th
century) in the likes of Bresse,
Ruhlmann, Tolkmitts, Schaffernak, Francis, Saint-venant, and so many others
have contributed immensely to the development of better techniques in the
analysis of gradually varied flow in open channels.
The theories developed follow the basic assumption that, the head loss at
any particular section is identical as for a uniform flow having identical
flow velocity and hydraulic radius as the section. The computation of
backwater curves based on this hypothesis has been proven satisfactory by
many experiments. The experimental verification indicates the validity of the
12
assumption for practical purposes and has been proven to be a reliable basis
for design.
The basic assumption has been extensively applied in the dynamic equation
of gradually varied flow. However, the computation of gradually varied flow
profiles involves basically the solution of the dynamic equation of gradually
varied flow. This equation here is developed and its methods of application
are also discussed. The solution methods of the dynamic equation include;
(i) Graphical integration method
(ii) Direct integration method
(iii) Numerical methods
2.3 ASSUMPTIONS FOR GRADUALLY VARIED FLOW
Important assumptions made in the analysis of gradually varying flow include:
The flow is steady
Streamlines are virtually parallel ensuring hydrostatic pressures
Friction losses at a section for gradually varied flow is the same as those
for uniform flow
Channel bed slope is small
There is no air entrainment
Depth of flow is equivalent to depth of section
The channel is prismatic with constant alignment and slope
Velocity distribution coefficient equals one (1)
The Mannings roughness coefficient is independent of the depth or length
of the channel considered.
2.4 THE EQUATIONS OF GRADUALLY VARIED FLOW
The basic assumption in the derivation of this equation is that the change in
energy with distance is equal to the friction losses.
f
S
dx
dH

13
The Bernoulli equation is:
H z
g
V
y + +
2
2

Differentiating and equating to the friction

f
S
dx
dz
g
V
y
dx
d

,
_

+
2
2

or
f o
S S
dx
dE
2.1
Where
o
S
represents the bed slope and f
S
represents the friction slope of
the channel under consideration
The equation below represents the change in specific energy with depth
2
3
2
1 1 Fr
gA
B Q
dy
dE

2.2
Where B represents the surface width of the channel under consideration;
The combinations of equations 2.1 and 2.2 yields
2
1 Fr
S S
dx
dy
f o

2.3
This is the basic equation of gradually varied flow. It describes how the depth,
y
, changes with distance
x
, in terms of the bed slope,
o
S
friction slope f
S
and the discharge,
Q
, and the channels shape (encompassed in Fr and
f
S
)
2.5 METHODS OF ANALYSIS
As already stated, there are three major ways in which the backwater
equation i.e. equation 2.3 can be analyzed
These three methods will be discussed in more details,
They include;
14
i) Graphical integration method
ii) Direct integration method
iii) Numerical methods
2.5.1 THE GRAPHICAL INTEGRATION METHOD
This method integrates the back water equation graphically. The method is
applicable to both prismatic and non prismatic channels. Considering two
channel sections at distances
1
L and
2
L respectively, from a chosen origin
and the corresponding depths of flow y1 and y2, the distance along the
channel floor can be given by

2
1
2 1
y
y
dy
dy
dx
L L L
2.4
15
16
practically impossible. Many attempts have been made to solve the equation
for a few special cases or to introduce assumptions that make amenable to
8mathematical integration. Most early methods used Chezys formula, but
have been dropped in favor of the Mannings formula in more recent times.
Among the early contributors of Bakhmeteff (1932), who carried out
integration by short by short range steps with the varied flow function. He
considered the critical slope in the small range of the varying depth in each
reach to be constant. This is true only for prismatic channels.
Monodobe (1938) made an effort to improve Bakhmeteffs method. He took
the effects of velocity change and friction head into account integrally without
the necessity of dividing the channel into successive reaches. His method
affords a more direct and accurate computation procedure where by results
can be obtained without recourse to successive steps. A demerit of this
method lies in the difficulty of using accompanying charts, which are not
sufficiently accurate for practical purpose.
Later, Lee (1947) and Von Seggern (1950) suggested new assumptions,
which resulted in more satisfactory solutions. Von introduced a new varied
flow function in addition to that of Bakhmettef, hence, an additional table for
the new function is necessary in the method unlike that of Lee.
Combining preceding considerations for a direct solution to the backwater
equation, Vent Te Chow (1955) came up with a procedure for the direct
solution of the equation.
N
n n
y C K
1
2
2.5
N
y C K
1
2

2.6
M
c c
y C Z
2
2
2.7
M
y C Z
2
2

2.8
Where;
17
N and M are hydraulic exponents for critical flow and uniform flow;
1
C and
2
C are coefficients.
T A Z /
3
2.9
Z represents the section factor for flow computation for discharge
Q
at
depth
y
2
2
2
Z
Z
Fr
c

2.10
2
2
K
Q
S
f
2.11
2
2
n
o
K
Q
S
2.12
2
2
K
K S
S
n o
f

2.13
K represents the conveyance for flow
Q
at depth
y
If these expressions are substituted into equation 2.3, the gradually varied
flow equation becomes
( )
( )
M
c
N
n
o
y y
y y
S
dx
dy
/ 1
/ 1

2.14
Equation 2.14 can be manipulated to yield
const J v F
N
J
y
y
N u F u A x
M
n
c
+
1
1
]
1

,
_

+ ) , ( ) , (
2.15
Where
o
n
S
y
A
2.16
N
J
y
y
B
M
n
c

,
_

2.17
n
y
y
u
2.19
J N
u v
/
2.20
1
1
+

M N
J 2.21
18

u
N
u
du
N u F
0
1
) , (
2.22

v
J
v
dv
J u F
0
1
) , (
2.23
Where
) , ( N u F
and
) , ( J v F
are varied flow functions.
According of him the length of flow profile between two consecutive sections 1
and 2 is given as
)]} , ( ) , ( [ )] , ( ) , ( [ ) {(
1 2 1 2 1 2 2 1
J v F J v F B N u F N u F u u A x x L +
2.24
The solution of the equation 2.24 can be simplified by the use of the varied
flow function table. The table gives us the value of
) , ( N u F
and
) , ( J v F
for
given values of N .
2.5.3 NUMERICAL METHODS
As earlier stated a direct solution of differential equation 2.3 for all sections is
practically impossible. The solution method developed by Chow is difficult to
model for computer programming. At best direct integration is only applicable
to prismatic channels. However, numerical techniques are available to proffer
an approximate solution of acceptable accuracy. By utilizing simple numerical
methods all types of gradually varied flow may be quickly and easily solved
using a microcomputer.
When equation 2.3 is applied to prismatic channels, the flow gradient is a
function of the flow depth y, since the section is uniform through out the length
of the channel; the equation is therefore in the form;
) ( y f
dx
dy
2.25
When the equation is applied to a non prismatic channel, the flow gradient
becomes a function of x and y, since the channel width and flow depth varies
along the length of the channel. The equation is in the form;
19
) , ( y x f
dx
dy
2.26
In this research equation 2.26 is of special concern.
There are two basic numerical methods that can be used
i. Direct step method (distance from depth)
ii. Standard step method (depth from distance)
2.5.3.1 THE DIRECT STEP METHOD (DISTANCE FROM DEPTH)
This method will calculate a distance for a given change in the surface height.
It involves the solving the reciprocal of equation 2.3
2.5.3.2 THE STANDARD STEP METHOD (DEPTH FROM DISTANCE)
The method will calculate a depth for a given distance up or down stream by
solving equation 2.3
2.6 CLASSIFICATION OF PROFILES
Before attempting to solve the backwater equation, a great deal of insight is
required of the profiles possible;
For a given discharge, f
S
and
2
Fr
are functions of depth
( )
3 / 10 3 / 4 2 2
/ A P Q n S
f

2.27
( )
3 2 2
/ gA B Q Fr 2.28
A quick examination of these two expressions reveal that they both decrease
with increase in A i.e. increase in y
Also for uniform flow,
f
S
=
o
S
and
y
=
n
y
So
f
S
>
o
S
when
y
<
n
y
f
S
<
o
S
when
y
>
n
y
And
20
2
Fr
>1 when
y
<
c
y
2
Fr
<1 when
y
>
c
y
From these inequalities it can be seen how the sign of
dx dy /
i.e. the surface
slope changes for different slopes and Froude numbers.
Taking an example of a mild slope, treating the flow as to be in three zones
i. Zone 1, above the normal depth
ii. Zone 2, between the normal and critical depth
iii. Zone 3, below critical depth
The directions of the surface inclination may thus be determined
Zone 1
y
>
n
y
>
c
y
f
S
<
o
S
2
Fr
< 1
dx dy /
is positive, surface rising
Zone 2
n
y
>
y
>
c
y
f
S
>
o
S
2
Fr
< 1
dx dy /
is negative, surface falling
Zone 3
n
y
>
c
y
>
y
f
S
>
o
S
2
Fr
> 1
dx dy /
is positive, surface rising
The condition at the boundary of the gradually varied flow may also be
determined in a similar manner
Zone 1
As
y
then f
S
and
2
Fr

0 and
dx dy /
o
S
Hence water surface is asymptotic to a horizontal line for its maximum
As
y

n
y
then f
S
o
S
0 and
dx dy /
0
Hence the water surface is asymptotic to the line
y
=
n
y
i.e. uniform flow
Zone 2
As for zone 1 as y approaches the normal depth:
21
As
y

n
y
then f
S

o
S

0 and
dx dy /
0
Hence the water surface is asymptotic to the line
y
=
n
y
But a problem occurs when y approaches the critical depth:
As
y

c
y
then
2
Fr

1 and
dx dy /
This is practically impossible but may be explained by pointing out that in this
region the gradually varied flow equation is not applicable because at this
point the fluid is in rapidly flowing regime.
Zone 3
As for zone 2 a problem occurs when y approaches the critical depth:
As
y

c
y
then
2
Fr

1 and
dx dy /
Again the same physical impossibility holds with the same explanation and
again in reality a very steep surface slope will occur.
As
y

0 then
dx dy /
o
S
the slope of the bed of the channel
The gradually varied flow does not hold here because it is clear what occurs
In general, normal depth is approached asymptotically and critical depth at
right angles to the channel bed.
The possible surface profiles within each zone can be drawn from the above
considerations.
All possible surface profiles for all the possible slopes of the channel are
shown in the figure below.
22
Figure 1
23
Figure 2
2.7 VELOCITY DISTRIBUTION IN OPEN CHANNELS
The measured velocity in the open channel will always vary across the
channel cross section because of friction along the boundary. Neither is the
24
velocity axisymmetric (as in pipe flow) due to the existence of the free surface.
It might be expected to find the maximum velocity at the free surface where
the shear force is equal to zero, but this is not the case. The maximum
velocity is usually found just below the surface. The explanation for this is the
presence of secondary currents which are circulating from the boundaries
towards the section center and the resistance of air/water interface. These
have both been found in both laboratory measurements and three
dimensional (3D) numerical simulations of turbulence.
The figure below shows typical velocity distribution across a rectangular
channel cross section
.
Figure 3
25
CHAPTER 3
SOLUTION TECHNIQUES
3.1 FOREWORD
This chapter deals with the numerical methods employed for the computer
analysis of surface profile of gradually varied flow. There are varieties of
numerical methods for solving the differential equations, which governs
backwater flow. Each of these methods must be critically studied in order to
arrive at the selection of the most suitable solution method.
In addition, the numerical computation of critical and normal depths is also
discussed. The determination of critical and normal depths serves as the
mean of knowing when flow approaches the critical and normal depths. This
knowledge helps in recognizing the condition at which the gradually varied
flow equation becomes unstable as earlier pointed out in chapter two, so that
necessary steps may be taken to forestall inaccurate computation.
Computer analysis of backwater involves a dynamic exchange of data
resulting from different computations, such that if these data are not properly
handled it may lead to accumulation of errors, which invariably will affect the
final output. It is therefore important at this juncture to highlight various errors
associated with the numerical solution of backwater equation.
3.2 INACCURACIES IN NUMERICAL ANALYSIS OF STEADY FLOW IN
OPEN-CHANNEL
Final results of numerical computations of unknown quantities are mostly by
approximations. They are exact but involve errors. Sources of errors include
machine round-off, experimental, truncating (local or global) and data-entry
(human error). Round-off errors result from the discarding of all decimals from
some decimal on, while the experimental errors are inherent in given data.
Truncating errors result from premature of breaking off. This is common when
double-precision data are treated as single-precision data. Truncating error is
26
local when it is related to the values of successive steps and global when it is
related to the true value. It is also possible to make mistake by supplying
wrong data to the computer from the input device. This is human error, and it
is almost inevitable in time shearing program.
Furthermore, methods of computation adopted for the analysis can as well
contribute to the error, which exists in the final results. Method associated
errors must therefore be taken into consideration in the selection of numerical
method of analysis and algorithm. The following are some error terms
associated with numerical solutions:
ABSOLUTE ERROR, |E0|: This refers to the absolute value of errors that
exists either between consecutive terms or between a term and the exact
value. The knowledge of the later helps in determining a criterion for the
stopping of iteration, especially when a method, which converges very slowly,
is used.
ERROR OF THE APPROXIMATION, EO: This is the value of the difference
between the exact value and the approximate value. This is mostly as a result
of the approximate nature of the physical law or principal upon which the
solution is based.
RELATIVE ERROR, RO: It is the error derived from dividing the error of the
approximation by the actual value. Relative error gives the percentage error,
and can be adapted to determine the number of significant digits in the
approximation.
ERROR DUE TO LOSS OF SIGNIFICANCE: There can be a loss of
significant digits due to round-off error. In this case the function value p(xk)
and the correction term p(xk) / p(xk) in Newtons method can fail to have
enough accuracy to make xk+1 closer to the root than xk.
27
ERROR DUE TO INSTABILITY: A problem is said to be unstable when small
changes in a parameter results in large changes in the computed root.
This instability relates to the particular way in which the calculation is made. In
essence it is dependent on the stability of the numerical method and algorithm
adopted.
In the case of algorithm stability an algorithm is admitted to be stable when it
terminates after a number of sequences of operation, otherwise it is unstable.
The general idea is that the accuracy of the answer increase with the number
of steps performed.
In the case of the stability of numerical methods, it is important to note that
there exist many stable problems that have unstable numerical solutions. This
arises from a solution involving auxiliary intermediate number calculated for
distance or depth, as the case may serve, which become input numbers for
subsequent steps in the calculation.
3.3 NUMERICAL SOLUTION TECHNIQUES FOR FIRST-ORDER
DIFFERENTIAL EQUATION
The equation that governs the analysis of gradually varying flow in non-
prismatic open channels is a first-order differential equation, as shown before.
It is of the form:
) , ( y x f
dx
dy

Owing to the fact that the channel properties (geometry and roughness) and
flow conditions are expected to vary along the channels, a direct model
solution for this equation is difficult. However, there abound a number of
numerical methods, which evolve fairly accurate solutions. These techniques
include:
I. Taylors Method,
28
II. Eulers Method,
III. Heuns Method
IV. Runge-Kutta Methods, and
V. Predictor-corrector Methods.
One of these methods is required to develop a model computer program for
the solution of the differential equation. Since each of these methods has its
peculiar weaknesses and merits as regards the present area of application,
there is need for a critical analysis of these methods. This analysis must be
based on the following key considerations:
i. Computational efforts required for each method
ii. The desired precision of computation
iii Program run-time, which also bears adequate check on speed of convergence
iv. Applicability of the method of non-prismatic channels of varying
characteristics
v. Stability of solution method
vi. Compatibility of the method and its algorithm with other parts of the same
program
3.4 ANALYSIS OF SOLUTION TECHNIQUES
The purpose of this analysis is to reveal the merits and the demerits of each
numerical method listed above as regards arriving at a satisfactory solution of
the gradually varied flow differential equation. This will result in the proper
selection of the solution method to apply.
3.4.1 TAYLORS METHOD
The method of Taylor is of general applicability and it gives a way to compare
the accuracy of various numerical methods for solving Initial Value Problem
(I.V.P.). It is of the form:
y(x+h) = y(x) + y (x)h + y(x)h
2
/ 2! + y(x)h
3
/ 3! +
For numerical purposes a finite series of order N is fixed to approximate
y(x+h). This gives:
29
y(x+h) = y(x) + y (x)h + y(x)h
2
/ 2! +y
(N)
(x)h
N
/N!
The method can be constructed to have a high degree of accuracy,
depending on the value of N. if the order N is fixed; it is possible to determine
a the step size h, so that the global truncating error will be as small as
desired. However, an N-order solution requires an N-degree of differential
equation. In the present case it is difficult to have a general model solution
when N>1. Therefore, for this method to be of any use here, it is required that
only the first two terms where N<2 be considered. This forms the basis for the
Eulers Method.
3.4.2 EULERS METHOD
The general step for the Eulers Methods is:
xk+1 = xk

+ h yk+1 = yk + hf(xk, yk)
For K=0,1,, M-1
The method has limited usage because of the larger error that is accumulated
as the process proceeds. To have an idea of how much computing effort
required to obtain an accurate approximation over an interval, the relationship
between the global truncating error and step size is required. The method
requires a longer time for convergence for a given interval.
Apart from the foregoing, it is the simplest among the methods under
consideration because it requires the least programming effort.
3.4.3 HEUNS METHOD
This method is otherwise known as the improved Euler Cauchy method. It is
sometimes referred to as second-order Runge-Kutta Method. The general
step for this method is as follows:
yk+1 = yk + h[f(xk,yk)+ f((xk,+h), (yk +hf(xk,yk))]/2
Heuns method is also regarded as a predictor-corrector method because the
Eulers method is used as its predictor, and then the trapezoidal rule is used
to make correction to obtain a more refined value. It is however a single-step
30
method like the methods of Euler, Taylor and Runge-Kutta because they only
use information from one previous point to compute the successive point. It is
therefore not as prone to stability problem as other predictor-corrector
methods. It has an edge over the Eulers method in terms of convergence
speed and accuracy.
3.4.4 RUNGE-KUTTA METHOD
The classical Runge-Kutta method of fourth order (N=4) provides the
accuracy desired for the computation of the gradually varied flow surface
profile. The method is quite accurate, stable and easy to program. Most
authorities proclaim that is not necessary to go a higher-order method
because the increased accuracy is offset by additional computational effort. If
more accuracy is required, a smaller step size can be used. Starting with (xo,
yo), four function evaluations per step are required to the discrete
approximations (xk, yk) as follows:
yk+1= yk + h(f1+2f2+2f3+f4)/6
Where:
f1 = f(xk,yk)
f2 = f(xk+h/2,yk+hf1/2)
f3 = f(xk+h/2,yk+hf2/2)
f4 = f(xk+h,yk+hf3)
The formula above computes flow depths from known channel distances.
When distance is to be determined for known depth, the equation becomes
xk+1= xk + h(f1+2f2+2f3+f4)/6
where:
f1 = f(yk,xk)
f2 = f(yk+h/2,xk+hf1/2)
f3 = f(yk+h/2,xk+hf2/2)
f4 = f(yk+h,xk+hf3)
One way to guarantee accuracy in the solution of I.V.P., with respect to
Runge-Kutta methods, is to solve the problem twice using step sizes h and
h/2 and compare answers at the mesh point corresponding to the larger size.
31
But this requires a significant amount of computation for the smaller step size
and must be repeated if it is determined that the agreement is not good
enough.
3.5 PREDICTOR-CORRECTOR METHODS
These include Adams- Bashforth-Moulton Method, Milne-Simpson Method
and the Hammings Method apart from the Heuns method. The major
problem these methods have is that they are not self-starting. They are mostly
used in pari-pasu with one-step methods, especially the Runge-Kutta
Methods, which are employed to obtain three initial approximate values
needed to start off. The predictor-corrector methods furnish better precision
than Runge-Kutta method because of refinement. However the methods
require enormous programming technique.
It is important to mention at this point that the list of numerical methods
discussed above does not exhaust all available solution methods, but those
ones considered were carefully selected. For instance, the Least-Square
Method and the Galerkin Method can as well be adapted. However, they are
considered unsuitable for this exercise because they involve elaborate
processes, which require unnecessarily intricate and enormous programming
effort.
3.6 SELECTION OF SOLUTION TECHNIQUE
The computation of gradually varied flow is a major part of the program to be
developed since the depth of flow at a number of points has to be determined.
Therefore, run-time plays a very important role in determining the selection of
the method. The method to be selected is expected to be simple while it
retains the properties of being stable and fast convergence to solutions.
For the purpose of this project the fourth order Runge Kutta method is
sufficiently accurate for the computation of profiles for arbitrary (both prismatic
and non-prismatic) sections. In the computation of profiles in prismatic
32
channels the Euler, Heuns and Runge-Kutta methods are used to enable
comparison of computation time time. In the computation non prismatic
channels Euler method is used to predict the flow depth in the next section,
this is then iterated till a prescribed error margin is reached. The cycle is then
repeated for the next section, all the methods used in the computation of
prismatic channels are also used.


3.7 ACCURACY OF SOLUTION AND CHOICE OF STEP INTERVAL
The choice of step interval affects the numerical solution of the differential
equation. The smaller the step interval, the greater the accuracy (this only
holds if computation has not yet converged; because upon convergence
further reduction in step size yields no further increase in the accuracy of the
solution). The step interval between cross sections depends on the uniformity
of the channel, desired accuracy and the speed of convergence of the
numerical method employed.
However, when distances are given and depths are to be computed the
reciprocal form of equation 2.3 will be used. The accuracy of the solution is as
well impaired as critical depth is approached. The distance at which the flow is
at critical depth is normally avoided by writing a check code that limits profile
computation as critical depth is approached to avoid overflow, so that
computation is done for depths below and above the critical depth. It is also
possible to fix the distance and use Equation 2.3.
3.8 COMPUTATION OF NORMAL AND CRITICAL DEPTH (PRISMATIC
CHANNELS)
The solve function in MATLAB is applied to the appropriate equations to
derive these values, this would be explained better in the next chapter.
33
CHAPTER 4
PROGRAMME DEVELOPMENT AND COMPUTATIONAL LOGIC
4.1 FOREWORD
This chapter expatiates on the programming language (MATLAB) employed
in the assembly of the project software; it further elucidates the computational
logic utilized in this project. Moreover, the variables required for data feeding
and the output for which they are required are stated in a tabular form. The
algorithms adopted for the software development are also outlined.
4.2 PROGRAMMING LANGAUGE
As stated earlier on in this write-up MATLAB was selected after a thorough
weighing of options on two main grounds:
i, it is a proficient programming environment, this allows for less voluminous
programs and thus enables faster development of the required program.
ii, it is an excellent means for production of programs where the display of
graphics is required.
4.2.1 What Is MATLAB?
MATLAB is a high-performance language for technical computing. It
integrates computation, visualization, and programming in an easy-to-use
environment where problems and solutions are expressed in familiar
mathematical notation.
Typical uses include:
Math and computation
34
Algorithm development
Data acquisition
Modelling, simulation, and prototyping
Data analysis, exploration, and visualization
Scientific and engineering graphics
Application development, including graphical user interface building
MATLAB is an interactive system whose basic data element is an array that
does not require dimensioning. This allows you to solve many technical
computing problems, especially those with matrix and vector formulations, in
a fraction of the time it would take to write a program in a scalar non-
interactive language such as C or FORTRAN.
The name MATLAB stands for matrix laboratory. MATLAB was originally
written to provide easy access to matrix software developed by the LINPACK
and EISPACK projects. Today, MATLAB engines incorporate the LAPACK
and BLAS libraries, embedding the state of the art in software for matrix
computation.
MATLAB has evolved over a period of years with input from many users. In
university environments, It is the standard instructional tool for introductory
and advanced courses in mathematics, engineering, and science. In industry,
MATLAB is the tool of choice for high-productivity research, development, and
analysis.
MATLAB features a family of add-on application-specific solutions called
toolboxes. Very important to most users of MATLAB, toolboxes allow you to
learn and apply specialized technology. Toolboxes are comprehensive
collections of MATLAB functions (M-files) that extend the MATLAB
environment to solve particular classes of problems. Areas in which toolboxes
are available include signal processing, control systems, neural networks,
fuzzy logic, wavelets, simulation, and many others.
35
4.2.1.1 The MATLAB System
The MATLAB system consists of five main parts:
4.2.1.2 Development Environment.
This is the set of tools and facilities that help you use MATLAB functions and
files. Many of these tools are graphical user interfaces. It includes the
MATLAB desktop and Command Window, a command history, an editor and
debugger, and browsers for viewing help, the workspace, files, and the search
path.
4.2.1.3 The MATLAB Mathematical Function Library.
This is a vast collection of computational algorithms ranging from elementary
functions, like sum, sine, cosine, and complex arithmetic, to more
sophisticated functions like matrix inverse, matrix eigenvalues, Bessel
functions, and fast Fourier transforms.
4.2.1.4 The MATLAB Language.
This is a high-level matrix/array language with control flow statements,
functions, data structures, input/output, and object-oriented programming
features. It allows both "programming in the small" to rapidly create quick and
dirty throw-away programs, and "programming in the large" to create large
and complex application programs.
4.2.1.5 Graphics.
MATLAB has extensive facilities for displaying vectors and matrices as
graphs, as well as annotating and printing these graphs. It includes high-level
functions for two-dimensional and three-dimensional data visualization, image
processing, animation, and presentation graphics. It also includes low-level
functions that allow you to fully customize the appearance of graphics as well
as to build complete graphical user interfaces on your MATLAB applications.
4.2.1.6 The MATLAB Application Program Interface (API).
36
This is a library that allows you to write C and FORTRAN programs that
interact with MATLAB. It includes facilities for calling routines from MATLAB
(dynamic linking), calling MATLAB as a computational engine, and for reading
and writing MAT-files.
4.3 COMPUTATIONAL LOGIC
This involves the ideology of reasoning applied in the development of the
project software. The logic employed for both arbitrary prismatic and non
prismatic channels are stated separately below.
4.3.1 ARBITRARY PRISMATIC CHANNEL
The irregular (Arbitrary) channel as defined in this project is meant to provide
the flexibility to define various shapes. It is not intended to be used as a
compound channel that might have extensive (horizontal) over bank flows.
Calculations in compound channels (varying roughness and channel bed
slopes) are not implemented in this project. Further, the width of the irregular
channel (of concern) does not decrease upward, the channel width increases
or remains the same as the height increases. The irregular channel is
basically a series of trapezoids stacked one on top of the other. In each of the
trapezoids, the area is the sum of the areas below plus a formula similar to
Equation 4.0 with y replaced by the depth in the trapezoid
an irregular channel idealized as a series of trapezoidal channels stacked one on top of the other
T
s
p
A
y
i
Figure 4.
37

,
_

+ +


mr ml
m
m m i m
s s
y y
y y T A y A
1 1
2
) (
) ( ) (
2
1
1 1
Equation 4.0

,
_

+ + + + +

2 2
1
1
1
1
1 ) ( ) (
mr ml
m i o m
s s
y y p T y P
Equation 4.1

,
_

+ +

mr ml
m m m
s s
y y T y T
1 1
) ( ) (
1 1
Equation 4.2
2
2 2
1
5
2
1
1 1
3
1
1
1
1 ) (
1 1
2
) (
) (

'

,
_

+ + + + +

'

,
_

+ +

,
_

mr ml
m i o
mr ml
m
m m i
o
s s
y y p T
s s
y y
y y T A
S
Qn
Equation 4.3

,
_

+ +

'

,
_

+ +

mr ml
m m
mr ml
m
m m i
s s
y y T
s s
y y
y y T A
g
Q
1 1
) (
1 1
2
) (
) (
1 1
2
2
1
1 1
2
Equation 4.4
In which Tm-1 is the top width of the m1 section (where T0=w is the width at
the base of the channel), ym-1 (where y0=0) is the elevation of the top of the m-
1 section, and sml and smr are the side slopes (vertical/horizontal) on the left
and on the right of the m section respectively. When calculating normal and
critical depths, the program first finds the flow rates for those depths at the top
of each section. It then locates the section for which Qi < Q < Qi+1 and solves
equations 4.2 and 4.3 to find critical and normal depths respectively within the
section.
For the calculation of profiles the program provides options for computing with
all the three stated methods (Euler, Heuns and Runge-Kutta) starting from a
known depth and location upstream of the channel to a known location
downstream of the channel. The irregular channel is defined by the user in a
38
table of X and Y coordinates. If a depth, either calculated or specified,
exceeds the maximum depth of the table, the program assumes that the
channel extends upward with the same side slopes as that of the last
specified section.
4.3.2 ARBITRARY NON PRISMATIC CHANNEL
In computing the flow profile, the following information is required:
1. The discharge for which the profile is desired.
2. The water surface elevation at the control section.
3. The geometric elements at various channel sections along the reach for
all depths of flow within the range expected. These data may be obtained
from hydrographic survey or contour map of the channel bottom.
4. The reach length and channel bed slope between sections (for the
purpose of simplicity equal reach lengths and bed slope are adopted in
this project)
It is assumed that channel cross sections vary gradually and linearly. Eddy
losses that might develop are catered for by increasing the Mannings
roughness coefficient n appropriately.
The program for the computation of depths in non prismatic sections also
provides options for calculating computing with Euler, Heuns and Runge-
Kutta methods. It can handle 101 sections (100 steps).
PREDICTOR CORRECTOR METHOD
A code for depth computation in non-prismatic channels using a predictor
corrector method is embedded in the program for prismatic channels. The
Euler method is used to predict the flow depth in the subsequent section this
is then iterated until an allowable error limit is attained. The same process is
then repeated for subsequent sections. The program can handle a maximum
of three sections at a time. The results from the preceding sections can be
39
recorded then the cumulative results plotted with the aid of MATLAB to
produce a complete profile plot.
4.4 DATA FEEDING
The input variables adopted for prismatic and non prismatic channels, their
descriptions and the output for which they are required are presented in the
tables below:
4.4.1 ARBITRARY PRISMATIC SECTIONS
S/No Input Variable Description Output for which variable is required
1 Q Flow rate Critical and normal depths and profile
2 n Mannings Roughness Critical and normal depths and profile
3 S Channel Slope Critical and normal depths and profile
4 X,Y Channel Coordinates Critical and normal depths, profile and
channel cross section plot.
5 y1 Upstream Control depth Profile plot
6 x1 Upstream Control depth location Profile plot
7 xn Downstream end location Profile plot, runtime
8 N Number of steps Profile plot, runtime
4.4.2 ARBITRARY NON-PRISMATIC SECTIONS
S/No Input Variable Description Output for which variable is required
1 X,Y Channel Coordinates Profile plot and channel cross section plot.
2 Q Flow rate Profile plot
3 n Mannings Roughness Profile plot
4 S Channel Slope Profile plot
5 h Step size (distance between
successive cross sections)
Profile plot, runtime
4.5 ALGORITHM DEVELOPMENT (flow charts)
For the purpose of this project, an algorithm can be defined as a finite list of
well-defined instructions for accomplishing some task that, given an initial
state, will proceed through a well-defined series of successive states,
eventually terminating in an end-state.
The concept of an algorithm originated as a means of recording procedures
for solving mathematical problems such as finding the common divisor of two
numbers or multiplying two numbers. A partial formalization of the concept
40
began with attempts to solve the Entscheidungs problem (the "decision
problem") that David Hilbert posed in 1928.
Algorithms are essential to the way computers process information, because
a computer program is essentially an algorithm that tells the computer what
specific steps to perform (in what specific order) in order to carry out a
specified task, thus, an algorithm can be considered to be any sequence of
operations that can be performed by a computer system. Flowcharts are often
used to graphically represent algorithms
For any computational process, the algorithm must be rigorously defined:
specified in the way it applies in all possible circumstances that could arise.
That is, any conditional steps must be systematically dealt with, case-by-case;
the criteria for each case must be clear (and computable). Because an
algorithm is a precise list of precise steps, the order of computation will almost
always be critical to the functioning of the algorithm. Instructions are usually
assumed to be listed explicitly, and are described as starting 'from the top' and
going 'down to the bottom'.
4.5.1 DEVELOPMENT OF THE NOTION ALGORITHM
The word algorithm comes from the name of the 9th century Persian
mathematician Abu Abdullah Muhammad ibn Musa al-Khwarizmi whose
works introduced Indian numerals and algebraic concepts. He worked in
Baghdad at the time when it was the centre of scientific studies and trade.
The word algorism originally referred only to the rules of performing arithmetic
using Arabic numerals but evolved via European Latin translation of al-
Khwarizmi's name into algorithm by the 18th century. The word evolved to
include all definite procedures for solving problems or performing tasks.
Before the algorithm (flow chart) was prepared the following important
questions were considered:
Has the problem been simplified and logically ordered?
41
Has a definite determination been made concerning a chosen technique,
methodology and organization?
Will the method selected be compatible with the computer to be used; are
the language, compiler, and interface problems resolved?
How much accuracy and precision is desired; have adequate checks and
tests been devised for program segment and totality?
Having answered these questions acceptably, the flow chart of the
programming procedures is formed as shown below:
42


43
INPUT X & Y COORDINATES
INPUT Q, S, n
COMPUTE Q
i
Q
i
< Q
SOLVE EQUATIONS 4.3 & 4.4 FOR NORMAL
DEPTH y
n
and CRITICAL DEPTH y
c
INPUT Y1,
X,Xn,N
COMPUTE y
n+1
FROM y
n
hf(x, y) >
0.02y
DOWNSTREAM
FINAL
LOCATION
REACHED?
DISPLAY PROFILE
DISPLAY PROFILE
PLOT CROSS SECTION
DISPLAY NUMERICALLY AND
GRAPHICALLY y
n
& y
c
STOP
START
NO
NO
YES
STOP
FIGURE 5. ALGORITHM FOR
PRISMATIC CHANNELS
NO
YES
44
START
INPUT X & Y COORDINATES FOR SECTION
1
PLOT CROSS SECTION 1
INPUT X & Y COORDINATES FOR SECTION
2
PLOT CROSS SECTION 2
INPUT X & Y COORDINATES FOR SECTION
3
PLOT CROSS SECTION
3
COMPUTE y
n+1
= y
n
+hf(x, y)
COMPUTE Sm and Fm
COMPUTE y
(n+1)m
= y
n
+hf(x
m
, y
m
)
x
m=
(x
n
+ x
n+1
)/2
y
m=
(y
n
+ y
n+1
)/2
|y
(n+1)m+1
- y
(n+1)m
|< e
COMPUTATION BETWEEN
SECTION 1& 2
SET FINAL y
(n+1)m
=y
n
YES
NO
YES
FIGURE 6. ALGORITHM FOR NON-
PRISMATIC CHANNELS
CHAPTER 5
SELECTED EXAMPLES
45
PLOT PROFILE
STOP
COMPUTATION BETWEEN
SECTION 2 & 3
NO
YES
Figure 7: The MATLAB command window
46
Figure 8: The
MATLAB Editor
window
Figure 9: Open Channel Application GUI (Graphical user interface)
QUESTIONS
47
1) Plot super critical flow and sub critical flow profiles (below normal depth) of a
prismatic channel with coordinates (0,2),(0.5,0.75),(1,0.5),(1.5,0),(3.5,0),(4,0.5),
(4.5,0.75),(5,2);
Flow rate =8 m
3-
/s, bed slope=0.001, Mannings roughness Coefficient=0.03
Also detect the normal and critical depths for the flow.
Answers
Normal Depth=1.9243m Critical Depth=0.96615m
48
Figure 10: Profile plot for super critical flow
Figure 11: Profile plot for sub critical flow (below normal depth)
2) Plot a sub critical flow profile (above normal depth) of a prismatic channel with
coordinates (0,3),(0.5,0.75),(1,0.5),(1.5,0),(3.5,0),(4,0.5),(4.5,0.75),(5,3);
Flow rate =10 m
3
/s, bed slope=0.001, Mannings roughness Coefficient=0.03
Also detect the normal and critical depths for the flow
49
Answer
Normal Depth=2.2658m Critical Depth=1.0835m
Figure 12: Profile plot for sub critical flow (above normal depth)
3) Computation Time Tests
Compare computation times for Euler, Heuns and Runge Kutta methods using a
prismatic channel with the following co-ordinates. (0, 6), (3, 0), (5, 0), (8, 6).
Flow rate = 20m
3
/s
Slope = 0.001
Mannings roughness=0.02
y (1) = 3.2m
x (1) = 0m
x (last) =
3000m
Answers:
50
FOR RUNGE KUTTA
N = 21
y (last) = 4.67771m
Runtime = 0.015625secs
FOR HEUNS
N = 559
y (last) = 4.67771m
Runtime = 0.078125secs
51
FOR EULER
N = 350000
Y (last) 4.67771m
Runtime = 7323.32secs
OBSERVATION
The Runge Kutta method converges after 20 steps with a computation time of
0.015625 seconds.
The Heuns method converges after 558 steps with a computation time of 0.078125
seconds.
The Euler method converges after 349999 steps with a computation time of
7323.32seconds.
From the above stated results it is easy to conclude that the Runge Kutta method offers
the best means for the computation of flow profiles in prismatic channels; as it
converges with the least number of steps for a given reach under the same flow
conditions with the shortest computation time.
52
4) Compute the flow profile for a non prismatic channel whose cross sectional
co-ordinates at 5 reaches are given below, using Euler, Heuns and Runge Kutta
methods and compare the results obtained.
Data for the channel properties are stated below;
CO-ORDINATES
Reach 1 X Y
0 15
50 10
50 3
150 0
250 0
600 3
1000 10
2500 15
Reach 2 X Y
0 14
40 10
55 3
120 0
500 0
600 3
1000 10
2500 14
Reach 3 X Y
0 15
45 10
42 2
120 0
250 0
500 2
1200 10
2400 15
Reach 4 X Y
0 15
45 10
45 2
120 0
250 0
500 2
1200 10
2400 15
Reach 5 X Y
0 15
45 10
45 3
120 0
250 0
500 3
1000 10
2500 15
n(1) = 0.02 So(1) = 0.001 h(1) = 100m
53
n(2) = 0.02 So(2) = 0.001 h(2) = 100m
n(3) = 0.025 So(3) = 0.001 h(3) = 100m
n(4) = 0.026 So(4) = 0.001 h(4) = 100m
n(5) = 0.026 So(5) = 0.001 h(5) = 100m
Profile obtained using Euler Method y end = 10.2674 m x end = 400m
Profile obtained using Heuns Method y end = 10.2637 m x end = 400m
54
Profile obtained using Runge-Kutta Method y end = 10.2636 m x end = 400m
Comparison of Results
This would be difficult to carry out because it would involve the inputting of a very large number
of reach data, which the written program cannot handle.
55
CONCLUSION AND RECOMMENDATION
CONCLUSION
Computer analysis of the surface profile of steady flow in open-channels has
been the main focus in this work. A computer program has been developed to
compute and plot backwater curves using MATLAB. Unquestionably, writing a
precision-sensitive program of this sort is rather challenging. On the other
hand, the hurdles encountered in the process of creating the program have
increased my interests in programming and hydraulics in general.
The major difficulty of the project lies in the proper arrangement and handling
of the geometric properties of the channel sections to achieve programmable
computations. For computational purposes natural courses are represented
approximately with smooth side lines of trapezoids and the irregularities of
sectional shapes are compensated for by adequate selection of Mannings
roughness coefficient.

The program developed can handle irregular (prismatic and non prismatic
sections (linearly and gradually varying)) channels provided horizontal
coinciding Y coordinates and equivalent Mannings roughness coefficients
can be provided.
The fourth order Runge Kutta method provides accuracy sufficient enough for
this project (prismatic channels, page 50).
56
MATLAB enabled a comparatively easier development of the required
software thanks to more sophisticated functions like the solve function which
was very useful in the computation of normal and critical depths.
Finally, this project has proven that the computation of the surface profile of
flow from a reach of known depth in both prismatic and non prismatic sections
can be done with good speed, ease and precision with the aid of a computer
program that is executable on a micro-computer.
RECOMMENDATION
This research deserves more probing, there is still a lot more to be done on it.
If given the opportunity I would love to explore all the possible conditions and
develop software that would be capable of handling these possibilities.
Data feeding was reduced to the minimum, future research can help in
discovering better ways of entering data (e.g. through satellite connections) to
provide more a flexible and efficient program
It is a project that is worthy of government sponsorship, as stated earlier on in
the write up it is useful in flood alleviation planning which is a major problem
we face in various locations of the country, development of software of this
nature will go along way in contributing to solving of this problem.
57
BIBLIOGRAPHY
Aiyesimoju. K.O.: River Engineering Lecture Notes, CEG 519, University of Lagos, Akoka, Yaba,
2006.
Boolos, George and Jeffrey, Richard: First Edition, Computability and Logic, Cambridge University
Press, London, 1999, 262 pages.
Boris A. Bakhmeteff: Hydraulics of Open Channels, McGraw Hill Book Company, Inc., New York,
1932. pp 143-215
Douglas J. F. John Gasiorek, John Swaffield, 5
th
Edition, Fluid Mechanics, Prentice Hall, Upper
Saddle River, New Jersey, (9 Nov, 2005), 992 pages.
Erwin Kreyszig 8
th
Edition Advanced Engineering Mathematics, John Wiley & Sons, Inc. Hoboken,
New Jersey, (11 march, 1999), 1296 pages
Irving H. Shames: 4
th
Edition, Mechanics of Fluid McGraw Hill Book Company, Inc., New York,
08/02/2002, 896 pages.
J.A. Ch Bresse: Cours de mecainque appliqu, 2e partie, Hydraulique, (Course in Applied
Mechanics, pt. 2, Hydraulics), Mallet- Bachelier, Paris, 1860.
Larry W: Mays First Edition, Water Resources Engineering John Wiley & Sons, Inc. Hoboken, New
Jersey, (11 Aug, 2000) 780 pages.
Markov A.A.: Theory of Algorithms.[Translated by Jacques J. Schorr-Kon and PST staff] Imprint
Moscow, Academy of Sciences of the USSR, 1954, 444 pages.
MATLAB Release 14, Help Files,
M. E. Von Seggern: Integrating the equations of non uniform flow, Transactions, Proceedings
American Society of Civil Engineers, vol 115. pp 71-88, 1950.
Microsoft Inc. Microsoft Encarta 2007,
Ming Lee: Steady gradually varied flow in uniform channels on mild slopes, Ph. D. thesis, University of
Illinois, Urbana, 1947.
58
Nagaho Monodobe: Back-water and Draw drop-down curves for uniform channels, Transactions,
American Society of Civil Engineers, vol 103. pp 950-980, 1938
Norman Bruton Webber: Metric 2 Revised Edition, Fluid Mechanics for Civil Engineers, Spon Press,
London, U.K. (1971) 372 pages.
Stroud K.A. Dexter J. Booth: 4
th
Revised Edition Advanced Engineering Mathematics, Industrial
Press, Inc. U.K. (April 2003) 1280 pages.
Tolkmitt G.: Grundlagen der Wasserbaukunst, ( Fundamentals of Hydrualic Engineering,), Ernst &
Sohn, Berlin, 1898.
Ven Te Chow: international Edition, Open-Channel Hydraulics Mc-Graw Hill Education, Colombus,
Ohio (31 Dec 1959) 692 pages.
http://en.wikipedia.org/wiki/Algorithm examples
APPENDIX
PROGRAM
LISTING
59
60