CHAPTER 6
GRAPHICAL SOLUTION OF
SEEPAGE PROBLEMS
6.1 Homogeneous Isotropic Flow
In case of homogeneous isotropic flow, the
total head h satisfies the equation (4.12) . In
the mean time, the velocity potential j and the
stream function y have the definitions given
by equations (4.12) and (4.17) respectively.
The functions j and y are orthogonal and
hence the streamlines and lines of constant
potential meat always at right angles.
As j = kh as defined in equation (4.11), then
for k= constant, the streamlines and the lines
of equal heads also meet at right angle. This
property can be used to obtain a graphical
solution of steady state flow through
homogeneous porous media.
flownet that includes a set of streamlines and
lines of equal potentials ( or equal heads). The
degree of accuracy of the solution depends on
the accuracy reached in the construction of
the flownet.
In the flownet, the following conditions are to
be satisfied:
1. all impermeable boundaries are
streamlines.
2. all water bodies are lines of constant head
(equipotential lines)
3. streamlines meet equipotential lines at right
angles
4. streamlines do not intersect one another
5. equipotential lines do not intersect one
another.
Using the previous conditions, it is
recommended to start the flownet
construction by drawing the streamlines taking
the advantage of the existing impermeable
surfaces in the domain of flow and making
sure that they meet water bodies at right
angles.
The equipotential lines are sketched in the
next step such that each cuts the streamlines
perpendicularly. The process may need a lot of
erasing and correction especially for beginners
until the flow net is completed to the persons
satisfaction. However, after that the
calculation of the flow rates and pressure
distribution becomes a routine matter.
A typical well drawn flow net is illustrated in
Fig. 6.1.a It represents the flow under a
masonry dam A horizontal impermeable layer
exists at a distance below the ground level.
The difference in head between the upstream
and downstream is “h”.
he flownet is constructed such that the dam
base and the impermeable layers are the
extreme stream lines while the highest and
lowest equipotential head lines are the ground
levels on the upstream and downstream
respectively..
It is always suggested to construct the flownet
such that all its elements are nearly squares.
This is not absolutely necessary but it is only
recommended to facilitate the calculation as it
will be noticed.
Around the corners, it may be difficult to
construct the flow net elements as squares. In
this case choose their shapes that the average
length of the two sides in the direction of flow
equal to the average length of the other two
sides.
In the flownet, the stream tube is defined as
the space between two neighboring
streamlines. It is obvious from the equation of
continuity that the flow through a stream tube
is always constant despite the change in the
width of the tube as the flow cannot cross any
of its sides since they are both stream lines.
Consider the three shaded elements A, B& C of
Fig. 6.1- a which are enlarged in Fig. 6.1-b. The
flow across element A equals the flow across
the element B as they both lie in the same
stream tube. Thus, From Darcy’s Law
qA = bA .k.(D hA)/lA =qB =bB. k.(D hB)/lB
(6.1)
in which, D hA and D hB are the head losses
through the elements A and B respectively.
As the elements are chosen to be squares,
then bA=lA and bB= lB. The use of this
property in equation 6.1 leads to the
conclusion that D hA=D hB.
This means that if the flownet is constructed
as squares the drop of head in all elements of
the streamtube are equal . In other words, if
the number of elements in a stream tube is
“Nd “ and the total difference in head between
the up stream and down stream is “H”, then
the drop in head in each element D h= H/Nd.
Thus, the rate of flow through the stream tube
equals (k. H)/Nd. Consider another
streamtube, say that containing the element
C, the flow qC through that tube must be [Link].
( D hC)/ lC .Since the elements are squares,
then bC=lC .. Moreover, the elements B and C
lie between the same two equipotential lines,
hence D hC =D hB which means that D hC =
H/Nd and the flow qC equals kH/Nd .
This means that the flow through all the tubes are
equal. Hence, if “Nf” is the number of streamtubes,
Then the whole amount of seepage under the dam
“Q” is given by the following relation:
Q= k. [Link] / Nd (6.2)
Also, the total head at the front of an element that lies
after m drops behind the upstream is given by the
relation:
hm = hu - m.H/Nd (6.3)
where hu is the total head at the upstream of the
dam. As a numerical example, Fig. 6.2 is
considered . It consists of a masonry dam of
length 20 m. The dam is supplied by two sheet
piles to reduce the uplift on the dam base as
well as the rate of seepage. One of the sheet
piles is located at the upstream end of the
dam base while the other is fixed to its
downstream end.
The water level in the up stream is 10 m above
the ground level while it coincides with the
ground level in the down stream. The
impermeable layer lies at a distance 20 m
below the ground level. The hydraulic
conductivity k=1.5x10-6 m/s.
From the constructed flownet ;
Nd = 22, Nf = 8
& Q= (1.5x10-6 )x.(10 )x(8)/22 = 5.45x10-6
m2/s
The uplift distribution under the dam is
illustrated in Fig 6.2 b
6.2 Flow through Multilayer Soil
It is clear from equation (5.18) and Fig. 5.3
that a streamline which crosses the interface
between two layers suffers a deflection that is
dependent on the ratio between the hydraulic
conductivity of the two layers. To study the
influence of the sudden change of the value of
hydraulic conductivity on the flow net
construction, Fig. 6.3 is considered.
Assume that the flownet is constructed such
that the elements are squares in region I
whose hydraulic conductivity is “k1“ (i.e. l1 =
b1). In that region, the flow through the drawn
At the interface between the two layers, the stream
tube suffers a deflection such that (k1/k2 ) = (tanq 1
/tanq 2). However, from the continuity equation, the
rate of flow through the stream tube remains
unchanged, or
k2 . b2 .(D h2/l 2) = k1.D h1 (6.4)
Choose the element dimensions in layer II such that D
h2 = .D h1 , then the substitution in equation (6.4)
leads to the following relation:
l2/b2 = k2/k1 (6.5)
This means that the flownet elements in
region II, will be rectangles whose length l2
equals the average breadth of the stream tube
at the element b2 multiplied by (k2 /k1 ).
To illustrate that, the trivial problem of uniform
flow through a multilayer soil is considered.
Assume that the flow occurs downwards
perpendicular to three parallel layers of equal
thickness 4 m. Assume further that the upper
surface of the top layer (layer I) has a total
head 2.00 m more than the lower surface of
the bottom layer (layer III). For k1 = 0.5 k2 = 2
k3= 2x 10-6 m/s, it is required to calculate the
amount of flow through 4.00 m breadth of the
layers. In Fig. 6.5, the flow net of the problem
is illustrated.
From the figure:
f =4, Nd = 14, Q= ([Link] .H)/Nd= (2x10-6.
4.2)/14 =1.143x10-6 m2/s
This problem can be solved using equation
(5.14). In this case:
kh = [(4x3)/( 4/2+4/4+4/1)]x10-6 = 1.714x10-
6
v = kh. H/B = [(1.714).(2)/(12)]x10-6 =
2.857x10-7 m/s
Q = 4.v= 1.143x10-6 m/s
which is the same result as that obtained from
the flow net method.
As another example, the flow through the
earth dam of Fig. 6.5 is considered. The dam
dimensions are shown in the figure. The water
level in the upstream is 16.00 m above the
bed level . In the down stream a drainage filter
with atmospheric pressure exists.
The dam is constructed using two types of
isotropic soil. The soil in zone 1 in the up
stream part has a hydraulic conductivity
1.00x10-7 m/s. The soil in zone 2 has a
hydraulic conductivity 4x10-7 m/s. The
interface between the two layers is an inclined
plane as shown in the figure.
Fig. 6.5 Seepage through an Earth Dam (from,
Cedergren, Seepage
Drainage and Flow Nets)
The flownet is constructed such that it’s elements in
zone 1 are squares. Since k2 = 4 k1, the elements in
zone 2, are rectangles with the length equals 4 times
the stream tube breadth.
From the figure, Nf = 3.5, Nd = 8
accordingly:
Q= (1x10-7) . (3.5).(16)/(8) = 7x10-7 m/s
6.3 Use of Transformations to solve
nonhomogeneous and anisotropic Problems
The steady state flow through porous media is
described in its general form by equation
(4.23). In two dimensions, the equation
reduces to:
{k (h/ x)}/ x+ {k( h/ y)}/ y =0
(6.6)
The definitions of potential flow of section 4. 3
do not apply in this case . However, if a
transformation of the dependent variable
h(x,y) can be found such that the equation
reduces to a form similar to that given by
equation (4.10), then a graphical solution in
the transformed domain may be possible.
This can be explained as follows. Assume that
there is a variable “H” which is a known
continuous single valued function of “h”, “x” &
“y “ whose first two derivatives with respect to
“x” and “y” exist and are continuous, or:
H= f(h,x,y) (6.7-
a)
The inverse of this function is also assumed to
be existing and is unique, i.e.
h =f-1 (H,x,y) (6.7-b)
Now the function H (h, x,y) can be chosen such
that it satisfies the Laplace equation, i.e
2H/ x2+ 2H/y2 = 0 (6.8)
Then the velocity potential and stream
function can be defined and a graphical
solution of “H” becomes possible. Knowing “H”
in terms of “x”,”y” , relation (6.7-b) can be
used to obtain the corresponding values of H.
However, it must be stated that finding a
variable “H” that satisfies equations (6.7) and
(6.8) together with the conditions of
uniqueness and continuity is not simple .A lot
of trial and error works may or may not lead to
The transformation can be also done in the
independent variables. This means that a new
domain is found such that the coordinates of
this new domain “X”, “Y” are related to the
original coordinates “x”, ”y” by unique single
valued continuous functions with continuous
first and second derivatives. These functions
may have the form:
X(x,y)= fx(x,y) (6.9-a)
Y(x,y)= fy(x,y) (6.9-b)
In the meantime, the inverse function of fx and
fy exist and are unique. If the actual h in the
new domain satisfies Laplace equation, then a
graphical solution in the (X,Y) domain
becomes also possible.
After completing the solution in the new domain, the
inverse of equations (6.9) can be used to transform
the solution in the original domain.
The transformation of the independent variable can be
applied to solve problems of flow through anisotropic
soil. In two dimensions, equation(4.33) takes the
following form:
kx 2h/ x2+ky 2h/ y2 =0 (6.10)
Assume that this equation is satisfied in the
domain “d” of (x,y) plane of Fig. 6.6-a. Define
the variables X, Y such that:
X=( ky /kx ).x (6.11-a)
Y=y (6.11-b)
The substitution from relations (6.11) in
equation (6.10) for kx ¹ 0& ky ¹ 0 implies that:
2h/ X2 + 2h/ Y2 =0 (6.12)
This means that “h” satisfies Laplace equation
in the (X,Y) plane and hence the flow net can
be constructed in this plane.
The solution of the problem can be conducted
as explained following these steps :
1. use equations (6.11) to draw the boundaries
of the considered domain in the (X,Y)-plane
2. draw the flownet in the (X,Y)-plane and
hence calculate Nf and Nd .
3. calculate the rate of seepage from the
equation:
Q= (ka .[Link]) /Nd
(6.13)
where,
ke= kx ky
(6.14)
Use the inverse of equations (6.11 ) to map
the flow net in the original (x,y)-plane. Note
that the flow lines and lines of equal head do
not necessarily meet at right angle in the (x,y)-
plane. This flow-net can be used to compute
the total head values and uplift forces.
An example of the explained procedure is
illustrated in Fig. 6.7. It is concerned with the
flow under a sheet pile in an anisotropic media
with kx=4 ky = 6x10-7 m/s. The water level in
level while it coincides with the ground level in
the downstream. The impermeable layer lies
at a depth of 12.00 m below the ground level
and the sheet pile extends 6.00 m through the
soil. It is required to draw the flownet and
calculate the rate of seepage under the sheet
pile.
Define the new plane (X,Y) with X=x. (ky/kx)
=1/ 6 x and Y=y and the domain of flow is
drawn as shown in Fig. 6.8-a. The constructed
flow net is shown in the same figure.
The total seepage can hence be calculated
using formula (6.13).
Q= 10-7[ 1.6 . 4.10]/8 = 1.22x10-6
m2/s
To draw the flownet in the actual (x,y) plane,
use the relation x= Ö 6 X, y=Y to map the
points on stream lines and lines of equal head
from the (X,Y) plane. For example the point
X=5, Y=0 on the down stream whose head= 0
will be mapped into the point x=12.24, y=0.
The whole mapped flow net is shown in Fig.
6.8-b.
6.4 Problems on Chapter 6
Sketch the flow net and compute the rate of
seepage under the dam with sheet pile shown
in figure for the 4 cases,a,b,c&d
a) kx1= ky1= kx2= ky2
b) kx1= ky1=2 kx2= 2ky2
c) kx1= ky1= 0.5kx2=0.5 ky2
d) kx1= 2ky1= kx2= 2ky2
Answer:( 5x10-6)
2)A masonry dam is underlain by a sandy
layer as shown in figure. Draw the uplift
stress distribution under the dam for the
following cases:
a = b =5 m , k1=k2= 0.000003 m/s
a = .5b =5 m , k1=k2= 0.000003 m/s
a = b =5 m , k1=2k2= 0.000003 m/s
a = b =.5 m , k1=0.5k2= 0.000003 m/s
a = .5b =5 m , k1=2k2= 0.000003
m/s