Comparison of SEEP - W Simulations With Field Observations For Seepage Analysis Through An Earthen Dam (Case Study - Hub Dam - Pakistan)

Comparison of SEEP/W Simulations with Field Observations for Seepage Analysis through
an Earthen Dam Case Study! "ub Dam # Pa$istan%

&'

(nternational )ournal of *esear+h ()*% ,ol#-. (ssue#'. August /0-1 ISSN 2348-6848
Comparison of SEEP/W Simulations with Field Observations
for Seepage Analysis through an Earthen Dam Case Study!
"ub Dam # Pa$istan%
Engr. Imran Arshad
1
and Dr. M. M. Babar
2

2
Professor, Institute of Water Resources Engineering and Management (IWREM), Mehran
University of Engineering and Technology (MUET), Sindh Pakistan. E-Mail:
[email protected]
2
M.E Research Student, Institute of Water Resources Engineering and Management (IWREM),
Mehran University of Engineering and Technology (MUET), Sindh Pakistan. E-Mail:
[email protected] (corresponding author)

Abstra&t
The present research work is designed to
model seepage analysis of an earthen dam
by using finite element approach. For this
purpose a research study was conducted on
Hub dam, which is a small earthen dam
located at about 35 km, north-east of
Karachi city, Pakistan. In the study the
amount of seepage through and under body
of the main dam is computed, profile of
phreatic line is simulated for different
scenarios and compared with the observed
data. For the purpose of this study, SEEP/W
the sub-program of Geo-Slope software is
used. Data pertaining to design parameters
and dam geometry are given as input to the
software to compute the unknown
parameters. Finally results are validated by
comparing them with the observed data. The
main dam is composed of three different
kinds of reaches; therefore in this research
only one reach with core wall i.e. Zoned
Embankment Section at CH: 48+75 is
studied. Computations are carried out for
three different scenarios, that is: maximum
pool level, normal pool level, and minimum
pool level.
Calibration of the material properties is
made on the basis of minimization of error
while comparing observed hydraulic heads
with the simulated ones. The flownet has
been drawn comprising of streamlines,
equipotential lines, velocity vectors showing
dominant flow (seepage) field and phreatic
line depicting seepage behavior of the Hub
dam. The seepage flux (discharge), exit
gradient and maximum seepage velocity for
the entire pond level scenarios and for all
the selected section are computed. At lowest

&2

pond level minimum seepage occurs at
highest pond level maximum seepage occurs.
It is also ascertained that the exit gradient is
within the permissible limits that is that less
than unity for all the scenarios; thus it also
conforms to safety criteria of the dam.
Seepage velocities for the entire pond level
scenarios and for the selected section are
computed; at lowest pond level minimum
seepage velocity is observed and at highest
pond level maximum velocity occurs.
Residual head dissipation trend is modeled
and predicted for all the sections of interest
for different scenarios. At selected section
i.e. Zoned Embankment Section at CH:
48+75 for low pond level slightly smoother
dissipation rate is followed, however, at
higher pond levels a somewhat rapid
dissipation of head occurs at sheet pile
positions; this of course signifies the
effectiveness of sheet pile. Initially dead
dissipation follows somewhat smoother
trend, however at the position of core wall
and sheet pile an abrupt rise in dissipation of
head is exhibited, which again signifies the
effectiveness of the two seepage control
devices.
Validation of any model is made by
comparing simulated results against the
observed ones; this is done to ensure model
applicability. If this comparison shows a
good coincidence, then the model developed
can be recommended for practice. Table 4
contains the data pertaining to observed
piezometeric heads and simulated ones and
the relative error. Performance of the model
is assessed evaluated on the basis of
statistical parameters, i.e. mean error, root
mean square error and model efficiency;
these results are presented in Table 6.
'eywords! Seepage Analysis, Phreatic
Line, Earthen dam, SEEP/W, Finite Element
Modeling.
INTRODUCTION
It is well known fact that excessive seepage
in any type of a dam is one the root cause to
destabilize the dam structure and thereby
bring economical havoc. Ensuingly
pragmatic efforts are employed to carryout in
depth study of the seepage analysis through
and beneath the body of a dam. Generally
designers by employing different techniques
augment safety of a dam and smear errors in
computation due to maintenance of water
storage, especially by focusing on hydro
structure of the dam. The seepage control of
any dam may be analyzed by virtue of
various available methods. Seepage is the
main aspect and its control enjoys main
position in designing, construction and
maintenance of any dam. Thus a dam
engineer must be well versed in
understanding seepage problems, their
solution and preventive measures
monitoring. The flow conditions of any
porous environment can be investigated by
using numerical techniques framed in the
form of a software solution, i.e. computer
program Kamanbedast et al. (2011). The
main difficulty in diagnosing the seepage
problems is fixing the location of phreatic
surfaces; which at initial stages cannot be
fixed and thus requires iterative processes
Kazemzadeh-Parsi et al. (2012). The seepage
analysis of a dam is essential for evaluation
of its safety and stability especially by using
numerical techniques; by doing this one can
analyze seepage field and make its

&3

comparison under different conditions. By
doing this the effect of core and other
assorted factors can also be investigated
Quanshu et al. (2010).
Many computer softwares have come in
general use, and any hard computations and
simulation can be carried out through them
by giving them appropriate inputs and data.
This results in less error frequency and more
detailed analysis when compared with field
observations. The numerical model SEEP/W
can be employed to carry out simulation of
seepage and phreatic surface. The SEEP/W
program is capable enough to simulate quite
effectively seepage rates and phreatic
surfaces in homogenous and non-
homogenous earthen dams Mohammed et al.
(2006). SEEP/W is a finite element computer
enabled (CAD) software which is capable
enough to solve groundwater flow, seepage
and excessive pore water pressure problems
within the porous media such as soil and
rock. The software is capable enough to
resolve the problems ranging from simple
saturated steady state issues to
saturated/unsaturated time dependent
problems. The software is also capable
enough to employ in designing of
geotechnical, civil, hydrological and mining
engineering problems. The principal quality
of the software is due to its ability to allow
seepage analysis as a function of time and
this process is considered as infiltration of
precipitation. Due to transient characteristic
of the system, it provides a window of
opportunity for researchers to analyze such
problems; for instance migration of a wetting
front and dissipation of excessive pore water
pressures Geo-Slope International (2007).
In view of all above facts, the present
research work is designed to model seepage
analysis of an earthen dam by using finite
element approach. For this purpose a
research study was conducted on Hub dam,
which is a small earthen dam located at
about 35 km, north-east of Karachi city,
Pakistan. The dam is the only dam built on
Hub river, which originates from the Kirthar
range mountains from an elevation of about
6000 ft. in the north of Arabian sea and the
river after covering a rocky terrain of about
220 km outfalls into the sea. The catchment
area comprises of about 3,410 square miles
till the dam site. This whole catchment area
consists of arid zones of Sindh and
Balochistan. The catchment area is confined
by Pab range of mountains on its right, while
Kirthar range is located on its left side. It is
stretched in north-south direction with its far
most tip highest elevation close to Khuzdar,
approximately at about 7000 ft. above sea
level.

40


Fig. 01: Zoned Embankment Section at CH: 48+75
The main dam is 15,640 ft long and 152 ft
high earthen embankment. Except for a
3,100 ft reach (between Ch. 32 + 00 to
63+00) of zoned section in river valley, the
entire length of the embankment is made of
homogenous section with a supplemental
downstream shell of pervious material. The
zoned embankment has silt (intermixed ML
and CL material) and clean river sand-gravel
in shoulder of the closure section.
The objective(s) of this research work was to
study the seepage behavior of earthen dam
by using Finite Element analysis, to develop
and calibrate a computer model for an
earthen dam using FEM based software i.e.
the SEEP/W, and to compare observed and
simulated data.
Materials and Methods
Steps for Modeling of Hub Dam
In first attempt, in order to achieve the
research objectives of the present study cross
sections at Zoned Embankment chainage
Section i.e. CH: 48+75 was selected for to
model by using SEEP/W software. In second
attempt the SEEP/W software is used to
generate FEM mesh to carry out the seepage
analysis. The up- and down-stream boundary
conditions are assigned as Dirichlet and
Neumann boundary nodes according to
given conditions. The nodes at bottom of the
foundation of dam are considered with zero-
flux (Nuemann) condition. When the model
is completely developed then it is verified by
the SEEP/W software and after acceptance
of the model by the software, it is ready for
computation. For the selected cross-section,
computation is carried out for different
scenarios of water levels. The material
properties for the materials used in dam
section are calibrated. Finally simulated
results obtained from the SEEP/W software
for the selected section was compared with
the observed data obtained from the
WAPDA.
GOVERNING EQUATIONS
In this research work, finite element
approach is employed to solve the governing
differential equations pertaining to seepage
through body of dam its foundation. The
SEEP/W software (program) is a sub-
program of the Geo-Slope (software)
computer, which is used to cater for seepage

4-

problems through porous soil media.
SEEP/W is a FEM based CAD type software
used to analyze seepage and groundwater
flow problems. Following partial differential
equation (PDE) is the governing equation
used for modeling of SEEP/W program:

..1
where
H- is hydraulic head, Kx- and Ky- are
hydraulic conductivity in x- and y-
directions, respectively, Q- is the applied
source or sink terms, t- is the time domain
and volumetric water content.
Eq. (1) is a two-dimensional non-linear
second order PDE and caters for transient
flow conditions; its derivation involves the
basic constitutive law of Darcy for
groundwater flow, given as below:

H K v =

..2
where

=
y
H
x
H
H

..3
in which v- is average velocity through soil
media known as the Darcian velocity; K- is
hydraulic conductivity of the soil material;
and
H
- is the gradient of hydraulic head in
x- and y- directions.
The Eq. (1) is time variant and states that
the difference between the flow entering an
elemental volume and leaving an elemental
volume at a point is equal to the change in
the volumetric water content in a particular
time. If the volume of influx equals to the
volume of out flux then the equation caters
for steady state conditions, thus the right
hand of the equation changes to zero.

..4
Changes in volumetric water content depend
upon properties of the soil and changes in the
stress state. Following set of two variables
essentially describe the state of stress under
saturated and unsaturated conditions, that is
(ua) and (ua - uw), where ua - represents
pore air pressure and uw - stands for pore
water pressure.
The SEEP/W programme is based on
constant total stress conditions i.e. no
loading and unloading of soil mass is
involved. Other aspect is that the pore air
pressure remains constant during transient
process i.e. ua - remains constant, which
implies that volumetric water content
remains unchanged. Volumetric water
content changes are dependent on changes in
(ua - uw). A change in volumetric water

4/

content in terms of change in pore water
pressure is represented by the following
equation:

..5
where mw- is slope of the storage curve.
The total hydraulic head is as under:

..6
where
uw- is pore water pressure, - is specific
weight of water, H Total hydraulic head,
and y- is elevation head.
Now equation (6) can be arranged as:
..7
Substituting equation (7) into equation (5) we get the following equation:
..8
Now by substituting the above equation in equation (3) we get the following expression:
..9
As the subject elevation is static, due to which the derivatives of (y) w.r.t time vanishes and
consequently the differential equation appended below will be the resultant differential equation:
..10
FEM Mesh Formation and Its
Verification by Using SEEP/W
Software
FEM meshes for the selected section are
developed by using the SEEP/W software.
The material properties for each section with
proper dimensions are made as input to the
software respectively and verification for
each cross section has been made
accordingly. The FEM mesh at the selected
Section is composed of four types of
elements, i.e. triangular, square, rectangular
and trapezoidal type of elements of different
sizes Fig. 02. The number of nodes is 2,299
and the total number of elements is 2,206.

45


Fig. 02: Mesh Formation for Zoned Embankment Section at Ch: 48+75.
After all the necessary inputs, the computer
program SEEP/W verified the mesh
development and delivered report that the
vertical and horizontal meshing is strong
enough and there is no error in formation of
mesh model. Thus the model is ready for
computation and analysis of the results.
Setting of Boundary Conditions
Computations are carried out for following
three different scenarios, viz: (i) Maximum
pool level (346 ft), (ii) Normal pool level
(339 ft), and (iii) Minimum pool level
(270 ft).
Boundary conditions are set as: (i) At fill
level up- and down-stream boundary
conditions are considered as Dirichlet
boundary conditions for all the above given
scenarios, and (ii) In foundation up-, down-
and bottom level are considered as with zero-
flux condition i.e. Neuman boundary
conditions for all the above given conditions.

Problem Considered for Analysis and
Computation
The following problems are to be considered
for analysis and computation:
Development of flow net by tracing
streamlines and equipotential lines
for different conditions.
Observation of velocity vectors and
thereby seepage behaviour for
different conditions.
Profile of the phreatic line for
Estimation of seepage quantity
through the dam profile and its
foundation for different conditions.
Computation of Exit gradient,
maximum velocity and Residual head
along the dam foundation under
Results and Discussion
Calibration of Material Properties of Hub
Dam Model
For calibration of material properties for the
selected section of the Hub dam, initially
identical guess values were specified for all
the sections. These guess values for different
types of materials used in the dam are
presented below in Table 1. Calibration of
the material properties is made on the basis
of minimization of error while comparing
observed hydraulic heads with the simulated
ones.

41

Table 1: Material Properties (Guess Values)

S.
No
Material
type
* Hydraulic
conductivity
(ft/sec)
01 Foundation 10
-4
to 10
-6

02 Shell 10
-5
to 10
-6

03 Core 10
-8
to 10
-7

04
Filter
Blanket
10
-2

* Source: WAPDA
Using SEEP/W software, the material
properties (hydraulic conductivities)
calibrated for the selected section are
presented in Table 2.
Table 2: Calibrated Values of Material
Properties
for Selected Section at CH: 48+75.
S.
No
Material
type
Hydraulic
conductivity
(ft/sec)
01 Foundation 3.015 x 10
-6

02 Shell 2.385 x 10
-5

03 Core 2.000 x 10
-8

04
Filter
Blanket
3.280 x 10
-2

Flownet with Stream- and Equipotential
Lines, Phreatic Line Behaviour and
Velocity vectors
The SEEP/W software is also used to get
seepage analysis through the dam and its
foundation for different pond level scenarios.
For this purpose, using the software flownet
has been drawn for all the selected sections
as shown in Fig. 3 5. The flownet
comprises of streamlines, equipotential lines,
velocity vectors showing dominant flow
(seepage) field and phreatic line depicting
seepage behavior of the Hub dam. From the
Figures it is revealed that the stream and
equipotential lines are normal to each other,
which conforms to seepage theory. The
effectiveness of filter blanket and core wall
at higher pond levels is more significantly
demonstrated at the selected section.

Fig. 3: Flownet for Selected Section at CH: 48+75 (Pond level = 270 ft.)

4&




Seepage Flux, Exit Gradient and
Maximum Seepage Velocity
Using the SEEP/W software, the seepage
flux (discharge), exit gradient and maximum
seepage velocity for the entire pond level
scenarios and for the selected section are
computed; these are listed in Table 3. At
lowest pond level minimum seepage occurs
that is of the order of 2.029 x 10
-4
(ft
3
/sec/ft);
at highest pond level maximum seepage
occurs which is of the order of 5.565 x 10
-4

(ft
3
/sec/ft). A graphical correlation of
seepage flux versus pond level is also shown
in Fig. 6.
Likewise it is also ascertained that the exit
gradient is within the permissible limits that
is that less than unity for all the scenarios
and at the selected sections for study; thus it
also conforms to safety criteria of the dam.
Fig. 7 shows a graphical relationship for exit
gradient as function of pond level. In this
case initially a linear behavior is followed;
however the exit gradient rises exponentially
as the pond level increases.
Table 3: Computed seepage flux, exit gradient and maximum
seepage velocity at Selected Section for different pond levels
Parameters
Upstream Pond Levels
Minimum 270 (ft.) Normal 339 (ft.) Maximum 346 (ft.)
Seepage flux
(ft
3
/sec/ft)
2.029 x 10
-4
5.250 x 10
-4
5.565 x 10
-4

Exit gradient 0.137 0.274 0.402
Max. seepage velocity
(ft/sec)
1.775 x 10
-6
2.678 x 10
-6
3.181 x 10
-6

44

Similarly seepage velocities for the entire
pond level scenarios and for the selected
section are computed; at lowest pond level
minimum seepage velocity is observed
which is of the order of 1.775 x 10
-6
(ft/sec);
and at highest pond level maximum velocity
occurs is of the order of 3.181 x 10
-6
(ft/sec).
Fig. 8 shows a graphical relationship for
maximum seepage velocity as a function of
pond level; under this case at first a linear
behavior is followed, however the velocity
rises exponentially as the pond level goes up
on increasing.

Fig. 6: Seepage flux vs. pond levels at
Selected Section at CH: 48+75

Fig. 7: Exit gradient vs. pond levels at
Selected Section at CH: 48+75

Fig. 8: Max. Seepage velocity vs. pond
levels Selected Section at CH: 48+75
Residual Head Dissipation Trend
Residual head dissipation trend is also
modeled and predicted for all the sections of
interest for different scenarios. From Fig. 9
through Fig. 11, it can be seen that initially
dead dissipation follows somewhat smoother
trend, however at the position of core wall
and sheet pile an abrupt rise in dissipation of
head is exhibited, which again signifies the
effectiveness of the two seepage control
devices.

Fig. 9: Head dissipation trend along the dam
foundation
Selected Section at CH: 48+75 (270 ft. pond
level)

4'


Fig. 10: Head dissipation trend along the
dam foundation
level)

Fig. 11: Head dissipation trend along the
dam foundation
level)
Model Validation
Validation of any model is made by
comparing simulated results against the
observed ones; this is done to ensure model
applicability. If this comparison shows a
good coincidence, then the model developed
can be recommended for practice. Table 4
contains the data pertaining to observed
piezometeric heads and simulated ones and
the relative error.
Table 4: Observed and simulated hydraulic heads
(normal pool level 339 ft)
Sections
X - distance
(ft.)
Observed
head
Ho (ft.)
Simulated
head
Hs (ft.)
Relative error (%)

C
H
:

4
8
+
7
5

435 305 306.47 -0.482
477 259 260.73 -0.668
527 236 238.19 -0.928
653 225 222.60 1.067
Performance of the model is assessed
evaluated on the basis of statistical
parameters. Following parameters that is
mean error (ME), root mean square error
(RMSE) and model efficiency (EF) are
assessed [Willmut, 1982]; their formulation
( )
100
H
" H
o
s o

=

42

is given below and computational steps are
performed in Table 5.9.
( )
=
=
n
1 i
oi si
H H
n
1
ME

..11
( )
5 . 0
n
1 i
2
oi si
H H
n
1
RMSE

=

=

..12

( )
( )
=
=
=
n
1 i
2
oa oi
n
1 i
2
oi si
H H
H H
1 EF

..13
where
Hsi is the ith value of simulated head,
Hoi is the ith value of observed head, and
Hoa is the average or mean of observed
head.
Table 5: Observed and simulated hydraulic heads with
statistical computational steps (pond level 339 ft.)
Sections
X-
distance
Observed
head
Ho (ft.)
Simulated
head
Hs (ft.)

C
H
:

4
8
+
7
5

435 305 306.47 1.47 2.161 1072.5625
477 259 260.73 1.73 2.993 175.5625
527 236 238.19 2.19 4.796 1314.0625
653 225 222.6 -2.4 5.760 2232.5625
The EF is another parameter to evaluate the
performance of the model. For the developed
simulation model, RMSE and ME values are
found 2.019 and 0.745 ft, respectively and
the maximum relative error amongst all the
data sets is 1.067 %. Thus it is found that the
performance of the model is good enough
with model efficiency of 99.60 %.

Table 6: Summary of statistical parameters
showing model performance
Statistical
Parameters
Values
Mean Error (ME)
0.735

Root Mean Square
Error (RMSE)
2.019

Model Efficiency (EF) 99.60 %
Maximum relative error 1.067 %.
( )
2
oa oi
H " ( )
2
oi si
H H ( )
oi si
" "

43


Fig. 12: Relationship between observed and
simulated hydraulic heads.
Additionally verifiability of the model is also
made by comparing observed and simulated
values of piezometeric heads; such graph is
illustrated in Fig. 12. The slope of the line is
observed to be approximately at 45 degree;
thus the Fig. indicates no considerable
difference between observed and simulated
head values. Consequently, it is concluded
that simulated values of piezometeric heads
are not much different than the observed
ones.
Summary and Conclusions
From FEM analysis of seepage through
earthen dam using SEEP/W software, we
evaluated that the phreatic line has been
simulated at the selected section for the three
scenarios i.e. Minimum, Normal and
Maximum pool levels and compared with the
actual data and the model demonstrates high
efficiency and good fitness. Through the
study it is observed that dam safety is not
endangered from the seepage point of view
since the phreatic line pattern follows
standard design criterion. For the three
scenarios of Minimum, Normal and
Maximum pool levels the exit gradient value
is within permissible limits (i.e. less than
1.0) for the selected section, which implies
that the dam is safe against piping for all the
scenarios and there is no any possibility of
internal erosion due to seepage. Estimated
seepage flux is minimum and maximum
seepage velocity is within safe limits. Cut off
wall exhibit substantive effect on dissipating
the residual head, and therefore its
effectiveness and of the core wall is
demonstrated significantly.
Acknowledgements
The authors wish to express their gratitude to
ACE Pakistan (Associated Consulting
Engineers), WAPDA Pakistan (Water and
Power Development Authority) officials
deputed at Hub dam and all other individuals
who have been source of help throughout the
research period.
References
1. M. J. Kazem Zadeh, F. Daneshmand
(2012) Unconfined Seepage
Analysis in Earth Dams Using
Smoothed Fixed Grid Finite Element
Method. International Journal for
Numerical and Analytical Methods in
Geomechanics Volume 36, Issue 6,
pages 780 797.
2. SEEP/W Engineering E-Book (May
2007), Seepage Modeling with
SEEP/W 2007 An Engineering
Methodology GEO-SLOPE
International Ltd.
3. A. Kamanbedast, M. Shahosseini
(2011), Determination of Seepage

'0

and Analysis of Earth Dams (Case
Study: Karkheh Dam). Iranica
Journal of Energy and Environment 2
(3): 201-207, ISSN 2079 2115.
4. LI Quanshu, LIU Jianjun (2010),
Numerical Analysis of the Seepage
Field in Core-Dam. School of Civil
Engineering and Architecture,
Southwest Petroleum University,
Chengdu, China, 610500, Pages 492
499.
5. T. A. Mohammed, B. B. K. Huat, A.
A. Aziz, H. Omar, S. Maail, (2006)
Seepage through Homogenous and
Non-Homogenous Earth Dams:
Comparison between Observation
and Simulation. Electronic Journal
of Geotechnical Engineering ;
www.ejge.com.
6. Geo-Slope Software Example
(2007), Seepage through A Dam
Embankment. Geo-Slope
International Ltd, Calgary, Alberta,
Canada.
7. WAPDA, (2009), 4th Periodic
Inspection Report of Hub Dam,
ACE WAPDA.
8. Williams, E., (1986), "Seepage
Analysis and Control for Dams",
Department of the Army U.S. Army
Corps of Engineers, Washington.
9. Salama Nasim Yousif Al-Labban
(May 2007), Seepage Analysis Of
Earth Dams By Finite Elements. A
Thesis Submitted To The College Of
Engineering Of University Of Kufa
In Partial Fulfillment For The
Requirements Of The Degree Of
Master Of Science In Civil
Engineering (Geotechnical
Engineering).
10. S. A. A. Khattab (2010) Stability
Analysis of Mosul Dam under
Saturated and Unsaturated Soil
Conditions. Al- Rafidain
Engineering Journal Vol. 18 No. 01.
11. Benmebarek, N., Benmebarek, S. and
Kastner, R., (2005), "Numerical
Studies of Seepage Failure of Sand
within a Cofferdam. Computers and
Geotechnics,
www.sciencedirect.com.

Comparison of SEEP - W Simulations With Field Observations For Seepage Analysis Through An Earthen Dam (Case Study - Hub Dam - Pakistan)

Uploaded by

Copyright:

Available Formats

Comparison of SEEP - W Simulations With Field Observations For Seepage Analysis Through An Earthen Dam (Case Study - Hub Dam - Pakistan)

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Comparison of SEEP - W Simulations With Field Observations For Seepage Analysis Through An Earthen Dam (Case Study - Hub Dam - Pakistan)

Uploaded by

Copyright:

Available Formats

Comparison of SEEP/W Simulations with Field Observations for Seepage Analysis through

an Earthen Dam Case Study! "ub Dam # Pa$istan%

You might also like