Professional Documents
Culture Documents
net/publication/280093677
Seepage and Slope Stability Analysis of Mandali Earth Dam, Iraq: A Case Study
CITATIONS READS
9 6,008
4 authors, including:
Bakenaz A. Zeidan
Tanta University
135 PUBLICATIONS 417 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Egyptian Coastal Lakes and Wetlands - I and II - Contributed book - published by Springer International Publishing during late 2018 or early 2019 View project
All content following this page was uploaded by Bakenaz A. Zeidan on 04 June 2017.
PRAGUE
2017
Irrigation and Hydraulic Engineering, Faculty of Engineering -Tanta University, Tanta, Egypt
ab.zeidan@f-eng.tanta.edu.eg
ABSTRACT
Failure of earthen dams may occur due to different reasons such as; structural instability, excessive
hydraulic gradients, seepage through dam body or dam foundation, rapid drawdown and seismic
excitations. The factor of safety for the dam slope stability, under different cases of operations, is a
vital indicator to ascertain the dam overall safety. In this paper, 2-D Finite Element model is employed
for simulating seepage and slope stability analysis of earth dam problems via GeoStudio code. Four
analytical methods are used to assess the stability of the dam side slopes. Phreatic seepage surface,
pore water pressure distribution and total hydraulic head variation of earth dams are obtained and
analyzed for three cases of operation under steady state conditions. Three typical cases of failed
earth dams are considered; Fontenelle dam (USA), Carsington dam (England), and Walter F. George
dam (USA). Benchmark safety regulation criteria of United States of the Army of Crops of Engineers
(USACE) and British Dam Society (BDS) are applied. The obtained results confirm that for typical
cases of failed dams, failure is attained due to either seepage or slope instability or combined
seepage and slope instability. The case of rapid drawdown is the most critical operating case;
compared to other cases of operation. Scenarios for redesign of failed dams to prevent failure are
presented and discussed.
July 3–7, 2017 th Annual Meeting
ICOLD – CIGB
PRAGUE
2017
1. INTRODUCTION
Dams are built for specific functions such as water supply, irrigation, flood control and hydroelectric
power generation. Most of the large dams in the world were built during the middle decades of the
twentieth century. There are two types of modern dams, namely: embankment dam and concrete
dam. Embankment dams can be classified into two main categories earth-fill dams and rock-fill dams.
Embankment dams represent about 85% of dams all over the world. There are several factors to be
considered in selecting an earth dam type such as: topography; foundation conditions; environmental
impacts, construction facilities and socio-economic studies. A feasible dam should be; built from locally
available materials; stable under all operating and loading conditions; watertight enough to control
seepage; have appropriate outlet works to crest dam overtopping [1-3].
2. PROBLEM STATEMENT
In the present work, 2-D Finite Element Modeling is employed for simulating combined seepage and
stability analysis of reservoir earth dam problems via GeoStudio software. Phreatic seepage surface,
pore water pressure distributed and total head hydraulic head variation of earth dams are obtained
and analyzed. Typical cases of study of well-known international earth dams are analyzed for three
cases of operation for slope stability analysis. Analytical methods are used to verify dams´ slope
stability. The developed model is verified with acceptable accuracy. Three typical cases of failed dams
are analyzed; Fontenelle dam (USA), Walter F. George dam (England). Benchmark safety regulation
criteria of (USACE and BDS) are applied. Figure (1) shows a schematic representation for the problem
statement and boundary conditions for a typical earth dam.
35
Reservoir
30
25
20
Γ2
elev
Γ1
15
H1 Γ3
10
Ω Dam
5
0
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130
dist
Γ4 Γ4
Foundation
Γ4
3. STATE OF ART
Seepage and slope stability failures are addressed via many authors since Henry Darcy, 1856 [4], who
gave the basic law of flow through porous media. Darcy’s law was based on series of experiments
conducted in a vertical pipe filled with sand. Fellenius (1936) [5], developed the Ordinary Method of
Slices (OMS) known as Felonious or Swedish method. It is assumed that the forces acting on sides of
any slice are neglected. Zienkiewiez and Chung, 1967 [6], published the first finite element simulation
to solve the Laplace equation for steady ground water seepage., Taylor and Chow, 1976 [7], recorded
that the Finite Element Method was used to assess the potential seepage flows and uplift pressure in
the foundation rock for Bannett Dam in Canada. Kratutich, 2004 [8], used thermal mode at ANSYS
computer code to simulate numerically the case of no stationary free surface in earth dams. Karjani,
2009 [9] used Geostudio computer software to analyze Maroon dam, estimate flow net at passing
condition and slope stability factor at overall stability for different operating conditions has been
calculated. Zomorodian and Abodollahzadeh, 2010 [10], used Geostudio software to investigate the
effect of horizontal drains on upstream slope of earth fill dams during rapid drawdown. Tatewar and
Laxman N. Pawade, 2012 [11], used Geostudio software to investigate the slope stability of the 21m
July 3–7, 2017 th Annual Meeting
ICOLD – CIGB
PRAGUE
2017
high Bhimdi earth dam, by changing different parameters such as changing berm width and changing
position of filter drains. Hasani et al., 2013 [12], studied the seepage analysis in Ilam earth dam for
four mesh size in order to assess the effect of meshing on results accuracy Kirra et.al, 2015, [13]
studied the effect of combined seepage and slope stability on the failure and stability of typical earth
dams.
4. THEORETICAL APPROACH
A feasible dam should be; built from locally available materials; stable under all operating and loading
conditions; watertight enough to control seepage; have appropriate outlet works to crest dam
overtopping. This research presents an investigation to stability criteria for the dam safety against
combined seepage and slope stability failure.
𝛛 𝛛𝐇 𝛛 𝛛𝐇 𝛛 𝛛𝐇 𝝏𝒉
[𝐊 𝐱 ]+ [𝐊 𝐲 ]+ [ 𝐤𝐳 ]=𝐒 In Ω ……………. (1)
𝛛𝐱 𝛛𝐱 𝛛𝐲 𝛛𝐲 𝛛𝐳 𝛛𝐳 𝝏𝒕
where, Kx, Ky and Kz are the coefficient of permeability in x, y, z directions, respectively, S is specific
yield and 𝐇 = p/γw + z = total fluid head, P = pressure, γw = unit weight of water and z = elevation
head. Equation (1) is known as Laplace’s equation which is considered as the governing equation for
groundwater there dimensions flow through aquifers. For an isotropic, homogeneous aquifer under
steady state conditions, Equation (2) can be simplified to the following equation:
𝝏𝟐 𝐇 𝝏𝟐 𝐇
+ =𝟎 In Ω ……………………………….................………………….. (2)
𝝏𝒙𝟐 𝝏𝒛𝟐
It is assumed that; the soil media is homogeneous, isotropic, and physically stable; the pressure is
atmospheric everywhere on the water table (phreatic surface); the flow of ground water through the
flow domain is steady and the hydraulic conductivity through the dam body is constant everywhere.
The boundary conditions as given by Fig. 1 are; (Γ1) Entrance Surface, (Γ2) Phreatic Surface, (Γ3) Exit
Surface (Seepage surface) and (Γ4) Dam Foundation Boundary. These boundary conditions can be
summarized s follows:
The upstream boundary surface (Γ1) is considered as an equipotential line which is known as Dirchlet
Condition for a prescribed head as:
The boundary surface (Γ2) is the phreatic surface of the flow through the dam, its location and its
profile are unknown a priori such that:
The boundary surface (Γ3) is the seepage surface which is considered as an isobar such that:
July 3–7, 2017 th Annual Meeting
ICOLD – CIGB
PRAGUE
2017
The boundary surface (Γ4) is the dam foundation boundary. This boundary is assumed to be
impervious such that:
∂𝐇
= 0 ………………………………………………………………………................… (6)
∂n
where n is the normal directions to the boundary (Γ4).
where, F.S. factor of safety, τf failure shear strength of the soil, and τ shear stress of the soil. The
stability of earth fill dam depends on its geometry, components, materials, properties of each
component and the forces to which it is subjected. There are some analytical tools for assessing the
stability of a slope by using simple failure models. These techniques are limited to slopes of
homogenous material, soils, fractured rocks which are behaving like soils and for plain strain (2-D)
problems only. Some of these tools are listed by [14]. For steady state seepage condition the factor of
safety using Fellenius method is [5, 15]:
𝐢=𝐧
∑𝐢=𝟏(𝐜𝐋𝐢 +(𝐖𝐢 𝐜𝐨𝐬 𝛂𝐢 −𝐮𝐢 𝐋𝐢 )𝐭𝐚𝐧𝛗)
F.S. = ……………………….........…..….....(8)
∑𝐢=𝐧
𝐢=𝟏 𝐖𝐢 𝐬𝐢𝐧𝛂𝐢
For steady state seepage condition the factor of safety using Bishop Method is:
𝐢=𝐧
𝟏
∑ (𝐜𝐛𝐢 +(𝐖𝐢 −𝐮𝐢 𝐛𝒊 )𝐭𝐚𝐧 𝛗)
𝐢=𝟏 𝒎𝜶𝒊
FS = ………………….………………......……...…..(9)
∑𝐢=𝐧
𝐢=𝟏 𝐖𝐢 𝐬𝐢𝐧𝛂𝐢
𝒕𝒂𝒏𝝋 𝐬𝐢𝐧 𝜶𝒊
where, 𝒎𝜶𝒊 = 𝒄𝒐𝒔𝜶𝒊 +
𝑭𝑺
c = effective soil cohesion, L = length of the bottom of the slice, b = width of the slice and equal to (L
cos𝜶 ), u = pore water pressure, W= weight of the slice, 𝛂 =inclination of the bottom of the slice and
𝛗 = effective internal friction angle.
Geostudio computer code [20] is employed in the present study to simulate seepage and stability
analysis of atypical earth dam. It is a Finite Element based code. The basic concept of the finite
element method is to divide the problem region into subdomains (finite elements) connected at their
common nodal points and that the unknown function of the field variable is defined approximately
July 3–7, 2017 th Annual Meeting
ICOLD – CIGB
PRAGUE
2017
within each element. The approximate solution of each element expressed by continuous function is
as follows [3]:
e 𝐞
𝐇 = ∑noe
e=1 [Ni ] {𝐇𝐢 } ……………………………………..................……..… (11)
where, 𝐇𝐢𝐞 = nodal value of (H) for ith node in element (e), noe = total number of elements and Nie =
shape function of element (e).
There are more different approaches to formulate the approximate solution of the problem. In the
present study, the standard weighted residual method with Galerkin’s criterion [6] is used to
approximate the solution of the unknown variable (H). Thus, equation (2) can be written in a matrix
form as [3]:
in which {𝐇} is the unknown nodal potential head vector, {𝐅} is the nodal external flux vector and [𝐊]
is the conductivity matrix given by:
where, V e = the domain of element (e) and Ni = interpolation or shape function. Derivations of the
above functions are given in detail by [3].
The solution of seepage problem with phreatic surface require a successive adjustment for the
location of the phreatic surface and the finite element mesh size till the desired degree of convergence
for the head H is achieved. In all iterative methods, the solution is started by using initial guess for the
unknown values and the solution is obtained by repeating the solution of the system of equations
successively through recurrence relations to update the old values until the solution converges closely
enough to the true values within some prescribed tolerance of error [3].
Figure (3) shows the finite element mesh employed for the model verification. Figures (4) present the
numerical results for water head variation, flow pattern and pore water pressure variations. Figures (5)
show the results of slope stability analysis under cases of operation.
340
320
300
July 3–7, 2017 th Annual Meeting
ICOLD – CIGB
PRAGUE
2017
260
240
420 220
elev
400 200
380 180
360 160
340
140
320
120
300 Figure (3) Finite element mesh
100
280
80
260
420
60
240
400
40
220
elev
380
30
45
50
10
40
20
20
35
5
200
25
15
360
0
180 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420
340
160 dist
320
140 Figure (4a) Water head variation through dam body and its phreatic surface
300
120
280
100
260
80
240
60
220
elev
40
200
20
180
0
0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420
160
dist
140
Figure (4b) Flow pattern through dam body
120
100
80
60
0 - 20
40
20
20
40
0
0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420
dist
Figure (4c) Pore water pressure through dam body and phreatic surface
Slope stability analysis
2.253
PRAGUE
2017
A summary of obtained results are tabulated in Table 1. Results show the factor of safety ((F.O.S))
simulation of upstream and downstream slope for all cases of operations. For upstream slope, the
absolute mean error (AME) is about 0.16 and the average absolute percentage difference (AAPD) is
about 8.8%. For downstream slope, (AME) is about 0.075 and (AAPD) is about 4.1%, a good
agreement is attained for all cases of operations.
Fontenelle dam is managed by US Bureau of Reclamation and is located on Green River southeast of
La Barge, Wyoming, USA. The dam was constructed in 1964 and is a 139 feet height zoned earth fill
embankment with a crest length of 5,450 feet. Figure (6) shows a plan of Fontenelle dam.
D.S
Figure (6b) Image for failure of Fontenelle dam (JD Rice, 2007)
July 3–7, 2017 th Annual Meeting
ICOLD – CIGB
PRAGUE
2017
An uncontrolled concrete spillway is located on the right abutment. There are three outlet works, a river
outlet is at the base of the embankment near to the center of the dam, the east canal outlet is near to
the left abutment, and the west canal outlet is near to the right abutment. The dam has a power plant
adjacent to the toe of the dam, (JD Rice, 2007). A cross section of the dam is shown in Figure (7) and
material properties of zones are given in Table (2). Zone 1 is the low permeability core which consists
of selected silt, gravel compacted in a 6 inch layer. Zone 2 is the shell material zone, which consists of
gravel, sand and cobbles which is compacted in an 8 inch layer. Zone 3 is the dam foundation layer, it
consists of layer of alluvium (silt) with thickness is 34m overlying bed rock (sandstone). The upstream
slope of the dam is inclined at 1 vertical to 3 horizontal and the downstream slope is inclined at 1
vertical to 2 horizontal. The hydraulic conductivity parameters are obtained based on the typical
conductivity parameters. Table (3) presents the dam geometrical properties. Seepage problems
became evident at Fontenelle dam when the dam nearly failed during first filling in September 1965.
The original Fontenelle Dam began to leak during its first year of operation in 1964. In 1965, there was
record winter snowfall and high spring runoff. Due to damage in the outlet works stilling basin, flows
were limited through the outlet works at this time and the spillway passed the majority of the spring
runoff event. The reservoir quickly rose five feet above the spillway crest and only two feet below the
design elevation. On the afternoon of September 3, 1965, the project construction engineer noticed
seepage exiting from about halfway up the embankment near the spillway on the right. Approximately
10,500 cubic yards of material had eroded from the embankment. Emergency measures were taken to
lower the reservoir through the outlets. A photograph of the collapsed area is presented in figure (8).
To improve the performance of the dam was the work of grout for joints of the foundation layer. The
performance of the dam remained good until January 1982, then the amount of seepage is increased
through the body of the dam and the foundation layer, which led to the use of seepage barrier deeply
into the foundation will be to know from seepage analysis of the dam.
90
80
70
60
50
elev
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280
dist
90
80
70
60
50
elev
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280
dist
90
80
70
60
50
elev
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280
dist
Figure (8a) Water head variation and flow lines through dam body
July 3–7, 2017 th Annual Meeting
ICOLD – CIGB
PRAGUE
2017
90
80
70 0
60 20
50
elev
40
40
30
20 60
10
80
0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280
dist
Figure (8b) Phreatic surface and Pore water pressure through dam body
90
80
70
1.6014e-004
60
50
elev
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280
dist
90
80
70
60
0.6
50
elev
40 0.4
0.4
30 0.2
0.2
0.2
20
10
0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280
dist
PRAGUE
2017
Table (3): Comparison Between Original Section of Fontenelle Dam and BDS Safety Limits.
80
70
60
50
elev
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280
dist
90
80
70
60
50
elev
40
30
20
10
50
85
80
55
65
0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280
dist
90
80
70
7.4079e-008
60
50
elev
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280
dist
The number of nodes is 1026 and the number of element is 1938. Figures (10) present water head
variation and pore water pressure through the dam body, respectively. Figure (11a) shows flow lines
July 3–7, 2017 th Annual Meeting
ICOLD – CIGB
PRAGUE
2017
near dam toe, in addition the water head contours near downstream slope which endanger stability of
downstream slope. Figure (11b) presents the pore water pressure through the dam body excessive at
dam downstream. Figure (11c) presents average flow rate through the dam body equal (1.6014*10-4
m3/sec/m). Table (4) and Figure (12) give the hydraulic safety criteria. It can be noticed that the
hydraulic safety criteria based on (USBR) is not acceptable, (USBR, 2014). Figures (12) present the
relationship between horizontal and vertical hydraulic gradient with the distance (X) from the base of
the dam. It can be noticed that the average value of horizontal hydraulic gradient in core zone = 0.53
and the average value of vertical hydraulic gradient in core zone = 0.52. The hydraulic gradient at exit
of the dam for horizontal and vertical direction are 0.4 and 0.3 respectively, these value are significant.
It is clear that the dam failure occurred at dam downstream due to excessive seepage through dam
body and dam foundation, so it is a seepage failure type.
7. CONCLUSIONS
In this work, 2-D Finite Element Modeling is employed for simulating combined seepage and stability
analysis of reservoir earth dam problems via GeoStudio code. Phreatic seepage surface, pore water
pressure distributed and total head hydraulic head variation of earth dams are obtained and analyzed.
Combined seepage and slope stability analysis of reservoir earth dam conducted in the present
study assess the following conclusions:
Rapid drawdown condition is the critical condition for dam slopes instability.
To prevent seepage failure in earth dams:
It is recommended to impose clay core or diaphragm wall to lower phreatic surface
significantly away from the downstream slope as well as flow lines.
It is recommended to employ drainage facilities, filters, to avoid concentration of high
pore water pressure near downstream slope.
The value of exit hydraulic gradient for failed dams are higher than the value of exit
hydraulic gradient for operating dams for (X/ B) ≥ 0.75, this reason lead to seepage or
piping failure.
The seepage quantity of Fontenelle dam is decreased from 13.84 m3/day/m to 0,0064
m3/day/m by using concrete diaphragm wall with penetration depth 34m through the
foundation layer.
8. REFERENCES
[1] M. Elshemy, R.I. Nasr, M.M. Bahloul and I.M. Rashwan, The effect of blockages through earth
dams on the seepage characteristics, Faculty of engineering, Tanta University, Egypt, (2002).
July 3–7, 2017 th Annual Meeting
ICOLD – CIGB
PRAGUE
2017
[2] M. A. M. Ismail, S. Min Ng and K. Gey, Stability Analysis of Kelau Earth-Fill Dam Design under
Main Critical Conditions, Malaysia, The Electronic Journal of Geotechnical Engineering (EJGE),
(2012).
[3] Bakenaz A. Zeidan, A Numerical (FEM) Study of the Effect of Anisotropy on Phreatic Seepage
Flows, PhD Thesis, Civil Engineering Department, Indian Institute of Technology IIT, Powai, Bombay,
India, (1993).
[5] Lambe, T. W., and Whitman, R. V., Soil Mechanics, John Wiley and Sons, Inc., SI Version, New
York, U. S. A, (1979).
[6] Zienkiewicz, O.C. and Taylor, R.L., The Finite Element Method; Volumes I, II”, 5th Edition, First
Published In (I967) By McGraw-Hill.
[7] National Water and Climate Center, Watershed Science Institute, EM 1110-2-1901, Sep., (1986).
www.wcc.nrcs.usda.gov/watershed/piedmont/a-a.pdf/
[8] Kratochvil, J., Numerical modeling of Non stationary Free Surface Flow in Embankment Dams,
Brno University of Technology CZ, (2004).
[9] A. Kamanbedast and A. Delvari, Analysis of Earth Dam: Seepage and Stability Using Ansys and
Geo-Studio Software, Iran, World Applied Sciences Journal 17 (9): 1087- 1094, (2012).
[10] S.M. Zomorodian and S.M. Abodollahzadeh, Effect of Horizontal Drains on Upstream Slope
Stability During Rapid Drawdown Condition, Shiraz University, Iran, International Journal of Geology,
Issue 4, Volume 4, (2010).
[11] S.P. Tatewar and Laxman N. Pawade, Stability Analysis of Earth Dam by Geostudio Software,
India, International Journal of Civil Engineering and Technology (IJCIET), Volume 3, Issue 2, July-
December (2012).
[12] H. Hasani, J. Mamizadeh and H. Karimi, Stability of Slope and Seepage Analysis in Earth Fills
Dams Using Numerical Models (Case Study: Ilam Dam),Iran, World Applied Sciences Journal 21 (9):
1398-1402, (2013).
[13] M.S. Kirra, M. Shahien, M. Elshemy, B. A. Zeidan (2015) “Seepage and Slope Stability Analysis
of Mandali Earth Dam, Iraq: A Case Study” International Conference on Advances in Structural and
Geotechnical Engineering, ICASGE’15, 6-9 April 2015, Hurghada, Egypt.
[14] M.S. Kirra, B. A. Zeidan, M. Shahien, M. Elshemy (2015) “Seepage Analysis of Walter F. George
Dam, USA: A case Study” International Conference on Advances in Structural and Geotechnical
Engineering, ICASGE’15, 6-9 April 2015, Hurghada, Egypt.
[16] FERC, (1991), Chapter IV, Embankment Dams, Federal Energy Regulatory Commission available
at: http://www.ferc.gov/industries/hydropower/safety/guidelines/eng-guide/chap4.PDF
[17] Ismaeel, Kh. S., Seepage and Stability Evaluation of Duhok Dam, M. Sc. Thesis, College of
Engineering, University of Duhok, Iraq, (2006).
[18] USBR, United State Department of interior Bureau of Reclamation, Design Standard DS-13(4),
Embankment Dams, Static Stability Analysis, Chapter (4), October (2011).
[19] NRCS, Technical Release No. 60, Earth Dams and Reservoirs, Natural Resources Conservation
Service, (2005); available at www.info.usda.gov/CED/ftp/CED/TR_210_60_Second_Edition.pdf
July 3–7, 2017 th Annual Meeting
ICOLD – CIGB
PRAGUE
2017
[20] ULDC, Urban Levee Design Criteria, Engineering criteria and guidance for the design,California
Department of Water Resources, May (2012).
[21] USACE, Slope Stability, Engineering Manual 1110-2-1902, Department of the Army, Corps of
Engineers, Washington DC, United States of America, (2003); available at
www.usace.army.mil/inet/usacoe-docs/eng-manuals/em1110-2-1902/entire.pdf
[22] BDS, The British Dam Society at the Institution of Civil engineers, Great George Street, London,
SW1P 3AA, (1994). http://britishdams.org/conferences
[23] GEO-SLOPE INTERNATIONAL . Ltd, Calgary, Alberta, Canada. T2p 2Y5, (2004). http:// www.
Geo-Slope. com
[25] Taylor & Francis / Balkema, Look, B. G. Handbook of Geotechnical Investigation and Design
Tables, London, UK, pp. 91, (2007).