See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/237376360

Effectiveness of seepage control elements for embankments on semipervious

foundations

Article in Canadian Geotechnical Journal · January 2011

DOI: 10.1139/t81-067

CITATIONS READS

8 101

4 authors, including:

Jean-Pierre Tournier
Hydro-Québec
25 PUBLICATIONS 137 CITATIONS

SEE PROFILE

All content following this page was uploaded by Jean-Pierre Tournier on 06 July 2016.

The user has requested enhancement of the downloaded file.

572 CAN. GEOTECH. J. VOL. 18, 1981

McCarthy, P.Eng. of Western Caissons Ltd., for PECK,R. B. 1969. Deep excavations and tunnelling in soft
information and assistance related to this study. ground. Proceedings, 7th International Conference on Soil
Mechanics and Foundation Engineering, Mexico, State-
of-the-Art Volume, pp. 225-290.
SANDEGREN, E., SAHLSTROM, P. O., and STILLE,H. 1972. Be-
MCROSTIE, G. C., and SCHRIEVER, W. C. 1967. Frost pressures haviour of anchored sheet pile exposed to frost action.
in the tie-back system at the National Arts Centre excava- Proceedings, 5th European Conference on Soil Mechanics
tion. Engineering Journal, 50(3), pp. 17-22. and Foundation Engineering, Madrid, Vol. 1, pp. 285-291.
PAPPAS,N. D., and SEXSMITH, D. P. 1968. Performance moni- STILLE,H. 1976. Behaviour of anchored sheet pile walls.
toring of a deep cofferdam in sensitive clay. Canadian Doctoral thesis, Royal Institute of Technology, Stockholm,
Geotechnical Journal, 5, pp. 80-96. Sweden, 192 p.

Effectiveness of seepage control elements for embankments on

semipervious foundations
GUYLEFEBVRE, LUPIEN
CLAUDE
Ddpartement de Gdnie Civil, Universitd de Sherbrooke, Sherbrooke, Quk., Canada J I K 2R1

Socidtk d'dnergie de la Baie James, 800 est, bod. de Maisonneuve, Montrdal, Qud., Canada H3L 4M8
AND

TOURNIER
JEAN-PIERRE
Lalonde, Girouard, Letendre et ass. Ltke, 1400 ouest, rue Sauvd, suite 214, Montrdal, Qud., Canada H4N 1C5
Received March 30, 1981
Accepted April 27, 1981

Results of seepage analyses for two dams on semipervious foundations obtained with finite
element techniques are presented in order to assess the relative effectivenessof seepage control
elements in reducing the downstream exit gradient. After presenting results to illustrate the
effect of permeability anisotropy, the effectivenesses of a downstream relief trench, an upstream
impervious blanket, a partial cut-off, and a more pervious drain material are compared in terms
of the factor of safety against piping at the downstream toe.

Des rCsultats d'analyse d'koulement obtenus par Cltments finis pour deux barrages construits
sur des fondations semi-permiables sont prCsentCs afin de comparer l'efficacitk relative de diffk-
rents ClCments pour rCduire le gradient de sortie au pied aval. Aprks la prksentation de rCsultats
pour illustrer l'effet de l'anisotropie de perrnkbilitk, l'eRicacitC d'une tranchCe de dicharge
aval, d'un tapis amont, d'une coupure partielle et d'une plus grande permtkbilitC des zones
drainantes est comparCe sur la base du facteur de sCcuritC contre la boulance au pied aval.
Can. Geotech. J., 18,572-576 (1981)

Introduction Modern methods of analysis using the finite element

The prediction of water pressure distribution asso- technique may remain questionable in view of the
ciated with seepage through embankments and their difficulty of correctly assessing the hydraulic con-
foundations is important in dam design in order to ductivity of the foundation and the various zones of
ensure structure stability against internal erosion. the embankment. These methods can, however, be
Conventional seepage analysis does not fully take used to evaluate parametrically the influence of the
into account the complexity of multi-zones or hor- various factors on seepage and can help, for example,
izons with high permeability contrasts. Considerable in selecting the most effective seepage control feature
experience has led to the development of numerous for a particular project.
empirical approaches for seepage analyses (Bligh The purpose of this note is to present some solu-
1910; (Lane 1935; Turnbull and Mansur 1959,1961). tions for seepage through and under embankments
0008-3674/8 1/040572-05~01.00/0 .:-
@ 1981 National Research Council of Canada/Conseil national de recherche~du Canada
 NOTES 573

to show the relative effectiveness of common design

solutions for the control of exit gradients, i.e., a
downstream relief trench, upgraded drain material,
an upstream impervious blanket, and a partial
cut-off.
Method of analysis
The finite element technique, developed and pro-
grammed by S. P. Neuman and P. A. Whiterspoon
(1970) for problems of steady state seepage with a
free surface, and applicable to heterogeneous porous
media with complex geometric boundaries and arbi-
trary degrees of anisotropy, has been used. The
method can handle problems where the free surface
is discontinuous and where portions of the free
surface are vertical or nearly vertical. The method
and the program had been extensively tested by
Neuman and Whiterspoon (1970) against analytical FIG.1. Case 1: (a) geometry and permeability assumptions;
(b) finite element mesh. NOTE:A = anisotropy ratio.
solutions, experimental models, and published re-
sults. The version of the program used for the present
study was also verified using published results.
The seepage has been analysed for two embank-
ments on semipervious foundations incorporating
different seepage control features. The cases analysed
are compared on the basis of the factor of safety with
conditions of zero effective stress at the downstream K, = 5 x 10-~crn/s
2 0 rn K V = KH/9
toe. For the purpose of the study, the factor of safety
is defined as I
t
v-
,\ ,
,
, \\\ *, ,,
IMPERVIOUS
F.S. = i,/i = l/i
where i, = the critical gradient (assumed to be unity
in this study), and i = the vertical component of the
exit gradient.
Geometry and soil properties
The first case consists of a homogeneous 9.5 m
FIG.2. Case 2: (a) geometry and permeability assumptions;
high dike with a chimney drain, a hydraulic head of (b) finite element mesh.
6.4m, resting on a relatively pervious two-layer
foundation. As often observed in natural deposits,
pervious horizon in the foundation was simulated.
the upper layer has been assumed to be more per-
Four analyses of case 1 were conducted using the
vious than the lower layer by a factor 5. The geometry
same mesh and the same properties (Fig. la,b) except
and the permeability assumptions are presented in
that the anisotropy ratio of the permeability in the
Fig. l a and the finite element mesh in Fig. Ib. The
upper layer of the foundation was varied from 1 to
second case consists of a 20m high zoned dam,
100. This was achieved by varying the vertical per-
resting on a 20 m thick till foundation. Geometry
meability while keeping the horizontal permeability
and permeability assumptions are presented in Fig.
constant. The minimum factor of safety against
2a and the finite element mesh in Fig. 2b.
heaving at the downstream toe of the embankment
SigniJicance of anisotropy has been evaluated for each analysis and is presented
Assumptions about the anisotropy ratio of the in Fig. 3.
hydraulic conductivity in the foundation are known Seepage characteristics of a semipervious founda-
to have a tremendous effect on seepage analysis tion condition (essentially similar to case 2) with an
results. The problem is illustrated by two series of upstream till blanket under the shell and a 1.5 m
analyses; first the anisotropy ratio was varied in the thick pervious lens or horizon located at mid-depth
upper part of the foundation and secondly a more in the foundation were also analysed and the results
 CAN. GEOTECH. J. VOL. 18. 1981

v
-
H
UPPER LAYER

/
/
+ISOTROPIC CONDITIONS IN T H E
/ UPPER LAYER
/

U
e I I I I I
10 2 0 3 0 4 0 5 0 6 0 7 0 8 0 90 100
I I I
R A T I O O F HORIZONTAL T O V E R T I C A L H Y D R A U L I C ANISOTROPY RATIO O F 2 5 I N T H E
CONDUCTIVITY I N T H E UPPER L A Y E R ( K H / K V )

FIG.3. Illfluence of anisotropy on permeability, case 1.

u
$ 1

0
I
10 20 30
I
40
I
50 60
I I I
70
I -(m)
UPSTREAM B L A N K E T LENGTH
L VARIABLE 0 1 2 3 4 5 6 7 8 9 1 0 1 1
KH = 5110~~ern15 I I , , I , I I I 1 ) ,L/H
K V = K H 19 K = I -x I
U P S T R E A M B L A N K E T L E N G T H (L)
HYDRAULIC HEAD (H)
LENGTH OF THE PERVIOUS LENS [ml FIG.5. Effectiveness of an upstream impervious blanket,
125 100 75 50 25 case 1.
I I I I I
Effectiveness of clrainage f e a t ~ ! r . ~tos control seepage
Design of a dam should incorporate one or several
elements to control seepage through the embankment
and foundation. A downstream drain of some kind
is inherent to any design while such features as an
upstream blanket, partial cut-off, or downstream
relief trench depend on local site conditions and
other design considerations. It is therefore important
for the designer to evaluate as accurately as possible
FIG.4. Influence of a more pervious lens in a homogeneous the effectiveness of all possible seepage control fea-
foundation, case 2. tures in order to arrive at the technically and eco-
nomically optimum solution.
Most of the results presented to illustrate the
are presented in Fig. 4. The length of the horizon, effectiveness of various drainage features, generally
starting under the downstream toe of the embank- used for seepage control, refer to case 1 (Fig: la).
ment and progressing toward the upstream side, has
been varied from 0 to 120 m. Figure 4 shows that the Upstreanz impervious blanket
factor of safety decreases significantly as the length The effectiveness of an upstream impervious blan-
of the pervious horizon increases until the horizon ket in controlling the exit gradients at the toe of a
reaches the upstream toe of the impervious zone of dam, for both isotropic and anisotropic conditions
the embankment. The analysis considered only one of the upper foundation layer, was analysed by vary-
depth for the pervious horizon and one should expect ing the length of the blanket. The factor of safety
more critical conditions for shallower horizons or against heaving at the downstream toe is presented in
for a horizon rising upwards near the downstream Fig. 5 as a function of the length expressed both in
toe. metres and as its ratio to the hydraulic head.
 NOTES

IH T" UPPER LAYER

*, ISOTROPIC CONDITIONS = B I I I I 1 -
.- /
/
I N T H E UPPER LAYER
20 40
DEPTH
60 80
OF T R E N C H
100

THICKNESS O F U P P E R LAYER (%'

B 8

* 3

2 - * -
/
. *, A N I S O T R O P Y OF 2 5
I N T H E UPPER LAYER

25 50 75 100 D/H (%I

PENETRATION R A T I O I N T H E
U P P E R L A Y E R D / H (%I

FIG.6. Effectiveness of a partial cut-off trench, case 1. u 2

0
[L
I I I I I
0
I- 20 40 60 80 100
u DEPTH OF TRENCH
2 THICKNESS OF FOUNDATION (%'
FIG.8. Effectiveness of a relief trench: (a) case 1 ; (b) case 2.

hydraulic conductivity of the drain to the foundation

varied from 1 to 10 000. An anisotropy ratio of 25
was assumed for the upper layer of the foundation
while the drainage blanket was considered isotropic.
Relief'trerzches at the clownstream toe
L I I I I I I -
I I0 100 I000The depth of a relief trench at the downstream toe
l0000
Kv D R A I N of case 1 was varied from zero to the mid-depth of
K v FOUNDATION the upper layer (5.5 m). The variation of the factor of
FIG.7. Influence of hydraulic conductivity of drain material, safety against heaving is presented in Fig. 8a as a
case 1. function of the ratio of the depth of the relief trench
to the thickness of the upper layer. An anisotropy
Partial c~rf-off ratio of 25 was assumed for the upper layer.
Figure 6 shows the influence of a cut-off trench in As previously mentioned, the upper layer of the
the upper pervious layer of the foundation both for foundation in case 1 is five times as pervious as the
isotropic conditions and an anisotropy ratio of 25. bottom layer. In order to show the influence of a
The depth of the cut-off trench has been varied fro111 relief trench in a one-layer foundation, analyses were
zero to the bottom of the upper layer. conducted for case 2, varying the depth of the trench
Pertnenbility of rhain n7nterinls from zero to 80% of the thickness of the foundation
layer. An anisotropy ratio of 9 was assumed for the
A downstream drainage blanket should be de-
foundation. The results are presented in Fig. 8b.
signed to be very pervious relative to the surrounding
materials. The influence of the permeability of the Relative eflectiveness o f ' s e e ~ ~ a control
ge eleinents
draining material is illustrated in Fig. 7 where the The results of analyses to compare the effectiveness
hydraulic conductivity in the downstream drainage of various elements of seepage control confirm the
blanket is varied in order to have the ratio of vertical extreme importance of anisotropy. To illustrate the
 576 CAN. GEOTECH. 1. VOL. 18, 1981

relative effectiveness, most analyses have been con- illustrated by the analyses presented here that as-
ducted on case 1, keeping all conditions similar except sumptions about the anisotropy ratio influence dras-
for the variation of the considered seepage control tically the results of seepage^analyses. Considering
element. the difficulties in assessing accurately the anisotropy
When the embankment of case 1 has, as seepage of hydraulic conductivity in natural foundations, the
control element, only a downstream drainage blanket results of any seepage analyses, irrespective of the
having a hydraulic conductivity of 1 X cm/s, so~histication of the methods used. will remain
the factor of safety against heaving at the downstream affected by a significant margin of error. The sensi-
toe is 1.6 for an anisotropy ratio of 25 in the upper tivity of the results to the assumptions can, however,
layer of the foundation. The factor of safety rises be evaluated as shown in Fig. 3. Under these condi-
only to 2.6 for an upstream blanket length of 65 m, tions, the results of seepage analyses are
i.e., 10 times the hydraulic head. A partial upstream likely to prove more helpful to the designer in
cut-off trench extending through the whole upper evaluating the relative effectiveness of various design
layer (5.5 m) increases the factor of safety to 2.9. The elements and thus optimizing his choice.
factor of safety increases from 1.6 to 2.6 when the
drainage blanket conductivity is increased by a factor
of 1000. However, a downstream relief trench ap- Acknowledgement
pears to be much more effective than other types of The results presented in this article were part of a
protection. In fact a relief trench only 1.8 m deep study conducted for the Socittt d'tnergie de la Baie
brings the factor of safety from 1.6 to 3.1, and to 3.7 James. The authors are grateful to SEBJ for their
for a trench 2.7 m deep. The effectiveness of the relief permission to publish this paper.
trench is, however, enhanced by the two-layer found-
ation, and to check the effectiveness in a so-called
homogeneous foundation, one should look at the BLIGH,W. G. 1910. Dams, barrages, and weirs on porous
foundations. Engineering News, 64, pp. 708-710.
analysis of case 2 (Fig. 8b), where the foundation LANE,
E. W. 1935. Security from under-seepage-masonry dams
consists of one 20 m layer with an anisotropy ratio on earth foundations. Transactions of the American Society
of 9. The factor of safety changes from 2.7 to 4.8 of Civil Engineers, 100, pp. 1235-1351.
when the depth of the relief trench is increased from LEFEBVRE, G., and LUPIEN, C. 1978. Analyse et interprktation
des relevis des piCzomktres au barrage Choinikre 2 Savage
1.5 to 6 m. Mills. Rapport GEO-78-02, DCpartement de genie civil,
UniversitC de Sherbrooke, presented to the Ministbre des
Conclusions richesses naturelles de la Province de Quebec.
Natural deposits most often show some kind of NEUMAN, S. P., and WHITERSPOON, P. A. 1970. Finite element
method of analysing steady seepage with a free surface.
bedding or stratification, allowing water to flow more Water Resources Research, 6, pp. 889-897.
easily in the horizontal direction. Back-calculation TURNBULL, W. J., and MANSUR, C. I. 1959. Design of under-
of seepage, using the same method of analysis, at a seepage control measures for dams and levees. ASCE Jour-
dam site where numerous piezometric readings were nal of the Soil Mechanics and Foundations Division,
available after reservoir filling, have indicated an 85(SM5), pp. 129-159.
1961. Investigation of under-seepage-Mississippi
anisotropy ratio as high as 100 in the foundation River levees. Transactions of the American Society of
(Lefebvre and Lupien 1978). It is well known and Civil Engineers, 126(1), p. 1429.

View publication stats