World Academy of Science, Engineering and Technology

International Journal of Civil and Environmental Engineering

Vol:11, No:8, 2017

The Behavior of Dam Foundation Reinforced by

Stone Columns: Case Study of Kissir Dam-Jijel
Toufik Karech, Abderahmen Benseghir, Tayeb Bouzid

 of low mechanical characteristics, which required treatment by

Abstract—This work presents a 2D numerical simulation of an stone columns [3].
earth dam to assess the behavior of its foundation after a treatment by The alluvial foundation is treated with stone columns under
stone columns. This treatment aims to improve the bearing capacity, the influence of the dike to consolidate the soil in order to:
to increase the mechanical properties of the soil, to accelerate the
• improve the bearing capacity of the foundation of the
consolidation, to reduce the settlements and to eliminate the
liquefaction phenomenon in case of seismic excitation. For the dam.
evaluation of the pore pressures, the position of the phreatic line and • increase the equivalent mechanical characteristics of the
reinforced soil.
Open Science Index, Civil and Environmental Engineering Vol:11, No:8, 2017 publications.waset.org/10008631/pdf

the flow network was defined, and a seepage analysis was performed
with the software MIDAS Soil Works. The consolidation calculation • reduce horizontal and vertical deformations of the
is performed through a simulation of the actual construction stages of foundation of the dam.
the dam. These analyzes were performed using the Mohr-Coulomb
• accelerate the consolidation of the main layer, thanks to
soil model and the results are compared with the actual measurements
of settlement gauges implanted in the dam. An analysis of the bearing the vertical draining effect of stone columns.
capacity was conducted to show the role of stone columns in • eliminate the liquefaction phenomenon of fine sand in
improving the bearing capacity of the foundation. case of seismic solicitation.
The treatment has affected a total area of 28,000 m2; the
Keywords—Earth dam, dam foundation, numerical simulation, theoretical depth of columns is 22.5 m for the central part, and
stone columns, seepage analysis, consolidation, bearing capacity. 17 m for the sides [5]. The water tightness of the alluvial
foundation is provided by a diaphragm wall; the sealing edge
I. INTRODUCTION is provided by a grout curtain.

C ONSTRUCTION of stone columns under an embankment

is a common economical ground improvement method
[1], [2], numerical modeling is a simple and efficient
The levels and main dimensions of the dam are as [3]:
• Peak level: 48.25 NGA,
• Highest water level: 47.90 NGA,
alternative for the approach of real behavior of soils reinforced • Full water level: 44.5 NGA,
by stone columns [2]. Representing a 3D stone column grid in • Maximum height on natural terrain: 48 m,
2D is usually represented by a series of parallel trenches. The • Peak width: 8 m,
stiffness as well as the permeability of both soft soil and • Crest length: 368 m.
coarse grained inclusion needs to be adapted in order to model
the deformation behavior and drainage conditions for III. GEOTECHNICAL DATA
consolidation correctly [1]. The purpose of this work is to The levee is an earth dike with a symmetrical clayey core;
verify if the results obtained by calculating at the level of the its height is 50 m; the outer slopes vary from 2H/1V to 3H/1V
reinforced foundation by ballasted columns are concordant upstream and 2H/1V to 2.5H/1V downstream.
with those measured in situ after the commissioning of the The alluvial foundation is composed of generally fine
dam and to reduce the settlements and the risk of liquefaction materials, whose thickness may reach 40 m. Analysis of
[3]. surveys and tests led to the consider the foundation as a
homogeneous material [3]. The alluvial foundation is treated
II. PROJECT DESCRIPTIONS AND GENERAL DESIGN by a network of stone columns) with a treated surface of
Kissir Dam is located 14 km west of Jijel (Algeria); it is 28.000 m2, the columns are arranged in a hexagonal mesh with
intended to increase the resources of drinking water, industrial a space step of 3 m. The theoretical depth of columns is 22.5
and irrigation, ensuring a supply of 48 million m3 per year [4]. m (or refusal of substrate) for the central part and 17.5 m (or
The levee consists of upstream and downstream refill of refusal of substrate) for the side parts. Fig. 2 presents a
alluvium, a central clay core protected by a filter upstream and geotechnical cross-section of the foundation.
a filter and a drain downstream. It is based on a thick alluvial The mechanical properties of the foundation and fill
filling of low mechanical characteristics [4]. The typical cross- materials are summarized in Table I [3].
section of the levee is shown in Fig. 1.
The foundation of the dam consists of a thick alluvial filling IV. CONSTRUCTION SEQUENCE
The works began in June 2006 and were completed in late
Pr Toufik Karech, Mr Abderahmen Benseghir, and Dr Tayeb Bouzid are
2010. The levee was constructed between May 2008 and July
with the University of Batna 2, Department of Civil Engineering, Algeria (e-
mail: [email protected]). 2009.

International Scholarly and Scientific Research & Innovation 11(8) 2017 1187 ISNI:0000000091950263
 World Academy of Science, Engineering and Technology
International Journal of Civil and Environmental Engineering
Vol:11, No:8, 2017
Open Science Index, Civil and Environmental Engineering Vol:11, No:8, 2017 publications.waset.org/10008631/pdf

Fig. 1 Typical cross-section of the levee (PK 200.000)

TABLE I
MECHANICAL PROPERTIES OF THE FOUNDATION AND FILL MATERIALS
Specific weight Friction angle Cohesion Young’s modulus Poisson’s ratio Hydraulic conductivity
γ (kN/m3) φ' (degree) C’ (kPa) E (MPa) ν K (m/s)
Alluvial foundation 20 25 0 10 0.3 10-5
Alluvium refills 21 38 0 40 0.3 10-3
Clayey core 19.50 25 0 20 0.3 10-9
Filter 21 38 0 40 0.3 10-3
Drain 21 38 0 40 0.3 10-2
Rockfill 21 38 0 40 0.3 10-2
Stone columns 21 38 0 60 0.3 10-2

Fig. 2 Geotechnical cross-section of the alluvial foundation

Alluvial embankments were compacted by 50 cm thick Fig. 3 Rise of embankment

layers; core clay materials were compacted by 30 cm thick
layers. The rise of the embankment is synthesized in Fig. 3. Besides the reinforcement due to the characteristics of the
Water filling began in October 2009 and the normal water stone columns, the vibratory action during the installation of
level was reached in May 2010 [4]. the columns densifies the soil and causes an increase of the
The volume of the columns used is measured in the Free mechanical properties of the surrounding soil and
State without compaction. According to the specifications of consequently the reduction of settlements [2], [5], [7], [8]. As
DTU 13.2 [6], the volume must be at least 1.5 times the this last effect is not yet clearly quantified, it is not taken into
theoretical volume of the column used. This is the case for account in the numerical model.
90% of columns as measured in the field [4], Moreover, due to
the ballast expansion at the end of compaction, a mean V. NUMERICAL ANALYSIS
increase of the column diameter of the order of 10% is
A. Presentation of the 2D Numerical Model
observed. This increase of diameter is taken into account in
the numerical model. The numerical models are built according to the typical

International Scholarly and Scientific Research & Innovation 11(8) 2017 1188 ISNI:0000000091950263
 World Academy of Science, Engineering and Technology
International Journal of Civil and Environmental Engineering
Vol:11, No:8, 2017

cross-section of the dam, considering two cases for the thick layer; the stone columns are 22.5 m long and 0.88 m
foundation: untreated foundation and foundation treated with thick, with a spacing of 3 m.
stone columns. The alluvial foundation is modeled by a 40-m

Fig. 4 2D Numerical model

Open Science Index, Civil and Environmental Engineering Vol:11, No:8, 2017 publications.waset.org/10008631/pdf

Fig. 5 Steady state – pore pressure

Fig. 6 Steady state - flow network - untreated foundation

Due to 2D modeling of the 3D network of stone columns, B. Seepage Analysis

the substitution ratio is increased from 8.6% in 3D case to This analysis allows to assess the pore pressures and the
29.33% in 2D model. One way of proceeding without determination of the saturation line and flow network in steady
affecting the 2D geometry is to reduce the rigidity of the state of full water level 44.5 m using Darcy's law to express
columns in the ratio of 8.6/29.33 = 3.4 to restore the rigidity the phenomenon of infiltration in saturated and unsaturated
equivalent to 3D. soils. Fig. 5 shows the pore pressure field and the position of
The numerical analyzes were performed using the software phreatic line.
MIDAS SoilWorks® for determining the position of the water Figs. 6 and 7 show the steady flow network and the
line and the software PLAXIS® for the hydro-mechanical permeability coefficient in the case of untreated foundation
analysis. and the treated foundation.
The 2D numerical model is shown in Fig.4.

International Scholarly and Scientific Research & Innovation 11(8) 2017 1189 ISNI:0000000091950263
 World Academy of Science, Engineering and Technology
International Journal of Civil and Environmental Engineering
Vol:11, No:8, 2017

Fig. 7 Steady state - flow network - treated foundation

It is observed that the flow takes the trajectories where

Open Science Index, Civil and Environmental Engineering Vol:11, No:8, 2017 publications.waset.org/10008631/pdf

permeability is significant (stone column, drain), namely, the

flow through the foundation is led by the stone columns.
C. Consolidation Analysis
A consolidation calculation was performed for the typical
cross-section. These analyzes were performed using the
elasto-plastic model of Mohr-Coulomb, by a simulation of the
actual construction of the levee, taking into account the time Fig. 9 Deformed mesh without stone columns
of realization.
The results of the analysis are compared with the site
measurements from the settlement gauges implanted in the
dam. The cut PK-200,000 -where the depth of the alluvial
foundation is maximum- contains five gauges: TASS4,
TASS5, TASS6, TASS7 and TASS8.
- Settlement gauges: TASS4, TASS5, TASS8 are located at
the contact dam-foundation.
- Settlement gauge: TASS6 is located in the foundation.
- Settlement gauge: TASS7 is located in the levee.
Note that the settlement gauges: TASS7, TASS8 have
unstable measurements, then they were not taken into
consideration.
Settlement gauges: TASS4, TASS7, TASS8 do not make
part of the comparison.
The position of settlement gages is shown in Fig. 1. Figs. 8 Fig. 10 Vertical displacement TASS5
and 9 show the deformed mesh of both models (with and
without stone columns). It is observed that displacements are
higher in the case of the untreated foundation; maximum
vertical displacements are on top of the models, giving 3.85 m
in untreated model and 3.42 m in treated model.

Fig. 11 Vertical displacement TASS6

Fig. 8 Deformed mesh with stone columns
Figs. 10 and 11 represent the vertical displacement curves
of numerical models and the actual measurements from
different settlement gauges. Measured settlements are lower

International Scholarly and Scientific Research & Innovation 11(8) 2017 1190 ISNI:0000000091950263
 World Academy of Science, Engineering and Technology
International Journal of Civil and Environmental Engineering
Vol:11, No:8, 2017

than the numerical ones with stone columns; this difference is mise en place. INVACO2: Séminaire International, innovation &
valorisation en genie civil & materiaux de construction n° : 5O-352
due to the installation effect of the columns which increases 2011.
the characteristics of the soil around the columns, and [8] Bouassida, M., Ellouze, S., Ben Salem, Z. & Znaidi, M.N., Parametric
consequently reduces the settlements. In Fig. 11, it is observed study of improved soft clay due to installation of a group of stone
column. International Conference on Numerical Computation in
that the settlement deference is reduced in depth. Geotechnical Engineering NUCGE'08, 2008.
D. Bearing Capacity Analysis
An analysis of the bearing capacity of the foundation was
performed for both foundation models (reinforced or
unreinforced). Fig. 12 shows the load-displacement curve that
shows the evolution of displacement according to the load
parameter. It is observed that treatment with stone columns
increased the bearing capacity of the foundation for a very
important value.
Open Science Index, Civil and Environmental Engineering Vol:11, No:8, 2017 publications.waset.org/10008631/pdf

Fig. 12 The load-displacement curve

VI. CONCLUSIONS
Using 2D modeling to represent a 3D grid of stone columns
needs an adjustment of parameters in order to preserve the
correct conditions and behavior of the foundation. The
displacement is more important in the case of untreated
foundation; this difference was not observed when the depth
increases. Settlements in real cases are less than the numerical
model with stone columns; this difference is due to the
installation effect of columns. Taking into account the effects
of setting up columns in the numerical model requires further
study.

REFERENCES
[1] T.M. Weber, S.M. Spring man, M. Gäb, V. Racansky, H.F. Schweiger,
Numerical modelling of stone columns in soft clay under an
embankment, Geotechnics of Soft Soils – Focus on Ground
Improvement – Karstunen & Leoni (eds) © 2009 Taylor & Francis
Group, London, ISBN 978-0-415-47591-4.
[2] Nguyen, N.T., Foray, P. & Etienne Flavigny, E., Prise en compte de
l’effet de la mise en place dans la modélisation numérique en 3D des
colonnes ballastées dans l’argile molle. 18ème Congrès Français de
Mécanique Grenoble, 27-31 août 2007.
[3] Agence Nationale Des Barrages et Transfert, Étude d’exécution et
surveillance des travaux de réalisation du barrage de Kissir dans la
wilaya de Jijel, Rapport de révision du projet. 2006.
[4] Agence Nationale Des Barrages et Transfert. Barrage de Kissir wilaya
de Jijel, Monographie. 2010.
[5] Dhouib, A. & Blondeau, F., Colonnes ballastées. Presses de l’ENPC,
2005.
[6] Document Technique Unifié 13.2, Travaux de fondations profondes
pour le bâtiment - Partie 1: Cahier des clauses techniques, Edition S1
44 - juin 2006.
[7] Rangeard, D., Phan, V.T.P., Lambert,S. & Martinez J., Densification de
sols fins compressibles par colonnes ballastées: influence du mode de

International Scholarly and Scientific Research & Innovation 11(8) 2017 1191 ISNI:0000000091950263