Safsaf Dam
Safsaf Dam
Safsaf Dam
SUMMARY
The hardfill symmetrical dam concept has been developed and applied in
dam construction in the last two decades. The advantages are numerous in terms
of stability, hydrological and hydraulic safety and construction. The paper
illustrates those advantages through the case study of SafSaf dam in Eastern
Algeria, 36 meter-high, constructed between 2007 and 2011 on a highly fractured
and weak limestone foundation.
1. INTRODUCTION
In 1992, Pierre Londe and Michel Lino detailed the FSHD concept in Water
Power and Dam Engineering in 1992: ”The faced symmetrical hardfill dam: a new
concept for RCC”.
It has also been addressed in the French research program BaCaRa for
RCC dams, in the ICOLD bulletin n°117, “The gravity dam - a dam for the future”,
in a paper for the ICOLD Barcelona congress, 2006, “Hardfill symmetrical dams -
10 year experience feedback”, Q.84, R.71, by Michel Lino and orally presented at
the ICOLD Brasilia congress, 2009, “Concrete faced hardfill symmetrical dams
and safety towards design and construction floods”, Q.88, by Thibaut Guillemot.
The concept has been adapted by the Japanese engineers after 2000 under
the denomination of cemented sand and gravel (CSG) dam and several dams
were built.
The symmetrical profile of the dam gives good stability conditions despite a
low density and reduced shear strength due to the mix design and the simplified
placement, respectively γ = 21,5 kN/m3, ϕ’ = 35° and c’ = 0 kPa for SafSaf dam.
The stability conditions are poorly affected by water levels above the
maximum water level, in terms of stress variations and sliding stability. As a rigid
dam, it can resist to overtopping. These advantages are important with regard to
embankment dams in case of design flood exceeding.
The rigid dam body is not subject to internal erosion as for embankment
dams. This failing mode was critical in the case of SafSaf dam with the karstic
foundation. The initial design did not take into account a systematic search and
treatment of sinkholes, which would have been a hard, expensive and possibly
inefficient work. The dam had to be founded on the alluvial deposit of the riverbed.
Sinkholes under the alluvial deposit would not have been identified before the
dam construction and the risk of piping into the foundation was high.
For structures across the dam body, the watertightness is made by linking
the concrete face of the dam to the concrete structure. The contact between the
structure and the dam body, which is a weak interface for internal erosion in
embankment dams, does not need special treatment or provisions.
1.2.5. Monitoring
The dam site is on a non-perennial rive with irregular flow rates. There is no
flow most of the time, with a deep groundwater level, and there are sudden floods
a few days a year.
Fig. 1
Rock samples of the dam foundation and right abutment during excavations
Such a foundation, with low shear strength, cannot bear a classical concrete
gravity dam and the initial design in the nineties planed a 30 meter-high zoned
embankment dam, founded onto the alluvial deposit of the riverbed. Thus, the
spillway, designed for a 2 660 m3/s flood, was settled on the right bank, next to the
embankment, which demanded large rock excavations.
In addition, the diversion canal, under the dam body, was designed for a
construction flood of 400 m3/s.
The initial design was considered critical towards hydrological safety: the
hydrological review showed that the initial study underestimated the floods
volumes and peak flows). The initial design flood of 2 660 m3/s and the
construction flood of 400 m3/s were far under what is to be considered for an
embankment dam: they roughly corresponded to the 500 year-flood and the
2 year-flood. The design flood was upgraded to 5 220 m3/s (10 000 year-flood).
The cost of the project would have been strongly affected. Thus, ISL
Ingénierie proposed to the owner a major change in the project to improve its
safety while keeping it financially viable: the embankment dam was replaced by a
FSHD, 36 meter-high, founded on the bedrock beneath the alluvial deposit of the
river and topped with a free overfall crest spillway, designed for a 7,5 meter-high
head. A stilling basin of 60 meter-long, 113 meter-wide and 7 meter-deep at the
downstream toe of the dam dissipates the energy of the flood discharges.
The costs figured by the contractor are 30 M€ for the embankment dam
(design flood: 2 660 m3/s) and 33 M€ for the hardfill dam (design flood:
5 220 m3/s).
The plan view and the typical cross section are shown in Fig. 2 and Fig. 3.
Fig. 2
Plan view of the dam and the spillway
Fig. 3
Typical cross section of the dam and the spillway (slopes: 0,7 H / 1 V)
Alluvial aggregates are taken in selected zones where they are clean and
well graded. They do not need any treatment but screening above 80 mm. The
quantity found in the reservoir area was large enough to ensure the total
production of hardfill for the project. On Fig. 4 is shown the type of aggregates
used for hardfill. The grading curves show the specified grading envelope (red),
the samples taken in zones fitting the specifications (blue) and in zones to be
avoided for aggregate use (grey).
100
90
80
Percentage (%)
70
60
50
40
30
20
10
0
0,001 0,01 0,1 1 10 100
Diametre (mm)
Fig. 4
Typical aggregates used for hardfill production
The majority of alluvial deposits proper for hardfill production show a lack of
fine particles to reach a percentage between 5% and 15% smaller than 0,08 mm,
specified for compaction efficiency.
The first mix formulations in laboratory were made with 8% of limestone filler
from a local cement factory and 80 kg/m3 of cement. This was giving good results
in terms of density and compressive strength, but as the filler might be unavailable
at some periods of the construction, it was replaced with fine sand (8%) and an
addition of 40 kg/m3 of cement. A loss of dry density of 3 to 5% was observed in
the samples, resulting in a loss of about 20% in compressive strength.
As the mix without filler, with fine sand and 120 kg/m3 of cement was
meeting the specifications (dry density of 21,5 kN/m3 and compressive strength at
90 days of 7 MPa), it was adopted for the dam construction.
3.2. PRODUCTION
The production is made with a concrete mixing plant of 120 m3/h capacity
(Fig. 5), producing an average of 15 000 m3 of hardfill a month during the
construction.
Fig. 5
View of the concrete mixing plant
3.3. PLACEMENT
Fig. 6
Hardfill placement on the dam
3.4. CONTROL
The method statement for placing hardfill was elaborated during a field test
at the beginning of the upstream cofferdam construction. Apart from the laboratory
tests on hardfill components, the quality control consisted in dry density
measurements with a radiation-type densimeter and control of compressive
strength at 90 days.
The samples for compressive strength were not made with a vibrating table
because the hardfill was too dry. A procedure similar to Proctor sampling had to
be developed to reach a sample density close to the specified density. The hardfill
samples were cylindrical with a diameter of 25 cm and a diameter over length ratio
of one.
In situ coring was tested on the site but was abandoned: the contrast
between hard gravels embedded in a smooth paste resulted in damaging the
sample surface within the coring device. The samples were not cylindrical after
extraction.
The crest length of the dam is 270 m and the base length is 120 m, without
any construction joint. The hardfill layers are continuous on the dam length and
width. Small lateral vertical cracks, a few 1/10th of millimeter opening, have been
observed on hardfill layers at regular interval of about 10 m. On the upstream
cofferdam, left without concrete cover during the construction, no leakage was
observed in the cofferdam body during floods, what tends to show that the cracks
can heal after opening.
The stability calculations have been carried out with a friction angle of 35°
and no cohesion into the dam body. As the watertightness of the dam is provided
by the upstream concrete face, the layer surfaces received a minimal treatment
consisting in dragging a harrow after compaction and cleaning with air before
spreading the next layer.
Fig. 7
Harrow for interlayer treatment and typical surface rendering
Fig. 8
Downstream face armoring with precast elements
The spillway is cast on the precast elements at the end of the hardfill dam
body erection.
5. CONCLUSION
SafSaf dam is a good example of a rigid dam that could be built on a weak
rock foundation, where a conventional gravity dam was not feasible, and as a
good alternative to an embankment dam. The remarkable points are:
Fig. 9
SafSaf dam during construction
Fig. 10
SafSaf dam after completion
Dams: SafSaf (Eastern Algeria)
FSHD, RCC