Design and construction of Chile’s Puclaro dam10 September 1999
Puclaro dam is founded on very pervious riverbed alluviums more than 100m deep. The design therefore included a vertical cut-off through the foundation, tied to the inclined upstream concrete face of the dam
There is a great amount of experience worldwide on concrete face rockfill dams (CFRDs) founded on rock, and on poorer quality materials such as weathered rock. There is less experience in designing them over alluvial foundations. The Santa Juana and Puclaro dams in Chile are both founded on deep riverbed alluviums, of 30m and 113m maximum depths, respectively.
Santa Juana was commissioned in 1995. The reservoir was filled in 1997 and its performance has been excellent, with total seepage lower than 50 l/s.
Puclaro dam, which is due for completion in October 1999, is very similar to Santa Juana.
To decrease seepage and for stability purposes a vertical cut-off was built. This impervious diaphragm wall was tied to the inclined upstream concrete face of the dam.
The riverbed plinth together with the cap of the vertical cut-off, form a joint with a very special design feature. This and other characteristics of Puclaro dam are presented in this article.
The Puclaro design
Puclaro is located on the Elqui river, 40km east of La Serena and 500km north of Santiago. It is 435m above sea level and its purpose is to improve the supply of water for irrigation purposes, feeding 21,000ha of the Elqui valley.
The dam is a concrete face gravel fill, 83m high, with a 640m crest length and a volume of 4,630,000m3 of gravel fill. The reservoir capacity is 200km3 and it will flood an area of 760ha. The spillway was designed for a 1 in 10,000 year flood, ie 2500m3/sec, and it can handle the calculated probable maximum flood (PMF) of 3800m3/sec.
The dam’s crest width is 8m and it has an ‘L’ parapet wall 4m high. The upstream concrete face varies in thickness from 0.45m in the riverbed plinth to 0.30m at the crest.
The embankment fill is largely provided by gravel from the bed of the Elqui river. Its mean particle size gradings are shown opposite and compared to corresponding figures for Santa Juana, Salvajina and Aguamilpa.
The zoning for different compaction requirements, and location of filters is shown in the diagram on p18.
The dam site area presents a stratified sequence of continental clastic and andesitic igneous rocks whose history can be characterised as follows. The scenery shows great volcanic activity assigned to the Cretaceous. Later in the early Tertiary the rock stratas were intruded by granitic bodies. During the Quaternary glacial action eroded a first valley, and fluvial action during the modern era has sculptured the valley to the present profile. Permeabilities range from 10-3m/sec in the upper 60m of sediment to 10-5m/sec in the lower 60-113m depth.
The sides of the dam site show a sequence of sedimentary volcanic strata where breccias, sandstones, tuffs and andesite are found in a horizontal attitude with slight variations in the direction and slope of the strata. The rock fracturing is medium (5-10 fractures/metre) with local zones in the upper 10m where it is high or very high (more than 15 fractures/metre). The rock shows a sound matrix, but in the upper 5-10m meteorization effects on fractures are frequently noted. The right abutment under el 460 is of medium permeability, with water admittance higher than 20LU. On the left abutment permeability is extremely variable and its highest values are 30LU at some levels.
Chile is located on the contact border of the Nazca and South American plates, which are converging at a speed of 10cm/year. This movement results in the subduction of the Nazca plate under the South American continent in an eastern direction. Almost all the earthquakes in the region where Puclaro is located are due to the relative movement of these plates and the resulting stress and deformation along the slide.
Since the Valparaiso earthquake, on 17 August 1906, instrumentation has been put in place to measure earthquakes in Chile. In the central part of the country, from an analysis of the seismic history, it has been concluded that during the 1535-1906 period there were at least 16 earthquakes at 7.0 or higher on the Richter scale, nine earthquakes at 7.5 or higher, and four at 8.3 or greater magnitude. From these figures, it can be concluded that the intervals between two events with strength equal or higher than the values shown in the table are: Based on the instrumental records available since 1906, the magnitude-frequency ratio (through Gutenberg and Richter formulae) and the soil maximum horizontal acceleration attenuation (using Martin’s expression, 1990) for the La Serena zone are: Design considerations
The feasibility studies considered basically two types of dam: gravel-fill with central impervious core; and rockfill with concrete face (CFRD). The conclusion was that CFRD was the best choice. The decision was based mainly on economic studies: materials adequate for a CFRD dam were to be found in the alluvium of the Elqui river, which could provide all the material from within a distance of 12km. On the other hand, suitable material for a central core was to be found 14km distant from the dam site. The decision was reinforced by the fact that CFRD dams have a better seismic behaviour and a shorter period of construction.
Several axes were analysed. The variables considered in the cost estimate of each alternative were the gravel fill volume, the area of the concrete face and the area of the diaphragm wall.
The general design of the dam is in agreement with most recent practice and its zoning is adjusted to the state of the art. Slopes were chosen of 1.5:1 (H:V) upstream and 1.6:1 (H:V) downstream. A parapet wall was considered in order to save gravel fill.
Static and dynamic stability analyses were made through finite element models adapted to the project characteristics. The state of stresses within the dam prior to an earthquake were first determined, as a result of the loads applied, fill placement, construction of the concrete face slab and reservoir filling. Total movements were then evaluated at the end of the face slab construction and of the diaphragm wall. Finally, the movements produced by the reservoir filling were determined.
The propagation of seismic movements applied to the contact surface between the base rock and the riverbed alluvium used for construction of the dam were analysed, taking into account the influence of the geotechnical dynamic properties of the materials through which the waves are propagated. The fundamental dynamic properties (shear module and damping ratio) were considered as a function of the effective confinement stress and of the deformation level that can be reached during an earthquake.
The design earthquake produces a maximum 0.39g acceleration at the top of the dam, with a 14MPa shear stress at the lower central third part of it. In the case of a maximum possible earthquake the values are 0.55g and 21MPa, respectively.
The safety factors were calculated, together with the acceleration distribution at the dam. The results indicate that for the design earthquake instability appears only at a high wedge of the dam, at a level located 4m under the top of the dam. The corresponding level is 27 m under the top of the dam in the case of the maximum possible earthquake.
The permanent horizontal deformations are lower than 0.03m, and no higher than 0.19m, in the case of the design and the maximum possible earthquakes, respectively. The dam was considered to be stable for this kind of event.
The geology and topography of the dam site support a design of several long straight sections, with few direction changes. Widths of 3m and 4m have been considered, with a unique 0.50m thickness. The reinforcement is 0.3%.
The perimetral joint has two defence lines: a lower copper water stop; and an upper mastic filler covered with Hypalon. A non-cohesive covering over the joint is available in case of a malfunction.
Regarding grouting, three drilling lines were considered — one impermeabilisation and two consolidation lines. The maximum depth of the grout curtain is 40m.
The positive experience of asphalt concrete face dams on alluvial material and the adequate performance of the CFRD’s plinth in non-conventional foundations (from Salvajina, which is partly on residual soils, and Khao Laem, which is on karst) have made it acceptable to consider the design of an articulated plinth on dense alluvium.
The possibility of founding a CFRD dam on riverbed alluvium was based on two considerations:
•The alluviums must be of low compressibility, so that no major deformation occurs when the reservoir is filled.
•A flexible structure must be designed so that the deformations experienced by the soil, as a result of the water load, can be absorbed.
Geotechnical exploration of the materials at Puclaro show that the existing alluvium is highly non-compressible. Numerical analysis shows that deformation is within the structure’s design tolerances. The structure has been conceived as a set of independent slabs joined only by impermeable elements, so it can adapt easily to any deformation of the soil. In addition, the design of the slabs’ copper water stops allows deformations much greater than would be produced in the perimetral joints of the abutments to be absorbed, by ensuring that the central part has plenty of space.
Between the diaphragm wall and the plinth of the concrete face of the dam, there are two slabs which allow some flexibility. These connection slabs are horizontal and are crossed by transverse joints with a 10m space between. The joint between the slab and the concrete face slab is designed in the same way as the perimetral joint. The diaphragm joint has the same impermeability elements than the rest of the joints, ie a copper water stop and a mastic filler. An anti-spalling reinforcement has been considered for these slabs.
The joint of the abutment plinth with the diaphragm wall and the horizontal plinth is accomplished using a special concrete structure founded on the rock.
The area to be impermeabilised at Puclaro is 16,500m2, so the concrete panel wall was a major item in the project cost. From the water loss point of view, seepage of up to 0.25m3/sec under the wall was tolerable. The cut-off specified was 58m deep and 0.80m thick. Grouting was only used at the cutoff wall where it was in contact with rock at each abutment.
The project uses a slab width of 15m, with a thickness t=0.30+0.002H. The specified concrete uses pozzolanic cement and air entrainment, and is specified at 21MPa resistance after 28 days. Reinforcement has been specified at 0.4% vertical and a 0.35% horizontal on the 12m slab next to the plinth, and 0.35%-0.30% in the upper zone of the slab.
The contract for construction of Puclaro was awarded to Consorcio Constructora Valle del Elqui after public bidding. This consortium is a joint venture of Mendes Junior of Brazil and Diwydag of Germany. The amount of the contract was US$60M.
Among the other contracts, the cutoff wall was subcontracted to bauer Spezialtiefbau Gmbh (Germany) and the electromechanical equipment was supplied by Voëst Alpine Ice Chile Ltd, with the support of Ecaso, Inamar and Klinger. The 11km road through the valley from La Serena to Vicuña was relocated under a separate contract.
Construction started in September 1996 and will be completed after 39 months, in November 1999. First filling of the reservoir has been scheduled for 15 October 1999.
As the valley at Puclaro is 350m wide, the river was handled so that two thirds of the embankment could be placed while the deviation tunnel was excavated in the left abutment. The 4.63M m3 required for the embankment was borrowed from river bed alluviums, which were excavated down to the water table at depths up to 5.5m deep. The mean distance from the borrow area to the dam was 2km.
The mean embankment fill placement volume was 156,000m3 per month, and the maximum was 334,000m3 per month. The following machinery was used: •Hitachi Ex 300 LC backhoes (six).
•Iveco 35 ton off-highway trucks (38).
•Cat D8R bulldozers (two).
•Ingersoll Rand 150 D vibratory roller (two).
The vertical cut-off wall had a surface area of 16,500m2 and was built in alternating 7m and 2.80m panel lengths. The minimum panel contact was 0.8m. For the excavation, Bauer used traditional clamshells, 6ton chisels and a rotary milling drill. The mean efficiency in two daily shifts was 43m2/day. The use of a rotary drill assured the wall was embedded to 50cm in the rock abutments.
The inclined concrete face was built in 15m widths using a sliding form. When completed, the total surface area amounted to 16,810m2. The sliding form progress amounted to 3m/hr. The excess concrete placed, compared to the theoretical volume, was 4.5% — equivalent to 2cm of extra thickness.
Construction and filling of the Puclaro reservoir had some adverse effects on the environment. In order of importance these were:
•The towns of Gualliguaica, San Carlos, Punta Azul, La Polvada and El Chape were flooded. This affected almost one thousand people, all working in agriculture. Families were relocated in upper areas of the valley, and two new towns were built by Dirección de Obras Hidráulicas, the owner of the project.
•Of the 760ha flooded, 340ha was used for agriculture and 250ha was used for vineyards.
•The road through the valley had to be relocated.
However, when these effects were weighed against the benefit of improving the water supply to irrigate a further 21,000ha of land, the Puclaro project was considered to be feasible.
Related ArticlesSpotlight on... South America Spotlight on... CFRD