Speedy construction is a pre-requisite for RCCs in tropical climates and two projects in south east Asia are making good progress. Report by Suzanne Pritchard
Earlier works at Yeywa before completion of integrated cofferdam (April 2007)
As RCC dam construction is centred upon a sequence of highly mechanised activities, the key to successful delivery of a fast yet high quality and economical project is a simple design that facilitates a smooth construction process. Two interesting RCC dam projects achieving excellent progress are under construction in south east Asia – the Yeywa hydropower project, located on the Myitnge river in central Myanmar, and the Son La scheme on the Da river in Vietnam.
Yeywa is Myanmar’s first RCC dam. The scheme comprises a 134m high RCC gravity dam with a total volume of 2.5M m3 of concrete. Other features include an ungated spillway for a design flood of 6,600m3/s and a 790MW (4 x 197MW) powerhouse at the toe of the dam on the left bank.
The 2,400MW Son La project is under construction approximately 360km north west of Hanoi. It is a 138m high structure with RCC volume of approximately 3.1M m3 (total volume is 4.6M m3), with a peripheral spillway with a capacity of 35,000m3/s. Located on the right bank, the spillway has eight gates in addition to 16 low level gates to control water levels during the flood season.
Son La is considered to be of national importance and will supply 9GWh annually to the grid. River regulation will also enable the 1980MW Hoa Binh plant, which is downstream, to operate fully. Son La is the second of four dams to be built on the river and is an integral part of the largest hydropower project currently under construction in south east Asia. Upon completion the whole scheme will have an installed capacity of 6,532MW and will provide flood control, water supply and regulation.
Preferred choice
Compared with conventional gravity and concrete face rockfill dams (CFRDs), RCC dams are generally the preferred choice of designers when working in the tropical conditions experienced in south east Asia, especially as the dry season can last only about six months. The effective scheduling of construction sequences involved with the RCC process helps to facilitate continuous progress, which is particularly advantageous during the wet season as it can reduce the cost of the river diversion. It can also reduce timescales and the cost of the project as a whole. The arrangements that facilitate such economies for the project are:
* Integrated Cofferdam: Construction of an integrated cofferdam as part of the main cross-section enables downstream construction works at a later stage.
* Intentional overtopping: The purposeful overtopping of RCC sections located in the river section, and continued RCC placement in the dam portions protected against floods.
Both of the above arrangements contribute to smaller diversion tunnels or culverts to secure the construction site against floods in the wet seasons, which help to reduce project costs. At Yeywa, a longitudinal separation wall (needed to separate the tailrace channel from the spillway) was constructed between the overtopping sections and left bank sections, which allowed for continued placement of RCC at the left bank during the wet seasons. At Son La, the same task is fulfilled by the diversion culverts.
The advantages of selecting RCC were seen at Yeywa, where some major setbacks have been experienced during construction. The most serious was in October 2006 with the occurrence of a 1:50 year flood at the end of the rainy season. The project’s integrated RCC cofferdam arrangement protected the downstream works in the river section against the floods. The 60m high cofferdam was designed for floods with return periods of 1:50 years.
The construction works themselves have remained relatively free from the major damage that can be caused by such occurrences, although there has been some delay. The intentional overtopping of the RCC sections already constructed in the river section can take place at the same time as continuing with RCC construction on the left bank section. Such a major flood security advantage is not to be underestimated, especially in countries where extreme power shortages combined with frequent shortages of fuel and pumping capacities are prevalent. It is argued, therefore, that this proves an advantage of RCC in relation to CFRD and rock fill methods of construction.
Construction of Son La dam is on a tight schedule, though, because of the need to improve flood control on the Da river. The significant number of flood events that occur each year during the wet season means that river diversions must be able to handle large floods. The floods in 2007 wet season reached some 12,000m3/s while in the previous year the flow rate reached even higher, at approximately 15,000m3/s.
Integrated Planning
An integrated plan is required early in the design process to help ensure the greatest opportunity to reap the full economic and quality benefits associated with a well-designed RCC dam. The plan must ensure that the sourcing, transportation, production and placing of the RCC can run smoothly, especially in the rainy season and when confronted by significant flooding. Other factors that need to be considered include the appropriate selection of construction methods and equipment for RCC transportation to, and application at, the dam.
As with any construction endevour, ensuring uninterrupted pace of construction is vital and so disruptive activities or structural interfaces have to be a minimised, possibly moreso for RCC dams. From a design point of view, this means that structures that would intercept the linear progress of the RCC equipment on the placement area must be kept to a minimum, if not banned. Such structures include: galleries, which should be reduced in number to only the essential; transverse galleries which connect inspection galleries but should be eliminated; and, likewise, vertical shafts (such as staircases and elevators) or other large chambers in the RCC sections should be eliminated. Where such structures cannot be avoided, the placement area can be increased to develop the full effect of the high degree of mechanisation involved in the RCC placement process.
At Yeywa, the power intake towers were designed as conventional reinforced concrete structures abutting onto the upstream face of the RCC dam. This enabled the contractor to build the four towers above the penstock inlets before the start of RCC construction.
This not only helped to minimise effects on RCC construction activities, but has also enabled the Department of Hydropower in Myanmar to construct these above the inlet bellmouths and closed gate positions in advance. Such methods have helped to avoid significant delays.
All in the mix
The desired high quality of RCC dams depends on accelerated rates of construction. The speed at which RCC is placed has a great influence on the quality of the horizontal lift joints, ie the bond between the 300mm thick RCC layers in the dam to ensure that the tensile strength and seepage across the horizontal lift joints are effectively identical to that of the parent RCC itself.
The mix design methodology is based on a high-cementitious approach, which enables the delivery of construction speed, lift joint quality and thereby simplified targets. The total cementitious contentof the mix, cement plus pozzolan, will not be less than 150kg/m3 of RCC. Admixtures are used to help retard the set time up to 24 hours, and enables fresh concrete bonding between layers.
High-cementitious RCC mixes with a high volume of pozzolan as a cement replacement is considered to be the norm for the majority of large RCC dams. Pozzolan can be obtained from natural sources such as volcanic or fly ash from the by-products thermal power plants. Good pozzolan contributes to the strength of the RCC mix and insitu lift joint properties, which enables further cement replacement and more favourable thermal conditions – maintaining hot joints reduces the need for time-consuming joint preparations at a later stage.
Locally sourced pozzolans offer significant cost benefits to projects. The search for suitable pozzolans for Son La dam resulted in the use of fly ash from the Pha Lai thermal power station, some 425km from the construction site. The pozzolans were a more effective cementitious material than Portland cement. Consequently, mix design trials indicated that a total cementitious content of 220kg/m3, comprising 60kg of cement and 160kg of fly ash, would produce the necessary characteristics for a good quality RCC dam. In addition, three full-scale trial embankments were constructed.
Fly ash, which contains a large proportion of unburnt carbon, can result in a higher LoI value which can have an effect on the strength and durability of the RCC structure. Therefore, the last of the three trials was undertaken not only for training purposes but also to prove that fly ash with a loss of ignition (LoI) value in the upper limit does not have a detrimental effect on RCC performance. LoI values for the fly ash from Pha Lai thermal power station were up to 25% but the third trial embankment used an RCC mix containing fly ash with a LoI of 12%.
LoI values from Pha Lai varied from a high of 30% to a low of 6% over a year, the use of ash with a LoI greater than 12% has never officially been recorded. Vietnamese regulations view the lower limit of 6% LoI as the appropriate standard to be used, even though 12% is permissible if sufficient tests have been carried out. The trial mix tests have shown that there is little difference in the strength and durability results after three years even with LoI values up to 20%. However, without more long-term results it would be difficult to justify the use of ash with LoI values above the 12% limit.
The ash from the Pha Lai ash lagoons is being processed in two facilities using a flotation method followed by drying. A new facility is coming onstream to significantly increase the quantity of ash processing to meet the 6% limit. However, this may still not be sufficient for the current construction schedule. Therefore, a method of producing more than sufficient quantities of ash with a LoI less than 12% has been proposed. The proposed method would help speed the construction process at Son La dam, and could potentially do so for other RCC projects.
An extensive trial mix programme was also necessary at Yeywa dam. Fly ash is not available in the country so unless a suitable economically efficient import ash was available, a natural pozzolan had to located, investigated and tested for its suitability for application in the RCC dam. One of the original possibilities in terms of import options was to bring fly ash from Mae Moh thermal power station in Thailand. However, there were uncertainties about the supply and transport routes which would be involved.
Instead, geological investigations to confirm the pozzolanic properties of materials from the different sources were undertaken to select the most suitable site for the development of milling facilities. Two natural pozzolans were located near Mount Popa and have exhibited exceptional performances when used with locally available Portland cements. After extensive tests, both in the laboratory and in the field, the optimum mixture proportions of the RCC was found to be 75kg/m3 of Portland cement and 145kg/m3 of natural pozzolan, which is an economic set of mixture proportions. The success of this clearly demonstrates the advantage of starting a trial RCC mix programme as early as possible during the design phase of the project.
The accelerated rate of construction that is needed for RCC dams is well illustrated in the case of Yeywa. RCC?placement began in February 2006. Within 14 months, which also of course included the rainy season, approximately 1M m3 of RCC was placed, with the monthly production rate reaching a maximum of more than 91,000m3. Compared to the original schedule, the dam is expected to be completed eight months early, by the end of 2008. The project’s turbines are set to be running by the end of 2009.
Such an accomplishment has been attributed to the minimal interference and cross structures, thereby ensuring continuity of RCC placement. In addition, the high-cementitious content approach to RCC and the highly efficient, 480m3/hr nominal capacity of the batching plant also have been important contributing factors.
Other factors include appreciation of the fact that speed will contribute to the lift joint quality by maintaining hot joints. This helped to release the contractor from time-consuming joint preparations which helped progress and enhance the overall quality of the dam.
Tribute has also been paid to CGGC Gezhouba, the RCC contractor from China. Its extensive experience of RCC dams has been described as being invaluable.
Son La, in comparison, is just getting underway with its RCC placement under a tight construction schedule. Given the high-cementitious RCC mix design (>150kg/m3) and retarding the set period of the concrete to up to a day for the bottom part of the dam the maximum daily volume of RCC?placed would be approximately 5,000m3 in layer volume. The overall average monthly RCC production at the dam rate is seen at approximately 84,000m3.
The dam was designed in accordance to international standards and checked against the Vietnamese /Russian standards. The design criteria developed for the project were formulated specifically as a Vietnamese standard by drawing upon the US Army Corps of Engineers’ Engineering Manuals (EM 1110-2-2200) and also the US Federal Energy Regulatory Commission (FERC) Guidelines, from 2002.
A two-stage structural analysis was performed involving rigid body and finite element (FE) modelling, and thermal modelling. The rigid body analysis examined sliding and overturning stability, the FE analysis calculating internal stresses within the dam structure. In addition, static and dynamic analyses were carried out, looking at tensile stresses under earthquake loading, to help in establishing the quality of RCC required and the mix design.
For the analyses, the assumed density of RCC was 2.5 tonnes/m3 and a compressive strength taken as 16MPa. From the models, the joint tensile strengths calculated were 0.8MPa for static loading and 1.2MPa under seismic loading.
Differential strain due to stresses were also calculated using the static FE analyses. The models helped determine the estimated differential deformation between the dam and powerhouse at the downstream toe, and between the dam and the penstocks down the downstream face. The maximum differential deformation of the penstock due to water load with the reservoir: at operating level were 4mm (vertical) and 21mm (horizontal); at maximum flood level were 4mm (vertical) and 23mm (horizontal).
In addition, the designers were also able to determine through the dynamic analyses athat the number of times the allowable stress would be exceeded during the maximum credible earthquake (MCE) was very small, and that it was expected that only micro cracking would be anticipated due to such a load.
In terms of the thermal analysis, data used were derived from laboratory tests, including specific heat, conductivity, diffusivity thermal expansion and adiabatic temperature rise. The model considered the proposed placement sequence within the monoliths including start and completion dates of the blocks and a linear rate of increase of the upper surface, modelled with 2m high steps. The foundation was modelled to 10m below foundation level. With a placing temperature of 22°C, the maximum temperature in the dam was calculated to be 40.7°C – less than generally accepted value of 45°C.
| Production and placement at Yeywa – the facts |
|
Crushing plants – Operating using a dry process; three have been installed with a similar arrangement of a primary jaw crusher, a secondary impact or cone crusher and a vertical shaft impact crusher. The specified production rate is 150,000 tonnes per month. The aggregate stockpile is 1M tonnes. |
| Son La schedule |
|
March 2005 – Construction of the project started with excavation of the 90m wide diversion tunnel and two 12m by 12m dry season culverts, plus excavation of the abutments. |
| Project name check: who’s who? |
|
Yeywa: |