SOLVING SOME LONG STANDING RCC CONCERNS Brian A.

Forbes, Manager Dams Engineering, Gutteridge Haskins and Davey Pty Ltd, Consulting Engineers, Australia The use of grout-enriched RCC, the sloped layer placing method and zoned cooling of RCC at the Tannur and Kinta Dams promises improved quality for these turn of the century RCC dams. The Tannur Dam is currently under construction in Jordan and a contract for the construction of Kinta Dam in Malaysia is about to be let. Both are the first RCC dams to be constructed in these countries. At Tannur, RCC placing for this 60 m high gravity dam commenced in January 2000. Placing the 230,000 m3 of RCC is expected to be completed by late August 2000. Over half will have been placed during the dry summer period (May onwards) when ambient temperatures frequently exceed 40oC during the day. The Tannur dam will be the first of three RCC dams to be completed for the Ministry of Water and Irrigations Jordan Valley Authority as part of its Southern Ghors Irrigation Scheme. Consultants, Mott MacDonald of UK, GHD of Australia and CEC of Jordan are providing engineering services for design review, construction supervision and administration of the contract let to BEC Freres of France in joint venture with Al-Jafar of Jordan. Seven tenders were received at the end of March 2000 for the construction of the 90 m high Kinta gravity RCC dam. Build-own-operate concessionaires, Metropolitan Utilities Corporation are constructing the Kinta Dam to augment their water supply to the city of Ipoh. The dam will contain approximately 900,000 m3 of RCC and cost US$32 million. Overall completion of the dam is expected in 2 years. Consultants, GHD of Australia and AngkasaGHD of Malaysia carried out the design and will supervise the construction. Both of these dams innovatively address three principle concerns related to the design and construction of RCC dams, namely provision of a durable, aesthetic impermeable low cost facing to the RCC, effective bonding of lift joints to provide a monolithic structure and reduction of thermal cracking potential. GE-RCC The use of grout-enriched RCC was first trialed in China(1). Subsequently, it was used with success in Australia and New Zealand. At Tannur, and as planned for Kinta, GE-RCC is used to form the upstream face, downstream steps including the spillway, transition zone between RCC and rock abutments, encasement of water stops and drains, walls of drainage galleries and anywhere else where internally vibrated conventional concrete would usually be required, including encasement of reinforcing steel. The process is simple and effective. RCC is produced in the mixing plant, and spread into position in layers as for the dam. Where GE-RCC is required the spread RCC is dosed with a cement water grout. The grout is simply poured over the area required prior to compacting the RCC and given a few minutes to soak down into the loose RCC. A poker vibrator can then be used to effectively distribute the grout and consolidate the GE-RCC as a low slump conventional concrete. A 15-20 mm thick layer of bedding mortar spread over the previous GE-RCC lift surface assists the process. Trials have shown it to work with high to medium paste RCC mixes down to 120 kg/m3 of cementitious material. GE-RCC strengths can be controlled to some degree by the water cement ratio of the grout. Well graded RCC mixes with high fines content require water cement ratios in the range 0.9 to 1.0. Typically GE-RCC strengths approximate that of the parent RCC. Results for Tannur Dam are shown in Table 1.

Document number: 28661 Job Number: 411\009012\60

Table 1- Tannur Dam RCC and GE-RCC 28-Day Compressive Strengths of 150mm Cube Samples RCC Mix#1 (125+75) Description No. of Samples Average Strength (MPa) Standard Deviation (MPa) Coeff of Variation (%) RCC 70 22.1 3.05 13.8 GE-RCC 64 21.8 3.36 15.3 RCC Mix #2 (125+55) RCC 56 21.0 2.56 12.2 GE-RCC 56 19.7 2.61 13.3

The parent RCC mix contains 125 kg/m3 of ordinary portland cement and 75 or 55 kg/m3 of pozzolan as shown in the table above. The contact between GE-RCC and adjacent compacted RCC is generally indiscernible as seen in cores, i.e. monolithic. The advantages include: Only one mixing, transporting and spreading plant is required - i.e. the RCC plant. Simple procedure, little training required Not a critical path item; crews and mixing/vibrating plant easily augmented. Overall quality enhanced, durable surface, little to no-shrinkage cracking, aesthetic typical off-form concrete finish. Monolithic with parent RCC, good adhesion to rock foundation and excellent encasement of vertical and horizontal waterstops. Costs available from a range of projects show GE-RCC to be in the order of US$4 to 7/m2 based on treating a 0.4 m wide zone, i.e. US$10.00 to 17.50/m3 additional to the supply and delivery cost of the parent RCC. SL METHOD AND RCC LIFT JOINTS Placing RCC lifts by means of sloped layers (SL) reduces the overall surface of the placement area so that 300 mm thick layers of RCC can be placed within 1-2 hours of each other, i.e. within the initial setting time of the lower layer. In this way it is possible to build up 3 m thick lifts of RCC as a continuous process. The method was conceived and developed at the Jiangya RCC dam (131 m high) in China 1996-99(2). The layers are sloped at 1:10 to 1:15 (V:H) across the dam from abutment to abutment over a height of 3 m or whatever lift thickness is adopted. The usual crossfall to the upstream of 1:30 to 1:50 is retained to facilitate clean down and drainage of wash water etc off the completed 3 m lift surfaces. As the construction advances across the dam the completed 3 m lift surface is cured/green cut and prepared as a cold joint, forms are lifted and prepared for the start of the next 3 m lift once the lower lift of sloped layers reaches the opposite abutment. Final lift surface preparation, placing of bedding mortar etc proceeds as a 4-5 m wide strip, being the contact area of the individual 300 mm layer with the cold lift joint surface of the lower 3 m lift. This method results in a monolithic 3 m lift of RCC, without the usual joints at 300 mm intervals and their inherent bond problems. At Jiangya Dam 150 mm diameter cores in continuous lengths of up to 6.7 m were recovered. At Tannur Dam the lower 15 m was placed using the usual horizontal 300 mm lift process. Where lift joints exceeded an age of 2 hours a full layer of 140-160 mm slump bedding mortar

Document number: 28661 Job Number: 411\009012\60

was spread 4-6 mm thick over the surface just ahead of the RCC lift to enhance bond and reduce the effects of any segregation of the RCC, should this occur. Above 15 m the SL method was adopted to speed up the overall production rate by as much as 50%. The SL method has reduced the requirement for lift joint surface preparation and permits the earlier lifting of upstream and downstream formwork. Lifts of 1.2 m have been adopted to match the downstream step height. The 300 mm layers commenced with an initial slope of approximately 1:12.5. The system is working well and each 300 mm compacted layer of RCC is being placed within 1 hour of the lower layer. The GE-RCC facing to the steps shows no sign of the joints of the sloping 300 mm layers. Production is up 50% and 150 mm diameter cores recovered with a 2 m long single tube core barrel are being retrieved intact over the 2 m cored. The 1.2 m lift surfaces are treated as cold joints, green cut, cured and bonded to the next 1.2 m lift through a layer of bedding mortar. Current production is equivalent to 11/3 300 mm thick layers per day, i.e. 2 x 1.2 m lifts/6 day week, which is an average daily placing rate of 2,000 m3/day from the 190m 3/hour continuous RCC mixing plant. At Kinta the Contractor will have the option of horizontal or the SL method of placement. It is expected that the lower and upper levels of the dam will need to be placed in horizontal layers with bedding mortar over each lift surface as at Tannur, but the SL method will be encouraged in the central zone as the overall placement area increases. Lifts of 3 m thick are anticipated. CONTROL OF THERMAL CRACKING At Tannur dam the original design and specification required the RCC to be less than 26oC at time of placing and summer placing was not permitted. In order to continue through summer (May to mid-September) the Contractor has been required to institute some simple pre-cooling actions, including: use of winter-produced and stockpiled aggregate evaporative cooling of aggregates by stockpile irrigation use of chilled mixing water reducing the day shift placing period as ambient temperatures increase With the low ambient humidity at the site evaporative cooling is proving very effective. Irrigation lines have been set up along the upstream and downstream forms with fog spray nozzles at regular intervals. These shroud the dam in a moist fog like blanket which creates a local micro-climate over the work area. Ambient temperatures are 7-8oC lower than elsewhere around the site. The RCC during dumping, spreading and compaction is kept moist on an as needs basis by hand air/water mist sprays and the completed 1.2 m SL lift surfaces are kept cool wet by irrigation sprinklers. The temperature of the RCC delivered to the dam is recorded at the rate of 1 per sloped layer (i.e.1/150 m3 approximately). The engineer is controlling the work through these records on a temperature-volume averaging basis. It is expected only night time placing will ultimately be permitted. The temperature of the sub-surface zone (top 25 to 75 mm) of the completed 1.2 m SL lift is measured twice daily. The sub-surface RCC temperature data taken in late April is used as a basis for control. Saturation and irrigation of the lift and evaporative cooling is keeping the RCC temperatures in the surface zone of the lifts at or less than those recorded in late April, i.e. heat gain into the body of the RCC from the hot ambient conditions is being monitored and controlled.

Document number: 28661 Job Number: 411\009012\60

In the design of Kinta a rigorous 3-D thermal stress analysis was undertaken using the finite element program ANSYS(3). RCC trial mix design studies also carried out tests on samples to determine early age strengths and modulus of elasticity from 12 hours through to 3 days, the time when hydration heat gains peaked. During the heat gain period the RCC in the dam expands, going into a compression condition. However, because of low modulus and high creep parameters at this young age the RCC is effectively in a plastic-elastic condition and the level of compressive stress reached is less than would be the case if the RCC was older and the elastic modulus was correspondingly higher. During the cooling phase the RCC shrinks, but it does not return to a zero stress condition when its temperature reaches that at which it was placed; because of the modulus difference it actually goes into tension. This effect was modelled with ANSYS and the surficial zone of RCC on the upstream face was shown to develop tensile stresses at an early age in response to cooling as it matched the ambient conditions. The stresses were not as high on the downstream sloping face. The upstream face stresses exceeded the early tensile strength capacity developed by the RCC at this age. The slower cooling RCC further in from the face was resisting surface shrinkage and the analysis confirmed that the surficial zone of RCC required pre-cooling to reduce the stresses to a point where they matched the tensile capacity of the RCC. In fact, it was shown that only the outer 0.5-1.0 m of RCC required pre-cooling. The 3-D analysis also showed that reducing the spacing of the transverse joints from 100 m to 30 m gave little reduction in the peak surficial tensile stress, only a reduction in the area over which it occurred. The thermal stress condition of the dam was then analysed over a period of 10 years as the body of the dam cooled. With transverse joints at 30 m centres the thermal tensile stresses developed in the RCC due to foundation restraint and static loads were at all stages less than those predicted to be attained by the RCC in the longer term. As a result of these studies the only cooling of RCC at Kinta will be for that RCC placed in a zone along the upstream face. For practical reasons, it is proposed that the upstream 4 m be pre-cooled. The RCC temperature at time of placing is required to be 20oC, i.e. about 10o cooler than the remainder of the RCC. For dams in climates less temperate than Malaysia a greater temperature reduction would probably be necessary. REFERENCES: (1) Forbes, B.A. Grout Enriched RCC : a History and Future, Int. Water Power and Dam Construction, June 1999. (2) Forbes, B.A. Lichen, Y., Guojin, T., and Kangning, Y., Jiangya Dam, China. Some Interesting Techniques Developed for High Quality RCC Construction, Int. Symposium on RCC Dams, Chengdu, April 1999. (3) Crichton, A.J., Benzenati, I., Qiu, T.J., and Williams, J.T., Kinta RCC Dam - Are Over Simplified Thermal - Structural Analyses Valid?, Australian National Committee on Large Dams 1999 Conference on Dams, Jindabyne, November 1999.

Document number: 28661 Job Number: 411\009012\60

Brian Forbes is a Civil Engineer with 36 years of experience in the investigation, design, construction, safety review and valuation of all types of dams. He is Manager of Dams Engineering with Gutteridge Haskins and Davey Pty Ltd, Consulting Engineers Australia, which he joined when he moved from Zimbabwe and South Africa in 1979. Since 1983 he has been involved internationally in the engineering of over 25 RCC dams either as designer, specialist advisor or consultant. These include Longtan, Pangue, Jiangya, Porce II and La Miel I, which are amongst the worlds highest. He has contributed to numerous conferences and RCC symposium on a range of related topics.

Document number: 28661 Job Number: 411\009012\60

Figure 1 - General view of Tannur Dam stepped spillway constructed in grout enriched RCC. Step height 1.2m. Figure 2 - Vertical and horizontal PVC waterstop encased in grout enriched RCC. No evidence of voids or unconsolidated GE-RCC. Figure 3 - Typical off-form surface finish to Tannur Dam stepped spillway constructed in grout enriched RCC. Sloped layer lift joints evident only by form paint marks staining the GE-RCC. Figure 4 - Typical view of sloped layer construction at Tannur Dam . Viewed upstream. Lift line for 1.2m thick lift seen painted on form. Figure 5 - Spreading next 300mm thick sloped RCC layer out over previous 1.2m thick lift surface which has first been covered with a 4-6mm thick layer of bedding mortar. Viewed downstream. Figure 6 - ANSYS model of one half a 30m block of Kinta Dam showing stress condition at 300 days on reservoir filling. Figure 7 - Same as Figure No. 6 but for one half of a 100m block. Shows proportionally greater area subjected to high tensile stresses of similar magnitude.

Document number: 28661 Job Number: 411\009012\60