CFRD practice in Turkey17 April 2007
Hasan Tosun, Serhat Batmaz and E.Orkun Ayvaz describe the evaluation of design principles adopted for CFRDs in Turkey and discuss problems encountered during construction
The concrete faced rockfill dam (CFRD) is now one of the most popular types of dam in the world. In the 1940's it was used for dams with moderate height, but it developed significantly in the 1970s with the application of vibratory roller. The progress of this type of dam was most prevalent in South America and Australia, where significant advances were achieved in the design and construction of dams. The Australian practice on CFRD is extensive, as there are many examples of these types of dams in this country (Fitzpatrick et al, 1985). Brazil also possess valuable experience on CFRDs. Foz de Areia, Xingo and Segredo dams are the important structures in this country. Shuibuya dam in China, which is 232m high, will be the tallest CFRD in the world when completed. Bakun dam in Malaysia, which is the world's second highest CFRD, is expected to be complete in 2007.
Design and construction of CFRDs seem very simple when compared with other types of embankment dams. The authors point out however that it is necessary to pay more attention to plinth construction, foundation preparation, rockfill selection and placement, upstream face control, handling of waterstops and upstream face construction by slipforming. Durability of concrete face slab is also an important factor which influences the long-term behaviour of this type of dam (Yoon et al, 2002). The time-dependent deformation, especially the separation between concrete face slab and cushion layer, seems one of the most significant problems of high CFRDs (Zhang et al, 2004). Therefore, selection of rockfill material should be made with care to prevent unpredicted settlement during water impounding. Well-graded rockfill or gravel (with free drainage) is commonly adopted in developed countries. Some design engineers state that weathered rockfill or dirty gravel can also be used, if zoning is taken into account and compaction for rockfill material is provided. Settlement behaviour is largely dependent on deformability characteristics of embankment materials.
Dam specialists consider the fact that well-compacted CFRDs have a high resistance to earthquake loading. Bureau et al (1985) state that this is based on several factors including acceptable past performance of similar dams, a recognition that the entire embankment is unsaturated and the fact that compacted rockfill develops high frictional resistance. Intensive investigations have been performed to explain their behaviour under dynamic loading. Uddin (1999) introduces a dynamic analysis procedure for CFRDs subjected to strong seismic excitation. Numerical solutions performed for CFRDs indicate that they are as safe as other embankment dams. However, their behaviour is questionable when they are subjected to seismic loading, which is caused by near energy zone.
CFRD projects in Turkey
The construction of CFRDs in Turkey began in 1997, and has been used with increasing frequency in recent years. Six dams ranging in height from 74 to 152m have been designed and constructed since 1997 (see Table 1). Two of these were completed in 2003 and 2005, while four are currently under construction. Three of them, namely Kurtun dam, Atasu dam and Torul dam, are constructed within the same basin in the northern part of Turkey. In the West, Gordes and Marmaris dams are constructed for water supply purposes, while the last dam is located in the south and has been designed as a multi purpose dam for water supply, irrigation, flood control and electricity production. Further studies for twenty-five dams of this type are in the early stages of design. In 2005 new guidelines for the design of CRFDs was developed in the country.
Kürtün dam, which impounds 108.2hm3 of water at maximum water level, is the first operational CFRD in Turkey. It is 133m high from the foundation and 300m long on the crest. The total storage capacity of the reservoir is 80.3hm3 in normal water level. It was originally designed as a rockfill dam with central core. However, it was revised to a CFR type following intensive geotechnical studies. Intrusive volcanic rocks, regarded as an appropriate rock for dam foundations, forms the base of Kürtün dam. The crest width is 14m and the side slopes of the main embankment are 1.4H:1V for upstream and 1.5H:1V for the downstream. The volume of embankment is 3.8hm3 for which construction was completed at the end of May 1999. A delay of 1.5 years followed to provide settlement before construction of the concrete slab started. The water filling stage began at the start of February 2002. The project produces 198GWh per year with an installed capacity of 80MW.
Torul dam is located on the Harsit river, 5km east of Kürtün reservoir in northern Turkey. It is 152m high with 4.6hm3 embankment volume. When the reservoir is at normal capacity, the facility impounds 168hm3 of water with a reservoir surface area of 3.62km2. Its construction began in 2000 and is expected to finish in 2007. The main embankment consists of four major zones. The impervious section of the embankment is a concrete slab and plinth structure on the upstream. The crest width and length of dam are 13.1m and 320m, respectively. The side slopes of main embankment are 1.4H:1V for upstream and 1.5H:1V for downstream. The main elements of embankment are the large-sized crushed rocks of which the most durable and high strength rocks are located on the outer part of the shell. The base rock is mainly composed of tuff and limestone, which are underlined by basalt formation. The impermeability on the upstream is provided by a plinth structure including a single-row curtain grouting with various depth (47-52m) and two-row consolidation grouting with 16m depth. It was designed to produce electricity of 322GWh per year with an installed capacity of 103MW.
Atasu CFRD is 118m high, with a total embankment volume of 3.8hm3. It is located 17km southeast of Macka Country and is designed as a multi-purpose dam to provide water supply to Trabzon city while producing electricity. Its construction started in 1998 and will be complete in 2008. When the reservoir is at operation stage with normal water level, the facility will impound approximately 37.5hm3 of water with a reservoir surface area of 0.83km2. The upstream and downstream shells are composed of large-sized crushed rocks. The upstream and downstream slopes of the main body are 1.4H:1V and 1.5H:1V, respectively. It has a 372m long crest which is 8m wide. A 6m thick alluvium, which is composed of sand, gravel, clay and silt mixtures, was entirely removed before construction of the main embankment began. Thus, the dam was based on good quality bedrock including basalt and andesite. The plinth structure has two-row consolidation grouting with 11m depth and single-row curtain grouting with maximum depth of 40m. Once complete it will produce 150GWh per year with an installed capacity of 45MW.
Gordes dam is located on Gordes river in western Turkey. It is a multi-purpose dam designed to provide drinkable water to Manisa city while irrigating an area of 14423ha. Its height from riverbed and foundation is 82.9m and 94.9m, respectively. It has the largest capacity of CFRDs in Turkey. When the reservoir is at normal capacity, the facility impounds 448.5hm3 of water with a reservoir surface area of 14.05 km2. Its construction started in 1998 and will be complete in 2008. The volume of total embankment is 5.5hm3. The upstream and downstream slopes are 1.4H:1V and 1.5H:1V, respectively. It has a 547m long crest which is 8m wide. The geological formation of the dam site consists of schist and marble. The plinth structure was designed with two-row consolidation grouting (10m depth) and single-row curtain grouting (maximum 37m depth) to provide the impermeability of the upstream toe.
Dim dam is designed as a multi-purpose structure in southern Turkey. This CFRD is 123.5m high from the riverbed and 134.5m high from the foundation. When the reservoir is at operation stage with normal water level, the facility will impound approximately 253hm3 of water with a reservoir surface area of 4.5km2. The upstream and downstream shells are composed of large-sized crushed rocks. The upstream slope of embankment is 1.4H:1V, while it is just 1.5H:1V for downstream face. The volume of embankment is 5hm3 . Its crest length and width are 365m and 7.8m, respectively. The embankment was founded on good quality bedrock including schist and limestone formations. On the plinth structure, a single-row curtain grouting and two-row consolidation grouting were constructed to obtain impervious conditions on the upstream toe. The depth of curtain grouting ranges from 26 to 44m, while that of consolidation grouting has a constant depth of 10m along the alignment of the plinth structure. It will produce 123GWh per year with an installed capacity of 38.3MW. It is expected to be complete in 2007, although it was the first CFRD to begin construction in the country.
Marmaris dam is the first CFRD in the country to be constructed by a municipality with international credit. The dam is 74m high from the foundation with an embankment volume of 1.3hm3. It is available to store 30hm3 of water at normal water level in the reservoir and provides drinkable water for Marmaris city and its vicinity. It was planned to supply water for the region up to 2040. The geological unit of foundation and abutments is entirely composed of peridotite rock. The upstream and downstream faces have the same slope value with 1.6H:1V. The plinth structure poses the consolidation and curtain grouting. The consolidation grouting was performed in two rows with 7.5m depth. The curtain grouting was constructed with maximum 40m depth.
Design and construction practice
CFRDs in Turkey were designed considering similar zonation materials. Their slopes are 1.0V:1.4H for upstream face, while it is 1.0V:1.5H downstream for five CFRDs in the country. However, the slopes of both sides is 1.0V:1.6H for Marmaris dam (see Table 2). A typical cross section of a CFRD constructed in Turkey is given in Figure 1. In general, the embankment is composed of quarry rock, processed-crushed and selected rock having a maximum grain size of 300mm as a cushion layer of concrete slab, impervious clay material to provide impermeable conditions on the heel of the embankment and natural river alluvium to protect the impervious zone compacted clay material upstream.
The main element of embankments is quarry rock. It is generally divided into two zones. The first one, which has a maximum grain size of 1000mm, was used on the downstream side, while the other part with a maximum grain size of 600 to 800mm is adopted in the upstream side. Both zones were compacted as 4 to 6 passes by a vibratory compactor having 18-20 tons of static loads. For the quarry rock located on the upstream portion of all dams, water jet (150L/m3) is systematically used during the compaction process. Fine content is limited to 2% for both zones. These materials are crushed from different rock types, which are originally composed of igneous and metamorphic rocks (see Table 2). For Dim dam only, the source of quarry rock is limestone.
Laboratory tests are performed on core samples of intact rock to obtain the physical and mechanical properties of embankment material. The long-term durability of these materials are evaluated by means slake durability tests. The limitations for some physical and mechanical tests are given in the related specification of State Hydraulic Works. More detailed data should be obtained before commencing final design (Brenner et al, 2005; Tosun et al, 1999).
The thickness of the concrete slab for all CFRDs in Turkey ranges from 0.3m to 0.8m. In fact, an empirical equation is used to calculate the thickness of concrete slab in design stage, as given below.
d= 0.30 + 0.003H
Where d is the total thickness of concrete slab and H is the water head at the related level. Both of them have a unit of meter. For more conservative design, the constant given in this equation can be taken as 0.35 instead of 0.30. The thickness of concrete face slabs is 0.35m at the crest level for Gordes and Torul dams, while it is approximately 0.3m for other dams at the same level. The plinth width changes from 4m to 10.5m for CFRDs in Turkey. The minimum and maximum values of plinth thickness are 0.4m and 0.7m for all CFRDs with exception of Marmaris. A typical cross section of a plinth structure constructed in Turkey is given in Figure 2.
Plinth structure is generally denoted by consolidation grouting and curtain grouting in order to provide impervious conditions on foundation and abutments. A single-row curtain grouting has been adopted with different depth as based on the nature of foundation rock. Consolidation grouting is generally applied on both sides of the curtain grouting as two-row with different depth (See Table 2).
For a CFRD in Turkey, an instrumentation programme is prepared to measure settlement and pressure under the dead load of embankment. This program should include at least three sections and four levels for each section. Figure 3 shows the locations for settlement gages and pressure cells for a section of a dam and the distribution of displacements measured from settlement gages. It is common in Turkey that for each location the change of settlement on embankments over time is evaluated to find a correlation with the predicted value by means of numerical methods (Tosun et al, 2006). For example, the total settlement value was measured as 232cm maximum for the case of end-of-construction for Kürtün dam. This was confirmed by the results of a finite element analysis (Tosun and Turkoz, 2006).
Seismic hazard and total risk analyses
The total risk for dam structures mainly depends on two factors: (1) the seismic hazard rating of the dam site and (2) the risk rating of the dam and appurtenant structures. The seismic hazard of a dam site can be based on the peak ground acceleration. This value is derived from the defined design earthquake which produces the main seismic loads. For preliminary study, the existing map of seismic zones can be used to estimate the seismic hazard of a dam site. The risk rating of the dam can be based on the capacity of the reservoir, the height of the dam, the evacuation requirements and the potential downstream damages. In general, the seismic and risk ratings are evaluated separately (icold, 1989). Recently, these two factors were combined to define the total risk factor for dam structure (Bureau, 2003).
The seismic sources were identified and the recurrence interval of earthquakes was estimated (Jiminez et al, 2001). As a result of an extensive survey and a search of available literature, several sources have been identified to help analyse the seismic hazard of dams in Turkey (Yucemen, 1982; Erdik et al., 1985; Saroglu et al., 1992). The data on historical and 20th century instrumentally recorded earthquakes for Turkey and its vicinity collected by the National Disaster Organization was considered as a basis for the seismic hazard analyses. Earthquakes that occurred within the last 100 years were also used for estimating seismic parameters. Due to the unavailability of strong motion records, various attenuation relationships were adopted (Ambraseys, 1995; Campbell, 1981) to calculate the peak ground acceleration (PGA) acting on dam sites. Throughout the study, seismic zones and earthquakes within the area having a radius of 100km around the dam site were considered. For all analyses, the peak ground acceleration was determined by considering the Maximum Credible Earthquake (Kramer, 1996). All procedures mentioned above can be executed by the DAMHA program, developed at the Earthquake Research Center of Osmangazi University-Turkey, that is working on the basis of geographic information system (GIS). The extensive studies have been realised for large dams of other types in Turkey (Tosun and Seyrek, 2005; Tosun and Savas, 2005).
For the study, six CFRDs located in different seismic zones were considered. The PGA and TRF values with risk classes obtained from seismic hazard and potential risk analyses by means of the program mentioned above are given in Table 2 for all dams. Two of them are very close to the energy source, while others are located on sites with low seismicity. The summary output of seismic hazard analyses for Gordes dam is given in Figure 4. The PGA values range from 0.01g to 0.20g. According to ICOLD classification, one of them will be subjected to moderate hazard rating. The others are identified as low hazard ratings (see Table 3).
Kurtun, Atasu and Torul dams are very close to each other and located in the same energy source. Kurtun dam is not located on active seismic region and its Total Risk Factor (TRF) value is not high. Therefore, it is identified as risk class II. This dam site is in excellent condition, but landslides at the reservoir area could be a problem for overall stability of project. Atasu dam is not one of the critical dams within the basin. It will be subjected to a peak ground acceleration of 0.05g with a ML of 5.3. Table 3 indicates that it seems safe for earthquake conditions. It is identified as class II with moderate risk. Torul dam will be subjected to a peak ground acceleration of 0.07g with a ML of 7.9 and it is not close to the energy source. The TRF value is high, even if it is located in an area with low seismic activity.
Dim dam will be subjected to a peak ground acceleration of 0.01g with a ML of 5.4 (Table 3). Its location has very low seismic activity. It is identified as class II with moderate risk. There could however be leakage problems in the foundation over the long-term for this structure.
There are two separate dams under the influence of near energy zone in western Turkey. One of them is Marmaris dam which is located in an active seismic zone. Its Peak Ground Acceleration (PGA) and Total Risk Factor (TRF) values, obtained from a detail seismic hazard and potential risk analyses, are 0.20g and 129.1, respectively. It is identified as risk class of III. Gordes dam, which is located on a shear zone, has a TRF value of 121.0 and it identified as risk class of II. It will be subjected to a peak ground acceleration of 0.09g with a ML of 6.7 (Table 3).
The settlement of embankment materials is a critical issue for the stability of CFRDs. The predicted values generally differ from measured ones. The authors state that this difference results from modelling limitation. Settlement on CFRDs depends on the nature and compaction degree of embankment material. In other words, the settlement predicted by numerical methods is largely dependent on the deformation moduli of embankment materials. Therefore, these parameters must be obtained from laboratory tests under actual stresses or close to these values. The durability characteristics of main embankment materials should be investigated with care to prevent settlement problems, which cannot be easily solved later on. Consequently, this study indicates that settlement analysis by a finite element model is a credible method when suitable parameters of embankment materials are used in the analysis.
The main requirement of an earthquake-resistant design of a dam is to protect public safety and property. Therefore, seismic criteria and analyses parameters for dams should be selected more conservatively than those for conventional structures, since failures can be more disastrous. The seismic hazard analyses performed on CFRDs in Turkey indicates that all dams with exception of Marmaris dam are not critical structures. Their hazard ratings are low (hazard class I). Marmaris dam is identified in hazard class II and its rating is moderate. As based on potential risk analyses, Torul and Marmaris dams are the most critical structures. Their TRF values are greater than 125 and they are classified into risk class III with high risk. The rest are classified as risk class II. For dams with a high risk rating, detailed further analyses should be considered in the future.
Table 1 Table 2 Table 3 View of Kurtun dam Kurtun dam Dim dam Dim Dam Torul dam Torul dam Atasu dam Atasu dam Figure 1: Typical maximum cross-section of Atasu dam Figure 1 Figure 2: The plinth structure of Dim dam Figure 2 Figure 3a: Location of settlement gages and distribution of settlement for a cross section of Kurtun dam Figure 3a Figure 3b: Location of settlement gages and distribution of settlement for a cross section of Kurtun dam Figure 3b Figure 4: Location of dams on the seismic-tectonic map of Turkey Figure 4 Figure 5: The summary output of seismic hazard analyses for Gordes dam Figure 5 Author Info:
Hasan Tosun, Professor, President of Dam Safety Association Ankara, Turkey; Serhat Batmaz, Section Director, State Hydraulic Works Ankara Turkey; and E.Orkun Ayvaz, Civil Engineer, State Hydraulic Works Eskisehir TurkeyTablesTable 1 Table 2 Table 3