Onwards and upwards4 June 2008
Construction of the Rogun hydroelectric project was suspended after the break-up of the former Soviet Union, but work at the site in Tajikistan is expected to be resumed in the near future. Roland Schmidt reports on the challenge of completing this complex project
The Republic of Tajikistan owns almost 4% of the world’s hydro power resources, ranking eighth in the world on an overall basis, and first on a per-capita basis. More than 95% of electricity in Tajikistan is generated by hydro plants. Rogun hydroelectric project is located on the Vakhsh river, and is the uppermost project in a cascade of plants which also comprises:
• Nurek (3000MW).
• Baipaza (600MW).
• Sangtuda 1 (670MW, under construction)
• Sangtuda 2 (220MW, in planning)
• Golovnaya (240MW).
• Perepadnaya (30MW).
• Centralnaya (15MW).
The Vakhsh river basin is located within the Pamir-Altai Mountains and, with a drainage area of some 39,000km3, it is fed from snowfields and glaciers. Because of its mountainous terrain, the Vakhsh is heavily loaded with sediments, especially in the upper stretch where Rogun is located. The annual runoff of the Vakhsh river at the Rogun site is approximately 20km3, corresponding to a long term average discharge of 635m3/sec and a mean annual power generation of up to 14.5TWh/yr (at a 335m dam height). At the dam site, the Vakhsh river flows through a narrow, 400-500m deep, V-shaped gorge, with slopes of both banks as steep as 50°. Rogun HEP will not only generate power but will also have positive effects on the existing downstream hydro power plants, particularly increasing the energy output during winter.
The Rogun project was conceived as a dual-purpose project for irrigation water management at Amu-Darya river and for hydroelectricity. The original project’s main components were:
• A 335m high clay core embankment dam.
• Reservoir with a volume of 13.3km3.
• Spillway comprising an intake shaft, a tunnel and an open chute.
• Underground power house with 6x600MW installed capacity.
• Outdoor switchyard and a 500kV transmission line system.
From 1965 to 1978, the first feasibility study and a design for construction of the Rogun project were developed by the Soviet design institute Hydroproject Tashkent. Preparatory construction works began in 1976 and the main construction activities started in 1982. In 1990, construction activities slowed down as a result of the collapse of the Soviet Union. The scope of work implemented from 1976 to 1990 constitutes a significant volume of construction, comprising both underground works and surface facilities. During a flood in 1993, the diversion tunnels were blocked, which caused overtopping of the 45m high upstream embankment cofferdam which was subsequently washed away.
In October 2004, Russian Aluminium (RUSAL) and the Government of Tajikistan reached an agreement to resume work with the aim of completing construction of the first stage of the project, primarily for the supply of power for existing and new aluminium smelters in Tajikistan.
In February 2005 RUSAL commissioned lahmeyer International of Germany to carry out a bankable feasibility study for Stage 1 completion of the scheme. The final report of this study was issued in December 2006. Work included a technical review of the project layout concepts and the elaboration of an optimal project layout. It was necessary to determine the extent to which existing construction works could be used in the completion project and to evaluate the residual value of these works. Lahmeyer’s initial task was limited to investigating the feasibility of Stage 1 of the project, with a required generation of 5.6TWh/yr, which corresponds to a dam height of 235m. Parameters for the staged extension were to be defined by the investor RUSAL and the Government of Tajikistan. Later, Lahmeyer was also commissioned to prepare a study on the optimal project parameters, to assess the final dam height and the related optimal installed capacity. This study identified the optimal height of the Rogun dam to be at or slightly above 285m, corresponding to Stage 2, with the installed capacity and related parameters shown in Table 1.
Scope of existing works
The existing Rogun scheme consists of multiple incomplete facilities erected during the construction period from 1976-90, part of which can be used for the new project after rehabilitation. The existing civil works comprise underground works, temporary and permanent roads, site utility systems (power and water supply, sewage), existing construction materials (e.g. more than 14Mm3 of stockpiled aggregates) and reservoir preparation works. On one side, the existing works, namely access roads and tunnels, provide immediate access to the construction site itself and thus will allow earlier diversion of the Vakhsh river than in the hypothetical case of the same site in virgin condition. On the other side, the existing works have deteriorated and are partly obsolete, requiring securing or demolishing. Most of the unused tunnels are to be backfilled and sealed off with concrete plugs. Such measures will incur additional costs but these can be offset by savings from the reduced construction time.
Hydraulic steel structures already onsite could partly be used in the completion project. In view of the outdated design and make of the existing electromechanical equipment, it is recommended that these parts be replaced with modern equipment.
Besides the extraordinary dam height, other challenges exist at the Rogun project. The heterogeneous bedrock at the dam site is composed of Jurassic and Cretaceous sedimentary systems, with some intercalations of gypsum. A wedge of salt fed by the Gaurdak salt formation is squeezing upward along the Ionakhsh fault, which crosses the reservoir and will therefore become submerged. Where this fault approaches the surface, the salt is leached, leaving a 10-12m wide zone of soft residue. With depth, the thickness of salt increases progressively. Hydrogeological and geotechnical modelling had indicated a risk of accelerated leaching of the Ionakhsh salt wedge with the raising of the reservoir level and, as a result, undesirable settlements at ground level. Therefore, at the river diversion stage of the original project, a system of protective measures were put in place, including local salt replacement, grout curtains acting as hydraulic barriers and injection wells to compensate hydrogeological and hydro-chemical gradients. Extensive grouting barriers are yet to be implemented at the construction completion project.
The territory of Tajikistan extends from the border of the Eurasian Plate with the Tienshan range in the north to the border of the Indopacific Plate with the Pamir in the southeast. Convergence of the two plates causes compression in the intercalated tectonic unit of the Tajik Depression. The compression results in an imbricate structure of thrust sheets in the sediments of the Tajik Depression.
From a seismo-tectonic point of view, the deep crustal Hissar-Kokshaal fault ranks foremost for the project. This is a reverse fault of the first order, which extends for several hundred kilometres and has accumulated a displacement of several kilometres. Another regional fault, the second order Illiak-Vakhsh fault, follows 1-2km to the south of the Hissar-Kokshaal fault. These thrust faults are rooted in a decollement surface, developed at 5-10km depth in the Jurassic salt formation.
The Rogun dam site is framed by two regional third order faults: Ionakhsh upstream and Gulizindan downstream of the site. Both are reverse faults of some 100km in length with a cumulated displacement in the order of 1km. Annual displacements in the order of 1-2mm were determined in the course of the project investigations. As could be established in the course of the very comprehensive microseismic monitoring during the filling of the nearby Nurek reservoir, the displacement on the Ionakhsh fault mainly occurs as aseismic creep. As a result of the lubricating effect of the Gaurdak salt, brittle events of small magnitude on minor faults in the adjacent strong sediments accompany this creep. Consequently, fairly large displacements can be predicted on the faults at the dam site, including possible sympathetic movements in the decimetre range on the higher order shears and faults in the rock mass forming the dam foundation.
Earthquakes occur in Tajikistan almost on a daily basis, mostly with magnitudes below 5.5 on the Richter scale. The strongest earthquakes recorded in the wider region were the Karatag earthquake of 1907 with M=7.3, and the Khait earthquake in 1949 with M=7.6. A peak ground acceleration in the horizontal direction of 0.56g was applied for the feasibility design of the dam type options studied. In view of the experience gained regarding reservoir induced seismicity at Nurek, the possibility of a triggered earthquake up to M=6.6 has been acknowledged.
Recommended project completion
The recommended project layout for completion comprises a rockfill embankment dam, in combination with an underground power house, on the left bank of the river, using the largely excavated cavern. The proposed layout of the dam, power house and river diversion is similar to the original design of 1978, whereas the design and layout of the spillways, mid-level outlet and tailrace tunnels were essentially modified by Lahmeyer. Another major modification introduced is the 175m high start-up dam, which is integrated in the main dam and allows commencement of power generation some 3.5 years prior to stage 1 completion.
Four different dam type options (rockfill embankment dam, CFRD, arch gravity and concrete arch dam), in combination with two powerhouse variants (surface and underground), were studied by Lahmeyer. Taking into account the heterogeneous bedrock with horizons of low shear strength, and in view of the present active faults, a conservatively designed embankment dam with clay core was identified as the preferred option, as this involves the least amount of uncertainties in design and construction. Embankment dams have a high resilience to foundation deformations and are therefore better suited than a brittle concrete dam or a CFRD. Furthermore, the Nurek dam located 70km downstream, with its world record height of 300m, constitutes a valid precedent for the selected dam type option. The Stage 1 dam is 235m high and requires a rockfill volume of 33.6Mm?. The main axis and the central clay core of the embankment dam remains at the originally planned location; thereby, maximum use is made of the already existing excavation for the core trench. The height of the dam can be increased up to el.1250 (or even higher) during Stage 2 of construction.
The optimal parameter study showed that the optimal installed capacity for Rogun is 2400MW. Three Francis turbines will therefore be installed in Stage 1, which are rated 600MW for the final stage. Provisions will be made for a fourth unit to be installed in Stage 2. During Stage 1, partial load operation will yield approximately 400MW per unit. In view of the large dimensions (inlet diameter 5.7m) at some 30 bar, gate valves will be applied as control valves, which will be integrated in the turbines. The main transformers will be located in a separate cavern 18.5m wide, which will also house the gas insulated switchgear (GIS). Power evacuation is by GIS and 500kV cables, through one out of the two existing cable galleries.
For each power generating unit, a separate power intake, headrace tunnel, pressure shaft and pressure tunnel is provided on the left bank. The three power intakes for Stage 1 are located at invert el. 1100, each equipped with trashracks and recesses for bulkhead gates. Different to the original design of 1978, new tailrace tunnels will be built, independent from the diversion tunnels, with power outlets integrated in the downstream RCC toe dam. In case of implementing the project with an integrated start-up dam, two turbine units with temporary power waterways will be commissioned on completion of the start-up dam, for early power generation.
Both partly existing diversion tunnels DT1 and DT2 on the left and right banks will be rehabilitated and used to divert the river during construction. The use of these tunnels, designed to cross the downstream dam embankment through a concrete culvert structure, requires extensive repair of large rockfalls. These had occurred at the intersections of temporary lined tunnels with the existing faults, particularly Fault 35.
The tunnels are each about 1480m long. In the 320m long part upstream of the gate section the tunnels run under pressure, except during very low flows, and have a D-shape cross-section (11x11m). Downstream of the control gates, each tunnel constitutes, hydraulically, an open channel, with a D-shape section of 14x17m. The intake invert of DT1 is about 10m lower than DT2, to facilitate river damming and diversion. By modifying the invert slope, maximum flow velocities downstream of the control gates kept below 24m/sec, for a 1 in 50-year flood.
One upstream and two downstream rockfill cofferdams are to be constructed, for diversion of river flow during construction. To divert the river through the left bank reach of diversion tunnel DT1 and the auxiliary tunnel, initially a rockfill diversion bund will be constructed up to el.1000. This will subsequently be extended up to el. 1044, which is sufficient to generate the head required for diverting the 50-year design flood of 3550m3/sec. Thus, a 70m high cofferdam will need to be constructed in one season, and will become an integral part of the main dam. In parallel, the rockfill cofferdam CD2 is built up to elevation 984m asl for protection of the dam site, which will allow for construction of the culvert section for river crossing of diversion tunnel DT2. In the subsequent dry season, the river will be diverted through DT2, which allows construction of the culvert section for DT1. In parallel, the auxiliary tunnel is closed and downstream cofferdam CD3 is constructed to seal the downstream dam foundation against the Vakhsh river.
The ability to pass large floods is crucial for the stability of the Rogun embankment dam, meaning a carefully elaborated and conservative design is applied for the spillway. In the 1978 design the spillway had a limited discharge capacity of 3400m3/sec, corresponding to almost 60% of the 10,000 year flood. Such a flood could only be passed by considering the full retention of the Rogun reservoir and ensuring six turbine units are in operation. From today’s standards this design philosophy is not suitable for a rockfill dam, which would not withstand overtopping for a longer period; hence the spillway constitutes the weak point in the original design.
The spillway concept proposed by Lahmeyer has therefore been designed for the PMF of 7500m3/sec, and it consists of a combination of the following structures:
• An open chute spillway on the right abutment, for at least 50% of the design discharge, corresponding to a specific discharge of about 83m
• Two tunnel spillways on the right bank, each for about 25% of the design discharge.
The free flow sections of tunnels and chutes each include a number of aerators to avoid damage on concrete surfaces by cavitation, which is a critical issue with flow velocities reaching up to 50m3/sec for Stage 2. Both Flip buckets at the exits of the chute and tunnels are located such as to ensure the jets will enter in a huge plunge pool, which is located downstream of the power outlets and the RCC toe dam, providing a sufficient water cushion depth for energy dissipation from heads of almost 300m. The large scale of the related excavation works requires detailed studies on work methodologies, during the next project stages.
A temporary mid-level outlet tunnel, with intake at el. 1070 on the left bank and aligned above the two diversion tunnels, crosses the downstream dam embankment and discharges into the plunge pool. The mid-level outlet will be temporarily used for controlled impoundment and to provide additional safety by being less exposed to bed loaded flow than the two diversion tunnels; it becomes redundant and will be permanently plugged after the completion of the final stage of the open chute spillway.
This spillway concept, which was proposed by Lahmeyer for the safe evacuation of floods from heads as high as 265m, necessitates a huge, partly pre-excavated plunge pool downstream of the dam. The extent of the related excavation and sloped protection works constitutes a major drawback of this proposal, which would be largely avoided if vortex tunnels were applied instead of the high level spillway tunnels with ski jump at their end.
Such a scheme, which could perhaps integrate the existing diversion tunnels, had been proposed and extensively tested by a hydraulic scale model by the Scientific Research Institute of Energy Structures (NIIES) in Moskau. The flows enters the vortex spillway through a vertical shaft with a crown section at its bottom, where the cross sectional area is reduced and accordingly the maximum velocity occurs. The eccentric connection to the adjacent tunnel section causes axial rotation of the flow and high energy dissipation due to wall friction of the vortex flow.
Up to now, only one prototype of a vortex-type spillway of comparable dimensions has been constructed – the spillway system of Tehri dam in India. This comprises an open chute of some 5500m3/sec capacity, four vortex tunnel spillways and two large open chutes. Experience gained during its operation, which started in 2006, was not available to Lahmeyer but should be used to confirm or to modify the design if this concept is considered a viable alternative for the Rogun project in the future.
Particular attention has to be paid to the risk of cavitation, which could be caused by instabilities of the vortex flow in the tunnel. Potential damages would be accelerated, if sediments enter into the shaft, eg via density currents. The high velocities near the tunnel wall cause abrasion, especially in the case of sediment transport. Further investigations would be required concerning the admissible operating conditions, the necessity to include a control gate, and the scale effects of the 1:50 model on the two-phase (water-air mix) flow.
If including vortex tunnels in the spillway concept it would be mandatory to maintain the open chute, which could, during an extreme flood, also be operated significantly above the designed discharge capacity, whereas the discharge of the vortex tunnel spillways is limited by the capacity mainly at the crown section.
To shorten the overall construction time of the project, the tendering period for the main civil works contract should be used for starting some preparatory works under pre-contracts. Some of these works are currently already underway, eg rehabilitation of access roads and selected administrative and residential buildings in Rogun City.
The filling works for the integrated start-up dam can only commence after the completion of the river diversion works. Then, after the integrated start-up dam has reached approximately el 1100m, the temporary turbine units can be commissioned. An estimate of about 8.5 years has been given for the duration of the Stage 1 construction works under the main contract for the recommended option of a clay core rockfill dam with integrated start-up dam, assuming monthly filling rates of up to 580,000m3 in total.
Heightening the dam to Stage 2 would require the additional backfilling of 8Mm3, which could be completed within less than two years.
The Stage 1 investment cost for the project, including contingencies, amount to about US$1.94B (2005 prices), of which 68% are civil works cost, 17% electromechanical equipment cost, and 11% mitigation costs for environmental and social impacts. The incremental investment cost for Stage 2, with almost double energy yield than Stage 1, only amounts to US$0.5B or about 25% of the Stage 1 investment cost.
An economic analysis was carried out for evaluation of the hydro power options studied from the viewpoint of the national economy in Tajikistan. This considered the project costs (without customs, taxes and other duties) and assessed the benefits of the project by the power and energy of Rogun, as well as the additional or reduced power and energy of the downstream cascade, and it valued benefits by the costs of the alternative (thermal) power generation.
The economic analysis clearly identified the layout option with rockfill embankment dam as the optimum solution, despite necessitating the highest investment costs and 10 months longer for construction compared with the option of a CFRD.
The dynamic unit cost (ratio of the cumulated benefit and cost streams, discounted over the disbursement and repayment periods) are US¢3.9/kWh at a discount rate of 8.5%; the dynamic unit cost reduces to US¢2.5/kWh in case of a combined development of Stage 1 and 2. Similarly, the project benefits increase, as shown by the economic internal rate of return (EIRR), which improves from 11% for the Stage 1 project to 16% for a combined development of both stages.
Evidently, the incremental capacity and energy (between Stage 1 and Stage 2) is highly attractive economically. For some US$430 per installed kW it would be feasible to double the energy output. These results clearly support the development of Rogun in a single stage to the optimum height of about 285m, provided that the demand for electricity in the local and regional power markets materialises in the near future.
Key Issues for completion
The first priority for project completion is a final decision on the ultimate height of the Rogun dam, which needs to be agreed among the Government of Tajikistan, riparian states involved in the project, and the investor and/or international financial institutes who may become involved in the scheme.
According to definitions by international financing institutions, the Vakhsh river, as a tributary of Amu-Darya, is a transboundary river, and as such, is beyond exclusive control by a single party. In line with regional and international inter-state agreements on the use of water resources, Tajikistan should notify and consult with the downstream states before modifying the Vakhsh river’s hydrology. The successful initiation of such a notification process is a prerequisite for safeguarding financing of the project and its implementation.
Normally, the optimal dam height (which corresponds to about 285m for Rogun), is chosen as the final dam height. In view of the favourable EIRR of the project, which declines only slowly when increasing the height above the optimum, and considering that the Rogun scheme and its size constitute a unique project opportunity, it is understandable that the Tajik Government opts to fully utilise the dam site potential. Also, an investor/financier may agree on a final dam height even slightly above the optimal one, provided that imminent transboundary impacts will be addressed and mitigated accordingly with the downstream riparian states of Uzbekistan, Turkmenistan and Afghanistan, which rely on the Amu-Darya river’s summer flows for irrigation. Obtaining the consensus by these states to the Rogun project would probably be facilitated, if the rule for the combined operation of the Rogun and Nurek reservoirs would essentially preserve the current discharge regime, as considered by Lahmeyer in the optimal parameters study.
After agreement has been reached concerning the final height, the Stage 1 feasibility study needs to be amended accordingly. Since RUSAL has in the meantime withdrawn as financier of the project, the Government of Tajikistan has invited international financing institutions to provide funds to update the study.
To safeguard the financing of the project it is crucial to justify the project economically. While RUSAL had planned to use a major portion of the generated energy for expanded aluminium production capacities, financiers such as the World Bank would opt for satisfying suppressed local demands and for energy exports.
This would require a detailed assessment of different power mode options (base, mixed, peak), in due consideration of the existing aluminium industry’s electricity demand and the corresponding proportion of electricity provided by Rogun to national and the most promising external markets. Within the frame of a further detailed demand-supply balance, the seasonal energy supply characteristics and the various energy qualities need to be modelled and assessed.
Dr.-Ing Roland Schmidt, Department Head of Hydropower and Hydraulic Structures Design, Lahmeyer International, Germany. Email: Roland.Schmidt@lahmeyer.de
|Environmental and social impacts|
The Stage 1 reservoir at Rogun would inundate about 1600ha of cultivated land, of which one-third is arable and the remainder is pasture. Biodiversity is comparatively low, and there are no important habitats or rare species which would be endangered. The population of the valley upstream of Rogun consists almost entirely of ethnic Tajiks who are generally farmers and herdsmen. A preliminary cost estimate of the environmental, social and reservoir preparation measures directly associated with Stage 1 of the project resulted in a figure of about US$250M, inclusive of the residual value of measures already implemented. The largest items are replacement of flooded transport, power and telephone infrastructure. The most complex issues are resettlement planning and implementation.