As they stand, the hydro power facilities on the Ukraine’s Dniester river have badly affected the river system and its wildlife. Planners had long aimed to introduce a pumped storage plant to the Ukrainian system, but the capital investment required was beyond the resources of the country. A possible source of additional finance was the World Bank, but before considering whether a new plant could be funded, it was imperative to find out its effect on the ecosystem. Would it mitigate the Dniester’s problems, or would it add to them?

Several studies of the Dniester were carried out to provide a full picture of the river system. They included design and environmental impact assessments, co-ordinated by the Hydroproject Institute in Kharkov. In addition, studies commissioned by the World Bank included a feasibility study on economic and engineering aspects of the pumped storage project, carried out by electrowatt Engineering of Switzerland, and safety studies completed by Canadian experts. Netherlands-based Delft Hydraulics won the task of modelling the behaviour of the river, and assessing the likely changes caused by the pumped storage plant.

The river basin was modelled using the company’s generic model, Ribasim (see panel on p32). The study focused on the water resources and hydrobiological aspects of the pumped storage project and its international impact (specifically impacts on the riparian rights of Moldova).

There were some limitations to the study: water quality could be studied only for the effects of dilution; and it was not a complete EIA, as Ukrainian regulations require this to be carried out by Hydroproject.

An integrated system

The Dniester river is home to three dams with hydro power plants. The furthest downstream is the Dubossary reservoir, located in Moldova, which has been in operation since 1956. This dam, with its four 11MW turbines and 164M m3 reservoir, is subject to severe sedimentation.

The Delft study focused on the Ukrainian part of the hydro system, some 340km upstream of Dubossary and close to the city of Novodnestrovsk. The most upstream part of the system is the Dniester hydroelectric power station (DHPS). The power station comprises six 119.5MW turbines, each employing a head of about 40m, and the associated reservoir has a storage capacity of around 2000M m3.

Downstream of DHPS, the Buffer reservoir stretches some 20km to the Buffer dam. This reservoir, which was originally planned for completion in 1987, has an active capacity of some 70M m3, while the Buffer dam is equipped with three 9MW turbines, taking advantage of a head of 13m. The Buffer reservoir and dam are to be essential in operating the pumped storage plant.

The pumped storage plant itself will have three major components:

•An upper reservoir, located on a plateau around 150m above the lower (Buffer) reservoir. The upper reservoir will have an area of around 2.5km2 at its maximum operation level of +229.5m, with a volume of 38.8M m3. At its minimum, the reservoir will be at +215.5m and will have a volume of 6.1M m3. The volume of water to be pumped and then sent through the turbines will be 32.7M m3 at its maximum.

•The lower reservoir, which will in fact be the Buffer reservoir, which is at a level of +77m.

•Seven pump-turbine units using Francis turbines. In the first phase three units will be installed whose capacity is 324MW in turbine mode and 408MW in pump mode. Flows through each pump will reach around 270m3/sec in turbine mode and 250m3/sec in pump mode.

Aspects of the river

The nature of the river was changed for the first time when the Dubossary dam was built in Moldova. This flattened out flow variations and, in the absence of fish pass schemes, divided the river into two separate ecosystems for fish.

The construction of the DHPS reservoir in the 1970s altered still more the character of the river. The reservoir has a large capacity and it drastically reduces the natural dynamics and flow of the stretch of river to the Dubossary reservoir.

As regards temperature the DHSP reservoir is distinctly stratified, with cold water layers at the bottom. Water can only be released via the turbines from the deep layers so temperatures downstream of the dam are consistently higher than before in winter and lower than before in summer — 47km downstream the effect can still be seen as a 6°C temperature reduction in summer. The deep water layers also contain less dissolved oxygen and water downstream of the DHPS has been shown to have depleted oxygen levels as a result.

Turbidity has been decreased downstream of the dam and this has promoted macrophyte growth and zooplankton.

Like the Dubossary dam, the DHPS dam also has no fish pass facilities, so since 1987 the river has in effect had three separate ecosystems for fish. In the operating regime of the DHPS the turbines are started up and shut down almost instantaneously two or three times a day to provide peak power, and no water is released from the dam at other times. This results in an irregular flow of pulses downstream of the dam, which can result in level changes of up to 1m. These water level changes are highly unnatural for the Dniester river, and hydrobiological damage is reported, especially in the areas immediately downstream of the Buffer dam. This is particularly important as these areas have become substitute spawning grounds since the dams were built, when traditional areas were no longer accessible to the fish. Daily variations of up to 0.3m are considered acceptable in this region and one purpose of the Buffer dam is to suppress the pulses to meet this criterion.

Pumped storage in the system

When the pumped storage plant is added to the Dniester system it will introduce great water level variations in the Buffer reservoir over a typical 24-hour day. These variations will be superimposed on those caused by releases from the DHPS reservoir.

As noted above, large changes in water level are usually unnatural and should be avoided as far as possible. However, water level variations in the Buffer reservoir must be accepted to minimise flow variations in the river downstream.

To verify the extent to which water levels downstream of the Buffer reservoir can be minimised a second model was set up to consider the daily operation of the entire hydro power complex. It examined the water balance of the system (calculating water volumes and flows, but not considering power produced or consumed or travel times of discharged water). The basic relations modelled were:

•Water level – volume relations for the three reservoirs.

•Flow – water level relations for the downstream river.

•Head – discharge relations for the turbines in the DHPS and Buffer dams.

•Head – discharge relations for the control gates in the Buffer dam.

The model was prepared using Windows 3.1/Microsoft Excel 5.0. It incorporated data on infrastructure and equipment provided by the Dnistrohydroenergo Power Utility in Novodnestrovsk, and made simulations over a 24-hour period using time intervals of 10min.

The characteristics of the turbines in the DHPS permit operation for heads of 25-55m. The model assumes that each turbine always operates at maximum capacity; only the net head determines efficiency and therefore generated power.The model allows five operating cycles each day, and for each the number of turbines operating and the startup and shutdown times must be specified. Turbine flow for a period is assumed to be fixed.

A continuous inflow of 12m3/sec is assumed from local drainage into the reservoir, including leakage from the DHPS.

The model allows for two phases of operation of the pumped storage plant, when three and seven pump-turbines, respectively, are in operation. The pump-turbine units are assumed to operate at a fixed 150m head, using fixed flows of 270m3/sec in the turbine mode and 250m3/sec in the pumping mode.

The input includes specification on how many of the three turbines in the Buffer dam are operational. The operating regime of the Buffer dam turbines is limited to discharges of 80-150m3/sec for individual turbines, depending on the available head. Discharge through the control gates is limited to 300m3/sec per gate, so in optimum conditions the Buffer dam can control flows up to around 1050m3/sec. The model works by considering how many of the Buffer dam turbines are operational for a given head and required Buffer dam flow. Each available turbine discharges its maximum admissible flow.

Two control gates in the Buffer dam are opened as soon as the discharge capacity of the turbines is exceeded.

The following combinations of conditions were simulated:

•Three pumped storage units operational, upper reservoir 38% of final area.

•Seven pumped storage units in operation, upper reservoir encompassing full area.

•Turbine flows from the main reservoir to produce average outflows of 120m3/sec, 350m3/sec and 500m3/sec from the Buffer reservoir, on a 24 hour basis.

Results and conclusions

According to the simulation results, with the Buffer reservoir in full operation water level differences over a day would not exceed 0.2-0.3m for average downstream flows up to 500m3/sec. For flows of 100-150m3/sec the level differences are lower — in the range 0.05-0.1m. Operating the Buffer dam correctly can therefore make it possible to change water level variations of up to 7m in the Buffer reservoir, when the pumped storage plant is fully in operation, to variations of a few decimetres in the river downstream.

Following these simulation results, the following conclusions were made. With the buffer dam in operation, operation of the pumped storage plant:

•Would not impair the riparian rights of users of the Dniester who are downstream of the DHC.

•Would not affect seasonal flow requirements for the hydrobiological system.

•May improve temperature, oxygen, and turbidity conditions for the hydrobiological system.

•Must be subjected to rules of operation to suppress daily level fluctuations downstream of the buffer dam to below 0.2m.

Following these conclusions, the authors recommended that:

•Electronic equipment should be installed to regulate the hydraulic system for the control gates in the Buffer dam. This will enable control gates to be regulated continuously and will eliminate the level variations caused by operating the gates in intervals of several hours.

•Water level monitoring equipment should be installed downstream of the Buffer dam. If this is installed less than a few kilometres away it can be used as a feedback signal to control the hydraulic gates.

The remaining problem is that of the temperature changes in the river downstream. Although the operation of the pumped storage plant causes some mixing of the water in the Buffer reservoir it does not greatly affect the temperature. Studies are now being carried out under the direction of the Hydroproject Institute into methods of changing the temperature of the water at the intake of the turbines of the DHPS reservoir..

Following on

Following the provision of the Dniester Water Resources Study the pumped storage plant received positive results from funding agencies. The project was costed at around US$250M. To meet this requirement the World Bank made a preliminary pledge of credits totalling US$200M, while Dutch and Swiss grants were made available for funding preparations.