A A Sotnikov* and I M Pylev** talk about their experiences of designing, studying and operating hydraulic turbines

An increasing number of hydroelectric plants with high-power turbines are being designed, constructed and put into commercial operation around the world, especially in North and South America, Russia and China. These turbines are up to 700MW in capacity, and to ensure that they are reliable and efficient, it is essential both to investigate important parameters during the design phase, and to draw on the experience gained from studying operational turbines.

The Sayano-Shushenskaya hydro power station went into service in December 1978 and has become a case study project for these high power turbines.

The project specifications for construction of the Sayano-Shushenskaya plant stipulated installation of 12 hydro units, each of 530MW power output. Further studies suggested that 650MW turbines could be developed for the plant; this would effectively reduce the length of the powerhouse by 20m. Investigating further, alternative designs for increasing unit power to 800MW and 1000MW were also considered.

While the alternative designs increased the power available from the plant they incurred additional costs. Firstly, the higher power turbines required more concrete in the construction, increasing from 0.24m3/MW (at 640MW) to 0.26m3/MW (at 800MW) and 0.3m3/MW (at 1000MW); secondly, there would be a cost incurred in gaining operating experience of the new turbine designs.

Finally, the following parameters were adopted:

•Maximum output 735MW (212-220m head)

•Rated head 194m

•Rated output 650MW

•Rotational speed 142.8rpm

•Number of units 10

A design incorporating a 20m wide spiral case with a 6m inlet diameter was adopted, after considering design and economic factors. The draft tube was 17m high and 27.3m long.

The unit mass of the hydro turbine equipment amounted to 1.9 kg/kW, with a maximum efficiency of about 96%. The general concept of the turbine design is shown in the diagram above. It has a spiral case of mixed design, where the load is borne partly by the steel shell and partly by reinforcement of the unit block (see diagram right). The design of the fastening between the spiral case and the stay ring, and of the reinforcing around the spiral case, reduced stress on the assembly and mitigated twisting of the stay ring shrouds and bending of the stay vanes.

The stay ring has four sections and is of welded construction. It has box-like shrouds and 19 cast stay vanes.The servomotors for the 20 guide vanes operate individually, which among other benefits makes it possible to use them for temporary runners, despite the larger travel required.

Using temporary runners

Temporary runners have been used because they make it possible to put the turbines into operation at 60m head, thus reducing the period required to recoup construction costs.

Among the characteristics that were required for the temporary runners were the following. They should:

•Fit into the water passages of the permanent turbines.

•Ensure turbine operation over the head range from 140-60m.

•Operate at the same speed as the permanent runners.

•Be reliable during the temporary operation period.

•Have minimal production costs.

•Only the runner and its components should be replaced and not the turbine itself.

The output of the temporary runners ranged from 155MW to 405MW, depending on the head. They were made of carbon steel and had protective facing plates of 4mm-thick stainless steel on the runner blade suction side.

The permanent runners would be made of stainless steel, and have 16 blades which are made of two parts of different thickness welded together. This construction considerably reduces the weight of the runner blanks and therefore the machining required.

Hydraulic studies at Sayano-Sushenskaya included developing new runners for both permanent and temporary operation. The guide vane mechanism was also thoroughly investigated as runners of different specific speeds would be in the same water passages and individual drive problems needed to be resolved. Experimental studies were supplemented with the analysis of quasi three-dimensional fluid flow and by solving the problems of dynamic interaction between rotating (runner) and stationary (guide vanes) cascades of profiles.

Experimental optimisation of the runners was carried out to minimise cavitation problems. The influence of operating conditions and submergence on cavitation ‘pitting’ was studied, using two-layer, easy-to-remove varnish coats on model equipment. A special data processing procedure enabled possible cavitation damage on the runner blades to be located and measured.

As dictated by the start-up sequences, operating time for the first two turbines, with temporary runners, should be 18000hrs and 12000hrs at heads ranging from 60m to 140m. Operating experience has confirmed the effectiveness of temporary runners during the long-term construction of large hydroelectric projects. Operation of the first temporary runner made up 36,900hrs, while the second one made up 26,200hrs, and this resulted in additional power generation of 10,400GWh and 7,400GWh, respectively.

Examining stress

Stress analysis was used to determine the stress–strain condition of the most heavily loaded elements: the spiral case and runner. The turbine block model intended for stress analysis was made to a scale of 1:64.

After 1000 cycles of loading by 2.86MPa hydrostatic pressure, the stresses in the spiral case structural elements had stabilised. Maximum tensile stresses at the inlet section amounted to 140MPa and bending stresses to 60MPa. At the junction of the spiral case shell and stay ring there were periodic stresses. The maximum stresses were at the stay vane leading edges (ratio between maximum and medium stresses was 1.5). Maximum stresses at the stay vane trailing edges were 140MPa. The level of stresses in the iron-concrete shell reinforcement did not exceed the permissible value.

Experimental investigations of the stressed condition of the permanent runners were carried out on a 630mm diameter model at 20m to 40m heads. When static and dynamic stresses in the runner blades were measured, maximum stresses were noted on the pressure side of the runner blade outlet edge, in the area of restraint at the runner band. Over the operational range of heads and loads, maximum static and dynamic stresses arise at 212m head and 735MW output: 120MPa and 14MPa, respectively. The main frequency of the dynamic stresses is the rotational speed frequency.

Operating experience

During trial operation, with the turbine running over a wide range of heads, a number of improvements have been introduced aimed at improving reliability and lengthening the time between overhauls.

Model tests carried out at the design stage showed that at heads below 145m the permanent runners would operate with excessive vibrations and pressure pulses. The first trial start-up of a permanent runner was performed at a reduced head of 100m. Because of the pressure pulses (with vortical frequency up to 370kPa and vibrations up to 600mm), further operation of the runner was recommended to be above 120m head.

Comprehensive field tests (efficiency, hydrodynamic, vibration and strength) of temporary and permanent runners were carried out in several stages, as the heads were increased. The turbines were tested with temporary runners at 60m, 100m, 120m and 140m heads. The tests showed that turbine performance characteristics (efficiency output) obtained by index methods are in agreement with the model test results. By considering amplitude and frequency characteristics (pressure pulsations, vibration of supporting components, shaft runout, power pulsations) the range of operating conditions was conventionally divided into three zones for Francis turbines. These were:

•Zone 1. Relatively low level of pressure pulses (up to 100kPa). The vibrational state regarded as satisfactory. Turbine operation was permitted, but not recommended due to low efficiency.

•Zone 2. There were excessive pressure pulses in water passages with vortical frequency 0.4-1.2Hz. Peak-to-peak amplitude of pulses in the draft tube at 62m head was as much as 220kPa (170kPa in the spiral case, 110kPa under the head cover). Power pulses reached 10%. The vibrational state was unsatisfactory and turbine operation was not recommended.

•Zone 3. The lowest level of non-steady processes (up to 80kPa), vibrational state is satisfactory and efficiency is at maximum. Vibrations of turbine bearing at heads ranging from 110 to 120m are 50mm, while at heads 60m and 140m they are up to 140mm. Long-term operation was recommended.

Under a head of 90m, the spider installed under the runner lowered the level of pressure pulses by a factor of 1.5 in zones 1 and 2, while in zone 3 the spider had no marked effect.

The turbines were tested with permanent runners at 110m, 120m, 140m, 150m, 175m and 190m heads. At 100m head, peak-to-peak amplitude of pulses in the draft tube and spiral case (almost at maximum output) reached 370kPa (frequency 22-25Hz) and 200kPa (vortical frequency 0.5-0.75Hz), respectively. Radial vibrations of the turbine bearing housing were 600mm and at the shaft runout were 0.9mm.

Installation of a six-way spider had the following effects: pressure pulses were reduced by a factor of 1.5-2; radial pulses of the bearing were decreased to110mm; vertical vibrations of the head cover were 200-270mm. Further investigations revealed that some pressure pulses in the spiral case can be identified as acoustic resonance of the system with shortened temporary penstock.

As the head was increased to 150m, the dynamic characteristics of the turbines were significantly improved.

As the head was increased from 150m to 175-190m, pressure pulses in the optimum operational zone did not exceed 30kPa, vibrations did not exceed 80mm, and the measured radial force acting on the turbine bearing amounted to 200-350kN.

Stresses in the runner blades were measured under steady-state (on-load) conditions at 196m head in the range of 0-670MW and under transient conditions (startup, shutdown, load rejections, changing over to synchronous condenser mode). Both dynamic and static stresses were determined.

It is the outlet edge on the pressure side, at the runner crown, that carries the highest load of the runner blade. The stresses here are as much as 130MPa under conditions of maximum output and 85 MPa under no-load conditions.

The stresses on the outlet edge at the runner band do not exceed 50MPa under all steady-state conditions (tension). The most unfavourable zone for turbine operation is 200-500MW, where the peak-to-peak amplitude of dynamic stresses on the outlet edge at the runner crown reaches 35MPa, and at the runner band it reaches 25MPa. Turbine start-ups are characterised by a high level of dynamic stresses (up to 100MPa) during the first 8-10sec.

At load rejections, the stresses reached 140MPa with the level of dynamic stresses being 80MPa (peak-to-peak amplitude) and main frequency being 2-4Hz.

Fatigue stress analysis based on the experimental data for 30 years of turbine operation (taking into account residual stresses) showed the reliability of the runner design.

Valuable experience has been gained in investigating, designing and operating the Sayano-Shushenskaya turbines. This has justified the feasibility of developing hydraulic turbines rated up to 1MW and confirms that the use of temporary runners at multi-unit hydro power stations is promising.