Jet quality and Pelton efficiency

Jet quality and Pelton efficiency T. Staubli, A. Abgottspon, P. Weibel Hochschule Luzern - Technik & Architektur Technikumstrasse 21 CH-6048 Horw

C. Bissel, E. Parkinson ANDRITZ Hydro Rue des deux gares 6 CH-1800 Vevey

J. Leduc, F. Leboeuf Laboratoire de méchanique des fluides et d'acoustique Ecole Centrale de Lyon 36 avenue Guy-de-Collongue Fr-69134 Écully

Abstract The paper gives a short overview on contents and goals of a research project on jets of Pelton turbines performed by the Hochschule Lucerne and the Ecole Centrale de Lyon in collaboration with ANDRITZ Hydro. Effects governing the quality of jets in Pelton plants are discussed, such as: the jet dispersion, which is defined by the widening of the jet with distance from the nozzle and includes the air water mixture on the jet’s surface, the jet deviation, which is defined by the deviation of the jets centerline from the theoretical axis, deformation described by the out of roundness of the jet, rotation of the jet, the turbulence level, secondary flows leading to streaks on the jet, entrapped air bubbles causing sudden expansions on the jet surface and finally droplets or splashing water impinging on the free jet in the housing of the turbine. With case studies in power plants the potential for improvement of efficiency is pointed out. The combination of thermodynamic efficiency measurement, flow visualization and CFD simulations allowed in the investigated cases good interpretation. The images taken of the jet in the power plants allowed quantitative evaluation of the jet diameter and its dispersion and eventually deviation. A clear relationship of increased jet dispersion and decreased efficiency could be found. Furthermore, a direct correlation of upstream bends and jet dispersion was found. This effect is induced by secondary flows.

1. Importance of jet quality for Pelton turbines Pelton turbines belong to the class of action turbines. The kinetic energy of the jet is reduced to a minimum when passing through the buckets of the runner. This deceleration occurs within a short distance within the buckets. The pressures needed for the deceleration drive the runner and effect the shaft torque for turbine operation. The order of magnitude of the deceleration may be typically 10 000 m/s2 or 1000 times the acceleration due to gravity. The long experience in developing Pelton turbines led to an optimization of this process and maximum efficiencies close to 92% and under favorable conditions even higher efficiencies are reached. Necessary condition for such high efficiencies is a high quality jet impinging on the Pelton buckets. The jet quality can be described with the uniformity of the velocity distribution within the jet, the absence of secondary flows and the compactness of the jet. Villacorta [1] was the first to measure velocity distributions within the jet and to investigate the jet dispersion with laboratory experiments and field tests. Recently pressure distributions within the buckets were analyzed by Perrig [2] Switzerland's hydro machinery park encompasses more than 196 Pelton turbines with power of more than 5 MW installed in a total of 107 hydro power plants. About 70 percent of hydro electricity generated in plants with storage capacity is produced with Pelton turbines, machines which are also especially suited for peak power generation (network regulation). With respect to the total hydroelectric generation in Switzerland Pelton turbines have a share of about 40 percent. 92 of the 107 Swiss Pelton hydro power plants were built before 1970. Many of them did undergo major refurbishment in the meantime - mostly only runner and control system replacement. In order to minimize the investment costs, housings and injectors of the turbines are usually not replaced. Thus, it can be stated that in most of the plants, the jet quality corresponds to the state of the art in Pelton design, before Villacorta [1] wrote his thesis in 1972. Before initiating the project, which will be described in the following, strong indications was found that in a great number of Pelton plants, efficiency gains in the order of 0.5 to 1 percent could be achieved by improving the jet quality as it is described in case studies by Staubli [3, 4] and Peron [5]. The state of the art of today’s procedures for Pelton rehabilitation projects focusing not only on runner and control system but also on jet quality is described by Parkinson [6]

2. The project and its topics In order to consolidate existing knowledge on jet properties, to acquire new understanding and to find ways to predict the quality of jets, a research project was initiated by ANDRITZ Hydro, in collaboration with the Hochschule Luzern, Switzerland, and the Ecole Centrale de Lyon, France. External funding of the research activities was provided by Swisselectric-Research. The project encompasses work packages with field tests, laboratory measurements, CFD and theoretical studies. The theoretical background showed to be rather complex since not only local Froude, Weber and Reynolds numbers have influence on the jet properties, but also the entire upstream flow history. In the upstream flow, secondary flow due to bends, bifurcations and elements within the flow (e.g. internal or external servomotors for the needle operation within the injector) hardly decay due to the usually high Reynolds number occurring in hydro power plants. These secondary flows influence the topology of the jet and lead to deformation (surface), deviation (jet axis) and rotation of the jet's deformations. Secondary flow also causes increased dispersion of the jet diameter. The closer the elements are to the nozzle exit, e.g. needle and mouth piece, the more important their influence on the jet development becomes. Old designs of injectors often show smaller needle and mouth piece angles compared to more recent designs, an indicator that the jet quality might not be optimum. Computational Fluid dynamics (CFD) shows to be a valuable tool for prediction of flow disturbances and secondary flows. With free surface models even jet deviations and surface deformations can be predicted. During the project, a serie of power plants were analyzed using CFD simulation. Upstream unsteadiness of the flow reflects on the jet surface. Such unsteadiness may be of stochastic nature, e.g. due to the flow turbulence generated in the penstock or due to deterministic phenomena such as pressure wave propagations or periodic vortex formations at cavities, bifurcations or support structures (Kármán vortices).

Fig 1: Definition of dispersion and deviation The jet dispersion is defined by the increase of the jet diameter with distance from the nozzle. Included in the definition of the jet diameter is the air water mixture surrounding the core of the jet. The deviation is defined by the difference of the jet centre line from the theoretical axis. Once the jet has exited the nozzle, further effects may negatively influence the jet and its dispersion, such as: ventilated droplets or water centrifuged from the bucket cut outs, splashing water being diverted within the housing or from deflectors onto the jet. Also interference with other jets or injectors may occur. Furthermore, highly pressurized air bubbles may eventually be enclosed in the oncoming water. When expanding after the nozzle exit this will lead to a kind of explosions observed on the jet surface. During the starting process of Pelton turbines such enclosed air and the resulting explosions are quite common. The effect of larger amounts of air can be easily heard outside of the housing during start-ups. During normal operation this noise disappears.

3. Investigated Pelton plants In the Swiss power plants listed in table 1, thermodynamic efficiency measurements and jet visualizations were performed during the project. The jet dispersion and deviation could be quantitatively determined from the visualization in most of the cases. Clear correlation between turbine efficiency and jet quality was observed.

Hydro Power Plant (HPP)

Sedrun Rabiusa-Realta Soazza

Type

Power

Discharge

Head

Theoretical jet diameter (at nominal and 100% needle openings)

[MW]

[m3/s]

[m]

[mm]

51.6

10.1

588

182

13

3

498

147

40

7

704

195

twin horizontal turbines, 2-nozzles each horizontal, 2-nozzles horizontal, 2-nozzles

Table 1: Main data of the investigated power plants The installation of equipment for prototype visualization is delicate since the best positioning of camera and lighting instrumentation can not be found on the basis of trial and error but must be based on experience due to the inaccessibility of the equipment. Furthermore, the mechanical forces of possible water impingement on the camera and lighting instrumentation requires a rigid installation. Housings for camera and lights have to be waterproof and measures must be taken to avoid condensation building up on the lenses. In order to achieve acceptable image quality under the adverse circumstances present in the housing of an operating Pelton turbine, special equipment is necessary. The camera housing and the stroboscopic lights were mounted within protecting housings in the shelter of the injector and cut-in deflector and could be adjusted at different distances from the nozzle exit with a stepping motor. All equipment could be operated from outside of the housing. 3.1 HPP Sedrun Figure 2 shows the result of the thermodynamic efficiency measurements which have been performed in the HPP Sedrun, with twin horizontal turbines, 2-nozzles each. The blue curve is representing the efficiencies measured for the 2 nozzle operation and the green and red points measurements with only one nozzle in operation. The needle opening of the one nozzle operation corresponds approximately to the opening at 0.9 of the maximum turbine output with two nozzles in operation. Thus efficiencies of 0.45 output (1 nozzle) have to be compared with the efficiency at 0.9 output (2 nozzles) and we observe that the green points of the lower injector reach almost the efficiency of the two nozzle operation. In contrast the upper injector lies about 0.5 percent lower.

0.30

0.5%

Turbine efficiency [-]

1-nozzle operation, lower injector

2-nozzle operation 1-nozzle operation, upper injector

0.40

0.50

0.60 0.70 Turbine output [-]

0.80

0.90

1.00

Fig.2: Turbine efficiency measured in the HPP Sedrun Figure 3 shows the result of the jet diameter measurements. The blue curve gives the theoretical jet diameter which can be calculated from the theoretical jet velocity and the flow rate. We observe that differences of the measured jet diameters between the upper and the lower nozzles are significant. By comparing with the efficiency measurements, we see that the larger diameter corresponds to the lower efficiency. This result gives a clear indication that the jet dispersion correlates with the efficiency. One possible reason for the larger dispersion of the upper injector might be the stronger bend feeding the upper injector.

200 theoretical jet diameter

180

Jet diameter [mm]

lower injector, z/D0=1.62 160 upper injector, z/D0=1.62 140 120 100 80 60 0

0.1

0.2

0.3

0.4 0.5 0.6 Turbine power [-]

0.7

0.8

0.9

1

Fig. 3: Jet diameter measurements in the HPP Sedrun at an axis distance of 1.62 times the nozzle diameter D0 of the injector. Typical jet images taken in the HPP Sedrun are shown in figure 4 for two different distances z from the nozzle exit (D0 = nozzle diameter).

z

z D0

= 2.33 (bucket impact)

100%

80%

40%

D0

= 0.68 (nozzle outlet)

Fig. 4: HPP Sedrun: Jet images of the lower injector at different turbine power and nozzle distances

1.5%

3.2 HPP Rabiusa-Realta The configuration of the inlet distribution pipes of two nozzle, horizontal Pelton of the Rabiusa-Realta HPP turbines are rather unusual. Each inlet is equipped with an individual spherical valve. To realize this, the lower injector is equipped with double bends in two planes. Furthermore, the angle between the upper and lower jets is of only 45 degrees leading to a high risk of interference between the two jets. The efficiency differences for the one nozzle operations is measured to be 1.5 percent for this turbine. The peak efficiencies measured with the lower injector are even higher that the one for two nozzle operation. This could be probably explained by the fact that for one nozzle operation, jet interference does not occur. However, also for the lower jet flow conditions are suboptimal. Also in the Rabiusa Realta HPP, the differences between the measured jet diameters of the upper and the lower are significant, the upper jet having a larger diameter. Here again the larger jet dispersion correlates with lower efficiency. A strong deviation of the lower jet can be observed, as it is visualized on figure 6. The ratio of the distances a/b is 1.18. This strong deviation of the jets centerline is caused by the upstream flow history dominated for the lower injector by the double bends in different planes just upstream from the nozzle.

Turbine efficiency [-]

1-nozzle operation, lower injector

0.20

2-nozzle operation 1-nozzle operation, upper injector

0.30

0.40

0.50 0.60 0.70 Turbine output [-]

0.80

0.90

Fig.5: Layout of the HPP Rabiusa-Realta (left) and measured efficiencies (right)

Fig. 6: HPP Rabiusa-Realta: Jet deviation at the lower nozzle visualized with distances of the jet surface from the bucket splitter

1.00

Fig. 7 HPP Rabiusa-Realta: CFD simulation of the secondary flows in the jet The secondary flows have been analyzed in detail based on a CFD study with the commercial code Ansys CFX and using the shear stress transport model (SST model). The simulations were performed at model scale, thus at lower Reynolds numbers than in the prototype. The resulting secondary flows predicted at the two nozzle exits is shown in the close ups of figure 7 that displays the velocity vectors projection in a plane perpendicular to the jet axis. The color scale presents the magnitude of the velocity vector projections. We observe a very unusual distribution of the secondary flow. The highest values of the secondary flow magnitudes correspond to about 3 percent of the jet axial velocity. The jet deformation due to the secondary flows was simulated with a separate approach solving the Reynolds equations with a homogeneous model for multiphase (water-air) simulation.

Fig. 8 Surface phenomena observed in Rabiusa-Realta (left) and Sedrun (right) Not yet fully explained are phenomena on the jet surface observed in the power plants of Rabiusa-Realta and Sedrun as captured in the images of figure 8. Obviously the surface is disturbed by sudden, singular effects, which could be splashing water impinging on the jet or entrapped air expanding out from the jet core. 3.3 HPP Soazza In the HPP Soazza again flow visualizations and thermodynamic efficiency measurements have been performed, but this time before and after refurbishment. Refurbishment included runner replacement and housing optimization based on full homologous model tests. Before refurbishment there was again a clear difference in the efficiencies between the lower and the upper injector of 0.8 percent. After refurbishment the efficiency level was increased by 2.5 percent and the efficiency difference between the lower and the upper injector disappeared almost completely. Visualization at the upper injector before refurbishment revealed that plenty of splashing water was hitting the upper jet from above. By improving the housing geometry, this water could be diverted and during the visualiza-

Upper injector, 1-nozzle operation, after refurbishment Lower injector, 1-nozzle operation, after refurbishment

0.8%

Turbine efficiency [-]

2-nozzle operation, after refurbishment

2.5%

tion after refurbishment the splashing water could not any more be observed. Accordingly, the efficiency of the upper injector increased to the same level as for the lower injector and also the jet diameter measurement after the refurbishment did no more show a difference between the upper and the lower jet.

2-nozzle operation, before refurbishment

Upper injector, 1-nozzle operation, before refurbishment Lower injector, 1-nozzle operation, before refurbishment

0.20

0.30

0.40

0.50 0.60 0.70 Turbine output [-]

0.80

0.90

1.00

Fig. 9: HPP Soazza: Turbine efficiency before and after refurbishment

Conclusions Thermodynamic efficiency measurement, flow visualizations and CFD simulations proved in their combination to be excellent tools for a fast detection of weak points of the turbines in several hydroelectric power plants and were extremely helpful for the technical optimization of the rehabilitation or refurbishment projects. The images taken of the jet allowed quantitative evaluation of the jet diameter, its dispersion and eventually deviation. A clear relationship between increased jet dispersion and decreased efficiency could be found. Furthermore, a direct correlation of upstream bends and jet dispersion was found. This effect is induced by secondary flows. Other phenomena as sudden disturbances observed on the jet surface need still further research.

Acknowledgements This study was made possible by a grant of Swisselectric-Research. Industrial funding was provided by ANDRITZ Hydro. The authors thank the staff and the owners of the power plants of Sedrun, Rabiusa-Realta and Soazza for their support during the measurements and for making these studies possible. References 1. Villacorta R., "Theoretische und experimentelle Untersuchungen an Einlaufdüsen von Freistrahlturbinen", Dissertation ETH Nr. 4678, 1972 2. Perrig, A., “Hydrodynamics of the free surface flow in Pelton turbine buckets“, These N° 3715, 2007 3. Staubli, T., Hauser Hp., “Flow visualization - a diagnosis tool for Pelton turbines“, IGHEM2004, Lucerne, Switzerland, 2004 4. Staubli T., "KW Fionnay, Grande Dixence SA, Strahlbeobachtung", Internal Report 2004 5. Peron, M., Parkinson E., Geppert L., Staubli T., “Importance of jet quality on Pelton efficiency and cavitation“, IGHEM2008, Milan, Italy, 2008 6. Parkinson E., Vullioud G., Richard P., Heimann A., Keck H., Hauser H.P., Keiser W., Rothenfluh M., “Systematic Approach of Pelton Rehabilitation Projects. Practical Experience from Case Studies”. Proceedings of the HydroVision 2008 Conference, HCI Publications, 2008.

The Authors Thomas Staubli graduated in Mechanical Engineering from the Swiss Federal Institute of Technology (ETH) in Zürich. After two years of post-doctoral research in the field of flow induced vibration at Lehigh University, Pennsylvania, he worked in experimental fluid mechanics at Sulzer Hydro (now ANDRITZ Hydro) in Zürich. He then headed the Hydromachinery Laboratory at the ETH Zürich. During this period he directed research projects in the field of hydraulic machinery. Since 1996 he is professor for Fluid Mechanics and Hydro Machines at the Hochschule Luzern. André Abgottspon graduated in mechanical engineering from Hochschule Luzern – Technik & Architektur. Since 2006 he is research assistant at the CC Fluid Mechanics and Hydro Machines. Pascal Weibel graduated in mechanical engineering from Hochschule Luzern – Technik & Architektur. Since 2007 he is research assistant at the CC Fluid Mechanics and Hydro Machines. Claude Bissel graduated in Fluid Engineering at the Electricity and Mechanical Engineering school of Nancy (France) in 1991. He joined ANDRITZ Hydro (Switzerland) in 2001 and now heads the hydraulic Reaserch & Development for Pelton turbines within ANDRITZ Hydro. Etienne Parkinson graduated in Fluid Engineering at the Hydraulic and Mechanical Engineering school of Grenoble (France) in 1987. Following a PhD at Ecole Centrale de Lyon (France), he worked as a research assistant in hydraulic machinery at the Swiss Federal Institute of Technology in Lausanne (Switzerland) until 1995 where he joined ANDRITZ Hydro. He is now head of the R&D department of Vevey (Switzerland). Julien Leduc graduated at the engineering school Ecole Centrale de Lyon (ECL) in 2007. He is working as PhD student on water flows in Pelton turbines and supervised jointly by the laboratoire de méchanique des fluides et d'acoustique LMFA and ANDRITZ Hydro (Grenoble-Vevey). Francis Leboeuf has been a professor in Ecole Centrale de Lyon since 1989. He is currently the head of the Turbomachinery team of the Laboratory of Fluid Mechanics and Acoustics. In the past, he has been the dean of research of Ecole Centrale de Lyon for 10 years. His main fields of research are the unsteady and unstable flows in turbomachines, with applications to aeronautics, astronautics and energy, including hydraulics and nuclear engineering. Since October 2006, he has been appointed by the board of the Ecoles Centrale to coordinate research activities between the five Ecoles Centrale in France and China.