STATE OF THE ART REVIEW IN PUMP-TURBINE HYDRAULIC DEVELOPMENT

 A.Nichtawitz, E. Parkinson, M. Sallaberger and A. Sebestyen

VA TECH HYDRO                  Telephone: + 43 70 6987 3322

Lunzerstraße                            Fax: + 43 70 6980 2359

4030 Linz, Austria           E-mail:  alois.nichtawitz@vatech-hydro.at

http: www.vatech-hydro.com

Abstract: The following discussion reviews the research and development state of the art in pump-turbine hydraulics. Both experimental and numerical aspects are discussed. Indeed, more and more high-tech developments, [Parkinson, 1996], have to be used to fulfil the requirements of pump-turbines specifications, not forgetting the needs of deeper flow insight knowledge involved by the systematic introduction of Computational Fluid Dynamics (CFD) in design operations.

The paper gives an overview on the current state of the art in the field of pump turbine hydraulic development. At the same time the limitations are also clearly identified. In order to tackle the problems, which are not resolved yet, research programs were launched including external resources from the Universities of Lyon, Lausanne but also from a team at Sulzer Innotec.

Finally, one representative pump-turbine project in China, which recently has been put into operation, illustrates as an example the application of these long-term research and developments activities.

1    INTRODUCTION

Following the potential of pump-turbine projects in the world, specific development work is needed for this type of machine which combines both turbine and pump requirements. Indeed, an optimum turbine is not necessarily the optimum pump. Special attention must be paid to all components and their interactions. Therefore, a new pump-turbine design always is a compromise between both operating conditions defined by the project specifications. More and more projects also address the upgrading and refurbishment of existing machines [Sallaberger, 2000]. Mostly only the key parts of the machine are replaced by new ones and often the modernisation is coupled with an increase in power output. Together with the geometrical restrictions of the existing machine and complex hydraulic conditions this makes high demands on the hydraulic design.

Following an overview of the state of the art in standard pump-turbine developments, some representative research projects are shown to illustrate the trends in the activities requested to understand the physics and finally to increase the global performances of pump-turbines. The physical behaviour of the instability domain of the runner in pumping mode requires a multi-domain approach where interactions play a large role. In such a case, in order to characterize and understand the physics of such a phenomenon, unsteady Laser Doppler Velocimetry (LDV) and Particle Image Velocimetry (PIV) measurements help in identifying the control parameters. The final goal is to deliver high performances for the machines with safe and smooth operation by taking advantage of all research activities undertaken in the pump-turbine field.

2    RESEARCH AND DEVELOPMENT

2.1    State of the art

2.1.1    Experimental Analysis

Experimental analysis is involved in various steps of a pump-turbine development. Measurement of the model performance of any design involve characteristics, efficiency and cavitation analysis. Steady state flow surveys, pressure pulsations and hydraulic torque on guide vanes are also part of a testing program.

Such requirements imply the use of a turbine test rig following IEC recommendations with a highly trained staff as it is directly involved in the analysis work and quality control of the considered design.

A continuous effort is devoted to enhance the quality, accuracy and operation of our test rigs, see Table 1. It involves the massive automation of their control, tending towards half-automatic test rigs.

Combined with the introduction of imaging systems, it contributes to a productivity increase of these rigs. Speed in the development is important factor therefore in order to minimise development time. New technologies like stereolithography for model runner fabrication were introduced to follow these constraints.

Table 1   Test rigs of VA TECH HYDRO used for pump-turbine developments

2.1.2    Computational fluid dynamics

In the following an example is given how capable tools of CFD already are but it also reveals the current limitations at the same time. In this example the recovery coefficient of the draft tube in turbine operation is calculated. The target of this work was to study the coupling between the runner and the draft tube [Bellet, 2000].

This recovery coefficient z of the draft tube is calculated from a coupled runner-draft tube simulation and by an isolated draft tube simulation with experimental inlet boundary conditions. These are obtained with a five-hole pressure probe.

The comparison is applied on different operating conditions, detailed in Figure 1, and compared to the experimental values, presented as lines, in Figure 2. The circled symbols show results obtained with isolated draft tube calculations. The black spots show the results obtained with coupled simulations.

For many operation points correct results appear for both approaches, however it was not possible to get reasonable results for point A2 in the isolated mode because of the simplified inlet conditions. These inlet boundary profiles do not provide sufficient information about the near-wall scalar fields such as turbulence. The need for high precision experimental information is obvious.

Fig. 1    Discharge-head chart in turbine operation

Fig. 2    Comparison of recovery coefficient of draft tube

Fig.3    Discharge-head chart in

pump operation

The previous comparisons are not sufficient to check internal flow behaviours, especially among the unsteady patterns within the pump instability domain for example, see Figure 3. 

Indeed, for example, using the stage rotor-stator assumption may be misleading in the region of instability if not analysed with great care and experience. 

Going further along the characteristics, i.e. taking into account the unsteady flow patterns requires more sophisticated rotor-stator interfaces such as transient simulations.

2.1.3    Current limitations

The techniques as described in the above chapters are sufficient to tackle standard design control or improvements. However, to develop our know-how, considering future research work, these can appear as insufficient. The most important domain to be further developed seems to be the development of extensive Navier-Stokes Computational Flow Dynamics analysis, steady or unsteady, with coupled components such as distributor/runner/draft tubes.

One question also still remains open when calculating a runner/distributor flow in pump mode in the vicinity of the instability onset as for the validity of the basic hypothesis of the k-e turbulence model.

Flow simulations ask for extensive experimental validation work [Muggli, 1999]. For instance, specific information is required to validate such coupled analysis. Especially, as different numerical coupling methods exist, with various respective costs in terms of computing time and memory requirements, the validity range of such options must be carefully established. This last point must be clearly emphasised as it represents a key factor in our CFD Quality Policy.

2.2    Research activities

A pump-turbine model of nq=66, see Figure 4, has been extensively studied in order to address the questions raised by the validation of the various rotor-stator interfaces over the entire pump and turbine operation characteristics [Riedelbauch, 1996].

Fig. 4    Pump turbine model

The experimental model is equipped with flow surveys sections and specific windows for laser measurements as described in [Ciocan, 1998].Velocity vectors fields were defined in different sections of the distributor height. Laser Doppler Velocimetry but also unsteady pressure measurements were taken in the gap between the runner and the wicket gate. The experimental set also included turbulence measurements.The model tests were conducted on test rigs of both CREMHyG (Grenoble, France) and VA TECH HYDRO (Vevey, Switzerland).

The unsteady flow interaction between the runner and the wicket gate is currently being studied within the European ESPRIT project HPNURSA (High Performance numerical Unsteady Rotor Stator Analysis). Examples of results obtained at the same discharge for different rotor-stator models are shown at Figures 5a and 5b.

Fig. 5    stage (left) and frozen-rotor (right) simulations of a runner and wicket gate at discharge 0,10 (see Fig.3).

Developments are directed towards the improvement of post-processing associated with the huge amount of information generated from these unsteady simulations. These new developments integrate animation, visualisation and data mining technologies. 

3    THE NEW  PUMP-TURBINE FOR TIAN TANG

3.1    Project description

The TIANTANG pump-storage power plant is located in the province of Hubei in China at the Tian Tang River. Two units of reversible pump-turbines are installed (see figure 6). The main figures of these units are detailed in the following Table 2.

Table 2    Main data of tian tang pump-turbine

3.2    Hydraulic design

3.3    Performance

Many elements of the development process as described were used in order to secure the guaranteed performance of the Tian Tang machines. In a first step a runner was designed and “tested” numerically, see Figure 7, to fulfil the requirements of the specification. Various alternatives of profiles were compared in this stage of development. After a successful development the client witnessed the final configuration in 1998 during the Model Acceptance Tests.

Most of the fabrication was done at Kvaerner Hangfa Works situated in Chengdu, Province of Sechuan, except the runner which was fabricated at the Machinery Works of former Voest MCE in Linz, Austria. The highly accurate fabrication is the best pre-condition for a very satisfying operation at site.

The smooth and reliable operation over the complete range of operation is reported. It is a very positive feedback for the development engineers in the working in the “numerical test rigs” and on the physical model as well.

4    CONCLUSIONs

The paper is started with a review of the state of the art in the development of pump-turbine technology. A description of the existing development tools like computational fluid dynamics and physical model testing is given. As documented by specific research data the limitations of the existing tools become quite obvious.

The research programs running now are clearly dedicated to overcome these limitations and to create modern tools capable to describe and predict phenomena that cannot be fully understood yet.

One of the key issues of the ongoing research work is the development of extensive Navier-Stokes computational fluid dynamics e.g. unsteady flow simulations with coupled components such as distributor/runner/draft tube. Of equal importance is the validation of various methods of coupling and various turbulence models by more sophisticated measurement techniques.

The paper is ended by a short presentation of a recent pump-turbine project in China for which various elements of the described development tools have been applied.

Acknowledgements

The authors would like to thank Professor Kueny from the Swiss Federal Institute of Technology of Lausanne, IMHEF-LMH, and also the involved team at Innotec for their valuable contributions to the whole research work. 

References

[1]    Bellet L., Garcin H., Parkinson E. - Analysis of runner draft tube numerical coupling on Kaplan and Pump-turbine cases - XX IAHR 2000 - Charlotte, USA.

[2]    Ciocan, G.D., Desvignes, J.F., Combes, J.F., Parkinson, E., Kueny, J.L. - Experimental and numerical unsteady analysis of rotor-stator interaction in a pump turbine - XIX IAHR Symposium - Singapore, 1998.

[3]    Muggli F., Eisele K., Zhang Z, Casey M. et al – Numerical investigation of the flow in a pump turbine in pump mode – ImechE – London, UK, 1999.

[4]    Parkinson, E., Neury, C., Vullioud, G., Walther, W. - An optimum combination of numerical experimental tools for pump turbine developments - Modelling, Testing & Monitoring for Hydro Powerplants - II, Lausanne, Switzerland 1996.

[5]    Riedelbauch, S., Klemm, D., Hauff, C. - Importance of interaction between turbine components in flow field simulation - XVIII IAHR Symposium - Valencia, Spain 1996.

[6]    Sallaberger, M., Staehle, M., Thoma, W., Kiedrowski, T., Krasicki, R., Lewandowski, S. – Major progress in upgrading of reversible pump turbines – Hydro 2000, Bern, Switzerland 2000.

[7]    Sebestyen, A., Jaquet, M., Keck, H., - CFD-design procedure for runner replacement of reversible pump turbines, - XIX IAHR Symposium 1998, Belgrade.