Experimental analysis of the flow pattern of a pump turbine model in pump mode

Abstract

Listen

Translate article

Reversible pump turbines are the only means to store primary energy in an highly efficient way. Within a short time their operation can be switched between the different operational regimes thus enhancing the stabilization of the electric grid. These qualities in combination with the operation even at off-design conditions offer a high flexibility to the energy market. However, pump turbines pass through operational regimes where their behaviour becomes unstable. One of these effects occurs when the flowrate is decreased continuously down to a minimum. This point is the physical limitation of the pump operation and is very difficult to predict properly by numerical design without a model test. The purpose of the present study is to identify the fluid mechanical phenomena leading to the occurrence of instabilities of pump turbines in pump mode. A reduced scale model of a ANDRITZ pump turbine was installed on a 4-quadrant test rig for the experimental investigation of unstable conditions in pump mode. The performed measurements are based on the IEC60193-standard. Characteristic measurements at a single guide vane opening were carried out to get a detailed insight into the instabilities in pump mode. The interaction between runner and guide vane was analysed by Particle Image Velocimetry. Furthermore, high-speed visualizations of the suction side part load flow and the suction recirculation were performed. Like never before the flow pattern in the draft tube cone became visible with the help of a high-speed camera by intentionally caused cavitation effects which allow a qualitative view on the flow pattern in the draft tube cone. Suction recirculation is observed in form of single vortices separating from each runner blade and stretching into the draft tube against the main flow direction. To find an explanation for the flow phenomena responsible for the appearance of the unstable head curve also characteristic velocity distributions on the pressure side were combined with high-speed visualizations on the suction side of the pump turbine model. The results enhance the comprehension of the physical background leading to the instability and improve the numerical predictability of the instability in pump mode.