Hydrodynamics of the free surface flow in Pelton turbine buckets

The design of Pelton turbines has always been more difficult than that of reaction turbines, and their performances lower. Indeed, the Pelton turbines combine 4 types of flows: (i) confined, steady-state flow in the piping systems and injector, (ii) free water jets, (iii) 3D unsteady free surface flows in the buckets, and (iv) dispersed 2-phase flows in the casing. The flow in Pelton turbines has not been analyzed so far with such detail as the flow in the reaction turbines, thus the understanding of the physics of key phenomena, i.e. the initial jet/bucket interaction or the jet cut process, is weak. Moreover, some machines present erosion damages that have not been satisfactorily explained so far. In the framework of this study, the flow in the buckets is investigated with 4 experimental and numerical approaches: (i) Unsteady onboard wall pressure measurements. 43 piezo-resistive pressure sensors are distributed on the bucket inner surface, backside and sides. (ii) High-speed flow visualizations (onboard and external). Small rigid endoscopes are fitted either in a bucket, to observe the relative flow, or in the casing to observe the jet/bucket interaction. (iii) Onboard water film thickness measurements. (iv) CFD simulations. The 2-Phase Homogeneous Model and the 2-Fluid Models are compared with the experimental data. The 2-Fluid Model appears to be more accurate, while the 2-Phase Homogeneous Model is too diffusive. The initial jet/bucket interaction evidences the probable occurrence of compressible effects, generating an outburst of the jet and leading to erosion damages. When the jet impacts the bucket inner surface, a high-pressure pulse, which amplitude is larger than the equivalent stagnation pressure, is generated, caused by compressible effects. The jet appears to remain attached to the backside of the buckets far in the duty cycle, and the separation to be dependent on the test head. The bucket backside acts as the suction side of a hydrofoil undergoing the Coanda effect, generating a depression, and in turn a lift force contributing positively to the bucket and runner torques. The depression nevertheless onsets cavitation, causing erosion damages. The main bucket flow is independent on the test heads. Mixing losses are put into evidence, either due to the crossing of streamlines or due to flow interferences. An analysis of the bucket power budget highlights the important contribution of the central area in terms of power transferred from the fluid to the bucket. The power signal of the whole runner shows important fluctuations that are modulated by the operating conditions. From the Momentum conservation equation, the respective influences of the different forces acting on the flow are evidenced. Even if the inertia forces, i.e. the deviation, Coriolis and centrifugal forces, globally dominate, the viscous and surface-tension forces outweigh the formers at the end of the evacuation process and in the jet separation process. Obtaining