Prediction of S-shaped characteristics in reversible pump-turbines using different numerical approaches

Abstract

The prediction of characteristics and flow phenomena in reversible pump-turbines becomes increasingly important, since operations under off-design conditions are required to respond to frequency fluctuations within the electrical grid as fast as possible. Fulfilling the requirements of a stable and reliable operation under continuously expanding operating ranges challenges the hydraulic design and requires ambitious developments. Beyond that, precise estimations of occurring flow phenomena combined with a detailed understanding of their causes and mechanisms are essential. This study aims at predicting the S-shaped characteristics of two reversible pump-turbines by using different numerical approaches. Therefore, measurements at a constant wicket-gate opening of Δγ = 10° were done. Based on these experimental data, unsteady flow simulations are performed under steady and transient operating conditions respectively: Starting from the best efficiency point in generating mode, through the runaway, along the S-curve, down to operation in reverse pump mode. The hydraulic machines are spatially discretized in model size with a near-wall refinement of y+mean ≤ 5 and y+mean ≥ 30. The application of two different solvers discloses deviations in underlying methods. The turbulence modeling is basically executed by the k-ω-SST and the standard k-ϵ model. Focusing on higher order numerics, the Explicit Algebraic Reynolds Stress Model (EARSM) is selected in the commercial code and extended with an approach for curvature correction (EARSM- CC). In the open-source software, the four-equation v2-f model assumes the role of higher order numerics. The temporal discretization errors are observed using three different time-step sizes. As a supplement, experimental data obtained from the HydroDyna pump-turbine are used as additional validation, providing integral quantities and local pressure distributions at an operating point set on the S-curve. To sum this work up, a methodology is developed to approximate S-shaped characteristics and local effects within the hydraulic machinery properly.