A numerical analysis is performed to assess the capability of common simulation methods, in particular Ansys CFX, to predict the performance and NPSH curve of a centrifugal pump at very low specific speed for both, design and off-design conditions.
In all cases, we use an entire numerical model containing the impeller, the volute casing, the side chambers as well as suction pipe and pressure pipe. A three-dimensional setup is used, testing the following numerical models: steady, i.e. frozen rotor model, unsteady model accounting for the impeller movement and the relative impeller-volute position, single-phase flow as well as cavitating flow conditions.
The global performance of the pump is assessed in terms of pressure head, power consumption and pump efficiency for single-phase flow. Furthermore, the drop of the pump head and Net Positive Suction Head (NPSH) characteristics are analyzed for cavitating flow conditions. Numerical results are validated against experimental data.
Regarding non-cavitating flow conditions, the trend of the characteristic curves is well predicted, while absolute performance values differ from measured data significantly. The results of steady and unsteady calculations deviate from each other by less than 2%. Concerning cavitating flow, unsteady simulations have to be performed in particular for overload conditions, in order to obtain convergence of the solver. The trend of the measured NPSH curve is well captured with default cavitation model parameters. For nominal and overload, the predicted NPSH curve underestimates the measured one significantly.
Abstract. The most common method for simulating cavitating flows is using the governing flow equations in a form with a variable density and treats both phases as incompressible in combination with a transport equation for the vapour volume fraction. This approach is commonly referred to as volume of fluid method (VoF). To determine the transition of the liquid phase to vapour and vice versa, a relation for the mass transfer is needed. Several models exist, based on slightly differing physical assumptions, for example derivation from the dynamics of single bubbles or large bubble clusters. In our simulation, we use the model of Sauer and Schnerr which is based on the Rayleigh equation. One common problem of all mass transfer models is the use of model constants which often need to be tuned with regard to the examined problem. Furthermore, these models often overpredict the turbulent dynamic viscosity in the two-phase region which counteracts the development of transient shedding behaviour and is compensated by the modification proposed by Reboud. In the presented study, we vary the parameters of the Sauer-Schnerr model with Reboud modification that we implemented into an OpenFOAM solver to match numerical to experimental data.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.