Cavitation prediction in centrifugal pumps relies on the use of multiphase computational fluid dynamics (CFD). Analytical models represent cost-efficient alternatives but remain largely confined to planar, simplified blade geometries. This study aims to introduce a novel, coupled analytical-numerical framework allowing a fast prediction of the performance drop caused by cavitation.
Conformal mapping is applied to extend the classical potential flow solution for a cavitating flat plate cascade to centrifugal impellers with logarithmic curved blades. The shape and radial span of cavities are analytically predicted as a function of the cavitation number in a rotating blade passage. In addition, the head drop associated with cavity growth is predicted by means of a newly developed potential flow-CFD coupling algorithm. This framework links blade loading distributions obtained from single-phase non-cavitating CFD with the analytically predicted cavity span, thus quantifying cavitation-induced performance losses. Validation is conducted against experimental results and multiphase CFD simulations.
The developed framework has allowed to reproduce the head drop behaviour of a commercial impeller, with excellent agreement at flow rates above the nominal working point. Computational cost is reduced by one order of magnitude with respect to multiphase CFD calculation routines with no detriment to accuracy in head drop prediction.
A computationally efficient, easily reproducible method has been introduced for cavitation prediction. This has the potential to streamline the design process of centrifugal pumps, making it suitable for rapid parametric exploration or integration into optimization routines.
