Modal decomposition analysis of flow characteristics with a pump model

Reversible pump-turbines are currently the only commercially available large-scale energy storage devices, serving as the core components of pumped storage power stations. The safety and stability of these units have always been of concern, as their operational instability is often closely related to internal flow mechanisms. The complex fluid motion contains the vast majority of information in high-dimensional data. Thus, accurately extracting flow structures from complex flow fields and delving into their intrinsic mechanisms remain challenging. This paper introduces two flow field modal decomposition methods, Proper Orthogonal Decomposition (POD) and Dynamic Mode Decomposition (DMD), to reconstruct transient complex flow fields under pump operating conditions in pump-turbines. The aim is to study the characteristics of flow field evolution in a low-dimensional form. The results indicate that both methods can effectively capture the main features of the flow field, particularly in the blade region, showing distinct separation and vortex structures. This validates the feasibility of POD and DMD methods in achieving dimensionality reduction and reconstruction of complex flow fields.