Comparative study of reduced-order modeling method for the cavitating flow over a hydrofoil

Abstract

As a high-dimensional complex nonlinear dynamic system, the analysis of the essence of flow has always been a difficult problem, especially in the flow including phase change. In recent years, it has become a feasible method to reduce the dimension of flow structure by reduced-order modeling (ROM) methods. In this paper, through the cavitation numerical simulation of NACA0015 hydrofoil, two ROM methods are used to reduce and restore three different cavitation respectively-proper orthogonal decomposition (POD) and dynamic mode decomposition (DMD). The applicability of two methods in cavitation is discussed and reasons are analyzed. The results show that for stable cavitation, POD, DMD methods can accurately restore the flow field of a few modes with high energy. For unstable cavitation, only POD method can restore real flow field well. This situation is mainly due to the fact that POD, DMD method are applicable to different energy ratios, and different main mode selection criterion of DMD will lead to different main mode. ROM can greatly simplify the complexity of flow. Selecting a reasonable ROM can improve the accuracy of a small amount of database, and provide a basis for intelligent prediction of flow analysis.