An Analysis On The Characteristics Of Cavitation Flow Around A Hydrofoil Based On Dynamic Mode Decomposition

Cavitation flow around a hydrofoil is a complex unsteady multi-phase flow problem,which includes many cavitation behaviors such as the cavity growth,shedding,collapse and fusion.The cause of this problem involves such factors as local low pressure,re-entrant jet motion and associated vortices.The analysis of flow characteristics of hydrofoil cavitation flow can help to understand various physics phenomena and their causes in the process of hydrofoil cavitation flow.This is also related to the solution of the key to the problems of underwater vehicles,with important theoretical significance and engineering value.In this paper,based on the numerical simulation of the flow field around a hydrofoil with different cavitation numbers and angles of attack,the dynamic mode decomposition method is used to decompose and analyze the numerical simulation results.The main research contents and results are as follows:Based on phase homogeneous flow model,Schnerr-Sauer cavitation model and6)- turbulence model are used to carry out the numerical simulation of cavitation flow around NACA66(MOD)hydrofoil under various conditions.The comparison between the numerical and experiment results verifies the validity of the numerical calculation method used in this paper.The cavitation evolution process is analyzed by combining hydrofoil lift coefficient curve,cavitation evolution image,fast Fourier transform curve of velocities of the flow field.The results show that there are complex vortices near the cavitation attached to the hydrofoil,which are closely associated to the re-entrant jet and have an important influence on the evolution of cavitation.The larger space the cavitation can develop,the more obvious the effects of the re-entrant jet and associated vortices are.By using self-compiled dynamic mode decomposition algorithm,a series of modes and their corresponding eigenvalues are extracted from the flow field data obtained by numerical simulation.The evolution frequency and growth rate of the modes are calculated from the eigenvalues.By analyzing the dominant modes occupying higher energy,mode reconstruction found that these modes all contain certain cavitation evolution characteristics.By comparing with the streamline diagram obtained from the mode reconstruction,it is found that the cavitation evolution characteristics in the mode are all related to the vortex structure in the flow field,and the main evolution frequency of voids is consistent with the cavitation shedding frequency.The dynamic mode decomposition method can obtain the characteristic frequency which is difficult to capture by conventional methods such as fast Fourier transform,which can also extract corresponding spatio-temporal characteristics.Modal analysis shows that with constant attack angle and inlet velocity to keep the Reynolds number constant,the cavitation number is adjusted by changing outlet pressure.It is found that the main evolution period of cavitation increases with the increase of cavitation number.And when the angle of attack changes,the evolution frequency of cavitation first decreases and then increases with the increase of the angle of attack.The dynamic mode decomposition algorithm based on similarity transformation and singular value decomposition used in this paper can decouple and decompose the spatio-temporal data of flow field to obtain the modes and corresponding characteristic value and screen out the dominant modes with higher energy proportion by a certain mode selecting principle.Through the application of dynamic mode decomposition method,it is found that when dealing with highly unsteady flow field,the appropriate time-space scale of data is required.In addition,it is found that the more modes decomposed,the more accurate the mode reconstruction is,but the various primary and secondary characteristics are more likely to be split and dispersed in different modes,which may weaken the relationship to certain physical phenomena in a single mode.