Investigation of Data Acquisition Conditions for Dynamic Mode Decomposition in Unsteady PSP Measurement

Abstract

Pressure-Sensitive-Paint (PSP) can measure a surface pressure field. Recently, fast-PSPs have been developed and applied to a visualization of the unsteady aerodynamic phenomenon. In general, the amount of unsteady aerodynamic visualization data is enormous. The Dynamic Mode Decomposition (DMD) technique can efficiently extract spatial and dynamic characteristics from enormous data. However, some data acquisition conditions for correctly discussing the flow field using DMD are still unclear. In particular, it is important to know the appropriate sampling frequency for specific aerodynamic phenomena. In this study, Karman vortexes in the wake of a square cylinder are visualized by the fast-PSP. Moreover, by using the unsteady PSP data, the time resolution of the input data is changed, and the influence on the DMD analysis for the Karman vortices is investigated. We find that the time resolution does not affect the flow structure of the DMD mode. On the other hand, the dynamic characteristics of the flow structure approach the true value when the time resolution is increased.