Surface pressure determination: a comparison between PIV-based methods and PSP measurements

Abstract

The present work investigates the challenges in measuring the surface pressure distribution using particle image velocimetry (PIV) in comparison with pressure sensitive paint (PSP) measurements. The aim of the investigation is to assess if PIV may be regarded as a complementary method to PSP, or even as a possible replacement at least for 2D flow analysis or at low speed, where PSP fails. The investigation is based on two different test cases: (1) the attached flow around a NACA-0012 airfoil and (2) the separated flow around a backward-facing step (BFS). Both cases are investigated at transonic flow conditions. For the airfoil, both PIV and PSP measurements were conducted for a direct comparison. At such high Reynolds number, the boundary layer thickness is very small, and therefore, the pressure in the flow field outside the boundary layer can be easily determined from the velocity data with an isentropic flow assumption. The results of the PIV-based surface pressure reconstruction and the one of PSP are compared with the static pressure taps installed on the model surface. For the BFS test case, synthetic PIV images are produced from a Large Eddy Simulation (LES). In this case, flow separation as well as strong viscous effects appear which inhibit the use of the isentropic flow assumption. Therefore, the Reynold averaged momentum equation is directly used and the pressure is determined solving the Poisson equation which is obtained by applying the divergence operator to the momentum equation. The results are compared with the pressure distribution on the surface given by the LES simulation. This case aims to analyze the accuracy of PIV-based surface pressure determination in case of massive flow separation. It will be shown that the deviations to the expected results are caused by the spatially discretization as well as by spatially averaging, which occurs in a standard PIV measurement. In addition, the effect of the velocity uncertainty on the estimated surface pressure distribution biases the result significantly. This implies that PSP will be hardly replaced by PIV.