Iterative domain partitioning in pressure from PIV

Abstract

A technique that generalizes pressure reconstruction from planar and volumetric particle image velocimetry (PIV) velocity data is proposed to eliminate the need for a user-specified choice of boundary conditions prior to pressure gradient integration. The method extends the dual model concept proposed by Jux et al (2020 Meas. Sci. Technol. 31 104001), which partitions the measurement domain into irrotational (IRR) and rotational (ROT) regions using the total pressure coefficient as a discriminant. The latter approach requires an educated choice of the IRR boundary condition location as well as a threshold (minimum) value set for the total pressure coefficient in the IRR region. Iterative domain partitioning (IDP) approaches the problem by evaluating IRR and ROT iteratively and thereby significantly reducing the result sensitivity to the user’s initial choice of IRR boundary and threshold value. Furthermore, a near-wall adaptation of the integration, based on iso-potential lines, is introduced to avoid excessive error propagation when integrating PIV data along the object boundaries. The method is demonstrated to reliably converge towards a unique IRR-ROT boundary and minimize error propagation. The verification is carried out by examining three-dimensional (3D) particle tracking velocimetry data, reduced to planar, from the steady flow over a NACA0015 airfoil at α = [0°, 10°, 20°]. Monte Carlo simulation with randomly varying boundary conditions yields the lowest dispersion of pressure distribution, as a metric to represent the method’s accuracy. The approach is generalized to data from 3D PIV around a wall-mounted cube where the surface pressure reconstruction reveals a distribution consistent with the flow field topology. Also, in 3D, variation of the boundary conditions yields minimal fluctuations of the reconstructed pressure field when using IDP.