Evaluation of FFT-based cross-correlation algorithms for PIV in a periodic grooved channel

The turbulent flow in a periodically grooved channel is investigated using four different particle image velocimetry (PIV) algorithms. Each of the four methods is based on a fast-Fourier transform (FFT) cross-correlation between two interrogation areas. The study compares the four most common PIV cross-correlation algorithms with the same set of images. The measured average velocity vector field and the derived flow statistics, including the turbulent fluctuations and Reynolds stresses, were very similar to previously published DNS calculations for each cross-correlation algorithm examined. The difference between the results produced with the interrogation area deformation FFT algorithm and the DNS calculations was consistently the smallest of the four cross-correlation algorithms evaluated.