Robust principal component analysis for modal decomposition of corrupt fluid flows

Robust principal component analysis (RPCA) is a powerful technique from robust statistics that can be used to extract dominant coherent structures from flow fields corrupted with outliers and missing measurements. The effectiveness of RPCA on flows acquired experimentally and by simulation is demonstrated. In all cases, RPCA is able to de-noise these fields and vastly improve subsequent modal analysis, showing that RPCA can be used to robustly process particle image velocimetry flow fields.