Computational Flow Physics Group

Oliver Schmidt and coauthors posted a new paper on arXiv on spectral analysis of jet turbulence.  Data from several LES (computed at Cascade Technologies ) across a range of Mech numbers was examined using spectral POD .  The data highlight that, from a global point of view, there are frequency/azimuthal-mode regimes where the jet exhibits either low-rank or non-low-rank behavior.  The low rank structures are Kelvin-Helmholtz wavepackets that have been observed in past and previously modeled with parabolized stability equations.  The non-low-rank behavior is associated with Orr-like wavepackets which dominate the downstream, fully developed jet and, as suboptimal modes, exist also in the potential core region.  The work helps explain previous successes and failures at modeling structures in jets, and points to resolvent-based models for future progress.

Aaron Towne et al. have posted a new paper on spectral proper orthogonal decomposition (SPOD) on arXiv .   SPOD refers to space-time POD modes in a statistically stationary process, which can be readily estimated from the eigensystem of the cross-spectral-density matrix (in frequency space).  This form of POD goes back to the early work of John Lumley, Bill George, Joel Delville, Mark Glauser and other researchers, but has been largely overshadowed by 'snapshot' POD modes based on a space-only decomposition.  The paper derives both old and new results about SPOD, and highlights some nice properties of SPOD including a close relation with DMD, Koopman modes, and resolvent analysis.