When a turbulent boundary layer separates from a surface and reattaches further downstream, the resulting recirculation region is known as a turbulent separation bubble (TSB). These flows exhibit strong unsteadiness over a broad range of frequencies, leading to detrimental effects such as structural vibrations, noise generation, mechanical fatigue, and fluctuations in thermal loads, all of which pose significant challenges in various engineering applications. Despite advances in both experimental and numerical research, the understanding of the low-frequency breathing motion of TSBs (i.e., the low-frequency expansion and contraction of the bubble) remains incomplete.
This project uses FELiCS for global linear stability analysis and resolvent analysis to investigate the driving mechanisms of TSB unsteadiness. Experiments are performed to obtain base flow and unsteady field data. FELiCS then applies linearized methods, incorporating assimilated eddy viscosity fields obtained via tuned RANS simulations, adjoint-based RANS optimization and Physics-Informed Neural Networks. Modal decomposition techniques (POD, SPOD) are used to analyze the spatio-temporal flow structures.The results from these analyses will inform the development of active and passive control strategy to mitigate low-frequency unsteadiness in aeronautical applications.
Recent advances in our research revealed standing wave dynamics as the driving mechanism behind the breathing motion in turbulent separation bubbles. Using time-resolved PIV data, we uncovered previously unreported spanwise standing wave patterns caused by sidewall reflections. We constructed a resolvent-based standing wave model with FELiCS that accurately reproduced both the spanwise structure and frequency content of the measured modes. Additionally, the modal analysis capabilities of FELiCS enabled us to identify zero-frequency eigenmodes with growth rates depending on spanwise wavenumber, originating from centrifugal instability. Both the resolvent and modal analysis tools in FELiCS allowed us to systematically investigate the physical mechanisms driving low-frequency dynamics. The animation below shows a low-order model of the breathing dynamics based on the resolvent modes, with the isosurface marking U=0 at the edge of the separation bubble. This showcases how FELiCS enables the development of simple yet powerful low-order models for insightful visualizations.
Links
https://www.tu.berlin/en/flow/research/projects/dynamics-of-turbulent-separation-bubbles
Publications
Sorry, no publications matched your criteria.