This thesis investigates physics-based modeling of coherent structures in fully turbulent swirling flows, with the overarching goal of guiding active and passive flow-control strategies. The work combines linear stability analysis (LSA), resolvent analysis (RA), linear receptivity and sensitivity concepts, and energy-budget analyses to uncover the physical mechanisms governing the growth, sustainment, and dissipation of coherent structures that are primarily driven by linear dynamics. The focus is placed on the validity of these models under turbulent conditions, addressing the role of eddy-viscosity modeling and the accurate representation of turbulent base or mean flow fields. All modeling results are systematically validated against experimental measurements. The framework is applied to flows representative of engineering configurations, where flow instabilities can adversely affect performance and operational safety. The studies are conducted within the scope of three different publications. In Publication 1, a global helical instability, called precessing vortex core, is investigated in a combustor-like turbulent flow. LSA and adjoint LSA, combined with structural sensitivity and energy-budget analyses, are used to identify the dominant production mechanisms and receptive regions of the instability. Open-loop periodic forcing experiments are performed to study synchronization behavior, using wall-pressure measurements and proper orthogonal decomposition extracted from velocity fields obtained by stereoscopic particle image velocimetry. The adjoint mode is shown to provide a reliable first-order prediction of receptivity. Nonlinear effects of the observed synchronization dynamics are motivated to be captured by a weakly nonlinear analysis. Publication 2 addresses convectively amplified swirl fluctuations in a mixing-tube configuration representative of gas turbine combustors. Using RA based on flow fields from large-eddy simulations, the dispersion relation and spatial structure of the amplified wave packets are accurately predicted compared to experimental measurements using particle image velocimetry. The analysis reveals that the dominant coherent structures correspond to inertial waves whose amplification is strongly enhanced by mean shear, and identifies potential control routes to diminish these structures using receptivity and sensitivity information. In Publication 3, the global vortex-rope instability in a hydro-turbine draft-tube flow is examined using LSA based on Reynolds-averaged Navier--Stokes simulations, shape and base-flow sensitivity analyses, and experimental validation via wall-pressure measurements. A comparison between frozen and perturbed eddy-viscosity models shows that, while the eigenvalues and mode shapes are only weakly affected, the predicted shape and base-flow sensitivities differ substantially. Only the perturbed eddy-viscosity model reproduces the experimentally observed sensitivity trends, as the dominant mechanisms governing the eigenvalue change act through base-flow modifications that are neglected in the frozen eddy-viscosity model. Overall, the thesis demonstrates that LSA- and RA-based frameworks can provide valid and physically interpretable predictions of linear coherent structures and their receptivity and sensitivity in fully turbulent swirling flows in order to guide flow control applications.
Jens Müller (Thu,) studied this question.