Session H06: Flow Instability: Control (5:45pm - 6:30pm CST)

Inspecting the fluid-Structure Interaction of wall-mounted rectangular plates.

Understanding the dynamics of wall-mounted flexible structures is instrumental for a variety of applications, including energy harvesting, structural design, control, and locomotion. Several non-linear processes are poorly understood. We performed Fluid-Structure Interaction (FSI) simulation of a thin plate attached to a flat wall exposed to a uniform inflow (Jin et al., Physics of Fluids,30, 2018) under various Reynolds and Cauchy numbers. The structure is modeled as Kirchhoff thin shell, and Large Eddy Simulation is used for turbulence modeling. The FSI coupling is performed with the immersed boundary method (Gilmanov, Le, Sotiropoulos, JCP 300, 1, 2015). Simulated flow fields and plate deformations show good agreement with the laboratory experiments; this is also the case of the spectral structure of the plate tip motions. Characterization of the flow in the vicinity of the plate revealed the strong linkage between the wake and vibrations of the plate tip.   

Acoustic and hydrodynamic responses of turbulent cavity flows

Flow-acoustic interaction produces a self-sustained feedback oscillation in a compressible cavity flow. We examine this interaction at Re=10,000 and Mach 0.6 and 1.4, and the effect of harmonic forcing, by decomposing resolvent response modes into acoustic and hydrodynamic components with Doak’s momentum potential theory. The forcing frequencies considered include those organic to the unforced flow, as well as higher non-organic values that trigger more substantial energy amplification. In the first scenario, the response acoustic modes emerge at the trailing edge and propagate forward, confirming the role of self-sustained feedback oscillation. For high-order forcing frequencies, the response acoustic modes concentrate in the shear layer region and weaken towards trailing edge, without impinging in it. This indicates that despite their larger response, the higher-order acoustic modes participate less in the feedback oscillation of the cavity flow. The acoustic response modes correlate highly with the hydrodynamic response modes. We also investigate the compressibility effect on the acoustic and hydrodynamic responses in the supersonic cavity flow. The resulting view of acoustic and hydrodynamic responses provides insights that can help guide control strategies.   

Stability of laminar supersonic flow on compression ramps

Flow characterized by separation linked to interacting shock systems and boundary layers on compression ramps has been the object of continuous research for decades, aiming at accurately predicting heat transfer through vehicle surfaces operating in this flow regime. A ubiquitous feature observed in laminar flows calculated by standard 2nd-order accurate schemes is the presence of small amplitude oscillations in the flowfield, which hinders obtaining true time independent solutions for the subsequent linear stability analysis. A recently developed high-order numerical method provides machine-precision time-independent base states, without these oscillations. Steady laminar basic flows were analyzed regarding their global modal and non-modal linear stability using the state-of-the-art LiGHT solver to solve the respective EVP and SVD problems. Results showed the importance of the separation region and recompression shock in the flows stability, as well as the connection of these regions through the global mode amplitude functions. Non-modal analysis revealed a significant transient growth at short time leading to the conjecture that such growth may be significant enough to cause by-passing the modal transition path.   

Stabilization of the fluidic pinball with gradient-based Machine Learning Control

We propose a fast and automated gradient-enriched machine learning control (MLC) from steady open-loop control to multi-input multi-output feedback laws. The framework alternates between explorative and exploitive gradient-based iterations, generalizing MLC and the explorative gradient method (EGM). We stabilize the flow past a cluster of three cylinders, known as the fluidic pinball. The control laws for the independent cylinder rotations have been optimized over successively richer search spaces: symmetric and general asymmetric steady actuation and feedback control. As expected, the control performance improves with each generalization of the search space. Surprisingly, a non-trivial asymmetric steady forcing outperforms symmetric steady forcing. Intriguingly, the optimal feedback controller is a combination of asymmetric steady forcing and phasor control. We hypothesize that asymmetric forcing is typical also for other pitchfork bifurcation attractors. We expect that gradient-based MLC will be employed in many future multi-input multi-output control experiments.