Closed-Loop Turbulence Control-From Human to Machine Learning (and Retour)

Noack, Bernd R.

doi:10.1007/978-981-10-7542-1_3

Cited by 14 publications

(11 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The learning time may also be contrasted with machine learning control (MLC) based on genetic programming requiring thousands of runs. Intriguingly, MLC simulations and experiments (Noack 2019) indicate that most feedback mechanisms and associated control are simple and react on well-defined events or on oscillatory coherent structures. These findings encourage cluster-based control also for experiments.…”

Section: Discussionmentioning

confidence: 99%

Cluster-based feedback control of turbulent post-stall separated flows

et al. 2019

Self Cite

View full text Add to dashboard Cite

We propose a novel model-free self-learning cluster-based control strategy for general nonlinear feedback flow control technique, benchmarked for high-fidelity simulations of post-stall separated flows over an airfoil. The present approach partitions the flow trajectories (force measurements) into clusters, which correspond to characteristic coarsegrained phases in a low-dimensional feature space. A feedback control law is then sought for each cluster state through iterative evaluation and downhill simplex search to minimize power consumption in flight. Unsupervised clustering of the flow trajectories for in-situ learning and optimization of coarse-grained control laws are implemented in an automated manner as key enablers. Re-routing the flow trajectories, the optimized control laws shift the cluster populations to the aerodynamically favorable states. Utilizing limited number of sensor measurements for both clustering and optimization, these feedback laws were determined in only O(10) iterations. The objective of the present work is not necessarily to suppress flow separation but to minimize the desired cost function to achieve enhanced aerodynamic performance. The present control approach is applied to the control of two and three-dimensional separated flows over a NACA 0012 airfoil with large-eddy simulations at an angle of attack of 9 • , Reynolds number Re = 23, 000 and free-stream Mach number M ∞ = 0.3. The optimized control laws effectively minimize the flight power consumption enabling the flows to reach a low-drag state. The present work aims to address the challenges associated with adaptive feedback control design for turbulent separated flows at moderate Reynolds number.

show abstract

Section: Discussionmentioning

confidence: 99%

Cluster-based feedback control of turbulent post-stall separated flows

et al. 2019

Self Cite

View full text Add to dashboard Cite

show abstract

“…After, the forced CNM (C 4)–(C 7) is identified, a regression solver can be employed to optimize the control law with respect to a cost function. Genetic programming has proven to be a powerful method for this method in dozens of turbulence control experiments (Noack 2019). We remark that the deterministic–stochastic network dynamics rules out ‘simple’ control design based on local linearizations, but requires the numerical solution of a non-convect nonlinear optimization problem.…”

Section: Figure 30mentioning

confidence: 99%

Section: Figure 30mentioning

confidence: 99%

Cluster-based network model

Fernex

Semaan

et al. 2020

J. Fluid Mech.

Self Cite

View full text Add to dashboard Cite

show abstract

“…These values are adopted from a previous MLC jet mixing study in the same facility with a single minijet (Wu et al 2018a). The parameters are identical or close to the ones employed in numerous experimental studies as summarized by and Noack (2019). Elitism is set to N e = 1, i.e.…”

Section: Parameters and Control Landscapementioning

confidence: 99%

“…A new model-free self-learning approach for general nonlinear control laws has been developed by Dracopoulos (1997) for commanding satellite motion and was rediscovered in fluid mechanics as machine learning control or MLC (Gautier et al 2015). A review of dozens of MLC experiments and simulations is provided by Noack (2019). The first MLC experiment was set to enhancing shear-layer mixing with 96 jet actuators driven in unison and 25 hot-wire sensors for feedback control (Parezanović et al 2016).…”

Section: Introductionmentioning

confidence: 99%