Bulletin of the American Physical Society
69th Annual Meeting of the APS Division of Fluid Dynamics
Volume 61, Number 20
Sunday–Tuesday, November 20–22, 2016; Portland, Oregon
Session L7: Flow Control: Feedback, System and Model Identification |
Hide Abstracts |
Chair: Jeff Borggaard, Virginia Tech Room: B115 |
Monday, November 21, 2016 4:30PM - 4:43PM |
L7.00001: The Stability Region for Feedback Control of the Wake Behind Twin Oscillating Cylinders Jeff Borggaard, Serkan Gugercin, Lizette Zietsman Linear feedback control has the ability to stabilize vortex shedding behind twin cylinders where cylinder rotation is the actuation mechanism. Complete elimination of the wake is only possible for certain Reynolds numbers and cylinder spacing. This is related to the presence of asymmetric unstable modes in the linearized system. We investigate this region of parameter space using a number of closed-loop simulations that bound this region. We then consider the practical issue of designing feedback controls based on limited state measurements by building a nonlinear compensator using linear robust control theory with and incorporating the nonlinear terms in the compensator (e.g., using the extended Kalman filter). Interpolatory model reduction methods are applied to the large discretized, linearized Navier-Stokes system and used for computing the control laws and compensators. Preliminary closed-loop simulations of a three-dimensional version of this problem will also be presented. [Preview Abstract] |
Monday, November 21, 2016 4:43PM - 4:56PM |
L7.00002: Modelling and Feedback Control of Bistability in a Turbulent Bluff Body Wake Rowan Brackston, Andrew Wynn, Juan Marcos Garcia de la Cruz, Georgios Rigas, Jonathan Morrison The turbulent wake behind many three-dimensional bluff bodies exhibits a bistable behaviour, the properties of which has been the subject of significant recent interest. This feature of the wake is known to contribute to the pressure drag on the body and is relevant for geometries representative of many road vehicles. Furthermore, due to its high visibility from surface mounted pressure measurements, it is a feature that may be observed and controlled in real-time. In Brackston et al (\emph{J. Fluid Mech.}, 2016) we have recently demonstrated such a feedback control strategy that aims to suppress the bistable feature of the wake. Starting from a stochastic modelling approach, we identify a linearised model for this mode of the flow, obtaining parameters via a system identification. The identified model is then used to design the feedback controller, with the aim of restoring the flow to the unstable, symmetric state. The controller is implemented experimentally at $Re \sim 2.3 \times 10^5$ and is found to both suppress the bistability of the flow and reduce the drag on the body. Furthermore, the control system is found to have a positive energy balance, providing a key demonstration of efficient feedback control applied to a 3D bluff body wake at turbulent Reynolds numbers. [Preview Abstract] |
Monday, November 21, 2016 4:56PM - 5:09PM |
L7.00003: ABSTRACT WITHDRAWN |
Monday, November 21, 2016 5:09PM - 5:22PM |
L7.00004: Towards DMD-Based Estimation and Control of Flow Separation using an Array of Surface Pressure Sensors Eric Deem, Louis Cattafesta, Hao Zhang, Clancy Rowley Closed-loop control of flow separation requires the spatio-temporal states of the flow to be fed back through the controller in real time. Previously, static and dynamic estimation methods have been employed that provide reduced-order model estimates of the POD-coefficients of the flow velocity using surface pressure measurements. However, this requires a “learning” dataset a priori. This approach is effective as long as the dynamics during control do not stray from the learning dataset. Since only a few dynamical features are required for feedback control of flow separation, many of the details provided by full-field snapshots are superfluous. This motivates a state-observation technique that extracts key dynamical features directly from surface pressure, without requiring PIV snapshots. The results of identifying DMD modes of separated flow through an array of surface pressure sensors in real-time are presented. This is accomplished by employing streaming DMD “on the fly” to surface pressure snapshots. These modal characteristics exhibit striking similarities to those extracted from PIV data and the pressure field obtained via solving Poisson’s equation. Progress towards closed-loop separation control based on the dynamic modes of surface pressure will be discussed. [Preview Abstract] |
Monday, November 21, 2016 5:22PM - 5:35PM |
L7.00005: Active flow control insight gained from a modified integral boundary layer equation Avraham Seifert Active Flow Control (AFC) can alter the development of boundary layers with applications (e.g., reducing drag by separation delay or separating the boundary layers and enhancing vortex shedding to increase drag). Historically, significant effects of steady AFC methods were observed. Unsteady actuation is significantly more efficient than steady. Full-scale AFC tests were conducted with varying levels of success. While clearly relevant to industry, AFC implementation relies on expert knowledge with proven intuition and or costly and lengthy computational efforts. This situation hinders the use of AFC while simple, quick and reliable design method is absent. An updated form of the unsteady integral boundary layer (UIBL) equations, that include AFC terms (unsteady wall transpiration and body forces) can be used to assist in AFC analysis and design. With these equations and given a family of suitable velocity profiles, the momentum thickness can be calculated and matched with an outer, potential flow solution in 2D and 3D manner to create an AFC design tool, parallel to proven tools for airfoil design. Limiting cases of the UIBL equation can be used to analyze candidate AFC concepts in terms of their capability to modify the boundary layers development and system performance. [Preview Abstract] |
Monday, November 21, 2016 5:35PM - 5:48PM |
L7.00006: Localized modelling and feedback control of linear instabilities in 2-D wall bounded shear flows Henry Tol, Marios Kotsonis, Coen de Visser A new approach is presented for control of instabilities in 2-D wall bounded shear flows described by the linearized Navier-Stokes equations (LNSE). The control design accounts both for spatially localized actuators/sensors and the dominant perturbation dynamics in an optimal control framework. An inflow disturbance model is proposed for streamwise instabilities that drive laminar-turbulent transition. The perturbation modes that contribute to the transition process can be selected and are included in the control design. A reduced order model is derived from the LNSE that captures the input-output behavior and the dominant perturbation dynamics. This model is used to design an optimal controller for suppressing the instability growth. A 2-D channel flow and a 2-D boundary layer flow over a flat plate are considered as application cases. Disturbances are generated upstream of the control domain and the resulting flow perturbations are estimated/controlled using wall shear measurements and localized unsteady blowing and suction at the wall. It will be shown that the controller is able to cancel the perturbations and is robust to unmodelled disturbances. [Preview Abstract] |
Monday, November 21, 2016 5:48PM - 6:01PM |
L7.00007: Sensitivity analysis of unstable periodic orbits in a weakly chaotic Kuramoto-Sivashinsky system Davide Lasagna Unstable periodic orbits (UPOs) often explain to a remarkable degree of accuracy global statistical features of the turbulent flow in which they are found. In other words, orbital averages, even for short-period UPOs, are good approximation of long-time averages computed over a chaotic, turbulent realisation. Here, we re-examine this property from a design perspective: Does the same degree of approximation exists between the \emph{sensitivity} of orbital averages with respect to design parameters and the sensitivity of the long-time average itself? Knowledge of this quantity is key in many fundamental design problems involving turbulent flows, most notably in control. In this work, we present an efficient, well conditioned adjoint algorithm derived from specialising well-known variational techniques to the inherent temporal periodicity of UPOs. Once an UPO is available, this algorithm computes the sensitivity of orbital averages with respect to many design parameters at once, regardless of the orbital stability properties. As a demonstration, we analyse the sensitivity to in-domain linear feedback of UPOs found for the Kuramoto-Sivashinsky system in a weakly chaotic regime. [Preview Abstract] |
Monday, November 21, 2016 6:01PM - 6:14PM |
L7.00008: Adjoint-optimization algorithm for spatial reconstruction of a scalar source Qi Wang, Yosuke Hasegawa, Charles Meneveau, Tamer Zaki Identifying the location of the source of passive scalar transported in a turbulent environment based on remote measurements is an ill-posed problem. A conjugate-gradient algorithm is proposed, and relies on eddy-resolving simulations of both the forward and adjoint scalar transport equations to reconstruct the spatial distribution of the source. The formulation can naturally accommodate measurements from multiple sensors. The algorithm is evaluated for scalar dispersion in turbulent channel flow ($Re_\tau = 180$). As the distance between the source and sensor increases, the accuracy of the source recovery deteriorates due to diffusive effects. Improvement in performance is demonstrated for higher Prantl numbers and also with increasing number of sensors. [Preview Abstract] |
Monday, November 21, 2016 6:14PM - 6:27PM |
L7.00009: What is the 'correct' formulation of the linearised Navier-Stokes equations for designing feedback flow control systems? Oliver Dellar, Bryn Jones The use of feedback control is looking increasingly attractive as a means of reducing the pressure drag which acts upon bluff body vehicles such as heavy goods vehicles, and thus reducing both fuel consumption and CO$_2$ emissions. Motivated by the need to efficiently obtain low-order models of such flows in order to utilise model based control theory, we consider the effect on system dynamics of basing the plant model on different formulations of the linearised Navier-Stokes equations. The dynamics of a single computational node's subsystem which arises upon spatial discretisation of the governing equations in both primitive variables and pressure Poisson equation formulations are considered, revealing fundamental differences at the nodal level. The effects of these differences on system dynamics at the full fluid flow system level are exemplified by considering the corresponding formulations of a two-dimensional channel flow, subjected to a number different of boundary conditions. This ultimately reveals which formulations of the governing equations are suitable for feedback control design, and which should be avoided. [Preview Abstract] |
Monday, November 21, 2016 6:27PM - 6:40PM |
L7.00010: Delaunay-based derivative-free optimization for efficient minimization of time-averaged statistics of turbulent flows. Pooriya Beyhaghi This work considers the problem of the efficient minimization of the infinite time average of a stationary ergodic process in the space of a handful of independent parameters which affect it. Problems of this class, derived from physical or numerical experiments which are sometimes expensive to perform, are ubiquitous in turbulence research. In such problems, any given function evaluation, determined with finite sampling, is associated with a quantifiable amount of uncertainty, which may be reduced via additional sampling. This work proposes the first algorithm of this type.\textasciitilde Our algorithm remarkably reduces the overall cost of the optimization process for problems of this class. Further, under certain well-defined conditions, rigorous proof of convergence is established to the global minimum of the problem considered. [Preview Abstract] |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700