Active Flow Control

In the present chapter I shall emphasize the frontiers of the field of flow control, pondering mostly the control of turbulent flows. I shall review the important advances in the field that took place during the past few years and are anticipated to dominate progress in the future. By comparison with laminar flow control or separation prevention, the control of turbulent flow remains a very challenging problem. Flow instabilities magnify quickly near critical flow regimes, and therefore delaying transition or separation are relatively easier tasks. In contrast, classical control strategies are often ineffective for fully turbulent flows. Newer ideas for turbulent flow control to achieve, for example, skin-friction drag reduction focus on the direct onslaught on coherent structures. Spurred by the recent developments in chaos control, microfabrication and soft computing tools, reactive control of turbulent flows, where sensors detect oncoming coherent structures and actuators attempt to favorably modulate those quasi-periodic events, is now in the realm of the possible for future practical devices. In this chapter, I shall provide estimates for the number, size, frequency and energy consumption of the sensor/ actuator arrays needed to control the turbulent boundary layer on a full-scale aircraft or submarine.

Time periodic Lorentz forces have been used to influence the separated flow on an inclined flat plate in deep stall at a Reynolds number of 10⁴. The influence of the control parameters effective momentum coefficient and excitation frequency as well as excitation wave form is discussed based on phase averaged PIV measurements. As expected, control authority depends strongly on momentum input and excitation frequency, but effects of the excitation wave form can be shown as well.

An experimental investigation of separation control using steady and pulsed plasma actuators was carried out on an Eppler E338 airfoil at typical micro air vehicle Reynolds numbers (20,000≤Re≤140,000). Pulsing was achieved by modulating the high frequency plasma excitation voltage. The actuators were calibrated directly using a laser doppler anemometer, with and without free-stream velocity, and this allowed the quantification of both steady and unsteady momentum introduced into the flow. At conventional low Reynolds numbers (Re>100,000) asymmetric single phase plasma actuators can have a detrimental effect on airfoil performance due to the introduction of low momentum fluid into the boundary layer. The effect of modulation, particularly at frequencies corresponding to F ⁺=1, became more effective with decreasing Reynolds number resulting in significant improvements in C _L,max. This was attributed to the increasing momentum coefficient, which increased as a consequence of the decreasing free-stream velocities. Particularly low duty cycles of 3% were sufficient for effective separation control, corresponding to power inputs on the order of 5 milliwatts per centimeter.

This is a fundamental study about the influence of plasma-actuators on boundary-layer flows, including both experimental and numerical investigations. The first set of experiments is conducted in quiescent air and these results are used to calibrate a numerical model which simulates the plasma-actuator in an existing RANS (Reynolds Averaged Navier-Stokes) code. The second set of experiments involves a flat-plate boundary layer at various free-stream velocities, where the actuator adds momentum to the boundary layer. The previously calibrated numerical model is used to simulate the influence of the actuator on the boundary layer. The agreement between simulation and experiment is very good and the simulations with the new model be considered a reliable predictive tool.

Unsteady zero-net-mass-flux (ZNMF) periodic excitation was proven to be significantly more effective than steady blowing and simpler to apply than steady suction for the control of boundary layer separation. Furthermore, it can utilize flow instability as efficiency magnifier. When generated by Piezo-fluidic actuators, it has a bandwidth that is suitable for a wide range of feedback control applications. However, the current state-of-the-art ZNMF actuators lack, for certain applications, sufficient control authority. Therefore, effective methods for coupling the excitation to the most unstable modes of the flow should preferably be sought after and utilized.

A new robust and simple actuator concept that combines steady suction and pulsed blowing is presented. It can generate wide band-width near sonic oscillations. Its performance was modeled and validated in several scales. The valve design allows highly efficient operation, not nullifying the favorable effects of future CLAFC schemes.

Three-dimensional (3D) excitation modes should be explored, as the flow naturally becomes 3D even if the baseline flow and the excitation are nominally two-dimensional (2D). To be for industrial applications, overall system efficiency should always be assessed, not only the improvement in aerodynamic performance.

Three performance based criteria for comparing different actuation concepts are presented and discussed. The first criterion evaluates the actuator based on its force or linear momentum generation capability as it operates in still fluid, while considering its weight, volume and power consumption. The second criterion is simply the actuator peak velocity Mach number relative to the free-stream Mach number, where it is rare to find any benefit from actuator with Mach ratio smaller than 0.1 and/or momentum coefficient smaller than 0.01%. The third, application dependent, criterion is the Aerodynamic figure of merit, an energy efficiency criterion, based on the improvement of the controlled performance (e.g., lift to drag ratio) of a certain application, when the power consumption (and also the weight) of the actuation system are taken into account.

For successful feedback flow control, an accurate estimation of the flow state is necessary. Proper Orthogonal Decomposition (POD) has been used to achieve this. However, if the POD modes are derived from a set of snapshots obtained from one flow condition only, the resulting modes will become less and less valid for a flow field that is for example altered by the effect of feedback flow control. In the past, a shift mode has been added to account for the change in the mean flow. Here, we present a new scheme that allows for the derivation of shift modes for all of the original POD modes. This DPOD mode set thus may span a range of flow conditions that are different in forcing, Reynolds number or other parameters affecting the modes. Artificial Neural Network Estimation (ANNE) allows for real time monitoring of the time coefficients associated with these DPOD modes.

We present a unified feature extraction architecture consisting of only three core algorithms that allows to extract and track a rich variety of geometrically defined, local and global features evolving in scalar and vector fields. The architecture builds upon the concepts of Feature Flow Fields and Connectors, which can be implemented using the three core algorithms finding zeros, integrating and intersecting stream objects. We apply our methods to extract and track the topology and vortex core lines both in steady and unsteady flow fields.

Vortex control concepts employed for slender, nonslender and high aspect ratio wings were reviewed. For slender delta wings, control of vortex breakdown has been the most important objective, which is achieved by modifications to swirl level and pressure gradient. Delay of vortex breakdown with the use of control surfaces, blowing, suction, high-frequency and low-frequency excitation, and feedback control was reviewed. For nonslender delta wings, flow reattachment is the most important aspect for flow control methods. For high aspect ratio wings, vortex control concepts are diverse, ranging from drag reduction to attenuation of wake hazard and noise, which can be achieved by modifications to the vortex location, strength, and structure, and generation of multiple vortices.

This contribution summarizes the flow control research results obtained at TU Braunschweig and their implication for control on high-lift devices. The superordinate aim of the examination is the control of leading-edge stall on a two-element airfoil by means of dynamic 3D actuators. This is of great practical interest in order to increase the maximum angle of attack and/or the lift coefficient and to alter the drag coefficient in takeoff and landing configuration of future aircrafts. To reach this aim, several pneumatic actuators were designed and systematically tested to determine their characteristics and impulse response on the input signal at first. Secondly, their potential for active flow control was investigated in a small wind tunnel. Thirdly, the interaction of promising actuator concepts was studied in detail to examine the benefit of actuator arrays. Finally, the experiences were combined to examine the potential of the actuators in delaying leading edge separation on a generic airfoil. Therefore, an appropriate airfoil was designed, built and equipped with the developed actuator technology and investigated in a large wind tunnel. The results illuminate the potential of dynamic actuators for high-lift applications and the effect of different actuator parameters (actuator design, orientation, spacing and position as well as amplitude, frequency and duty-cycle). Furthermore, practical actuator designs and operation rules are deduced from the examination. In the future they may be assistant to assign the results on real aircraft configurations.

This paper gives an overview of numerical flow control investigations for high-lift airfoil flows carried out by the authors. Two configurations at stall conditions, a generic two-element setup with single flap and a second configurationwith slat and flap of more practical relevance are investigated by simulations based on the Reynolds-averaged Navier-Stokes equations and eddy-viscosity turbulence models. For both cases flow separation can be delayed by periodic vertical suction and blowing through a slot close to the leading edge of the flap. By simulating different excitation modes, frequencies and intensities optimum control parameters could be identified. Comparison of aerodynamic forces computed and flow visualisations to experiments allows a detailed analysis of the dominant structures in the flow field and the effect of flow control on these. The mean aerodynamic lift can be significantly enhanced by the active flow control concepts suggested here.

The purpose of this manuscript is to address one of the many questions plaguing the application of fluidic active flow control for performance enhancement over wings and airplanes. Specifically, what mode of Active Flow Control (AFC) is most effective; steady suction, steady blowing, or a periodic variation of both? The tilt rotor model is chosen because it represents very demanding requirements over a wide range of incidence angles, α, varying from −90°<α<+20° and flap deflections 0°<Δ_f<85° as it transitions from hover to cruise.

Measurements were carried out on a V-22 airfoil that has a simple flap (contrary to the flap being used currently on the airplane) with AFC emanating from a single slot carved in the flap for the purpose of download alleviation. The same slot was used to improve the performance of the airfoil in cruise and determine its dependence on the method of flow control. Levels of momentum input and the frequency of the periodic actuation were investigated as well. Steady and oscillatory suction and pure (zero mass flux) periodic perturbations proved to be very effective at low momentum coefficients while steady blowing required a large threshold value be exceeded before proving its effectiveness. These observations apply in cruise and hover alike. Download measurements were also carried out on a three-dimensional, 1/10^th scale model of the airplane whereupon the two dimensional parametric studies were confirmed. The use of suction reduced the download on the model by 30% while weak periodic excitation reduced it by 16%.

The results of an ongoing research activity in the development and implementation of reduced-order model-based feedback control of subsonic cavity flows are presented and discussed. Particle image velocimetry data and the proper orthogonal decomposition technique are used to extract the most energetic flow features or POD eigenmodes. The Galerkin projection of the Navier-Stokes equations onto these modes is used to derive a set of ordinary nonlinear differential equations, which govern the time evolution of the modes, for the controller design. Stochastic estimation is used to correlate surface pressure data with flow field data and dynamic surface pressure measurements are used for real-time state estimation of the flow model. Three sets of PIV snapshots of a Mach 0.3 cavity flow were used to derive three reduced-order models for controller design: (1) snapshots from the baseline (no control) flow, (2) snapshots from an open-loop forced flow, and (3) combined snapshots from the cases 1 and 2. Linear-quadratic optimal controllers based on all three models were designed and tested experimentally. Real-time implementation shows a remarkable attenuation of the resonant tone and a redistribution of the energy into various modes with much lower energy levels.

The response of a supersonic cavity to open-loop forcing with a pulsed-blowing actuator is explored experimentally. It is shown that excitation at frequencies near the Rossiter modes are amplified, while frequencies between the first two Rossiter modes are attenuated. The results clearly demonstrate that the Rossiter modes in the supersonic cavity are weakly damped (not self-excited) modes. The supersonic modes are not saturated, and do not show the kind of nonlinear interactions with the forcing modes observed in subsonic flow. These differences between supersonic and subsonic flows are consistent with previously developed models of cavity oscillations, and the results suggest that linear techniques for the design of closed-loop controllers may be particulary effective for supersonic flows. For the flow regime studied, the oscillatory component of open-loop forcing does not play a significant role in the suppression mechanism in supersonic cavity flows.

Experimental and numerical investigations were carried out aiming at the reduction of the total aerodynamic drag of a generic car model by means of active separation control. For two different configurations separate control approaches were tested, taking into account the differences in the wake topology of the models. The targeted excitation of the respective dominant structures in the wake region leads to their effective attenuation. The experiments as well as the numerical simulations showed that a weakening of a spanwise vortex in the separated flow over the slant is strongly coupled with the occurrence of stronger streamwise vortices along the slant edges and vice versa.

In the current study, a hierarchy of control-oriented Galerkin models is proposed targeting least-dimensional representations at different operating conditions. These models are employed for passive as well as active actuation. In passive control, a linearised model is shown to reproduce a wake stabilization experiment of Strykowski & Sreenivasan (1990). In particular, the effect and optimal position of control wires are accurately predicted. In active closed-loop control, focus is also placed on experiment. Here, control design requires a model which has on the hand a sufficiently broad dynamic range and is on the other hand low-dimensional enough for online computation. POD Galerkin models have a desirable mathematical structure and dimension for an online capable control design but tend to be over-optimised for the reference conditions. The resulting limited dynamic bandwidth is associated with the underlying expansion modes which change their shape at different operating conditions. To increase that model bandwidth, a novel continuous mode interpolation technique is proposed. The mode interpolation smoothly connects not only different operating conditions, but also stability and POD modes and even flows at different boundary conditions. In addition, the extrapolation of modes outside the design conditions is illustrated. The interpolated modes enable ‘least-order’ Galerkin models keeping the dimension from a single operating condition but resolving several states. These models are well suited for control design. The mode interpolation technique is demonstrated for three benchmark problems, the flow around circular cylinder, a NACA airfoil and an Ahmed body.

A pilot experimental investigation was conducted to study the active generation and management of streamwise vortices in an incompressible jet flow. The lip of the jet was equipped with small flaps (flaplets) deflected away from the stream at 30°, that incorporated flow control slots through which steady suction and zero mass-flux perturbations were introduced. Control via suction, reduced the pressure on the flaplets, thereby entraining fluid from the surrounding fluid and generating streamwsie vortices. Flaplet length had very little effect on the vortex formation but exerted a profound influence on the dissipation of the vortices further downstream. Jet momentum could also be increased by up to 25%. Preliminary experiments using zero mass-flux control indicated that entrainment of the surrounding fluid was combined with a reduction of momentum in the jet.

The secondary flow over the impeller blade tips of axial fans is important for the aerodynamic and acoustic performance of the fans. Pressure coefficient and efficiency drop, and the usable range of the performance characteristic is diminished as the rotor flow stalls at higher flow rates. By applying active flow control in the tip region of the impeller, it is possible to reduce the negative effects of the tip clearance flow. In the present paper, steady air injection into the tip clearance gap is investigated. Three different air injection configurations are studied. Two configurations use up to 24 separate slit nozzles, which are evenly distributed over the circumference of the casing. In the third injection configuration, the individual nozzles are replaced by a continuous circumferential slit.

In this paper we investigate the optimal control approach for the active control of the circular cylinder wake flow considered in the laminar regime (Re = 200). The objective is the minimization of the mean total drag where the control function is the time harmonic angular velocity of the rotating cylinder. When the Navier-Stokes equations are used as state equations, the discretization of the optimality system leads to large scale discretized optimization problems that represent a tremendous computational task. In order to reduce the number of state variables during the optimization process, a Proper Orthogonal Decomposition (POD) Reduced-Order Model (ROM) is then derived to be used as state equation. Since the range of validity of the POD ROM is generally limited to the vicinity of the design parameters in the control parameter space, we propose to use the Trust-Region Proper Orthogonal Decomposition (TRPOD) approach, originally introduced by Fahl (2000), to update the reduced-order models during the optimization process. Benefiting from the trust-region philosophy, rigorous convergence results guarantee that the iterates produced by the TRPOD algorithm will converge to the solution of the original optimization problem defined with the Navier-Stokes equations. A lot of computational work is indeed saved because the optimization process is now based only on low-fidelity models. The key enablers to an accurate and robust POD ROM for the pressure and velocity fields are the extension of the POD basis functions to the pressure data, the introduction of a time-dependent eddy-viscosity estimated for each POD mode as the solution of an auxiliary optimization problem, and the inclusion in the POD ROM of different non-equilibrium modes. When the TRPOD algorithm is applied to the wake flow configuration, this approach converges to the minimum predicted by an open-loop control approach and leads to a relative mean drag reduction of 30% for reduced numerical costs (a cost reduction factor of 1600 is obtained for the memory and the optimization problem is solved approximately 4 times more quickly).

For the model-based active control of three-dimensional flows at high Reynolds numbers in real time, low-dimensional models of the flow dynamics and efficient actuator and sensor concepts are required. Numerous successful approaches to derive such models have been proposed in the literature.

We propose a software environment for a comfortable and performant testing of control, actuator and sensor concepts which may be based on such models. It is realized by providing an easily manageable Matlab control interface for the κ-ε-model from the Featflow CFD package. Potentials and limitations of this tool are discussed by considering exemplarily the control of the recirculation bubble behind a backward facing step.

In this paper we discuss an appropriate choice of a cost functional for vortex reduction for unsteady flows described by the Navier-Stokes equations. We discuss different possibilities for a definition of a vortex and for a corresponding cost functionals. The resulting optimal control problem is analyzed and numerical experiments are provided.

We consider the distributed optimal control of the Navier-Stokes equations in presence of pointwise state constraints. A Lavrentiev regularization of the constraints is proposed and a first order optimality system is derived. The regularity of the mixed constraint multiplier is investigated and second order sufficient optimality conditions are studied. In the last part of the paper, a semi-smooth Newton method is applied for the numerical solution of the control problem and numerical experiments are carried out.

In the present study the flow around a 2D bluff body with blunt stern is investigated experimentally and theoretically. The goal is to decrease and stabilize drag by active control. Low-dimensional vortex models are used to describe actuation effects on the coherent structures and the pressure field. Open-loop actuation as well as feedback control is applied using robust H _∞-controllers and slope-seeking feedback for a range of Reynolds numbers based on the height from 20 000 to 60 000. As expected, a decreased drag is observed to be related to delayed vortex shedding, i.e. an extended recirculation zone. Intriguingly, a control which mitigates the natural coupling between separating upper and lower shear-layer and the vortex street serves that purpose.

In conventional active noise control experiments, loudspeakers are used to generate the secondary anti-phase sound field to be superimposed destructively with the sound waves radiated from the primary source. In the present study, aerodynamic sound sources are used to actively control the tonal noise of an axial fan. This is achieved by disturbing the flow field around the blade tips in such a way that additional periodic forces are set up which in turn form the secondary sound sources. To disturb the flow, air is blown into the blade tip region through the casing wall. The resulting aerodynamic sound sources are adjustable in both amplitude and phase. Results for automatic control of the blade passing frequency are presented along with flow measurements in the interstage regime between rotor and stator.

Phase-shift control was applied to an atmospheric swirl-stabilized premixed combustor. Two different actuators were tested in the control scheme. An on-off valve was used to modulate secondary pilot fuel and a loudspeaker provided excitation of the air mass flow upstream of the burner. In individual mode, both actuators were able to successfully control a low-frequency combustion instability with similar levels of suppression. The pilot valve could also be triggered at subharmonics of the dominant oscillation frequency without loosing control performance. Using the two actuators simultaneously gave an even better suppression compared to individual operation. It was further shown that the pulsed pilot fuel could be used to assist purely acoustic control in the case of limited actuator authority.

This paper reviews recent progress concerning development of point-vortex reduced-order models for feedback stabilization of the cylinder wake flow. First, we recall briefly some earlier results related to the design of linear feedback stabilization strategies based on the Föppl system. Then we present derivation of a higher-order Föppl system bas on solutions of the Euler equations which desingularize the original Föppl vortices. We argue that such higher-order Föppl systems possess important advantages over the classical Föppl system which are relevant from the control-theoretic point of view. In particular, we present computational results indicating that a higher-order Föppl system can be stabilized completely in contrast to the classical Föppl system for which this is not possible owing to the presence of a stable center manifold spanned by uncontrollable modes.