Session BE: Flow Control II

Transitory Control of a Separated Flow

The dynamics of the flow transients associated with controlled attachment and separation over a stalled airfoil are investigated with the objective of enhancing the aerodynamic performance. Transitory response to pulsed actuation on time scales that are an order of magnitude shorter than the characteristic convective time scale is assessed. Actuation is effected by momentary [O(1~msec)] jets generated by an integrated spanwise array of combustion-based actuators. The flow field in the cross stream plane above the airfoil and in its near wake is computed from multiple high-resolution PIV images that are phase-locked to the actuation, and enable continuous tracking of vorticity concentrations. The brief actuation pulse leads to a remarkably strong transitory change in circulation about the entire airfoil that is manifested by severing of the separated vorticity layer and the subsequent shedding of the separated flow vortex. It is shown that the shedding of the severed vortex results in a momentary decrease in circulation which is followed by the formation of a new surface vorticity layer that leads to a longer increase in circulation that can last 5 to 10 convective time scales. The surface vorticity layer ultimately lifts off in a manner that is reminiscent of dynamic stall.   

Active Control of Flow Separation and Structural Vibrations of a Wind Turbine Blade

The feasibility of using arrays of synthetic jet actuators to control flow separation and blade vibrations of a wind turbine blade (S809 airfoil) model was explored in wind tunnel experiments. Using this technique, the global flow field over the finite span blade was altered such that at high angles of attack flow separation was mitigated. This resulted in a significant decrease in the vibration of the blade. In addition, flow control was implemented at low angles of attack using a spanwise distribution of active Gurney flaps, instrumented with synthetic jet actuators. The moments and forces on the blade were measured using a six component wall-mounted load cell. In addition, PIV technique was used to quantify the flow field over the blade. The structural vibrations were quantified using strain gauges, placed near the blade's root, and accelerometers, located near the blade's tip. Using synthetic jets, the flow over the blade was either fully or partially reattached, depending on the angle of attack, the spanwise location along the blade, and the Reynolds number. This resulted in a significant reduction in blade's vibrations, where the amplitude of the tip deflection was found to be proportionally controllable by either changing the momentum coefficient, the number of synthetic jets used, or their driving waveform.   

The Flow Field on Hydrofoils with Leading Edge Protuberances

The agility of the humpback whale has been attributed to the use of its pectoral flippers, on which protuberances are present along the leading edge. The forces and moments on hydrofoils with leading edge protuberances were measured in a water tunnel and were compared to a baseline NACA 63(4)-021 hydrofoil revealing significant performance differences. Three protuberance amplitudes and two spanwise wavelengths, closely resembling the morphology found in nature, were examined. Qualitative flow visualization techniques were used to examine flow patterns surrounding the hydrofoils, and Particle Image Velocimetry (PIV) was used to quantify these patterns. Flow visualizations have revealed counter-rotating vortices stemming from the shoulders of the protuberances. These streamwise vortices are a result of the spanwise pressure gradient brought about by the varying leading edge curvature. PIV was used to quantify the strength of these vortices as a function of angle of attack and leading edge geometry. At low angles of attack, these vortices are symmetric with respect to the protuberances; however, the symmetry is lost at high angles of attack. The loss of symmetry can be correlated with the separation point location on the hydrofoil.   

Modeling of tangential synthetic jet actuators used for pitching control on an airfoil

Pitching moment control in an airfoil can be achieved by trapping concentrations of vorticity close to the trailing edge. Experimental work has shown that synthetic jet actuators can be used to manipulate and control this trapped vorticity. Two different approaches are used to model the action of tangential-blowing synthetic jet actuators mounted near the trailing edge of the airfoil: a detailed model and Reynolds stress synthetic jet (RSSJ) model. The detailed model resolves the synthetic jet dynamics in time while the RSSJ model tries to capture the major effects of the synthetic jet by modeling the changes in the Reynolds stress induced by the actuator, based on experimental PIV data and numerical results from the detailed model. Both models along with the CFD computations in which they are embedded are validated against experimental data. The synthetic jet models have been developed to simulate closed loop flow control of the pitching and plunging of the airfoil, and to this end the RSSJ model is particularly useful since it reduces (by an order of magnitude) the cost of simulating the long-term evolution of the system under control.   

Inter-turbine Duct Flow Separation Control with SDBD Plasma Actuators: Experiment

Inter-turbine ducts (ITDs) need to diffuse the flow between the high and low pressure turbines in the shortest possible length to avoid an unacceptable weight penalty. Significant length reductions however can lead to separated flow regions that require flow control techniques to correct. The present research focuses on the use of single dielectric barrier discharge (SDBD) plasma actuators to prevent flow separation in aggressively expanding annular ITDs. An experimental test facility has been developed to study different ITD designs. It is designed to provide a range of inflow conditions with inlet Mach numbers from 0.4 to 0.6, turbulence intensities from 0.08 to 0.2, and different degrees of mean-flow swirl. A transparent ITD wall segment provides optical access for Laser-Dopper anemometer measurements. Other flow diagnostics include surface static pressure distributions, and surface flow visualization that are used to identify flow separation regions. The characteristics of the ITD flow and sensitivity to inflow conditions will be presented. These will be compared to flow simulations.   

Inter-turbine Duct Flow Separation Control with SDBD Plasma Actuators: Simulation

Reducing the duct length between turbine stages provides a weight savings, and can effectively increase overall performance. Significant length reductions however can lead to separated flow regions that require flow control techniques to correct. This research utilizes FLUENT CFD analysis to simulate the flow in diffusing inter-turbine ducts, and include the effect of SDBD plasma actuators to control flow separation. Two duct geometries are presented; both are annular and divergent in shape but differ in axial length. The shorter duct, featuring a more aggressive bend, exhibits a flow separation, whereas the longer one does not. For the separated duct case, the effect of plasma actuators placed azimuthally around the duct at a specified streamwise location are modeled by a body force distribution along the duct wall. To save computation time, each duct is meshed in 3D as a quarter of the full 360$^{\circ}$ duct using periodic boundary conditions on the radial edges. At the inlet, the flow Mach number is 0.5 with no swirl and a low turbulence intensity. Solutions are obtained using a coupled, unsteady, RANS solver with a standard $k-\epsilon$ viscous model and enhanced wall treatment. Computational results are compared with experimental results in a specially designed facility at the University of Notre Dame.   

Active Flow Control on an Aggressive Serpentine Duct Inlet

For military applications, inlet designs are constrained by low observability requirements, which call for the use of a serpentine inlet. The inlets purpose is to limit the line-of-sight to the compressor and decelerate the incoming flow while minimizing total pressure loss, distortion, and unsteadiness. In addition, in unmanned aerial vehicles, the inlet length can determine the overall size of the aircraft. For this reason, aggressive inlets can have a large impact on overall system efficiency. Experiments utilizing active flow control to mitigate separation in a highly aggressive serpentine duct (L/D=1.5), at Mach numbers up to 0.45, were conducted. Specifically, steady and unsteady flow control techniques were compared by measuring the static pressures along the inlet walls, the pressure recovery and distortion at the AIP, and the velocity field inside the duct using Particle Image Velocimetry. Through these experiments a better understanding of the highly three dimensional flow interactions was formulated.   

Linear proportional-integral control of flow over a circular cylinder

In the present study, we apply a linear proportional-integral (PI) control to the flow over a circular cylinder for mean-drag and lift-fluctuation reductions. Park \textit{et al.} (PoF 1994) controlled the flow over a circular cylinder using a linear proportional control, but this control is very sensitive to the sensing location and works only for limited parameter ranges. In our study, we introduce a PI control to increase the controllability. The cross-flow velocity component at the centerline in the wake region is measured and is used for the determination of the control input (blowing/suction at the cylinder surface) through the PI control method. The actuation is given at the upper and lower slots on the cylinder surface near the separation point satisfying the zero-net mass flow rate. For two different Reynolds numbers, $Re=80$ and $100$, we vary the sensing location $(x_s)$ from $1d$ to $8d$ and the optimal proportional and integral gains for each sensing location are obtained using the surrogate management framework. With the PI control, the mean drag and lift fluctuations are significantly reduced and the effective control parameter ranges are widened.   

Compressible Large Eddy Simulation of Flow Control over a Wall-Mounted Hump

A compressible large eddy simulation (LES) is used to simulate flow over a wall-mounted hump geometry. The flow is characterized by a turbulent, unsteady flow separation, recirculation bubble, and reattachment. The LES results are compared with experimental data over a range of subsonic compressible Mach numbers from 0.25 to 0.6. Control is applied just before the separation location by modeling both steady suction and zero net-mass flux actuation at the wall boundary. Adding control shortens the average separation bubble length, but steady suction is shown to be more effective at reducing drag than the oscillatory forcing. For a given momentum coefficient, the control is shown to be less effective at compressible mach numbers than incompressible flow.   

On the Oseen flows in 2D

This investigation concerns numerical evaluation of solutions of the Oseen equation in two--dimensional exterior domains. The significance of the Oseen approximation consists in that it captures the correct asymptotic behavior at infinity of the ``physically reasonable'' solutions of the steady--state Navier--Stokes system. Thus, such Navier--Stokes solutions can be computed as perturbations about the Oseen solution. A formal closed--form (series) solution of the Oseen problem was derived by Tomotika \& Aoi (1949), and features an infinite number of coefficients which have to be determined by matching against the boundary conditions. While in the past this solution was used to compute very coarse approximations only of the the Oseen flow, the present investigation seeks to determine such approximations with an arbitrary accuracy. We show that such a task is in fact nontrivial, as it leads to algebraic problems with extremely poor conditioning. We will discuss different methods of getting around these difficulties, and will present Oseen solutions for a range of Reynolds numbers. In addition, we will also address the structure of the Oseen wake in the infinite Reynolds number limit.