See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/230258931

Aerodynamic Flow Control

Chapter · December 2010

DOI: 10.1002/9780470686652.eae019

CITATIONS READS
12 2,926

3 authors, including:

David Greenblatt
Technion - Israel Institute of Technology
188 PUBLICATIONS 4,801 CITATIONS

SEE PROFILE

All content following this page was uploaded by David Greenblatt on 17 June 2018.

The user has requested enhancement of the downloaded file.

Aerodynamic Flow Control
David Greenblatt1 , Israel J. Wygnanski2 , and Christopher L. Rumsey3
1
Faculty of Mechanical Engineering, Israel Institute of Technology, Technion City, Haifa, Israel
2
University of Arizona, Aerospace and Mechanical Engineering, Tucson, USA
3
Computational Aerosciences Branch, NASA Langley Research Center, Hampton, Virginia, USA

via relatively small diameter pipes (Baumann, 1921). The

1 Introduction 1 advent of jet propulsion and lift augmentation requirements
2 Scaling Parameters 2 of high-speed military jets catapulted blowing to the fore-
front of BLC in the mid-1950s. The massively produced
3 Scope of AFC Applications 4
Lockheed F-104 and the Mikoyan Mig-21 both used blowing
4 Computational Approaches to AFC 6 over the surface of their flaps for landing. Although inte-
5 Conclusions and Outlook 9 gration of propulsion with aerodynamics did not proceed
List of Symbols 9 beyond the laboratory stage, it gave rise to new concepts
References 10 such as super-circulation and the “jet-flap”. The apogee of
active BLC research was reached before the appearance of
“Sputnik” whereupon a changeover of focus to space research
had occurred. The vast scope of international research con-
1 INTRODUCTION ducted during the first half-century is summarized in the
classic volumes edited by Lachmann (1961).
Flow separation and its control dominated the thoughts After a partial hiatus, boundary layer control or, more
of Prandtl when he formulated the boundary layer theory currently, active flow control (AFC) re-emerged as a research
(see Tollmien, 1955). In developing the basic theory, he area with wide application to aerodynamics. This interest
simultaneously demonstrated that flow separation is delayed was re-ignited as a result of two main factors: the one
by applying suction through a narrow slot cut through the physical and the other technological. The physical driver
surface of a cylinder. Although slot suction acted as a cat- was that separated flows could be manipulated and attached
alyst for future developments, research on slot suction for to surfaces, with attendant increases in lift, by means of
lift enhancement waned because the quest for high-speed periodic zero mass-flux (ZMF) perturbations whose input
flight necessitated thin wing sections. Nevertheless, suction momentum was a small fraction of that required by steady
through pores in the surface proved to be valuable in stabiliz- blowing (e.g., Nishri and Wygnanski, 1998; Seifert, Darabi
ing the boundary layer and delaying transition to turbulence and Wygnanski, 1996). Perturbations could be introduced
(e.g., Bussmann and Muenz, 1942). The potential for the via a narrow slot or any other device, which is capable of
practical application of boundary layer control (BLC) was producing an oscillatory momentum source. The efficacy
demonstrated by blowing highly-pressurized air, conveyed of this method was explained by the fact that a boundary
layer flow on the verge of separation contains an inflected
velocity profile that is susceptible to periodic perturbations
and amplifies them (see Flow Instabilities and Transition).
This creates an array of spanwise vortices that periodically

Encyclopedia of Aerospace Engineering, Online © 2010 John Wiley & Sons, Ltd.
This article is © 2010 John Wiley & Sons, Ltd.
This article was published in the Encyclopedia of Aerospace Engineering in 2010 by John Wiley & Sons, Ltd.
DOI: 10.1002/9780470686652.eae019
 2 Incompressible Flows and Aerodynamics

sweep the surface transferring streamwise momentum from the other hand, with steady increase in computer power, com-
the outer flow to the surface, thus enabling the boundary putational fluid dynamics (CFD) is increasingly becoming a
layer to remain attached and overcome larger adverse powerful tool for the prediction of the unsteady turbulent
pressure gradients (Darabi, 2002). This is in contrast to flows (Rumsey et al., 2006).
traditional BLC where suction removes the low-momentum This chapter provides a broad overview of modern active
fluid from the boundary layer while blowing re-energizes the flow control. Passive flow control techniques such as vor-
same boundary layer. Spanwise-uniform periodic excitation tex generators and turbulators (Chang, 1970), riblets (Walsh
exploits the instability of the boundary layer that amplifies and Anders, 1989), and Gurney flaps (van Dam, Yen and
the input, thereby leveraging its effects. Vijgen, 1999) are not reviewed here. Section 2 deals with
The technological driver and challenge behind AFC devel- the basic assumptions of actuation and includes empirically
opments was, and still is, the ability to deploy autonomous determined scaling laws. Some representative examples are
actuators and sensors within, or attached to, aerodynamic sur- described in Section 3. Section 4 describes CFD method-
faces. These are sometimes called micro-electro-mechanical ologies and their application to AFC, and Section 5 provides
systems (MEMS), but the most common length scales are conclusions. This chapter does not explicitly address the con-
O(10−3 ) to O(10−1 ) m. Two of the most common actuator trol of free shear layers, jets, and wakes. A full treatise of these
types are fluidic actuators (e.g., ZMF, Nagib et al., 2006) flows can be found in Joslin and Miller (2009).
that are deployed behind the slot to produce oscillatory per-
turbations, and surface-mounted mechanical oscillators (e.g.,
Viets, Piatt and Ball, 1987; Neuburger and Wygnanski, 1987; 2 SCALING PARAMETERS
Bar-Sever, 1989). Recent decades have witnessed significant
advances in zero and nonzero mass-flux devices including Flow excitation, actuation, or forcing is a critical aspect
piezoelectric (Chen et al., 2000; Glezer and Amitay, 2002) in transitioning active control from the laboratory to real-
valve-type (Seifert, Darabi and Wygnanski, 1996; Bachar, world applications. On the one hand, actuators with sufficient
2001; Seifert and Pack, 1999) and pulsed combustion or authority must be developed that simultaneously provide a net
detonation-driven (Crittenden et al., 2001) devices. The same system benefit. On the other hand, the correct location, fre-
is true for surface-mounted actuators that include piezoelec- quency, orientation, type of actuator, etc., must be determined
tric (Seifert et al., 1998), plasma-based (Sosa et al., 2006; and here theoretical studies are only partially helpful.
Post and Corke, 2004), arc filament (Samimy et al., 2004),
shape memory alloys (Wlezien et al., 1998), and Lorentz
force (Weier and Gerbeth, 2004) actuators. Perturbations 2.1 Steady suction and blowing
produced by the actuators may be small relative to a char-
acteristic velocity or vehicle dimension and thus exploit Initially, mass-flux was used to assess the efficacy of sepa-
boundary layer instability; but they may also be much larger ration control whenever suction or blowing was employed.
and hence “force” the flow, for example, by high-frequency This parameter was replaced by Cµ in the case of blowing,
alternating blowing and suction. Modern active flow control because the latter eliminated the dependence of lift on the
encompasses both steady and periodic approaches, as well as width of the slot from which blowing emanated (Poisson-
combinations of the two. Quinton and Lepage, 1961). Large variations in Cµ were
Our ability to understand and analyze these flows from investigated (0 < Cµ < 1), but the effects of slot width, or its
first principles is severely limited for two reasons. First, location, were never systematically analyzed. Practical values
the pre-existing turbulent flow complexity is exacerbated by of Cµ < 0.1 required the inclusion of mass flow to improve
a periodic unsteadiness that is often driven by instability the correlation of the data (Attinello, 1961) whenever the
mechanisms. Second, the wide variety of actuation devices jet and stream velocities were comparable. Neglecting local
listed above can have significantly different effects on the pressure gradients and compressibility effects suggests the
same basic flow field. Thus, theoretical methods (Gaster, Kit combined parameter Cµ [1–(U∞ /Uj )] = Cµ – 2CQ where CQ
and Wygnanski, 1985; Reau and Tumin, 2002) can at best is a mass flow coefficient. The above-mentioned modification
describe only qualitative trends and generally have a lim- enables the subtraction of a scalar quantity based on mass
ited predictive capability. The main advances to date have flow (CQ ) from a vector quantity based on the jet momen-
been empirical, or semiempirical, and these form the basis tum (Cµ ), which is appropriate if the inclination of the jet to
of our rather superficial present understanding. Neverthe- the oncoming flow is negligible. Since aeronautical applica-
less, experiments that isolate controlling parameters remain tions focused on the reduction of landing speed, the effect of
an indispensable approach for advancing our knowledge. On blowing on drag was not considered important and was left

Encyclopedia of Aerospace Engineering, Online © 2010 John Wiley & Sons, Ltd.
This article is © 2010 John Wiley & Sons, Ltd.
This article was published in the Encyclopedia of Aerospace Engineering in 2010 by John Wiley & Sons, Ltd.
DOI: 10.1002/9780470686652.eae019
 Aerodynamic Flow Control 3

unexplored. The focus on lift enabled many researchers to momentum components cannot be predicted from first princi-
model the “jet flap” by assuming the flow to be inviscid and ples – although lumped-element or reduced-order modeling
to predict a complete thrust recovery that was independent of is used (e.g., Gallas et al., 2002) – and the actuator must be
the inclination of the jet leaving the trailing edge of the air- calibrated by directly measuring the velocity field and hence
foil (e.g., Woods, 1958). Such independence was observed its momentum components. A notable exception is Lorentz
experimentally by Davidson (1956) and mostly verified later force-type actuators (e.g., Weier and Gerbeth, 2004).
up to jet deflection angles that were almost orthogonal to the Although the above definition can readily be extended to
free-stream (Hynes, 1968). include all devices that produce oscillatory momentum, the
Since many of the older investigations relied on force bal- parameter C µ  is not without its limitations. For example,
ance results, it was suggested that the lift increment attained the suction phase of the oscillatory cycle does not directly
by blowing encompassed two effects that sequentially depend add momentum, and calibrations performed under quiescent
on the level of Cµ . At low levels of Cµ , flow separation conditions (U∞ = 0) are not necessarily valid under test con-
was prevented by energizing the boundary layer (providing ditions (U∞ = / 0). To date, various experiments have been
BLC) and generating a lift increment
CL ∝ Cµ (Attinello, conducted, which indicate that this parameter is not univer-
1961; Poisson-Quinton and Lepage, 1961). In this case, the sally valid, and alternate forms have been suggested (see
jet momentum enabled the generation of a flowfield that could Section 2.4).
be approximated by a potential flow solution around an air- The preferred method of pulsed control is to superimpose
foil or wing (Williams, 1961). For Cµ exceeding a
threshold net positive or negative steady mass flux onto a nominally
level Cµ,crit , the lift increment is smaller,
CL ∝ Cµ due zero mass-flux device; certain pulsed valves have the disad-
to super-circulation (Spence, 1956). Cµ,crit separating the two vantage that the relative proportion of momentum cannot be
flow regimes depends on the flap deflection and its size, but it varied. Some actuators operate in resonance at frequencies
is also sensitive to wing angle-of-attack (α), thickness, shape, that are very much higher than those required for exciting
sweep, and aspect ratio. On a typical 25% flap, Cµ,crit ≈ 0.015 a useful instability. In these instances, some form of low-
tan(δf ), where δf is the flap deflection and provided α = 0◦ frequency modulation is employed and in addition to Cµ  ,
(Poisson-Quinton and Lepage, 1961). Recent observations the duty cycle (% of time that the actuator operates) plays an
(Cerchie et al., 2006) suggest that the assumptions giving rise important role.
to Cµ,crit might be erroneous because a wing may increase its
lift due to the upstream entrainment of the fluidic actuation, 2.3 Frequency scaling
be it blowing, suction, or ZMF forcing, without attaching the
flow downstream of the slot. When momentum is introduced in an oscillatory manner, the
correct frequency scaling parameters are not known without
empirical input. In general, different objectives – for exam-
2.2 Periodic perturbation and forcing ple, attaching an otherwise separated flow, the avoidance of
separation of an attached flow, or maximization of CL,max or
The introduction of frequency as a parameter, with or with- L/D – will result in different numerical values of a selected
out a steady mass-flux component, correct amplitude scaling parameter. The attachment of an initially separated shear
becomes an even more difficult task. Moreover, with the vast layer to an inclined solid surface by means of small amplitude
and ever growing range of actuation methods, a key task is perturbations is directly related to the shear-layer receptivity
to establish a common “output” parameter so that their rela- to the perturbations and their amplitude. Forcefully reat-
tive effects on the flow or performance parameters of interest tached flow (Figure 1) indicates that the vortices amplified
can be compared. Generalizing the steady blowing parame- by the instability reinforce and regulate the eddies that would
ter to include a periodic component (see Seifert, Darabi and have been shed if the flow would have been separated creating
Wygnanski, 1996) is a common approach and produces the a free mixing layer near the surface (Darabi, 2002).
combined momentum coefficient: Nishri and Wygnanski (1998) established that for a fully
turbulent upstream boundary layer separating from a generic
J J flap (i.e., an inclined surface) and in the absence of surface
Cµ,tot = Cµ + Cµ  = + (1)
q∞ L q∞ L curvature, the optimum reduced frequency, that is, requiring
the least momentum Cµ min , was F + ≈ 1.3, where F + ≡
where J and J represent the steady and unsteady momentum fe XTE /U∞ (Figure 2). However, the optimum F+ required
addition, respectively (e.g., Seifert, Darabi and Wygnanski, to prevent the separation of an initially attached flow was
1996; Greenblatt and Wygnanski, 2000). In most cases, the somewhat higher, in the range 2 ≤ F+ ≤ 4.

Encyclopedia of Aerospace Engineering, Online © 2010 John Wiley & Sons, Ltd.
This article is © 2010 John Wiley & Sons, Ltd.
This article was published in the Encyclopedia of Aerospace Engineering in 2010 by John Wiley & Sons, Ltd.
DOI: 10.1002/9780470686652.eae019
 4 Incompressible Flows and Aerodynamics

Figure 1. Forcefully reattached boundary layer by periodic excitation emanating from a slot showing a phase-averaged PIV “snapshot”
(arbitrary phase). Shaded regions represent spanwise vorticity contours; black curves represent streak lines. Reproduced from Darabi
(2002).

2.4 Extended parameter range and comparison (Figure 3). The results of aft control (Figure 4) represent the
of techniques dependence of
CL on the product of the blowing velocity
ratio and the slot width (h) for a given slot location and ori-
A number of different parameters have been proposed and entation at α = 0◦ . When the exponent β = 1, the abscissa
assessed for the characterization of airfoil performance, (Uj /U∞ )2 (h/c)β is equivalent to Cµ . The exponent β depends
usually lift coefficient. Nagib et al. (2006) conducted exper- on h/c and is approximated by a polynomial function. For
iments on a highly deflected simple flap and concluded that h/c < 0.005, the parameter is approximately equal to Cµ , and
the velocity ratio is a more appropriate parameter. They com- the significance of the slot width increases for h/c > 0.005.
bined this with Strouhal number to form the new parameter The flow is independent of Reynolds number, and the cor-
√
H = (UJ /U∞ )/ fL/U∞ . Stalnov and Seifert (2008) con- relation renders a reasonable dependence of
CL on the
sidered five different scaling parameters for control using blowing parameter, although for (Uj /U∞ )2 (h/c)β > 0.03,
high-frequency and pulsed actuation and concluded that the there is considerable scatter. The deleterious effect of blow-
Reynolds number-scaled momentum coefficient provided the ing on lift for (Uj /U∞ )2 (h/c)β < 0.003 should also be noted
best scaling for their data set. (see the upper inset). Good correlations were achieved for
A thick elliptical airfoil was tested by Cerchie et al. pitching moment (not shown), but drag data were seen to be
(2006), where slot thickness and location could be widely highly dependent on the slot width. Steady suction generates
varied and AFC applied at both the leading and trailing edges lift at very low levels of mass flow without the deleterious
effects observed in conjunction with steady blowing, and the
lift coeffcient data correlated well with the empirical param-
50 eter CQ (h/c)γ Re0.4 , where γ = f(h/R). When zero mass-flux
forcing was employed, a correlation was discerned between
40 lift coefficient and the empirical parameter Cµ  (θF+ )0.5 .
Cµ,min Leading-edge steady blowing and suction were also consid-
(x105)
30 ered. Blowing required large momentum coefficients (Cµ ≈
0.1%) to be effective while suction always produced a bene-
20
fit; for example, at Cµ = 0.018%, stall was not observed up to
Re Lf α = 25◦ with CL,max exceeding 2.
200 000 320 mm
10 300 000 320 mm
300 000 740 mm
700 000 740 mm
0 3 SCOPE OF AFC APPLICATIONS
0 1 2 3 F+ 4

An overwhelmingly large emphasis has been placed on two-

Figure 2. The minimum momentum coefficient required to attach
an otherwise separated shear layer to a deflected surface as a function dimensional airfoil studies, with the objective of producing
of reduced frequency. Reproduced with permission from Nishri and “simplified high-lift” (i.e., use of simple flaps/slats along
Wygnanski (1998)
c A. Darabi, 1998. with AFC, rather than complex Fowler flaps/slats). When the

Encyclopedia of Aerospace Engineering, Online © 2010 John Wiley & Sons, Ltd.
This article is © 2010 John Wiley & Sons, Ltd.
This article was published in the Encyclopedia of Aerospace Engineering in 2010 by John Wiley & Sons, Ltd.
DOI: 10.1002/9780470686652.eae019
 Aerodynamic Flow Control 5

Figure 3. An elliptical airfoil designed to test the effects of var-

ious parameters used in active separation and circulation control.
Reproduced with permission from Cerchie et al. (2006)
c AIAA.

parameter F + is used, the vast majority of investigations cit-

ing effective control produce 0.3 ≤ F + ≤ 4 at conventional
low Reynolds numbers (Re < 106 ), provided an amplitude
corresponding to Cµ  ∼ 0.1% is exceeded (Greenblatt and Figure 5. Lift coefficient as a function of reduced frequency on a
NACA 0015 airfoil with ZMF slot and a flap at 75% chord deflected
Wygnanski, 2000). It is generally accepted that the maximum
at angle δ (300 000 ≤ Re ≤ 600 000. Modified from Seifert, Darabi
lift attainable on airfoils and wings will depend on prevent- and Wygnanski (1996).
ing separation from an initially attached boundary layer, and
additional parameters such as boundary layer state and cur-
vature are important. The majority of airfoil data indicate mechanisms simultaneously, resulting in a corrugated span-
that maximum increases generally occur at lower reduced wise vortex structure (Neuendorf and Wygnanski, 1999). On
frequencies, typically at 0.5 < F + < 1 (see Figure 5). airfoils, the leading-edge radius can place limitations on con-
Streamwise curvature can have a profound effect on the trol strategies and significantly affect the leading parameters
efficacy of control, as well as the range of optimum frequen- (Greenblatt and Wygnanski, 2003). The relatively “simple”
cies and amplitudes. Perturbations introduced on a curved classical problem of a cylinder placed in cross flow intro-
surface may be amplified by Kelvin–Helmholtz and Görtler duces additional parameters because the separation line is not

Figure 4. Empirical correlation of the lift generated by steady blowing from the trailing-edge region at α = 0◦ ; Re and slot width were
varied. Reproduced with permission from Cerchie et al. (2006)
c AIAA.

Encyclopedia of Aerospace Engineering, Online © 2010 John Wiley & Sons, Ltd.
This article is © 2010 John Wiley & Sons, Ltd.
This article was published in the Encyclopedia of Aerospace Engineering in 2010 by John Wiley & Sons, Ltd.
DOI: 10.1002/9780470686652.eae019
 6 Incompressible Flows and Aerodynamics

precisely (geometrically) determined, it is flow-state depen- approximately uniform across the wingspan but remains
dent and multiple instabilities are present simultaneously effective to high angles of attack only near the tip. When
(Naim et al., 2007). sweep is introduced, a significant effect is noted inboard,
AFC has been demonstrated for a larger Reynolds number but this effect degrades along the span and produces virtu-
range, namely, 3 × 103 ≤ Re ≤ 4 × 107 , under significant ally no meaningful lift enhancement near the tip, irrespective
compressibility effects Ma ≥ 0.55, and in flight tests. At of the tip configuration (Greenblatt and Washburn, 2008).
Reynolds numbers below 5 × 105 , where transition does not In the former case, control strengthens the wing tip vortex;
occur naturally and cannot be passively forced, active sepa- in the latter case, a simple semiempirical model, based on
ration control may be the only effective method for delaying the trajectory or “streamline” of the evolving perturbation,
separation and generating useful lift. Dielectric barrier dis- served to explain the observations. In the absence of sweep,
charge plasma actuation has been demonstrated at Reynolds control on finite-span flaps is slightly more effective near
numbers as low as 3 × 103 (Greenblatt et al., 2008). At the the flap edges (Greenblatt and Washburn, 2008; Kiedaisch,
other end of the scale, periodic perturbations are effective up Nagib and Demanett, 2006). Control over a highly deflected
to Reynolds numbers of 4 × 107 and are essentially indepen- tip flap (40◦ ) produced dramatically larger loads on the flap
dent of Reynolds number for Re > 8 × 106 (Figure 6; Seifert consistent with a strong vortex rolling up over the flap edge.
and Pack, 1999). The effects of compressibility in the absence This effect, combined with negligible changes to the loads
of shocks are weak, and undesirable effects accompanying upstream of the flap, has the potential to produce signifi-
separation, such as vortex-shedding and buffet, can be signif- cant increases in yawing moments or controlled aerodynamic
icantly reduced or completely eliminated (Seifert and Pack, braking (Greenblatt, 2009).
2001). Separation resulting from shock wave/boundary-layer Leading-edge vortices on delta wings at high angles of
interaction can be ameliorated, providing that excitation is attack have been a significant research area because they are
introduced upstream of separation. primarily responsible for the lift generation at low speeds
Significant work has also been performed on three- (e.g., Mitchell and Délery, 2001). As the incidence angle
dimensional flows, where standard sweep laws are increases, the vortex “breaks down”, that is, it expands into a
appropriate for an infinitely swept cylinder (Naveh et al., highly fluctuating structure with flow stagnation in the cen-
1998). On finite unswept wings, the effect of control is tral part and is associated with loss of lift and unsteady loads.
A number of studies have illustrated the profound effect of
periodic perturbations on vortex breakdown and the resulting
increase in delta wing loads (e.g., Margalit et al., 2005). In
general, the data are consistent with those on swept infinite
cylinders, but the mechanism of control has still not been
fully elucidated.
The above discussion relates to “time-invariant” active
control, that is to say that the flow remains attached in a
time-mean sense. The process by which the flow separates
from, or attaches to, a surface is important when we want
to control a process whose characteristic timescale is much
larger than the typical period of eddy passage when periodic
actuation is applied: O(Xsep /U∞ ) (see Darabi and Wygnanski,
2004a, 2004b). Typical examples include the response of
vehicle control-surface flow (Amitay et al., 2004), dynamic
stall control (Greenblatt, Neuburger and Wignanski, 2001),
and control of wake vortices (Greenblatt et al., 2005).

4 COMPUTATIONAL APPROACHES
TO AFC
Figure 6. Effect of control on a NACA 0015 airfoil for Ma = 0.2
and Cµ  = 0.05%. Uncorrected (subscript u) wind tunnel data.
Reproduced with permission from Seifert and Pack (1999)
c A great deal of computational work has been performed dur-
AIAA. ing the past 20 years or so in the field of AFC, with much

Encyclopedia of Aerospace Engineering, Online © 2010 John Wiley & Sons, Ltd.
This article is © 2010 John Wiley & Sons, Ltd.
This article was published in the Encyclopedia of Aerospace Engineering in 2010 by John Wiley & Sons, Ltd.
DOI: 10.1002/9780470686652.eae019
 Aerodynamic Flow Control 7

of it accomplished using numerical solutions to forms of optimum filter width has been shown to be flow dependent,
the Navier–Stokes equations. A brief overview of the main a so-called dynamic methodology is used, where cs is a vari-
methodologies is given here, along with a few examples of able rather than a constant (Germano et al., 1991). Although
applications in the literature. explicit modeling of the subgrid scale effects is very com-
mon, the method of implicit LES, or ILES, has emerged over
the last several years as a popular alternative. With ILES, no
explicit modeling is employed. Instead, the numerical method
4.1 CFD methodologies is selected such that the numerical error fulfills desired prop-
erties and effectively acts like a subgrid model (Grinstein and
In direct numerical simulation (DNS), the Navier–Stokes Fureby, 2007).
equations are solved directly: all scales of motion are LES methodologies have been applied successfully to
resolved, and there is no modeling of turbulence. DNS AFC applications (Dandois, Garnier and Sagaut, 2006a,
requires that the grid should be fine enough to resolve fea- 2006b; Dejoan and Leschziner, 2004; Chang, Collis and
tures on the order of the Kolmogorov dissipation length scale Ramakrishnan, 2002; Rizzetta and Visbal, 2003; You, Wang
 1/4
η = ν3 /ε , and the time step needs to be fine enough and Moin, 2006). Most used either some form of dynamic
to resolve motions whose time scales are on the order of subgrid modeling or ILES. As it is with DNS, LES typically
τ = (ν/ε)1/2 . The simulation also has to be run long enough requires the use of low-dissipation numerical schemes so as
so that the time and phase-averaged properties become tem- not to excessively smear the flow features of the method it
porally converged, and the numerics must be accurate enough is attempting to resolve. The grid requirements are some-
(e.g., high-order, pseudo-spectral, or spectral) so that exces- what less restrictive for LES. As it is with DNS, ensuring
sive dissipation does not corrupt the small-scale resolved turbulent conditions within the computational domain can be
features. Even with today’s computers, it is impossible to problematic.
achieve these resolutions at reasonably high Reynolds num- Although DNS and LES are rapidly becoming usable tools
bers, particularly for complex geometries. However, it is in the CFD arsenal due to increase in computer memory and
relatively common to perform “under-resolved” or “coarse- speed, Reynolds-averaged Navier–Stokes (RANS) methods
grid” DNS, with the justification that the larger resolved are still by far the most commonly used to compute flow
scales have the major influence in the flow control problem of control problems. With RANS, the flow variables of the
interest. The disadvantage of this method is that it is difficult Navier–Stokes equations are decomposed into mean and fluc-
to assess the influence of under-resolving the finer scales. tuating components and then time-averaged. It turns out that
To date, there have been only a limited number of applica- the RANS equations are identical in form to the spatially fil-
tions of DNS or coarse-grid DNS to AFC applications (e.g., tered equations used for LES. As it is for LES, the unknowns
Barwolff, Wengle and Geggle, 1996; Wengle et al.,2001; τik and Qk must be modeled. The turbulent stresses can
Postl and Fasel, 2006; Kotapati, Mittal and Cattafesta, 2007). be modeled in many ways, including using second-moment
When performing flow control computations with DNS, most closure modeling, where a transport equation is solved for
researchers simplify the problem in the spanwise direction the turbulent dissipation rate (or related quantity) as well as
by solving over a finite slice, with periodic boundary con- for each of the six stress components. However, the most
ditions. The issue of inflow boundary conditions or initial common method employed is with linear or nonlinear eddy
conditions (ensuring turbulent flow) can be problematic; this viscosity models (see Gatski and Rumsey, 2002).
is sometimes handled via an additional source term added RANS and LES methodologies are derived differently and
to the equations or by inserting geometric features or blow- their variables have different meanings, but from a coding
ing/suction to trip the flow. standpoint, they are identical except that the models used for
In large eddy simulation (LES), a low-pass filter is obtaining the turbulent viscosity are very different. RANS
applied to the Navier–Stokes equations. As a result, two turbulence models do not involve a filter width and always
unknown quantities emerge that must be modeled: the sub- have an influence across the entire energy spectrum. Although
grid scale stress τik and the subgrid scale heat flux Qk . A gross unsteady motions can often be captured with RANS
commonly used explicit subgrid model is the Smagorin- solved time accurately (sometimes referred to as unsteady
sky model (Smagorinsky, 1963), for which the turbulent RANS or URANS), small-scale turbulence flow features are
viscosity that goes into the computation
of τik and Qk never resolved as the grid is refined because the turbulent
     
is given by νt = cs
¯ 2 S̄ , where S̄  = 2S̄ ik S̄ ik , S̄ ik = viscosity does not go away. This is by design: Reynolds
(∂ui /∂xk + ∂uk /∂xi ) /2, and
¯ is typically defined by some averaging is attempting to represent the mean effects of the
average measure of the local grid spacing. Because the turbulence through the additional turbulent viscosity term.

Encyclopedia of Aerospace Engineering, Online © 2010 John Wiley & Sons, Ltd.
This article is © 2010 John Wiley & Sons, Ltd.
This article was published in the Encyclopedia of Aerospace Engineering in 2010 by John Wiley & Sons, Ltd.
DOI: 10.1002/9780470686652.eae019
 8 Incompressible Flows and Aerodynamics

A particular consideration for RANS applied to AFC is boundary-layer cross flow, and nominally 2D separation-
its ability to capture unsteady flows. The averaging pro- flow-control over a wall-mounted aerodynamic hump shape.
cess could be problematic, but it is generally held that if The workshop summary and initial results were published
the timescale of any gross unsteady motion is much greater by Rumsey et al. (2006), and a follow-up survey on subse-
than the physical time step employed, which in turn is much quent results for the workshop cases was given by Rumsey
greater than the time scales associated with the turbulence, (2009). The lessons learned from the workshop and sub-
then the use of RANS is fully justified (Anderson, Tanehill sequent investigations, mainly regarding separation control,
and Pletcher, 1984). Beyond this issue, given a reasonably are briefly summarized here.
accurate numerical scheme of at least second order and fine For both the nominally 2D and 3D circular synthetic jets,
enough grid and time step, the issue of turbulence mod- the boundary conditions at the slots were shown to play a key
eling always surfaces as a likely reason for discrepancies role. Correct modeling of the membrane and internal orifice
between CFD and experiment. As mentioned earlier, there was necessary for capturing the complex nature of the flow
are many RANS modeling choices, although arguably the field. For the wall-mounted hump case, RANS results con-
Spalart and Allmaras (1994) one-equation SA model and sistently predicted too little eddy viscosity in the separation
the Menter (1994) two-equation k − ω shear-stress trans- region and consequently yielded too long a separation bubble.
port (SST) turbulence model have become the most widely An example is shown in Figure 7 for the hump case with syn-
used over the past 15 years for aerodynamic flow predic- thetic jet (oscillatory control in and out of the slot). Figure 7a
tions. Another consideration is the difficulty associated with shows time-averaged streamlines near the back of the hump
faithfully reproducing experimental boundary conditions and model. Results are shown using three different RANS turbu-
geometric features or irregularities. lence models, including a nonlinear explicit algebraic stress
Hybrid RANS/LES models take advantage of the fact that (EASM-ko) model (Rumsey and Gatski, 2001). All mod-
the LES and RANS implementations only differ in their els yielded similar results, with too big a bubble compared
treatment of τik . It is a relatively simple matter in CFD to experiment. In Figure 7b, the models’ underprediction
codes to blend the two different types of models so that the of turbulent shear stress (−u v ) magnitude in the separated
RANS type is active near walls and the LES type is active in region is evident throughout the blowing/suction cycle (four
wakes and separated regions. For example, in the detached representative phases are shown). This problem with RANS
eddy simulation (DES) method (Spalart et al., 1997), the models also shows up for other cases with separation and
baseline Spalart–Allmaras (SA) one-equation RANS model remains a key roadblock. If a physics-based correction could
(Spalart and Allmaras, 1994) merely modifies the distance be found, it would dramatically improve RANS capabilities
variable in its destruction term to be min (d, cdes
), where not only for AFC cases involving separation, but also for

= max(
x,
y,
z). When cdes
is invoked, the SA equa- general aerodynamic configurations at off-design conditions.
tion behaves in a similar manner to the Smagorinsky  model, As described in detail by Rumsey (2009), it turns out
in the sense that νt becomes proportional to
2 S̄ . In spite that by resolving the large-scale turbulence features in the
of its success in many applications, hybrid RANS/LES mod- separated region using hybrid RANS/LES, LES, or under-
eling still has many unresolved issues related to the interface resolved DNS, much better results can be obtained for
region, gridding, and wall modeling that are being actively this case. (There is one caveat to this conclusion: hybrid
debated and researched (Piomelli et al., 2002; Spalart, 2009). RANS/LES models typically have trouble predicting cases
For all of the methods – DNS, LES, RANS, and hybrid with smaller bubbles, because it is difficult to generate enough
RANS/LES – oscillatory blowing and suction flow control eddy content quickly enough when transitioning from RANS
is most often achieved via transpiration boundary conditions mode to LES mode in the bubble.) The question now is
applied at a wall located either directly on the aerodynamic whether these more expensive simulations can be used to help
surface or within an internal duct or slot included in the com- improve RANS turbulence models in the separated region.
putational geometry. Limited CFD work has been done with Given that RANS could not predict separation extent cor-
moving or distorting walls (e.g., Xia and Qin, 2006). rectly for the hump configuration, Rumsey and Greenblatt
(2007) conducted an investigation to determine how well
the trends could be captured for variations in control magni-
4.2 CFD validation for active flow control tude, control frequency, and Reynolds number. For the steady
suction type of control, CFD appeared capable of predicting
A flow control validation workshop conducted in 2004, titled variations due to Reynolds number and suction strength to a
CFDVAL2004, involved three cases: nominally 2D synthetic reasonable degree of accuracy, but oscillatory control trends
jet into quiescent flow, 3D circular synthetic jet into turbulent were not predicted well.

Encyclopedia of Aerospace Engineering, Online © 2010 John Wiley & Sons, Ltd.
This article is © 2010 John Wiley & Sons, Ltd.
This article was published in the Encyclopedia of Aerospace Engineering in 2010 by John Wiley & Sons, Ltd.
DOI: 10.1002/9780470686652.eae019
 Aerodynamic Flow Control 9

0.2 phase =80 deg.

SA

y/c
0.1 SST
EASM-ko
exp
exp 0
0.1 0 0.02 0.04 0.06
y/c

0.05 −u 'v'/U∝2
Slot
0.2
0 phase =170 deg.
0.7 0.8 0.9 1 1.1 1.2 1.3
x/c

y/c
0.1
SA
0.1
y/c

0
0.05 0 0.02 0.04 0.06
Slot
0 −u 'v'/U∝2
0.7 0.8 0.9 1 1.1 1.2 1.3 0.2
x/c phase =260 deg.

SST

y/c
0.1 0.1
y/c

0.05 Slot
0
0 0 0.02 0.04 0.06
0.7 0.8 0.9 1 1.1 1.2 1.3 −u 'v'/U∝2
x/c
0.2
EASM-ko phase =350 deg.

0.1
y/c

0.1
y/c

0.05 Slot
0 0
0.7 0.8 0.9 1 1.1 1.2 1.3 0 0.02 0.04 0.06
x/c −u 'v'/U∝2

Figure 7. RANS results for the hump model with synthetic jet (a) time-averaged streamlines; (b) turbulent shear-stress profiles in separated
region at x/c = 0.8. Reproduced with permission from Rumsey (2007).

5 CONCLUSIONS AND OUTLOOK these methods, effort should also be expended in improv-
ing RANS predictions. Carrying out dedicated experiments
Active flow control has reemerged as an important research intended specifically for CFD validation will be invaluable
area in aerodynamics. The control of large coherent struc- for future progress.
tures and the mechanical means developed to achieve
this have been at the heart of these developments. As
such, great potential exists for the effective control across LIST OF SYMBOLS
wide Reynolds number and Mach number ranges. Research
areas with potentially large payoffs include simplified c airfoil chord length
high-lift systems, three-dimensional configurations, and CQ slot suction coefficient, hUs /cU∞
vortex-dominated flows. The performance of experiments Cµ slot mean momentum coefficient, J/c q∞
that elucidate new aspects or isolate controlling parameters Cµ  slot oscillatory momentum coefficient
should be considered a priority. Concurrently, the devel- fe separation control excitation frequency
opment of effective and system-efficient actuation methods F+ reduced excitation frequency, fe Lf /U∞
remains a central challenge to the ultimate success of AFC. h slot width
CFD has emerged as the primary tool for prediction J time mean slot momentum
of these flows and much remains to be done because the J time mean slot momentum
commonly used RANS methods often fail to predict exper- L flap length
imental trends. It is expected that the recent progress in the Ma Mach number
arena of simulation methods such as DNS, LES, and hybrid XTE distance from slot to trailing-edge
RANS/LES will continue. However, due to the expense of q∞ freestream dynamic pressure

Encyclopedia of Aerospace Engineering, Online © 2010 John Wiley & Sons, Ltd.
This article is © 2010 John Wiley & Sons, Ltd.
This article was published in the Encyclopedia of Aerospace Engineering in 2010 by John Wiley & Sons, Ltd.
DOI: 10.1002/9780470686652.eae019
 10 Incompressible Flows and Aerodynamics

R radius of curvature Cerchie, D., Halfon, E., Hammerich, A., Han, G., Taubert, L.,
Re Reynolds number based on chord-length Trouve, L., Varghese, P. and Wygnanski, I. (2006) Some cir-
Uj peak jet slot blowing velocity culation and separation control experiments, in Progress in
Astronautics and Aeronautics, vol. 214 (eds R. Joslin and G.
Us slot suction velocity Jones), AIAA, Reston, pp. 113–166.
U∞ freestream velocity
Crittenden, T., Glezer, A., Funk, R. and Parekh, D. (2001)
ui mean velocity component Combustion-driven jet actuators for flow control. AIAA Paper
xi position vector 2001-2768, 31st AIAA Fluid Dynamics Conference & Exhibit,
β empirically determined exponent 11–14 June 2001, Anaheim, CA.
γ empirically determined exponent Dandois, J., Garnier, E. and Sagaut, P. (2006a) Unsteady simulation
ν kinematic molecular viscosity of synthetic jet in a crossflow. AIAA J., 44(2), 225–238.
ε dissipation per unit mass Dandois, J., Garnier, E. and Sagaut, P. (2006b) DNS/LES of active
θ momentum thickness separation control by synthetic jets. AIAA Paper 2006-3026, 3rd
AIAA Fow Control Conference, 5–8 June 2006, San Francisco,
 oscillatory component
CA.
Darabi, A. (2002) On the mechanisms of forced flow reattachment.
PhD Thesis. Tel-Aviv University.
REFERENCES Darabi, A. and Wygnanski, I. (2004a) Active management of nat-
urally separated flow over a solid surface. Part 1: The forced
Amitay, M., Washburn, A.E., Anders, S.G. and Parekh, D.E. (2004) reattachment process. J. Fluid Mech., 510, 105–129.
Active flow control on the stingray uninhabited air vehicle: tran- Darabi, A. and Wygnanski, I. (2004b) Active management of nat-
sient behavior. AIAA J., 42(11), 2205–2215. urally separated flow over a solid surface. Part 2. The separation
Anderson, D.A., Tanehill, J.C. and Pletcher, R.H. (1984) Com- process. J. Fluid Mech., 510, 131–144.
putational Fluid Mechanics and Heat Transfer, Hemisphere Davidson, I.M. (1956) The jet flap. J. Roy. Aer. Soc., 60(541).
Publishing, Washington. Dejoan, A. and Leschziner, M.A. (2004) Large eddy simulation
Attinello, J.S. (1961) Design and engineering features of flap of periodically perturbed separated flow over a backward-facing
blowing installations, in Boundary layer and Flow Control. step. Int. J. Heat Fluid Flow, 25, 581–592.
Its Principles and Application, vol. 1 (ed. G.V. Lachmann), Gallas, Q., Mathew, J., Kaysap, A., Holman, R., Nishida, T., Car-
Pergamon Press, New York, pp. 463–515. roll, B., Sheplak, M. and Cattafesta, L. (2002) Lumped element
Bachar, T. (2001) Generating dynamically controllable oscillatory modeling of piezoelectric-driven synthetic jet actuators. AIAA
fluid flow. US Patent 6,186,412, 13 Feb. Paper 2002-125, 40th AIAA Aerospace Sciences Meeting and
Bar-Sever, A. (1989) Separation control on an airfoil by periodic Exhibit, 14–17 Jan. 2002, Reno, NV.
forcing. AIAA J., 27(6), 820–821. Gaster, M., Kit, E. and Wygnanski, I. (1985) Large-scale structures
Barwolff, G., Wengle, H. and Geggle, H. (1996) Direct numerical in a forced turbulent mixing layer. J. Fluid Mech., 150, 23–39.
simulation of transitional backward-facing step flow manipu- Gatski, T.B. and Rumsey, C.L. (2002) Linear and nonlinear eddy
lated by oscillating blowing/suction, in Engineering Turbulence viscosity models, in Closure Strategies for Turbulent and Transi-
Modelling and Experiments 3 (eds W. Rodi and G. Bergeles), tional Flows (eds B.E. Launder and N.D. Sandham), Cambridge
Elsevier Science, Amsterdam, pp. 219–228. University Press, Cambridge, pp. 9–46.
Baumann, A. (1921) Tragflugel fur Flugzeugemit Luftraustrittsoff- Germano, M., Piomelli, U., Moin, P. and Cabot, W.H. (1991) A
nungen in der Ausenhaut. Deutsches Reichs Patent 400806. dynamic subgrid-scale eddy-viscosity model. Phys. Fluid. A.,
Bussmann, K. and Muenz, H. (1942) Die stabilitaet der laminaren 3(7), 1760–1765.
reibungsschicht mit absaugung. Jb. dt. Luftfahrtforschung I. Glezer, A. and Amitay, M. (2002) Synthetic jets. Annu. Rev. Fluid
36–39. Mech., 34, 503–529.
Chang, P.K. (1970) Separation of Flow by Paul, Pergamon Press. Greenblatt, D. and Wygnanski, I. (2000) Control of separation by
Chang, Y., Collis, S.S. and Ramakrishnan, S. (2002) Viscous effects periodic excitation. Progr. Aerosp. Sci., 37(7), 487–545.
in control of near-wall turbulence. Phys. Fluid., 14(11), 4069– Greenblatt, D., Neuburger, D. and Wygnanski, I. (2001) Dynamic
4080. stall control by intermittent periodic excitation. J. Aircraft, 38(1),
Chen, C., Zakharin, B. and Wygnanski, I. (2008) On the parameters 188–190.
governing fluidic control of separation and circulaton, AIAA Greenblatt, D. and Wygnanski, I. (2003) Effect of leading-
Paper 2008-629, 46th AIAA Aerospace Sciences Meeting and edge curvature on airfoil separation control. J. Aircraft, 40(3),
Exhibit, Reno, NV, January 7–10. 473–481.
Chen, F.-J., Yao, C., Beeler, G.B., Bryant, R.G. and Fox, R.L. (2000) Greenblatt, D., Melton, L., Yao, C. and Harris, J. (2005) Active
Development of synthetic jet actuators for active flow control control of a wing tip vortex. AIAA Paper 2005-4851, 23rd AIAA
at NASA Langley. AIAA Paper 2000-2405, Fluids 2000, 19–22 Applied Aerodynamics Conference, Westin Harbour Castle, 6–9
June 2000, Denver, CO. June 2005, Toronto, Ontario.

Encyclopedia of Aerospace Engineering, Online © 2010 John Wiley & Sons, Ltd.
This article is © 2010 John Wiley & Sons, Ltd.
This article was published in the Encyclopedia of Aerospace Engineering in 2010 by John Wiley & Sons, Ltd.
DOI: 10.1002/9780470686652.eae019
 Aerodynamic Flow Control 11

Greenblatt, D., Göksel, B., Rechenberg, I., Schüle, C., Romann, Experiments – 5 (eds W. Rodi and N. Fueyo), Elsevier Science,
D. and Paschereit, C.O. (2008) Dielectric barrier discharge flow Amsterdam, pp. 307–316.
control at very low flight Reynolds numbers. AIAA J., 46(6), Poisson-Quinton, P. and Lepage, L. (1961) Survey of French
1528–1541. research on the control of boundary layer and circulation, in
Greenblatt, D. and Washburn, A.E. (2008) Influence of finite span Boundary Layer and Flow Control. Its Principles and Applica-
and sweep on active flow control efficacy. AIAA J., 46(7), 1675– tion, vol. 1 (ed. G.V. Lachmann), Pergamon Press, New York,
1694. pp. 21–73.
Greenblatt, D. (2009) Active control of tip-flap loads. AIAA J., 47(3), Post, M.L. and Corke, T.C. (2004) Separation control on high angle
783–788. of attack airfoil using plasma actuators. AIAA J., 42(11), 2177–
Grinstein, F.F. and Fureby, C. (2007) On flux-limiting-based implicit 2184.
large eddy simulation. J. Fluid. Eng., 129(12), 1483–1492. Postl, D. and Fasel, H.F. (2006) Direct numerical simulation of
Hynes, C.S. (1968) The lift, stalling and wake characteristics of a jet turbulent flow separation from a wall-mounted hump. AIAA J.,
flapped airfoil in a two dimensional channel. Stanford University 44(2), 263–272.
report, SUDAAR No. 363. Reau, N. and Tumin, A. (2002) On harmonic perturbations in a
Joslin, R.D. and Miller, D.N. (2009) Fundamentals and applications turbulent mixing layer. Eur. J. Mech. B/Fluid., 21(2), 143–155.
of modern flow control. Progr. Astronaut. Aeronaut., 231, AIAA Rizzetta, D.P. and Visbal, M.R. (2003) Large-eddy simulation of
Press ISBN 978-1-56347-983-0. supersonic cavity flowfields including flow control. AIAA J.,
Kiedaisch, J., Nagib, H. and Demanett, B. (2006) Active flow control 41(8), 1452–1462.
applied to high-lift airfoils utilizing simple flaps. AIAA Paper Rumsey, C.L. (2007) Reynolds-averaged Navier–Stokes analysis
2006-2856, 3rd AIAA Flow Control Conference, 5–8 June 2006, of zero efflux flow control over a hump model. J. Aircraft, 44(2),
San Francisco, CA. 444–452.
Kotapati, R.B., Mittal, R. and Cattafesta, L.N. III. (2007) Numerical Rumsey, C. L. (2009) Successes and Challenges for Flow Control
study of a transitional synthetic jet in quiescent external flow. J. Simulations, International Journal of Flow Control, 1(1), March
Fluid Mech., 581, 287–321. 2009, 1–27 (ISSN 1756-8250). Also AIAA 2008-4311, 4th AIAA
Lachmann, G.V. (1961) Boundary Layer and Flow Control. Its Flow Control Conference, 23–26 June 2008, Seattle, WA.
Principles and Application, vols 1 and 2, Pergamon Press, New Rumsey, C.L. and Gatski, T.B. (2001) Recent turbulence model
York. advances applied to multielement airfoil computations. J.
Margalit, S., Greenblatt, D., Seifert, A. and Wygnanski, I. (2005) Aircraft, 38(5), 904–910.
Delta wing stall and roll control using segmented piezoelectric Rumsey, C.L., Gatski, T.B., Sellers, W.L. III, Vatsa, V.N. and Viken,
fluidic actuators. J. Aircraft, 42(3), 698–709. S.A. (2006) Summary of the 2004 computational fluid dynam-
Menter, F.R. (1994) Two-equation eddy-viscosity turbulence mod- ics validation workshop on synthetic jets. AIAA J., 44(2), 194–
els for engineering applications. AIAA J., 32(8), 1598–1605. 207.
Mitchell, M.A. and Délery, J. (2001) Research into vortex break- Rumsey, C.L. and Greenblatt, D. (2007) Parametric study of flow
down control. Progr. Aerosp. Sci., 37(4), 385–418. control over a hump model using an unsteady Reynolds-averaged
Navier–Stokes code. NASA/TM-2007-214897, September.
Nagib, H., Kiedaisch, J., Reinhard, P. and Demanett, B. (2006) Con-
trol techniques for flows with large separated regions: a new look Samimy, M., Adamovich, Kim, J.H., Webb, B., Keshav, S. and
at scaling parameters. AIAA Paper 2006-2857, 3rd AIAA Flow Utkin, Y. (2004) Active control of high speed jest using local-
Control Conference, 5–8 June 2006, San Francisco, CA. ized arc filament plasma actuators. AIAA Paper 2004-2130, 2nd
AIAA Flow Control Conference, 28 June–1 July 2004, Portland,
Naim, A., Greenblatt, D., Seifert, A. and Wygnanski, I. (2007) OR.
Active control of a circular cylinder flow at transitional Reynolds
numbers. Flow Turbul. Combust., 78(3–4), 383–407. Seifert, A., Darabi, A. and Wygnanski, I. (1996) Delay of airfoil
stall by periodic excitation. J. Aircraft, 33(4), 691–698.
Naveh, T., Seifert, A., Tumin, A. and Wygnanski, I. (1998) Sweep
effect on parameters governing control of separation by periodic Seifert, A., Eliahu, S., Greenblatt, D. and Wygnanski, I. (1998) Use
excitation. AIAA J., 35(3), 510–512. of piezoelectric actuators for airfoil separation control. AIAA J.,
36(8), 1535–1537.
Neuburger, D. and Wygnanski, I. (1987) The use of a vibrating
ribbon to delay separation on two-dimensional airfoils: some pre- Seifert, A. and Pack, L.G. (1999) Oscillatory control of separation
liminary observations. Presented at the Workshop on Unsteady at high Reynolds numbers. AIAA J., 37(9), 1062–1071.
Separated Flow, Air Force Academy, July 1987. Seifert, A. and Pack, L.G. (2001) Oscillatory control of shock-
Neuendorf, R. and Wygnanski, I. (1999) On a turbulent wall jet induced separation. J. Aircraft, 38(3), 464–472.
flowing over a circular cylinder. J. Fluid Mech., 381, 1–25. Smagorinsky, J. (1963) General circulation experiments with the
Nishri, B. and Wygnanski, I. (1998) Effects of periodic excitation primitive equations: part I, the basic experiment. Mon. Wea. Rev.,
on turbulent separation from a flap. AIAA J., 36(4), 547–556. 91, 99–164.
Piomelli, U., Balaras, E., Squires, KD. and Spalart, P.R. (2002) Inter- Sosa, R., Artana, G., Moreau, E. and Touchard, G. (2006) Flow
action of the inner and outer layers in large-eddy simulations with control with EHD actuators in middle post stall regime. J. Braz.
wall-layer models, in Engineering Turbulence Modelling and Soc. Mech. Sci. Eng., 28(2), 200–207.

Encyclopedia of Aerospace Engineering, Online © 2010 John Wiley & Sons, Ltd.
This article is © 2010 John Wiley & Sons, Ltd.
This article was published in the Encyclopedia of Aerospace Engineering in 2010 by John Wiley & Sons, Ltd.
DOI: 10.1002/9780470686652.eae019
 12 Incompressible Flows and Aerodynamics

Spalart, P.R. and Allmaras, S.R. (1994) A one-equation turbulence Weier, T. and Gerbeth, G. (2004) Control of separated flows by time
model for aerodynamic flows. La Rech. Aerospatiale, 1, 5–21. periodic Lorentz forces. Eur. J. Mech. B/Fluid., 23, 835–849.
Spalart, P.R., Jou, W-H., Strelets, M. and Allmaras, S.R. (1997) Wengle, H., Huppertz, A., Barwolff, G. and Janke, G. (2001) The
Comments on the feasibility of LES for wings, and on a hybrid manipulated transitional backward-facing step flow: an exper-
RANS/LES approach, in Advances in DNS/LES, First AFOSR imental and direct numerical simulation investigation. Eur. J.
International Conference on DNS/LES (eds C. Liu, Z. Liu, L. Mech. B, 20, 25–46.
Sakell), Greyden Press, Columbus, pp. 137–147. Williams, J. (1961) A brief history on British research on boundary
Spalart, P.R. (2009) Detached-eddy simulation. Annu. Rev. Fluid layer control for high lift, in Boundary Layer and Flow Con-
Mech., 41, 181–202. trol. Its Principles and Application. vol. 1 (ed. G.V. Lachmann),
Spence, D.A. (1956) The lift coefficient of a thin, jet-flapped wing. Pergamon Press, New York, pp. 21–73.
Proc. Roy. Soc., A 238(1212), 46–68. Wlezien, R.W., Horner, G.C., McGowan, A.R., Padula, S.L.,
Stalnov, O. and Seifert, A. (2008) On amplitude scaling options of Scott, M.A., Silcox, R.J. and Simpson, J.O. (1998) The air-
active separation control. 2nd International Conference on Jets, craft morphing program. AIAA Paper 1998-1927, 39th AIAA/
Wakes and Separated Flows, Berlin, Germany. ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and
Materials Conference and Exhibit, and AIAA/ASME/AHS
Tollmien, W. (1955) 50 Jahre Grenzschichtforschung, ihre Entwick- Adaptive Structures Forum, 20–23 April 1998, Long Beach, CA.
lung und Problematik. Henry Görtler/Walter Tollmien (Hg.): 50
Jahre Grenzschichtforschung, Braunschweig, S.1–12. Woods, L.C. (1958) Some contributions to jet-flap theory and to the
theory of source flow from aerofoils. A.R.C. Current Paper 388.
van Dam, C.P., Yen, D.T. and Vijgen, P.M.H.W. (1999) Gurney
flap experiments on airfoil and wings. J. Aircraft, 36(2), 484– Xia, H. and Qin, N. (2006) Comparison of unsteady laminar and
486. DES solutions of synthetic jet flow. AIAA Paper 2006-3161, 24th
AIAA Applied Aerodynamics Conference, 5–8 June 2006, San
Viets, H., Piatt, M. and Ball, M. (1987) Boundary layer control by Francisco, CA.
unsteady vortex generation. J. Wind Eng. Ind. Aerod., 7, 135–
144. You, D., Wang, M. and Moin, P. (2006) Large-eddy simulation of
flow over a wall-mounted hump with separation control. AIAA J.,
Walsh, M.J. and Anders, J.B. Jr. (1989) Riblet/LEBU research at 44(11), 2571–2577.
NASA Langley, 46(3), 255–262.

View publication stats

Encyclopedia of Aerospace Engineering, Online © 2010 John Wiley & Sons, Ltd.
This article is © 2010 John Wiley & Sons, Ltd.
This article was published in the Encyclopedia of Aerospace Engineering in 2010 by John Wiley & Sons, Ltd.
DOI: 10.1002/9780470686652.eae019