Journal of Fluids Engineering.

Received March 28, 2016;

Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

Experimental study of suction flow control

effectiveness in a serpentine intake
M. C. Keerthi Abhijit Kushari1
Department of Aerospace Engineering, Department of Aerospace Engineering,

d
Indian Institute of Technology Kanpur, Indian Institute of Technology Kanpur,

ite
Kanpur, Uttar Pradesh, Kanpur, Uttar Pradesh,
India – 208016 India – 208016

ed
mckee@iitk.ac.in akushari@iitk.ac.in
Member ASME

py
Valliammai Somasundaram

Co
Aeronautical Development Agency,
Bengaluru, Karnataka,
India – 560017

ot
valli@jetmail.ada.gov.in
tN
rip

ABSTRACT
sc

The intakes of modern aircraft are subjected to ever-increasing demands in their

nu

performance. Particularly, they are expected to carry out diffusion with the highest
Ma

isentropic efficiency while subjected to aggressive geometry requirements arising from

stealth considerations. To avoid a penalty in engine performance, the flow through

ed

intake needs to be controlled using various methods of flow control. In the present
pt

study, a serpentine intake is studied experimentally and its performance compared with
ce

and without boundary layer suction. The performance parameters used are non-
Ac

dimensional total pressure loss coefficient and standard total pressure distortion

descriptors. The effect is observed on surface pressure distributions and inferences are
1
Corresponding author: Abhijit Kushari, Department of Aerospace Engineering, Indian Institute of
Technology Kanpur, Kanpur, Uttar Pradesh, India 208016.

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

made regarding separation location and extent. A detailed measurement at the exit

plane shows flow structures that draw attention to secondary flows within the duct.

Suction is applied at three different locations, spanning different number of ports along

each location, comprising of ten unique configurations. The mass flow rate of suction

d
ite
employed ranges from 1.1 to 6.7 % of mass flow rate at the inlet of the intake. The

ed
effect is seen on exit total pressure recovery as well as circumferential and radial

distortion parameters. This is examined in the context of the location of the suction

py
ports and amount of suction mass flow, by the deviation in surface pressure

Co
distributions, as well as the separation characteristics from the baseline case. The

ot
results show that applying suction far upstream of the separation point together with a
tN
modest amount of suction downstream results in the best performance.
r ip
sc

INTRODUCTION
nu

An intake is required to deliver uniform flow to the compressor at an acceptable velocity,

and also have a high isentropic efficiency, low installation drag, and be of minimal length.
Ma

Additionally, there could be constraints from the airframe integration. For civil aircraft, an
ed

offset in the axis between the engine and the intake opening causes the intake to have an S-
pt

shaped curvature along its centerline, such as in the case of Lockheed L1011 Tristar and Boeing
ce

727. For military aircraft, there are additional requirements such as low radar cross-section [1]
Ac

and low sensitivity to entry air angle of attack. Popular examples of aircraft possessing such

intakes include General Dynamics F-16, McDonnell-Douglas F-18, Lockheed F-117 Nighthawk

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

and Lockheed Martin F-22 Raptor. S-shaped intakes have also been suggested for aircraft

intakes as they can be designed to avoid foreign object damage to the engines [2].

Each of these requirements makes the design of an intake challenging as they have

contradictory requirements. Further, the designers of present-generation unmanned combat

d
ite
aerial vehicles prefer serpentine intakes due superior stealth characteristics. Essentially, the

ed
presence of flow separation limits the aerodynamic performance of a serpentine intake. During

critical parts of the flight regime, such as sudden out of plane maneuvers and slips, where the

py
flow through the intake is affected, it is especially essential to ensure that the compressor does

Co
not stall. At these times, it may be necessary to apply flow control to prevent potentially

ot
destructive occurrences like surge or forced excitation of rotor blades, even at a slight loss of
tN
the primary output of the engine namely thrust. This serves as a motivation to study and
ip

develop flow control methods that result in the designed operation of an intake.
r
sc

It is well-known that as a flow with non-zero transverse vorticity component enters a bend, a
nu

streamwise component of vorticity is generated. This is a consequence of the radial pressure

Ma

gradient balancing the centripetal force. Squire and Winter [3, 4] carried out an inviscid analysis

to derive a formula that relates the magnitude of the streamwise vorticity generated along a
ed

bend, as a function of the inlet transverse vorticity and the total deflection angle. Bansod and
pt

Bradshaw [5], in an experimental study of multiple S-duct geometries, have demonstrated that
ce

the secondary flow creation is an inviscid phenomena, provided a velocity gradient such as a
Ac

boundary layer is present at the inlet. Experiments were conducted by Taylor et al. [6] on an S-

duct with square cross-section to study the development of secondary flow along the length of

the duct. It was seen that the thickness of the boundary layer at the entrance has a substantial

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

influence on the transverse velocity components along the duct. Further, the second bend

caused the original vorticity due to the first bend to be nullified and sustained a vortex of the

opposite sense. However, the flow near a small region near the outer wall of the second bend

did not conform to this, as the secondary flow upstream had reversed the sign of the transverse

d
ite
vorticity, resulting in sustaining that vorticity according to the Squire and Winter expression.

ed
Flow control methods can be classified based on whether or not they consume additional

energy [7]. Methods that involve the application of blowing [8], suction [9], and vortex

py
generator jets [10] are classified as active methods, and methods that employ devices such as

Co
vortex generators [11], roughness elements [12] are classified as passive. The functioning of

ot
vortex generators is based on the principle that the vortices produced by them re-energizes the
tN
slow-moving boundary layer fluid by drawing energy from the freestream flow. In certain
ip

studies [13], the vortex generators have been used to redirect the flow within the boundary
r
sc

layer in a direction opposite to that of the naturally occurring secondary flow. Although this
nu

method consumes the least amount of additional mass flow rate, the effectiveness of flow
Ma

control is very sensitive to the location of the jets and the jet velocity ratios, thus requiring

considerable efforts in optimizing.

ed

The benefits of boundary layer suction as a flow control method is well-recognized since the
pt

experiment by Prandtl [14], who applied suction through slits on the trailing side of a circular
ce

cylinder to suppress separation. Suction is also applied in laminar flows to delay its transition to
Ac

turbulence, commonly with the ultimate intention to delay separation. A study [15] on the

effect of suction on turbulent boundary layer shows that at suction momentum rate ratio of

about 2.6 or more, relaminarization occurs downstream of the strip, followed by a transition to

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

a turbulent self-preserving state. Kirk et al. [16, 17] have conducted a numerical and

experimental study on a geometry comparable to the one employed in the present paper. The

flow control method involved the use a fluidic actuator, which was a compact centrifugal

blower that extracts fluid from one region of the duct and ejects into another region. With this

d
ite
configuration, pure suction, suction with steady blowing and suction with pulsed blowing were

ed
applied. It was seen that the application of suction beyond 1.25 % of the total mass flow did not

have an effect on the location of separation point, but it greatly improved the total pressure

py
recovery at the exit plane. Further, the difficulties of numerical turbulence models in capturing

Co
all the phenomena, especially with regard to the secondary flow due to the first bend, were

ot
stated. Ball [9] performed tests on highly offset diffusers with suction and blowing. The amount
tN
of suction and blowing was about 2 % of the inlet mass flow. At about 2.3 % the total pressure
ip

recovery was very high and increasing the suction flow rate to 4.7 % caused negligible
r
sc

improvement. The uniformity of the flow at the inlet of an axial compressor has a great effect
nu

on its performance. Particularly, the loss of surge margin is a usual consequence to poor inlet
Ma

flow total pressure distribution. In a study by Reid [18], the effect of circumferential total

pressure distortion on surge margin loss was found to be more severe than the radial
ed

variations, although other researchers have pointed out significant importance of radial
pt

distortion also.
ce

Several studies have been undertaken to study the effect of flow control methods in simple
Ac

geometries such as flat plates to understand their effects in detail or to validate analytical

models. But the number of studies available in open literature wherein flow control is applied

on real complex geometries are relatively less. The flow within a diffusing, highly three-

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

dimensional duct with two bends, such as the one considered in the present study, may not

necessarily be compliant with flow control schemes that work well in simpler geometries.

Although, it may be possible to design an intake with better pressure recovery and distortion

properties, the modified design will usually have a larger length-to-diameter ratio. In such a

d
ite
case, an optimization should be carried out to verify which of the following results in minimal

ed
aircraft weight: the original intake with the associated flow control penalties, or the longer

intake which further increases the length of the airframe.

py
In the present study a typical serpentine intake is characterized and the effectiveness of

Co
suction as flow control is studied. The results reported in this paper show that substantial

ot
improvement of the distortion parameters can be achieved by employing controlled suction in
tN
the diverging portion of the diffuser, but, the location of the suction and the mass fraction of
ip

suction play important roles in the effectiveness of the flow control scheme.
r
sc
nu

EXPERIMENTAL SETUP
Ma

The experiments were performed on an open-circuit wind tunnel with a rectangular exit of
ed

dimensions 305 mm × 381 mm. The wind tunnel consists of a honeycomb mesh and two wire-
pt

mesh screens ensuring flow of low turbulence and a contraction section of area ratio 21. The
ce

average turbulence intensity at the tunnel exit was measured (using hotwire anemometer) to
Ac

be less than 0.5 %. The intake model has a bean-shaped elliptical inlet and a circular exit. In

order to interface its inlet to the exit of the wind tunnel, a transition section was used. Fig. 1

shows the schematic of the wind tunnel and the intake duct. An outline of the components of

the wind tunnel is shown in Fig. 1(a). Fig. 1(b) shows the intake with the transition section and

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

Fig. 1(c) shows the intake with various features labelled. The transition section and the intake

were attached at the end of the contraction cone marked “test section inlet” in Fig. 1(a). The

transition duct is a converging duct with an inlet-to-exit area ratio of 1.3. The duct is designed

by interfacing the rectangular tunnel exit area to the bean-shaped serpentine intake area. As

d
ite
the geometry is symmetric about the vertical plane, three circular arcs (one in the horizontal

ed
plane, and two in the vertical plane) are used to define the surface, all three starting from the

mid-points of three of the sides of the rectangular tunnel exit area, and continuing up to the

py
intake area. The arcs were slightly modified to ensure tangency requirements and converted to

Co
splines. The total length of the transition section is 1.00 m. A CFD study was carried out for

ot
different combinations of curvatures of the arcs, and the set that produced the most uniform
tN
exit axial velocity distribution (more than 95% of the area) was chosen for fabrication. The
ip

finalized radii of curvature of the upper, lower and the horizontal-plane arcs, all concave from
r
sc

inside, are 1.63 m, 2.32 m and 2.18 m respectively. The mid-plane velocity at the inlet of the
nu

intake was about 36.5 m/s for all measurements. This results in a Reynolds number of 8.4 × 105
Ma

based on the exit diameter. Wall measurements were carried out along the upper and lower

wall centerlines. The measurement stations were located at 18 equidistant (as measured along
ed

the surface of the duct) points on the lower wall centerline. A total of 25 stations were chosen
pt

along the upper wall centerline with lesser spacing towards the exit, since preliminary
ce

investigations indicated separation in that region (a result also reported shortly in this paper).
Ac

Each measurement station consisted of a static tap and a Preston tube, both made from steel

tubes of inner diameter 0.8 mm. While the static tap was used to measure the wall static

pressure, the pressure difference between the Preston tube and static tube was used to

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

evaluate an indicative skin friction coefficient (cf*) that was used primarily to identify the

separation location [19, 20]. They consist of an L-shaped tube to measure stagnation pressure

adjacent to a static tap on the wall. The geometry of the Preston tubes used here is described in

a previous study [21]. It is to be emphasized that cf* will not provide a value of the skin friction

d
ite
coefficient, as the method of using Preston tubes work only for unidirectional flows parallel to a

ed
flat plate, in the absence of pressure gradients. However, past studies [22, 23] on similar

internal flow geometries have shown that such an arrangement of static and Preston tubes can

py
be used to identify if the flow has locally separated, with a value of zero indicating separation.

Co
Due to the geometrical arrangement of the tubes, the Preston tube will always read a value

ot
higher than the static tube as the streamwise component of velocity is dominating, except in
tN
the presence of a recirculating region, where the differential pressure between static and
ip

Preston tubes is zero over a sufficiently long average.

r
sc

The exit plane of the intake is commonly referred to as the aerodynamic interface plane
nu

(AIP), and this plane coincides with the compressor inlet plane in the installed engine. The exit
Ma

plane measurements were made using a five-hole probe connected to two pressure scanners.

The static and Preston tubes were also connected to the pressure scanners. The five-hole probe
ed

had a tip diameter of 3.18 mm and was factory calibrated (Aeroprobe Inc.). The pressure
pt

scanners (Pressure Systems, model ESP 32 HD) had 32 differential ports each, with ranges 995
ce

kPa (4” water column) and 13.8 kPa (2 psi) and the acquisition was made sequentially at a rate
Ac

such that the sampling rate per port was 256 samples/s. Each measurement was made for two

seconds and the mean of the 512 samples was recorded for all the measurements.

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

Suction was applied at three streamwise locations, near the upper wall centerline (as

the separation region was detected in that region, as described in the measurements for the

baseline performance in Section III A) each containing five, three and four holes respectively, as

shown in Fig. 2. Each suction port was angled at 60° to the local mean freestream velocity

d
ite
vector, and a flexible tube of fixed length connected it to a large-diameter pipe, which led to

ed
the inlet of a blower through a venturimeter. The pipe acts as a settling chamber to ensure the

same pressure is subjected to the ends of the tube. When suction was in use, care was taken to

py
arrange the tubes uniformly such that the pressure losses due to bends are minimal. As the

Co
blower had no speed control and to prevent choking in the line, a T-junction was installed in the

ot
line, with one end drawing air from the atmosphere. A pair of valves located in each stream
tN
before mixing was used in combination to set the required mass flow rate.
ip

The performance parameters used for evaluating the performance are presented here. The
r
sc

wall static pressure coefficient cp at a given station is defined in the usual manner as:
nu

p s  p1 (1)
cp 
q1
Ma

The inlet dynamic pressure (q1) measured by a Pitot-static probe located at the inlet plane
ed

centroid, is equal to the difference between the inlet total (p01) and static pressure (p1). At a
pt

given location along the intake, the static pressure (ps) is the value measured by the static wall
ce

pressure tap. The indicative skin friction (cf*) coefficient is defined as:
Ac

p p  ps
cf *
q1 (2)

Here, pp is the pressure measured by the Preston tube as described previously. The total

pressure loss ω measured at each point in the exit plane is quantified by taking the difference

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

between the measured exit total pressure (p02) the inlet total pressure, non-dimensionalized by

the inlet dynamic pressure, as:

p01  p02

q1 (3)

d
In order to compare the total pressure distortion, a consistent set of parameters as defined

ite
by the SAE ARP 1420 (SAE International’s Aerospace Recommended Practice) is used here [24,

ed
25]. For this, the measurement grid at the exit plane is a set of 40 points, on eight radial rakes

py
and the center of five rings. The radii are chosen to keep each annulus area constant. The

Co
circumferential distortion intensity (CDI) at the k-th radial location (‘ring’) is defined as:

PAV  PAV  LOW

ot
CDI k  (4)
PAV
tN
where, PAV is the average of the set of eight total pressure values of the ring and PAV-LOW is
ip

the subset of only those values that are less than PAV. Similarly, the circumferential distortion
r
sc

extent (CDE) of a ring is defined as the angular extent over which the PAV-LOW values lie, as
nu

shown in Fig. 3. The radial distortion intensity (RDI) is defined at a given radius as, the ratio of
Ma

the difference between the face averaged total pressure and the ring-averaged total pressure

to the face averaged total pressure.

ed

PFAV  PAV (5)

RDI k 
pt

PFAV
ce

A standard parameter for calculating the distortion characteristic for the entire plane is the
Ac

DC60 parameter defined as:

PALL  PWORST 60

DC 60 
q1 (6)

10

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

where, PALL is the average total pressure of all the 40 points and PWORST-60 is the average of

the points in the sector with the lowest values of total pressure, and q1 is the duct inlet dynamic

pressure. Due to the nature of the flow in the present intake, the worst region was confined to

the radial line joining the upper wall centerline. As a result, the worst 60° sector was

d
ite
consistently the region centered about this radius, with the PWORST-60 effectively being the

ed
average total pressure of these five points, as shown by the inset figure in Fig. 3. A more

elaborate description of these parameters may be found in references 24 and 25.

py
The uncertainties of the measured and calculated parameters were evaluated from a set of

Co
ten repeated measurements at a steady condition for the baseline duct. Standard rules of

ot
propagation of uncertainty were employed to estimate the absolute uncertainties at 95%
tN
confidence interval and are listed in Table 1.
ip

RESULTS AND DISCUSSION

r
sc
nu

A. Baseline performance

The performance of the intake was first evaluated without the application of flow control.
Ma

The performance is adequately evaluated by measuring the properties of the air at the exit
ed

plane. The measurements of the flow along the length of the duct indicate the flow
pt

characteristics within the duct. The measurements so obtained, referred to as baseline

ce

measurements, also serve as a reference to evaluate the changes produced by each flow
Ac

control configuration. The variation of static pressure coefficient (cp) on the wall, along the

upper and lower wall centerline (as labelled in Fig. 1(c)) is shown in Fig. 4. Due to the gradient in

pressure at the inlet plane, the arithmetic average of the upper and lower wall pressure is taken

11

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

as the reference for all cp values, and is equal to p1 as mentioned earlier. The plots are overlaid

on an image of the duct along its plane of symmetry. Figure 4 also shows the variation of the

“ideal cp”, which is the one-dimensional, inviscid static pressure variation for a straight diffusing

duct with an identical area variation along the length. This distribution is representative of

d
ite
simplified duct geometry in non-dimensional form. Since the purpose of ideal cp is to

ed
understand the effect of area variation, the reference pressure for the ideal cp is conveniently

taken as the average of the experimentally measured static pressures on the upper and lower

py
wall centerlines of the present serpentine intake. A significant difference can be seen in the

Co
actual static pressure variation along the walls compared to the theoretical variation. This is a

ot
direct consequence of the effects such as curvature, viscous boundary layer growth and three-
tN
dimensionality. It can be seen that the value of cp for upper wall centerline is positive while the
ip

lower wall centerline is negative at the intake entrance (x/L = 0). Noting again that the
r
sc

reference for static pressure is taken as mean static pressure at the intake inlet area, this
nu

implies that the velocity near the upper wall is lower than that near the lower wall. This is a
Ma

result of unequal accelerations over the upper and lower walls of the transition section, which

itself has curved surfaces to transition the flow to the intake entrance shape. The lower value of
ed

cp for the lower wall centerline suggests that the flow has accelerated faster over the lower wall
pt

along the transition section. Within the duct, the pressure drops below the inlet value (i.e., cp is
ce

negative) for certain regions of the upper and lower walls. This is due to the combined effect of
Ac

curvature of the duct and the ensuing secondary flow, which is discussed next. Ultimately, the

average cp is seen to be higher at the exit (x/L = 1.0) when compared to the inlet, denoting that

the duct is decelerating the flow, and hence performing as a diffuser. This approximate analysis

12

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

was seen to be consistent with the higher resolution measurements made at the exit plane and

discussed later.

The difference between the upper and lower wall centerlines pressure trend is also

suggestive of other features that are discussed below. Similar variations exist in some of the

d
ite
earlier studies on serpentine ducts [16, 21]. The variation of wall static pressure follows the

ed
local flow acceleration and deceleration unless the effects of a loss-creating region, such as a

separated region, exist. Therefore, it is usually not possible to identify the separation point

py
from the static pressure distribution along a line. The increase in lower wall static pressure in

Co
the initial region of the duct up to x/L = 0.3 can be expected to be primarily due to the effect of

ot
flow deceleration over the curved walls in that region, as the variation in area, as shown by the
tN
ideal cp plot, is not significant in that region. The high curvature in that region has also caused a
ip

drop in the upper wall static pressure. After x/L = 0.3, a region of adverse pressure gradient
r
sc

begins along the upper wall centerline, indicating likelihood of boundary layer separation.
nu

Although the lower wall shows an adverse pressure gradient in the region between x/L = 0 to
Ma

0.3, it is unlikely that the boundary layer would have undergone separation here, as the velocity

would still be high due to the inertia of the fluid at the intake entrance. Further, it can be seen
ed

that there are two points of inflection, one in each “S-shape” of the duct. Each turning results in
pt

a radial pressure gradient causing secondary flows. Since the slow-moving boundary layer fluid
ce

and the high momentum core fluid are both subjected to the same pressure gradient, the flow
Ac

must correct itself to balance the centripetal force due to pressure gradient. The correction is

achieved by the core flow moving towards the outer wall, increasing its radius proportionally

with the square of its velocity. The boundary layer fluid moves towards the inner wall along the

13

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

walls of the duct to satisfy continuity. This sets up a pair of streamwise vortices, the strength of

which will be largely dependent on the radius of the turn and the nature of pressure gradient.

Due to the viscous dissipation within a vortex [26], a large total pressure defect is present at the

core of each vortex. This local defect will eventually be dissipated, provided it is convected

d
ite
downstream for a sufficiently long distance. Although it is possible to speculate about the

ed
separation point by examining the wall static pressure, it is not a definitive indicator of

separation, especially in a complex geometry such as the one under study.

py
The variation of the indicative skin-friction coefficient along the duct is shown in Fig. 5. As

Co
mentioned previously, a zero value implies that the local velocity is zero on an average,

ot
signifying flow separation. In the figure, it is clearly seen that the lower wall centerline has fully
tN
attached flow throughout the length of the duct. But the upper wall centerline shows near zero
ip

and a slight negative value after about x/L = 0.8. Considering that the wall static pressure was
r
sc

adverse for an extended distance up to this point, it can be concluded that flow has separated
nu

along the upper wall. The exact location of the separation point was seen to be at x/L = 0.86,
Ma

corresponding to a near-zero value of cf*. The cf* value increases downstream of this point due

to the effect of ambient pressure outside the duct.

ed

The variation of exit flow properties at the AIP is examined next. Fig. 6(a) shows the variation
pt

of the non-dimensional total pressure loss coefficient at the AIP, as measured over the 40-
ce

points SAE ARP 1420 grid. A significant loss of total pressure can be seen near the region
Ac

corresponding to the upper wall centerline, towards the end of which separation was identified

(Fig. 5). A majority of the region below that region shows low total pressure loss of about 0.2. It

is clear that the boundary layer separation at the upper wall centerline has contributed to a

14

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

massive total pressure defect, whereas as the flow is attached in the region near the lower wall,

the entire lower hemisphere shows very little loss of total pressure. The maximum value of ω

was found to be 0.76 and the average of all values was 0.26. Fig. 6(b) shows the variation of

axial component of velocity as measured by the five-hole probe. A deficit in axial velocity can be

d
ite
seen here, in the same region corresponding to high total pressure loss. From these

ed
measurements, the total pressure distortion descriptor DC60 was found to be 0.40, which is

substantially higher than what is acceptable for most modern aero-engines.

py
The distortion due to the total pressure variation in the AIP is now examined quantitatively

Co
using the distortion descriptors as defined in Section II. The circumferential distortion extent

ot
and intensity for the baseline case are shown in Fig. 7. The distortion intensity is seen to be
tN
slightly higher near the tip (maximum non-dimensional radius of 0.71, near the tip of the
ip

compressor blade downstream in an axial-flow gas turbine engine) compared to the hub. The
r
sc

distortion extent is significant where the distortion intensity is higher than acceptable. For the
nu

baseline case, distortion extent is seen to be moderately high for all the radii except very near
Ma

the tip. The reason for such a variation can be understood by looking at the total pressure loss

contours (Fig. 6(a)), which shows a large total pressure defect in that range of radius.
ed

The SAE ARP 1420 40-point grid is convenient to describe the distortion quantitatively.
pt

However, the grid is too coarse to extract small-scale flow structures, as there are just 10 points
ce

of measurements along a diameter. Therefore, a separate fine-grid measurement was made,

Ac

with a uniform rectangular grid of sides’ 0.0135D, which resulted in about 73 points along the

diameter. Due to the time-consuming nature of this measurement and the heating of the wind

tunnel blower air, the measurements were restricted to one hemisphere at a time, leading to a

15

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

total of 2155 points in each hemisphere. The probe sampling time was reduced to 1-second

equivalent to 256 samples. The total pressure loss coefficient contours for the fine-grid

measurement in the upper hemisphere is shown in Fig. 8. The in-plane velocity components are

overlaid in the form of streamlines in the AIP plane. It is to be noted that the velocity

d
ite
components depicted have a dominant (by at least an order of magnitude) freestream

ed
component of velocity that is not represented by the in-plane streamlines. As a result, the

streamlines are representative of only the transverse components of velocity; with an implied

py
streamwise mean flow velocity component. The presence of a distinct pair of counter-rotating

Co
vortices can be seen, in the form of concentric streamlines as well as a region of defect in total

ot
pressure caused by the viscous dissipation. A similar fine-grid measurement in the lower semi-
tN
circle showed no distinct vortices and had features that are captured in Fig. 6(a). This provides
ip

important information regarding the origin of the secondary flows. The direction of the vortices
r
sc

indicates that the core flow is moving towards the lower wall, which happens when the duct is
nu

turning towards the upper wall. As the upper wall shows an adverse pressure gradient from x/L
Ma

= 0.3 onwards, which also corresponds to the beginning of an upward turn, it is clear that the

high curvature bend on the upper wall at x/L = 0.4 is responsible for the counter-rotating
ed

vortices at the AIP. Further, the prolonged region of adverse pressure gradient along the upper
pt

wall has resulted in a flow separation at x/L = 0.86.

ce
Ac

B. Duct performance with flow control

Having identified the separation location, a choice was made to apply suction at a decisive

location to prevent it. As it is most beneficial to apply flow control before the flow deteriorates,

a location just upstream of the separation point on the upper wall was identified to make the

16

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

suction slots. In addition, two more locations were identified to apply suction. In total, there

were three rows of suction slots at different streamwise locations, as follows: (a) location A, at

x/L = 0.70, consisting of five slots, (b) location B, at x/L = 0.77, consisting of three slots, and (c)

location, at x/L = 0.87, consisting of four slots. The reason for the choice of these locations is

d
ite
the following: from the streamwise pressure variation at the upper wall, it can be seen that the

ed
recovering pressure trend that starts from x/L = 0.55 is arrested at x/L = 0.85, after which there

is a reduction of pressure gradient. Also, in between this region, the upper wall cf* shows a

py
decline after x/L = 0.67. A location was chosen after this point, at x/L = 0.70 (location A) and

Co
another at x/L = 0.77 (location B) roughly halfway to the separation point, with slight

ot
adjustments to ensure no more than one measurement point is lost at each location.
tN
Additionally, location C was chosen just downstream of the separation point to check for its
ip

effectiveness in distortion at the AIP despite the presence of separation. The details of the port
r
sc

locations are shown in Fig. 2(b). The nomenclature of each case name is such that, application
nu

of flow control through a slot is denoted by “1” and a slot that was unused as “0”. Therefore,
Ma

case 6A, which is denoted as 01110 + 11 + 0110, indicates that suction is applied at the middle

three ports at location A, the two outer ports at location B, and the two inner ports at location
ed

C. For clarity, Fig. 2(c) shows the suction port arrangement for all cases, where a hollow circle
pt

represents an unused port, and a filled circle port used for suction. For locations B and C,
ce

placement of a suction slot along the upper wall centerline was avoided as far as possible, so
Ac

that the wall static pressure and indicative skin-friction coefficient measurements can be made

at that location. Due to this, only for cases 8 and 10, a hole was drilled in the center at location

B, as it turned out to provide some benefit at the location based on the interim results.

17

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

Table 2 shows the complete list of flow configurations considered in this study. The total

mass flow rate from all ports relative to the flow rate at the entrance of the duct is expressed as

suction mass flow percentage. The calculated velocity of suction with respect to the inlet

velocity is shown as velocity ratio. The additional energy expended for suction is quantified as a

d
ite
ratio of the kinetic energy of the suction mass flow to that of the mean flow entering the

ed
intake. This is included as suction kinetic energy flux ratio in the same table.

The results of all the flow control cases are presented first. The variation of the wall static

py
pressure coefficient along the lower and upper wall centerlines are given in Fig. 9 and Fig. 10(a).

Co
The three vertical dashed lines denote the position of locations A, B and C. The legend from Fig.

ot
9 is used for all subsequent line plots corresponding to the flow control cases. From Fig. 9, it can
tN
be seen that the effect of application of suction is to reduce the wall static pressure at locations
ip

before the suction location, and to increase static pressure beyond the location. This can be
r
sc

explained by noting that, as suction removes the boundary layer, the flow within the duct
nu

accelerates creating a drop in the pressure. Downstream of the location, as some mass is
Ma

removed, the velocity is reduced resulting in additional diffusion and hence an increased

pressure compared to the baseline case. A similar trend can be seen along the upper wall
ed

centerline also, as depicted in in Fig. 10(a), except for the region close to and downstream of
pt

the suction locations. In order to better observe the details in this region of interest, an
ce

enlarged plot is shown as Fig. 10(b). The variation of the indicative skin-friction coefficient along
Ac

the upper wall centerline for all the cases within the region of interest is shown in Fig. 11. From

this figure, it is clear that the flow control cases show an altered boundary layer separation

characteristics with varying degree of success. The application of suction at the first location

18

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

(location A) is seen to be always accompanied by a sharp rise in cf*, denoting an increase in

streamwise velocity near the wall.

The effect of all the flow control cases on total pressure loss at the AIP for all the flow

control cases is given in Fig. 12. Although the qualitative changes are seen in the total pressure

d
ite
contours, the distortion descriptors, as shown in Figs. Fig. 13-Fig. 15, are useful for a

ed
quantitative assessment. The circumferential distortion intensity plot (Fig. 13) shows an

identical trend for all cases except two, which are discussed in the subsequent sub-section.

py
The circumferential distortion extent plot (Fig. 14) shows considerable variation near the

Co
hub compared to the mid-radial locations and the tip. This suggests that different flow control

ot
configurations have a pronounced effect near the hub, as only few configurations are capable
tN
of influencing the near-hub regions. The relative difference in variation of radial distortion (Fig.
ip

15) is seen to be minimal for all flow control cases considered. This is perhaps due to the type
r
sc

of distortion produced by the intake considered. From the definition of radial distortion
nu

intensity, this implies that the ratio of the mean total pressure in a ring to the mean of total
Ma

pressure in all rings is the same for all flow control cases.
ed

i. Effect of suction location:

pt

To consider the effect of the streamwise location of suction, cases 1 and 2 may be first
ce

considered. The significance of the locations were that one is before the separation point at x/L
Ac

= 0.70 (location A) and another is downstream of it at x/L = 0.87 (location C). Since the

percentage mass flow of suction is nearly the same for the two cases, the two cases are

comparable. From Fig. 10(b), for both the cases, the value of cp for suction cases are lower than

19

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

baseline up to the point of application of suction, beyond which, it is higher than baseline.

However, the magnitude of change in cp is higher for case 1 compared to case 2. This difference

can also be seen in Fig. 11, where the values of cf* are greater for case 1, even after location C.

This suggests the effectiveness of applying suction before the separation occurs. For case 1, the

d
ite
separation seems to have been eliminated as suggested by the large and positive cf*. The

ed
overall performance of the intake for the two cases can be understood from the total pressure

loss coefficient contour plots as shown in Fig. 12(a) and Fig. 12(b). Comparing case 1 with the

py
baseline case, we can see a reduction in the loss coefficient in the region near the upper wall

Co
centerline. Comparing case 2 with the baseline case, the reduction of losses is confined to the

ot
region very near to the upper wall. The circumferential distortion descriptors (Fig. 13 and Fig.
tN
14) for case 1 show reduced distortion intensity and a predominantly reduced distortion extent
ip

across the radius. Case 2 shows significant worsening near the hub (low radius) region and a
r
sc

superior improvement near the tip (maximum radius). The value of DC60 for cases 1 and 2 are
nu

0.33 and 0.37 respectively, compared to the baseline DC60 of 0.40. It is clear that case 1 has
Ma

resulted in a better performing intake compared to baseline and case 2. From the cp, cf* and ω

plots it can be said that application of suction upstream of the separation point results in a
ed

higher cp for a longer distance along the length of the duct, due to the absence of flow
pt

separation. This has resulted in sufficient pressure recovery to improve the flow at the exit. This
ce

has also improved the total pressure distortion characteristics of the intake. Case 2 has shown
Ac

an improvement in a small region near the tip, due to the higher value of cp locally. But due to

the inability to remove separation, the improvement in cp is negligible for a longer length and

hence case 2 performs as the baseline case in the other regions.

20

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

ii. Effect of additional suction ports:

With an understanding of the effect of streamwise location, a variety of suction port

configurations were tried as shown in Table 2. It has to be noted that cases 3, 4, 5 and 6b have

d
ite
an approximately same suction flow rate of about 4.1 % whereas cases 7, 8, 9 and 10 have

ed
higher values due to higher number of ports. All further cases are discussed below:

Case 3: In this case, suction is applied at locations predominantly upstream of the separation

py
point. The effect of adding more ports slightly downstream may be seen by comparing case 1

Co
(location A) and case 3 (location A+B). The improvement in cp and cf* for the latter case can be

ot
observed from Fig. 10(b) and Fig. 11, in the form of higher cp and cf* values. The result of this
tN
improvement can also be seen in the AIP (Fig. 12(c)). This case also shows lower distortion
ip

intensity, especially near the hub. However, this improvement comes at an expense of higher
r
sc

suction mass flow rate.

nu

Case 4: This case considers the possibility of applying suction both upstream and downstream
Ma

of the separation point. This proves to be better than applying suction at an upstream point

alone (case 1) but slightly worse than applying suction at the two upstream locations (case 3),
ed

as seen from the cp and total pressure loss plots (Fig. 12(d)). But the distortion extent and
pt

intensity values at most rings are worse than case 3 but better than case 1.
ce

Case 5: Here, suction is applied predominantly at the two downstream locations, B and C, and
Ac

just through one slot at location A. Compared to case 3 (suction only at the upstream

locations); a noticeable worsening can be seen. This can be expected from the comparison of

cases 1 and 2 in the sub-section “Effect of suction location”. The ω contours (Fig. 12(e)) show a

21

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

high loss region in the mid-radial location. This can be clearly seen in the circumferential

distortion intensity plot Fig. 13, where the relative worsening at mid-radius is high compared to

other radii. The DC60 value for this case is 0.31.

Cases 6A and 6B: In these cases, suction is applied evenly at all the locations, but confined to

d
ite
the ports that are near the upper wall centerline. Two different suction flow rates are

ed
considered in this configuration. Case 6A uses a low suction flow rate of 1.8 %, which clearly

shows lower cp and cf* values, and consequently a slightly poorer total pressure loss (Fig. 12(f)).

py
The circumferential distortion intensities (Fig. 13) show consistent increase for the low flow

Co
rate case, with a higher increase near the hub. The DC60 for this case is 0.32. Case 6B is

ot
comparable to case 4 due to similar suction mass flow rates, and the fact that case 4 also has
tN
suction applied at locations A and C. So far it has been observed that a higher upstream suction
ip

is beneficial whereas downstream suction worsens the flow. Accordingly, compared to case 4,
r
sc

an improvement should be expected in this case. This is clearly seen in the ω contours (Fig.
nu

12(g)). There are only slight improvements in the circumferential distortion intensities leading
Ma

to a DC60 of 0.28. Among the cases that consume low mass-flow rates, case 6A is seen to

provide the best DC60. However, the additional suction mass flow and suction kinetic energy
ed

flux consumed in this configuration for the higher mass flow rate case, namely case 6b, is seen
pt

to provide only a slight improvement of 12.5 % in DC 60, and none in mass-averaged total
ce

pressure loss. The value of maximum total pressure loss is improved, implying that a higher
Ac

mass flow rate of suction is only causing a local improvement in total pressure recovery.

Cases 7A and 7B: In these two cases, suction is applied at all locations, and is equivalent to

cases 6A and 6B, but with wider circumferential extents of suction at locations A and C. These

22

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

two cases show a distinctive behavior in the cp and cf* trends. A consistently high value of wall

pressure for all suction locations after the original separation point indicates that the maximum

amount of separated flow removal is achieved in this configuration. The total pressure loss

contours show relative improvement over each other (Fig. 12(h) and Fig. 12(i)). The DC60 values

d
ite
are 0.29 and 0.25, the latter being the lowest for all the flow configurations considered. Similar

ed
improvements in circumferential distortion plots can be seen. A comparison of the sub-cases

within cases 6 and 7 reveal that additional suction flow rate, where it can be afforded, always

py
improves the performance of the intake. The increase in energy is particularly significant, as the

Co
suction kinetic energy flux ratio increase is proportional to the square of suction velocity.

ot
Case 8: This is identical to case 6B but for the addition of a suction port in location B on the
tN
outer wall centerline. From the cp plot, it appears that this addition has improved pressure
ip

recovery beyond location B. The value of cf* is also high after location B, suggesting better
r
sc

separation removal. The improvement of distortion intensity near the hub can also be seen
nu

clearly. The total pressure loss (Fig. 12(j)) does not show any apparent improvement, although
Ma

the DC60 for this case is 0.26. It is interesting to note that such a low value of DC60 is achieved

in this case with a suction flow rate of 5.1 %, compared to the best performing case, case 7B
ed

with DC60 of 0.25, requiring 6.7 %. This configuration of flow control is evidently better when a
pt

lower distortion is to be ensured notwithstanding a high mass flow rate during critical parts of
ce

the flight regime. A fine-grid measurement was made for this case, and the result is shown in
Ac

Fig. 16. Compared to Fig. 8, the absence of strong secondary flows in the form of counter-

rotating vortex pair is apparent. Also, as a consequence of this, a reduced intensity of total

pressure loss can be observed. The asymmetry can be due to the slight variations in the flow

23

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

conditions between the suction ports, suggesting a sensitivity of the flow conditions to the

accurate placement of such devices.

Case 9: This is very similar to case 5, except that no suction is applied at location A. The total

pressure loss contours (Fig. 12(k)) show a further worsening compared to case 5. The distortion

d
ite
intensity plot also shows a remarkably poor value near the hub, along with case 5. The value of

ed
DC60 is the same as that for case 5. This case further emphasizes the benefit of applying suction

upstream, as even a small amount of upstream suction appears to have an effect on the

py
distortion characteristics near the tip.

Co
Case 10: This considers the effect of applying suction through all upstream ports. Compared to

ot
case 3, this has suction via one additional port at location B. Compared to case 3; a proportional
tN
increase in cp as well as reduction in distortion intensities in the region near the hub can be
ip

observed. There is no perceptible improvement in the total pressure loss (Fig. 12(l)), although
r
sc

DC60 shows a slightly better value of 0.26. Compared to Case 7B, the reduction is DC60 is
nu

better in this case even while consuming 1.7 % less amount of mass flow. Due to the smaller
Ma

mass flow rate, the kinetic energy flux is also 33.3 % lesser. Considering this has the least value

of DC60, this case may be considered as the best performing one among the higher mass flow
ed

rate cases. This case shows that even a marginal increase in the amount of suction upstream
pt

can provide a noticeable benefit in the duct performance.

ce

The values of the average and maximum total pressure loss coefficients and the DC60 values
Ac

for all cases are shown in Fig. 17 (a), (b) and (c) respectively. The values are also tabulated in

Table 3. For cases 1, 7A, 7B, 9 and 10, it can be seen that the average ω is higher than the

baseline case. With the exception of case 10, all the others include suction at downstream

24

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

locations, which explain the worsening. The maximum value of ω is seen to be reduced in all

the flow control cases, except the downstream suction cases that still show somewhat large

values. The DC60 values, which are the better indicators for distortion, show values that are in

line with the expectations from the suction locations and flow rates.

d
ite
CONCLUSIONS

ed
Experiments were conducted on a serpentine intake that is typical for a stealth aircraft. The

py
deficiency in performance was quantified in terms of total pressure loss coefficient and

Co
distortion descriptors. Boundary layer separation and secondary flows were identified for the

baseline case from the indicative skin-friction measurements and exit plane measurements

ot
tN
respectively. Subsequently, flow control was applied via three rows of suction slots, two of

which were located slightly upstream and one row downstream of the separation point. While
r ip

the mass-averaged total pressure loss in the AIP was found to improve marginally, with about 4
sc

% for the best performing cases, the maximum local total pressure loss showed an average
nu

reduction of about 9 % when control was applied in all the cases except when suction is applied
Ma

only downstream of the separation point (case 2). The DC60 parameter showed a strong
ed

correlation with suction mass flow rate, where the configurations are identical. But for those
pt

cases that correspond to a configuration where more suction is applied through the
ce

downstream locations than the upstream (cases 2, 4, 5 and 9), the improvement in DC60 is less
Ac

for the suction mass flow required. The instance when moderate amount of suction was

applied spanning a large stream-wise but a small span-wise area (case 6A) showed the best

DC60 improvement of 21 % relative to the baseline, while consuming a mass flow of 1.8 %.

Among the higher range of suction mass flow rate cases, the case with maximum upstream

25

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

ports utilized (case 10) showed the best improvement in DC60 of 35 % while consuming 4.9 %

of mass flow and 2.4 % of flow energy.

From this study, it was observed that even a slightly downstream point of application of

suction showed a worsened performance in terms of total pressure recovery and

d
ite
circumferential distortion. Different suction configurations yielded notably different results. The

ed
general trend was that a higher number of ports distributed upstream of the separation point

results in a better performing duct. The circumferential total pressure distortion plots showed

py
that the low flow rate can improve the flow as much as the high flow rate suction case near the

Co
tip. But near the hub, the high suction case was appreciably better.

ACKNOWLEDGMENT
ot
tN
This project was sponsored by the Aeronautical Development Agency, Bengaluru, India, and
r ip

the support received is gratefully acknowledged.

sc
nu
Ma
ed
pt
ce
Ac

26

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

NOMENCLATURE

AIP Aerodynamic Interface Plane

ARP Aerospace Recommended Practice

cp wall static pressure coefficient

d
ite
cf* indicative skin friction coefficient

ed
CDEk distortion extent of the k-th ring

CDIk distortion intensity of the k-th ring

py
D diameter of the exit plane of the intake

Co
ω total pressure loss coefficient

ot
k index of the radial location starting from the hub (k = 1, 2… 5)
tN
L rectilinear length of the duct

q dynamic pressure
ip

p static pressure
r
sc

p0 total pressure
nu

P exit total pressure referenced with the inlet freestream static pressure
Ma

RDIk radial distortion intensity of the k-th ring

x rectilinear distance along the duct from its inlet plane

ed

Subscripts
pt

1 duct inlet
ce

2 duct exit
Ac

AV averaged over points along a ring (constant radius)

AV-LOW averaged over points whose values are less than that of the ring average

FAV averaged over all points in the AIP

27

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

p Preston tube

s wall static tube

REFERENCES

[1] Wong, S. K., Riseborough, E., Duff, G. and Chan, K. K., “Radar Cross-Section Measurements

d
ite
of a Full-Scale Aircraft Duct/Engine Structure,” IEEE Transactions on Antennas and

ed
Propagation, Vol. 54, No. 8, pp. 2436–2441, 2006.

py
http://dx.doi.org/10.1109/tap.2006.879223

Co
[2] Gerhold M. N., “Aerodynamic performance tests of S-curved diffusers incorporating

bypass cavities,” Hawker Siddeley Aviation Int. Rep., 1977, HSA-KGT-B-GEN-O0655.

ot
[3] Squire, H. B., and Winter, K. G., “The Secondary Flow in a Cascade of Airfoils in a
tN
Nonuniform Stream,” Journal of the Aeronautical Sciences (Institute of the Aeronautical
ip

Sciences), Vol. 18, No. 4, pp. 271–277, Apr. 1951.

r
sc

http://dx.doi.org/10.2514/8.1925
nu

[4] Hawthorne, W. R., “Secondary Circulation in Fluid Flow,” Proceedings of the Royal Society
Ma

A: Mathematical, Physical and Engineering Sciences, Vol. 206, Vo. 1086, pp. 374–387, May

1951.
ed

http://dx.doi.org/10.1098/rspa.1951.0076
pt

[5] Bansod, P., Bradshaw, P. “Flow in S-shaped ducts,” Aeronautical Quarterly, Vol. 23, No.
ce

131, May 1972.

Ac

[6] Taylor, A. M. K. P., Whitelaw, J. H., and Yianneskis, M., “Developing flow in S-shaped ducts

1: Square cross-section duct,” Final Report Imperial Coll. of Science and Technology,

London (England), 1982.

28

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

[7] Gad-el Hak, M., and Bushnell, D.M., Separation control: Review, Journal of Fluids

Engineering, Vol. 113, No. 12, pp. 5–30, 1991.

http://dx.doi.org/10.1115/1.2926497

[8] Ball, W. H., “Tests of wall blowing concepts for diffuser boundary layer control,” AIAA

d
ite
20th Joint Propulsion Conference, 1984.

ed
http://dx.doi.org/10.2514/6.1984-1276

[9] Ball, W. H., “Tests of wall suction and blowing in highly offset diffusers,” Journal of

py
Aircraft, Vol. 22, No. 3, pp. 161–167, Mar. 1985.

Co
http://dx.doi.org/10.2514/3.45102

ot
[10] Johnston, J. P., and Nishi, M., Vortex generator jets - Means for flow separation control,
tN
AIAA Journal, Vol. 28, No. 6, pp. 989–994, Jun. 1990.
ip

http://dx.doi.org/10.2514/3.25155
r
sc

[11] Sullerey, R. K., Mishra, S., and Pradeep, A. M., “Application of Boundary Layer Fences and
nu

Vortex Generators in Improving Performance of S-Duct Diffusers,” Journal of Fluids

Ma

Engineering, Vol. 124, No. 1, pp. 136, 2002.

http://dx.doi.org/10.1115/1.1436096
ed

[12] Tani, I., “Boundary-layer transition,” Annual Review of Fluid Mechanics, Vol. 1, No. 1, pp.
pt

169–196, Jan. 1969.

ce

http://dx.doi.org/10.1146/annurev.fl.01.010169.001125
Ac

[13] Reichert, B., and Wendt, B., Improving diffusing S-duct performance by secondary flow

control, AIAA 32nd Aerospace Sciences Meeting and Exhibit, Jan. 1994.

http://dx.doi.org/10.2514/6.1994-365

29

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

[14] Prandtl, L., “On Motion of Fluids with Very Little Friction,” Verhandlungen des dritten

internationalen Mathematiker-Kongresses in Heidelberg, 1904. English trans. in: Ackroyd,

J. A. K., Axcell, B.P., Ruban, A.I. (Eds.), “Early Developments of Modern Aerodynamics,”

Butterworth-Heinemann, Oxford, UK, 2001, pp. 77–87.

d
ite
http://dx.doi.org/10.1017/s002211200222337x

ed
[15] Antonia, R. A., Zhu, Y., and Sokolov, M., “Effect of concentrated wall suction on a

turbulent boundary layer,” Physics of Fluids, Vol. 7, No. 10, p. 2465, 1995.

py
http://dx.doi.org/10.1063/1.868690

Co
[16] Kirk, A. M., “Active flow control in an advanced serpentine jet engine inlet duct,” M.S.

ot
Dissertation, Texas A&M University, TX, 2006.
tN
[17] Kirk, A. M., Gargoloff, J. I., Rediniotis, O. K., and Cizmas, P. G. A., Numerical and
ip

experimental investigation of a serpentine inlet duct, International Journal of

r
sc

Computational Fluid Dynamics, Vol. 23, No. 3, pp. 245–258, Mar. 2009.
nu

http://dx.doi.org/10.1080/10618560902835558
Ma

[18] Reid, C., “The Response of Axial Flow Compressors to Intake Flow Distortion,” American

Society of Mechanical Engineers 1969 Gas Turbine Conference and Products Show, Mar.
ed

1969.
pt

http://dx.doi.org/10.1115/69-gt-29
ce

[19] Keerthi, M. C., Kushari A., Somasundaram, V., “Study of serpentine intake performance
Ac

with and without flow control,” Paper No. ISABE 2013-1424, XXI International Symposium

on Air Breathing Engines, Busan, Korea, Sep. 9-13, 2013.

30

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

[20] Preston, J. H., “The determination of turbulent skin friction by means of pitot tubes,”

Journal of the Royal Aeronautical Society, Vol. 58, pp. 109–121, 1953.

[21] Keerthi, M. C., Kushari, A., “Effectiveness of vortex generator jets and wall suction on

separated flows in serpentine-duct diffuser,” Aerospace Science and Technology, Vol. 34,

d
ite
pp. 12-19, Feb. 2014.

ed
http://dx.doi.org/10.1016/j.ast.2014.01.013

[22] Sullerey, R. K., and Pradeep, A. M., “Secondary flow control using vortex generator jets,”

py
Journal of Fluids Engineering, Vol. 126, No. 4, pp. 650–657. 2004.

Co
http://dx.doi.org/10.1115/1.1779661

ot
[23] Pradeep, A. M., and Sullerey, R. K., “Detection of Separation in S-duct Diffusers using
tN
Shear Sensitive Liquid Crystals,” Journal of Visualization, Vol. 7, No. 4, pp. 299-307, 2004.
ip

http://dx.doi.org/10.1007/bf03181534
r
sc

[24] SAE Aerospace Recommended Practice, “Gas turbine engine inlet flow distortion
nu

guidelines,” S–16 Turbine Engine Inlet Flow Distortion Committee, ARP 1420, 1978.
Ma

[25] Goldsmith, E. l., and Seddon, J., “Practical Aerodynamic Intake Design,” Blackwell

Scientific publications, Oxford, 1993, Chap. 2.

ed

[26] Rott, N., “On the viscous core of a line vortex II,” Zeitschrift für angewandte Mathematik
pt

und Physik ZAMP, Vol. 10, No. 1, pp. 73–81, Jan. 1959.
ce

http://dx.doi.org/10.1007/bf01637199
Ac

31

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

Figure Captions List

Fig. 1 (a) Schematic of the wind tunnel (b) Schematic of the serpentine intake

d
diffuser with the transition section (c) Details of the intake with reference

ite
locations labelled

ed
Fig. 2 (a) Photograph of the downstream-end of the intake model with flow
control ports (b) Schematic of the same section showing the arrangement of

py
ports (c) Configuration of suction ports for different cases

Co
Fig. 3 Circumferential distortion intensity and extent definitions (inset: choice of
area for PWORST-60)

ot
Fig. 4 Variation of wall static pressure coefficient along the upper and lower wall
tN
centerlines
Fig. 5 Variation of indicative skin fiction coefficient along the upper and lower wall
ip

centerlines
r
sc

Fig. 6 (a) Total pressure loss coefficient contours, and (b) streamwise velocity
magnitude contours in the AIP for the intake without flow control
nu

Fig. 7 Circumferential distortion intensity and extent for each radial station for the
Ma

baseline case
Fig. 8 Variation of total pressure loss coefficient for the baseline case, with in-
ed

plane streamlines overlaid

Fig. 9 Variation of wall static pressure coefficient along the lower wall centerline
pt

for all the cases

ce

Fig. 10 Variation of wall static pressure coefficient along the upper wall (a) for the
Ac

entire length of the duct (b) expanded for the region of interest between x/c
= 0.6 and 1.0. Legend specifications are identical to those in Fig. 9.
Fig. 11 Variation of indicative skin-friction coefficient along the upper wall
centerline within the region of interest. Legend specifications are identical to

32

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

those in Fig. 9.
Fig. 12 Total pressure loss coefficient contours in the AIP for all flow controlled
cases: (a) Case 1 (b) Case 2 (c) Case 3 (d) Case 4 (e) Case 5 (f) Case 6A (g)
Case 6B (h) Case 7A (i) Case 7B (j) Case 8 (k) Case 9 (l) Case 10
Fig. 13 Circumferential distortion intensity at different radial stations for all the

d
cases. Legend specifications are identical to those in Fig. 9.

ite
Fig. 14 Circumferential distortion extent at different radial stations for all the cases.

ed
Legend specifications are identical to those in Fig. 9.
Fig. 15 Radial distortion intensity at different radial stations for all the cases. Legend

py
specifications are identical to those in Fig. 9.

Co
Fig. 16 Variation of total pressure loss coefficient for the case 8, with in-plane
streamlines overlaid

ot
Fig. 17 The performance of various suction configurations compared using the
tN
parameters (a) average total pressure loss coefficient, ωAV (b) maximum
total pressure loss coefficient, ωMAX, and (c) DC60
r ip
sc
nu
Ma
ed
pt
ce
Ac

33

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

d
ite
ed
py
(a)

Co
ot
tN
r ip
sc
nu
Ma

(b)
ed
pt
ce
Ac

34

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

Upper wall centreline

d
ite
ed
py
Co
Inlet plane

Lower wall centreline ot

tN
r ip
sc

Exit plane
(with SAE-ARP 1420 grid)
nu

(c)
Ma

Fig. 1 (a) Schematic of the wind tunnel (b) Schematic of the serpentine intake diffuser with
ed

the transition section (c) Details of the intake with reference locations labelled
pt
ce
Ac

35

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

d
ite
ed
py
Co
(a)

ot
tN
r ip

Location A
sc

Location B
nu

Separation location
Ma

Location C
Exit plane
ed

(b)
pt
ce
Ac

36

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

d
Case 0 Case 1 Case 2

ite
ed
py
Co
Case 3 Case 4 Case 5

ot
tN
r ip
sc
nu

Case 6 Case 7 Case 8

Ma
ed

Blocked port
Port used for suction
pt
ce

Case 9 Case 10
Ac

(c)
Fig. 2 (a) Photograph of the downstream-end of the intake model with suction ports for flow
control (b) Schematic of the same section showing the arrangement of ports (c) Configuration
of suction ports for different cases

37

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

700

Total pressure at the selected ring (a.u.)

600
PAV
500 Distortion extent

d
ite
400
PAV-LOW

ed
300 Worst 60

py
200

Co
100

0
ot
tN
0 45 90 135 180 225 270 315
Angle ( )
ip

Fig. 3 Circumferential distortion intensity and extent definitions (inset: choice of area for
r
sc

PWORST-60)
nu
Ma
ed
pt
ce
Ac

38

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

Streamwise variation of c p

0.5

d
ite
ed
0

py
-0.5 Lower wall c p

Co
Upper wall c p
Ideal cp

ot
-1
0 0.2 0.4 0.6 0.8 1
tN
Non-dimensional distance, x/L

Fig. 4 Variation of wall static pressure coefficient along the upper and lower wall centerlines
r ip
sc
nu
Ma
ed
pt
ce
Ac

39

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

2
Lower wall cf*
Streamwise variation of c f* Upper wall cf*
1.5

d
ite
1

ed
0.5

py
0

Co
-0.5
0 0.2 0.4 0.6 0.8 1

ot
Non-dimensional distance, x/L
tN
Fig. 5 Variation of indicative skin fiction coefficient along the upper and lower wall
centerlines
r ip
sc
nu
Ma
ed
pt
ce
Ac

40

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

d
ite
ed
py
Co
(a)
ot
tN
r ip
sc
nu
Ma
ed
pt
ce
Ac

(b)
Fig. 6 (a) Total pressure loss coefficient contours, and (b) streamwise velocity magnitude
contours in the AIP for the intake without flow control

41

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

0.8 90

Circumferential distortion extent (in °)

Circumferential distortion intensity 0.7 85

d
0.6 80

ite
0.5 75

ed
0.4 70

py
0.3 65

Co
0.2 Circ DI 60
Circ DE
0.1 55

ot
0.2 0.4 0.6 0.8 1
Non-dimensional radius, 2r/D
tN
Fig. 7 Circumferential distortion intensity and extent for each radial station for the baseline
case
r ip
sc
nu
Ma
ed
pt
ce
Ac

42

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

d
ite
ed
py
Fig. 8 Variation of total pressure loss coefficient for the baseline case, with in-plane
streamlines overlaid for the upper hemisphere

Co
ot
tN
r ip
sc
nu
Ma
ed
pt
ce
Ac

43

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

0.6

0.4

0.2

d
0 baseline

ite
c along lower wall

1: 11111 + 00 + 0000
-0.2 2: 00000 + 00 + 1111

ed
3: 11111 + 11 + 0000
-0.4 4: 01110 + 00 + 1111

py
5: 00100 + 11 + 1111
-0.6
6A: 01110 + 11 + 0110 (low)

Co
p

-0.8 6B: 01110 + 11 + 0110 (high)

7A: 11111 + 11 + 1111 (low)
-1 7B: 11111 + 11 + 1111 (high)

ot
8: 01110 + 111 + 0110
tN
-1.2 9: 00000 + 11 + 1111
10: 11111 + 111 + 0000
ip

0 0.2 0.4 0.6 0.8 1

Non-dimensional distance, x/L
r
sc

Fig. 9 Variation of wall static pressure coefficient along the lower wall centerline for all the
cases
nu
Ma
ed
pt
ce
Ac

44

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

cp along upper wall 0.6

0.4

d
ite
0.2

ed
0

py
-0.2

Co
-0.4

0 0.2 0.4 0.6 0.8 1

ot
Non-dimensional distance, x/L
tN
(a)
ip
r

0.6
sc
cp along upper wall

0.4
nu

0.2
Ma

0
ed

-0.2
pt

-0.4
ce

0.6 0.7 0.8 0.9 1

Ac

Non-dimensional distance, x/L

(b)
Fig. 10 Variation of wall static pressure coefficient along the upper wall (a) for the entire
length of the duct (b) expanded for the region of interest between x/c = 0.6 and 1.0. Legend
specifications are identical to those in Fig. 9.

45

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

cf* along upper wall 0.5

0.4

d
ite
0.3

ed
0.2

py
0.1

Co
0

-0.1
0.6 0.7 0.8 0.9 1

ot
Non-dimensional distance, x/L
tN
Fig. 11 Variation of indicative skin-friction coefficient along the upper wall centerline within
the region of interest. Legend specifications are identical to those in Fig. 9.
ip
r
sc
nu
Ma
ed
pt
ce
Ac

46

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

d
ite
ed
py
Co
ot
tN
r ip
sc
nu
Ma
ed
pt
ce
Ac

47

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

Fig. 12 Total pressure loss coefficient contours in the AIP for all flow controlled cases: (a) Case
1 (b) Case 2 (c) Case 3 (d) Case 4 (e) Case 5 (f) Case 6A (g) Case 6B (h) Case 7A (i) Case 7B (j)
Case 8 (k) Case 9 (l) Case 10

d
ite
ed
py
Co
ot
tN
r ip
sc
nu
Ma
ed
pt
ce
Ac

48

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

0.7
Circumferential distortion intensity
0.6

d
0.5

ite
0.4

ed
0.3

py
0.2

Co
0.1

ot
0
0.2 0.4 0.6 0.8 1
tN
Non-dimensional radius, 2r/D
Fig. 13 Circumferential distortion intensity at different radial stations for all the cases. Legend
ip

specifications are identical to those in Fig. 9.

r
sc
nu
Ma
ed
pt
ce
Ac

49

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

Circumferential distortion extent (in °)

100

90

80

d
ite
70

ed
60

py
50

Co
40

30
0.2 0.4 0.6 0.8 1

ot
Non-dimensional radius, 2r/D
tN
Fig. 14 Circumferential distortion extent at different radial stations for all the cases. Legend
specifications are identical to those in Fig. 9.
r ip
sc
nu
Ma
ed
pt
ce
Ac

50

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

0.3

Radial distortion intensity

0.2

d
ite
0.1

ed
0

py
-0.1

Co
-0.2
0.2 0.4 0.6 0.8 1

ot
Non-dimensional radius, 2r/D
tN
Fig. 15 Radial distortion intensity at different radial stations for all the cases. Legend
specifications are identical to those in Fig. 9.
r ip
sc
nu
Ma
ed
pt
ce
Ac

51

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

d
ite
ed
py
Co
Fig. 16 Variation of total pressure loss coefficient for the case 8, with in-plane streamlines

ot
overlaid tN
r ip
sc
nu
Ma
ed
pt
ce
Ac

52

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

0.29 8
 AV
% suction mass
0.28 6

% suction mass
d
ite
AV

0.27 4


ed
py
0.26 2

Co
0.25 0
0 1 2 3 4 5 6a 6b 7a 7b 8 9 10

ot
Case no.
(a)
tN
0.8 8
 MAX
ip

% suction mass
r
sc

0.75 6

% suction mass
nu
MAX

0.7 4
Ma


ed

0.65 2
pt

0.6 0
ce

0 1 2 3 4 5 6a 6b 7a 7b 8 9 10
Case no.
Ac

(b)

53

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

0.4 8
DC60
0.38 % suction mass
0.36 6

% suction mass
0.34

d
DC60

ite
0.32 4

0.3

ed
0.28 2

py
0.26

Co
0.24 0
0 1 2 3 4 5 6a 6b 7a 7b 8 9 10
Case no.

ot
(c) tN
Fig. 17 The performance of various suction configurations compared using the parameters (a)
average total pressure loss coefficient, ωAV (b) maximum total pressure loss coefficient, ωMAX,
and (c) DC60
r ip
sc
nu
Ma
ed
pt
ce
Ac

54

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

Tables caption list

Table 1 Absolute uncertainty of relevant parameters

Table 2 Flow control attributes for all the test cases

d
Table 3 Summary of performance parameters for all the test cases

ite
ed
py
Co
ot
tN
r ip
sc
nu
Ma
ed
pt
ce
Ac

55

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

Table 1 Absolute uncertainty of relavant parameters

Absolute uncertainty at
Parameter
95% confidence

d
ite
Static pressure coefficient, cp 0.03

ed
Indicative skin friction coefficient, cf* 0.03

Suction mass flow percentage 0.05

py
Total pressure loss coefficient, ω 0.02

Co
DC60 0.02

ot
Circumferential distortion intensity 0.02
tN
Circumferential distortion extent 3°
ip

Radial distortion intensity 0.02

r
sc
nu
Ma
ed
pt
ce
Ac

56

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

Table 2 Flow control attributes for all the test cases

No. of ports used at Suction
locations A, B and C Percentage Suction kinetic
Case
Case name suction velocity energy
no.
A B C mass flow ratio flux

d
ratio

ite
0 baseline 0 0 0 0.0 0.0 0.0
1 11111 + 00 + 0000 5 0 0 1.3 2.9 0.1

ed
2 00000 + 00 + 1111 0 0 4 1.1 3.1 0.1

py
3 11111 + 11 + 0000 5 2 0 4.1 6.7 1.9

Co
4 01110 + 00 + 1111 3 0 4 4.1 6.6 1.8
5 00100 + 11 + 1111 1 2 4 4.2 6.8 1.9

ot
6A 01110 + 11 + 0110 3 2 tN 2 1.8 2.9 0.1
6B 01110 + 11 + 0110 3 2 2 4.1 6.7 1.9
7A 11111 + 11 + 1111 5 2 4 3.2 3.3 0.3
ip

7B 11111 + 11 + 1111 5 2 4 6.7 6.9 3.2

r
sc

8 01110 + 111 + 0110 3 3 2 5.1 7.2 2.6

9 00000 + 11 + 1111 0 2 4 3.5 6.7 1.6
nu

10 11111 + 111 + 0000 5 3 0 4.9 7.0 2.4

Ma
ed
pt
ce
Ac

57

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-

Journal of Fluids Engineering. Received March 28, 2016;
Accepted manuscript posted May 24, 2017. doi:10.1115/1.4036827
Copyright (c) 2017 by ASME

Table 3 Summary of performance parameters for all the test cases

Percentage
Case Mass- Maximum
Case name suction DC60
no. averaged ω ω
mass flow
0 baseline 0.0 0.26 0.76 0.40

d
1 11111 + 00 + 0000 1.3 0.27 0.70 0.33

ite
2 00000 + 00 + 1111 1.1 0.26 0.77 0.37

ed
3 11111 + 11 + 0000 4.1 0.26 0.67 0.27

py
4 01110 + 00 + 1111 4.1 0.25 0.71 0.30
5 00100 + 11 + 1111 4.2 0.26 0.74 0.31

Co
6A 01110 + 11 + 0110 1.8 0.26 0.71 0.32
6B 01110 + 11 + 0110 4.1 0.26 0.65 0.28
7A 11111 + 11 + 1111 3.2 0.27
ot 0.69 0.29
tN
7B 11111 + 11 + 1111 6.7 0.27 0.65 0.25
ip

8 01110 + 111 + 0110 5.1 0.26 0.66 0.26

r

9 00000 + 11 + 1111 3.5 0.27 0.74 0.31

sc

10 11111 + 111 + 0000 4.9 0.27 0.67 0.26

nu
Ma
ed
pt
ce
Ac

58

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/0/ on 06/04/2017 Terms of Use: http://www.asme.org/about-