Content uploaded by Emmanuel Benard
Author content
All content in this area was uploaded by Emmanuel Benard
Content may be subject to copyright.
NOMENCLATURE
Symbols
Aarea
Ccost
CDbody drag coefficient
CLbody lift coeffcient
Cxdistance between the holes in piccolo tube
Djet diameter
Ddiameter of holes in piccolo tube
Flearning factor
ffrequency
Hconvection coefficient
Hjet height
Kthermal conductivity of air at film temperature
Kcoefficent in cost estimation
M Mach number
maverage mass flow rate, (ρjAjVj)
Nu Nusselt number
qinternal heat flux
PrPrandtl number
Re Reynolds number (ρVD/µ)
ABSTRACT
Customer requirements and vision in aerospace dictate that the next
generation of civil transport aircraft should have a strong emphasis
on increased safety, reduced environmental impact and reduced cost
without sacrificing performance. In this context, the School of
Mechanical and Aerospace Engineering at the Queen’s University of
Belfast and Bombardier have, in recent years, been conducting
research into some of the key aerodynamic technologies for the next
generation of aircraft engine nacelles. Investigations have been
performed into anti-icing technology, efficient thrust reversal, engine
fire zone safety, life cycle cost and integration of the foregoing with
other considerations in engine and aircraft design. A unique corre-
lation for heat transfer in an anti-icing system has been developed.
The effect of normal vibration on heat transfer in such systems has
been found to be negligible. It has been shown that carefully
designed natural blockage thrust reversers without a cascade can
reduce aircraft weight with only a small sacrifice in the reversed
thrust. A good understanding of the pressure relief doors and
techniques to improve the performance of such doors have been
developed. Trade off studies between aerodynamics, manufacturing
and assembly of engine nacelles have shown the potential for a
significant reduction in life cycle cost.
THE AERONAUTICAL JOURNAL MAY 2006 265
Paper No. 2976. Manuscript received 22 April 2005, revised paper received 2 September 2005, accepted 26 February 2006.
Key aerodynamic technologies for
aircraft engine nacelles
S. Raghunathan, E. Benard, J. K. Watterson, R. K. Cooper, R. Curran, M. Price, H. Yao, R. Devine and B. Crawford
Centre of Excellence for Integrated Aircraft Technologies
School of Mechanical and Aerospace Engineering
Queen’s University Belfast
Belfast, UK
D. Riordan, A. Linton, J. Richardson and J. Tweedie
Bombardier
Belfast, UK
.
.
decreased environmental impact and reduced cost, without at the
same time sacrificing performance. This may be seen in some of
the designs that have either been proposed or actually come to
fruition. Perhaps the prime example is the debate between larger
aircraft and faster aircraft epitomised by the Airbus A380 and
Boeing Sonic Cruiser designs. Only the former has progressed
beyond the conceptual design phase. Other antitheses exist:
transonic flight (M0⋅85) with a traditional swept wing versus high
subsonic flight (M0⋅70) with a lighter, cheaper unswept wing. The
latter entails a cruise time penalty but brings environmental
benefits and cost savings. In the search for a viable very large
aircraft design to supersede the Airbus A380, it is possible that
blended wing body designs may challenge the supremacy of the
traditional configuration that has dominated flight for almost ten
decades. Incremental steps, syntheses or paradigm shifts in aircraft
design will be facilitated by technological developments in
materials, manufacturing, structures, propulsion and aerodynamics.
The square-cube rule presents a serious challenge to the devel-
opment of very large aircraft, but the development of new
materials and fabrication processes will open up new possibilities
in the same way that GLARE has been an integral element of the
design of the A380. Decades of aerodynamic research in drag
reduction (passive shockwave/boundary layer control, active
suction, LEBUs, riblets, MEMS) have yet to provide truly signif-
icant results, and need to be more fully integrated into the whole
aircraft design process. Environmental considerations may drive
aircraft operating altitudes and Mach numbers down, and we may
yet see the use of turboprops on very large aircraft. Meanwhile,
Boeing is developing the 787, which will operate more systems
electrically, with significant weight savings.
A nacelle (Fig. 1) is a key system of aircraft and typically
consists of an inlet cowl, fan cowl, thrust reverser, core cowl and
primary exhaust nozzle. Its primary purpose is to provide a smooth
aerodynamic fairing for the power plant, while also ensuring a
smooth airflow into the engine. A nacelle also houses several
safety and environmental protection systems. In this context, the
School of Mechanical and Aerospace Engineering at the Queen’s
University of Belfast (QUB) and Bombardier have, in recent years,
been conducting research into some of the key aerodynamic
technologies and cost analyses for the next generation of aircraft
engine nacelles. Investigations have been performed into anti-icing
technology, efficient thrust reversal, engine fire zone safety, noise
attenuation, integration with the wing and life cycle cost optimi-
sation with other considerations in engine and aircraft design.
Some of the findings from these research activities are reported in
this paper.
2.0 ANTI-ICING(1-34)
For an aircraft to meet safety regulations (expressed as certification
requirements)(1,2) for flight into icing conditions, it must be
protected against the formation of ice. Ice can cause impact
damage to the aircraft structure and systems and/or performance
degradation effects due, for example, to its influence on the wing
aerodynamics. Lift can be reduced dramatically in conjunction
with very large increase of the drag. The requirements for wing and
engine installation on an aircraft, as set out in the relevant US
Federal Aviation Regulations (FAR) and European Aviation Safety
Agency (EASA) Certification Standards (CS), are typically met by
providing the forward lip skin of each vital surface with a hot air
anti-icing system (Fig. 2). A hot air system ensures that even a
minimum thickness of ice can be prevented from forming on the
external surface; a performance level that cannot be reached with a
pneumatic boot system. Other systems present some limitations in
the form of performance loss or energy cost(3).
The general principle of the hot air system is that it takes high
temperature air from the engine compressor and directs it forward
Sarc length
s jet width
St Strouhal number (fD/V)
Tair temperature
Uvelocity in the boundary layer at y
Uevelocity at the edge of the boundary layer
Vvelocity
VR velocity ratio
y distance in the boundary layer normal to the surface
y+wall units (uτy/ν)
δboundary-layer thickness
µdynamic viscosity
ρdensity
Subscripts
aaverage
Amr amortisation
Anti anti icing
Asm assembly
Fab fabrication
Iinternal
Jjet exit
Mat material
Misc miscellaneous cost
NC nose cowl
Sup support
ssurface area
SIZE
wwall
Abbreviations:
ACARE
BPR engine by pass ratio
CANICE simulation code for ice accretion
CS certification standards
DEMARDOC design for manufacture of aerodynamic profiles to
reduce aircraft direct operating cost
DFMA design for manufacturing and assembly
DFR discharge flow ratio
DOC direct operating cost
EASA european aviation safety agency
EBU engine built unit
FAA federal aviation administration
FAR federal aviation regulations
GLARE glass-reinforced fibre metal laminate
LEBU large eddy breakup device
MEMS micro electro-mechanical machine
pdf probability density function
PRD pressure relief door
RCM rapid cost model
RPR ram pressure recovery
sfc specific fuel consumption
SOAMATAS simultaneous optimisation of aerodynamic and
manufacture tolerances to reduce life cycle cost
VG vortex generator
1.0 INTRODUCTION
Aerospace design is driven by customer requirements, where the
customer may be thought of in the widest context as including
aircraft operators, fare paying passengers, national transport
systems and international treaty organisations. Current trends
suggest that the design of the next generation of civil transport
aircraft should place a strong emphasis on increased safety,
266 THE AERONAUTICAL JOURNAL MAY 2006
onto the inner lip skin surface to evaporate impinging water or melt
accreted ice on the outer side of the skin, limiting finally the
formation of ice. A preferred method of directing the hot air onto
the inside skin of the forward lip is through the use of a piccolo
tube (Fig. 2), though other methods, having varying degrees of
effectiveness, may be used. The piccolo system utilises the effect
of multiple steady hot jets impinging on the surface and interacting
in various ways, thereby efficiently heating the inner surface.
In an effort to support the objectives of the FAA Icing Plan and
facilitate Bombardier in the certification process, the School of
Mechanical and Aerospace Engineering at QUB in conjunction
with Ecole Polytechnique, Montreal has been involved with the
development of reliable ice accretion and anti-icing prediction
methods. The effect of engine vibration and possibility of
enhancement of heat transfer by pulsejets were also considered in
these studies.
2.1 Heat transfer correlation for piccolo systems, based
on experiments(3-10)
The performance of the piccolo anti-icing system at a given flight
condition depends on several factors. These include the mass flow
rate, temperature drop between the engine compressor and piccolo
tube, the amount of water catch, the impinging limits on the nacelle
surface and the conditions for thermal equilibrium at the nose cowl
surface. The critical aspect of design of an anti-icing system is the
prediction of the heat transfer of the impinging jets from the
piccolo tube.
There is evidence to indicate that for a single jet impinging on a
surface the maximum heat transfer at the stagnation point occurs
on the impingement surface when the distance between the jet and
the plate is approximately equal to the length of the potential core
of the jet (5-7 times the jet diameter) and heat transfer decreases
for distances greater than this length. Experiments(3-10) on hot/cold
air jet impingement on flat surfaces address key issues such as the
effects of jet diameter, orientation, Reynolds number, jet nozzle-to-
nozzle distance and jet nozzle-to-surface spacing on flow and heat
transfer on flat plates. The flat plate arena has been thoroughly
investigated and empirical relations have been developed that
address the key issues mentioned above. However, it is very rare to
find studies that have been dedicated to more complex situations
such as the interaction of an impingement jet and a curved
surface(7,8).
Experiments were conducted in the QUB 1⋅14m × 0⋅85m wind
tunnel of a full scale anti-icing system of an engine nacelle. The
external airflow velocity and temperature were maintained at
40ms–1, and 300K, respectively. The piccolo tube had three rows of
holes. The mean airflow flow temperature and velocity at the jet
exit were typically 600K and 340ms–1, respectively. Tests were
performed for several values of jet hole diameter, the spacing
between the holes and the distance between the jet exit and the
impinging surface. Based on these experiments a unique corre-
lation for impingement heat transfer was developed (Fig. 3). The
correlation is independent of hole diameter, number of holes, and
the distance between the holes and the impingement surface. It
depends mainly on mass flow per unit area and weakly on jet
spacing, expressed as a multiple of hole diameter. (The values
defining the correlation have been withheld for commercial
reasons.)
2.2 Simulation of ice accretion(11-34)
Simulation codes for ice accretion in both two-dimensional(13-19) and
three-dimensional(26) versions (CANICE) have been developed at
the Ecole Polytecnique, Montreal. The development of CANICE
has been geared towards the specific needs of Bombardier.
Presently, a basic model for a hot-air anti-icing system is being
used in CANICE. In this model, hot air from the engine is assumed
RAGHUNATHAN ET AL KEY AERODYNAMIC TECHNOLOGIES FOR AIRCRAFT ENGINE NACELLES 267
Inlet Cowl
Fan Cowl
Thrust
Reverser
Plug
Exhaust
Nozzle
Figure 1. A nacelle.
(a) Piccolo and jets.
(b) Bleed system.
Figure 2. Piccolo anti-Icing system.
drain on the engine power as a result of operating the hot-air anti-
icing system.
The first step towards achieving an optimum design of a hot-air
anti-icing system is the development of a reliable database, experi-
mental and numerical, related to the heat transfer from a single jet or
an array of jets impinging on a curved surface. Recent studies(27, 28) on
the usefulness of the empirical relations developed for hot-air jet
impingement on a flat surface for a highly curved two-dimensional
(2D) surface showed that the flat-plate empirical relations are inade-
quate for predicting heat transfer on curved surfaces. Figures 4 and 5
present some results of the numerical simulation of a hot-air jet
impingement on a concave surface for a 2D case. Here, Resis the
Reynolds number based on the arc length s
s, and Hand Sare the jet
height and width, respectively. As evident from Fig. 5, the most
important conclusion drawn from these studies is that the flat plate
correlations are not reliable enough for predicting heat transfer on a
concave surface. To overcome this limitation, a numerical corre-
lation was developed for the 2D case(28) and is now being used in the
2D version of the CANICE code.
to impinge upon the inner surface of the aerofoil leading edge or
the slat (Fig. 4). The inner region is modelled with a local internal
convection coefficient hanti which is considered to be known a
priori. The heat flux qfrom this region is then evaluated with the
help of the internal hot-air temperature TIand the local wall
temperature Tw
A limitation of this method is that the internal heat flux qor the
convection coefficient hanti and temperature Tiare based on empirical
relations(6) for a hot-air jet impinging on a flat plate. This has been a
commonly accepted practice in studies related to anti-icing system
modelling(13-25). It must be pointed out that this local distribution of
internal heat flux qior the convection coefficient hanti is based purely
on the local distribution on a flat plate and, therefore, does not account
for the curvature of the internal wall region of an aerofoil leading-edge
or the wing slat. Another limitation of such a model is that it fails to
provide an accurate estimate of the hot-air flux requirements and the
268 THE AERONAUTICAL JOURNAL MAY 2006
()
anti i w
qh TT=−
. . . (1)
Figure 3. Unique correlation for heat transfer in anti-Icing system (band c are indices).
Figure 4. Numerical simulation of a hot-air jet impingement inside a slat.
Figure 5. Comparison of the heat flux distribution (q) on the
surface of a slat and a flat plate ReS= 1,757.
.
⋅
⋅
⋅
2.3 Effect of normal vibration on heat transfer(35-50)
Vibration of anti-icing systems can arise from the vibration of the
aircraft structure, possibly due to unsteady flow caused by ice
formation. In spite of the obvious importance of potential effects on
heat transfer on boundary layers, there is still only a limited experi-
mental database available with respect to the effects of vibration on a
turbulent boundary layer.
An early study by Izzo(35) suggests minor changes in mean velocity
profile when compared with a non-vibrating case and that the
excitation of the surface in a fundamental mode only disturbs a small
portion of the energy spectrum rather than resulting in a broad
disturbance of turbulent properties. A possible interpretation of this
would be a strong correlation between the surface motion and the
activity of the high/low velocity streaks and the coherent structures
responsible for turbulence production in the near wall region of an
unperturbed boundary layer(36).
Studies have been conducted into the effects of a spanwise wall
vibration(37-39) on near wall turbulence. These show a decrease in near
wall turbulence due to an interaction between longitudinal vortices
and the near wall high/low velocity streaks.
The influence of longitudinal vibrations, which include travelling
waves, has also been investigated in relation to boundary layer
transition control by compliant walls(40,41). In this case a combination
of stream-wise and longitudinal wall motions produced a very
complex near wall turbulence structure. However, no significant
effect on skin friction was observed. From an aero-acoustic
viewpoint, several similar investigations have been carried out
ranging from studies of the acoustic near field from a jet engine to
turbulent boundary-layer excitation and induced skin vibrations(42-45).
It is understood from these investigations that the structure of the
turbulent boundary-layer may depend on the directions of vibration
and it is important to conduct controlled experiments to fully under-
stand the effect of vibration in a particular direction on the structure
of the boundary layer. This implies a strict control regime over any
form of surface motion in order to isolate the specific flow effect due
to an individual vibration mode.
The purpose of the investigation performed at QUB was to assess
the specific effect of a wall normal vibration on a turbulent boundary
layer while minimising longitudinal, spanwise and travelling wave
motions. In order to assess the effect of vibration excitation in a
single direction on turbulent boundary layers and heat transfer,
experiments were conducted with a zero pressure gradient turbulent
boundary layer on a flat plate in a low speed wind tunnel (Fig. 6)(50).
The turbulent boundary layer over the relatively short flat plate
(690mm) was triggered by sand paper placed over the first 150mm
of the plate. Hot wire anemometry was used to survey mean and
fluctuating stream-wise velocity components and the post-processing
of data involved the use of a triple decomposition algorithm. The
unsteady component represents the input disturbance frequency
imposed on the plate through a shaker and the fluctuating
component. After the input signal has been decomposed a phase
averaged cycle can be rebuilt which allows velocity profiles, turbu-
lence profiles and frequency information to be extracted for given
phase locations.
Figures 7 and 8 show typical results for velocity and turbulence
profiles for the above case. The results shown here are for plate
vibrating frequencies of 30Hz and 80Hz and at an amplitude of
vibration of 0⋅76 times the boundary-layer thickness. The velocity
profiles (Fig. 7) over a phase averaged cycle ‘converge’ into one
profile and this is not significantly different from a typical turbulent
boundary-layer profile(46-48). Identical behaviour is also observed for
the turbulence intensity data (Fig. 8). These results suggest that the
effect of normal vibration is negligible.
Figure 9 shows a typical power spectrum for the plate, derived
from an accelerometer mounted beneath the measurement position,
and velocity spectra from boundary layer measurements. The results
shown here are for a plate vibration of 80Hz. The accelerometer
measurements on the plate show the excitation frequency of 80Hz
RAGHUNATHAN ET AL KEY AERODYNAMIC TECHNOLOGIES FOR AIRCRAFT ENGINE NACELLES 269
Figure 6. Test set up for vibrations.
Normalised Velocity Profile Comparisons
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
00.20.40.60.81
U / Ue
y /
⎝⎝
Exp Data 80 Hz
1/7 th Powe r Law
Normalised Velocity Profile Comparisons
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
00.20.40.60.81
U / Ue
y /
Exp Data 30 Hz
1/7 th Power La w
Figure 7. Normalised Mean Velocity Profile Plots for a 30Hz and
80Hz sinusoidal vibration with 2mm pk-pk amplitude, over one rebuilt
phase averaged cycle. Both figures contain 300 separate boundary
layer profiles representing 1⋅2° intervals from 0° to 360°.
270 THE AERONAUTICAL JOURNAL MAY 2006
Figures 8. Turbulence intensity plots for a 30Hz and 80Hz sinusoidal vibration with 2mm pk-pk amplitude, over one rebuilt
phase averaged cycle. Both figures contain 300 separate turbulence profiles representing 1⋅2° intervals from 0° to 360°
Figure 9. Boundary-layer response to plate vibrations. Figure 10. Schematic diagram of the
experimental test set-up form pulse jets.
Figure 11. Velocity profile for pulse frequencies of 10 and 20Hz.
Azevedo et al(56) investigated impingement heat transfer using a
rotating cylinder valve for a range of pulse frequency. The results
showed that heat transfer degraded for all frequencies. In their exper-
iments, the velocity profile of the pulse jet shows the existence of a
two-peak region for every flow cycle. This results in disturbance to
the pulse flow and affects the flow structure and heat transfer. The
dependence of pulse characteristics on convective heat transfer was
discussed by Mladin and Zumbrunnen(55).
Farrington and Claunch(57) carried out a test to determine the
influence of flow pulsations on the flow structures of a planar jet at
Re = 7,200 and 0 < St < 0⋅324. The results of the test were captured
using infrared imaging and smoke-wire visualisation. They
concluded that for pulsating jets, the vortices were larger than the
steady jet and occurred closer to the nozzle. These larger vortices
resulted in an increased entrainment and led to a wider angle of the
potential core. Jets with large amplitude of pulsations entrained
surrounding fluid more rapidly and decayed more quickly than
steady jets. An increase in turbulence intensity can be associated
with the pulse decay.
Thus results of investigations to date on the effect of pulsing a jet
on heat transfer are conflicting. When compared with steady jets,
some of the work shows that a pulse jet increases the heat transfer
whereas other tests show that the pulse jets reduce heat transfer. In
order to understand further the effect of pulse jets on heat transfer
and to address the contradictions in results to date, experiments were
performed(59) for both steady jets and pulsating jets at QUB.
Figure 10 shows the schematic diagram of the experimental test
set-up. A vortex flow meter is placed just downstream of the
pressure regulator and is used to measure the mass flow of the air jet
impinging on the surface of the lip-skin. A Secomak model 15/2 air
heater is used to heat the air jet. The maximum air temperature was
570K. The pulse air jet is generated using a rotating cylinder valve
driven by an electric motor controlled by an electronic motor
controller. Heat transfer coefficients were calculated from the value
of temperature drop between the jet exit air and the lip-skin surface.
Time-averaged velocity of the centreline jet exit air close to the
nozzle was measured with a calibrated hot wire anemometer. A
sampling frequency of 1kHz was considered adequate to capture the
time variations.
The heat flux of the heated air jet impinging on the plate was
measured using a heat flow and integral thermocouple sensor from
RdF Corporation. The values of heat flux and plate temperature for
the stagnation point and for local measurements at different radial
positions were monitored and recorded by the data acquisition
system. Local heat transfer measurements were recorded at radial
distances from one to six nozzle diameters. The local Nusselt
number was calculated using Equation (3):
. . . (3)
Where q" is the stagnation point heat flux measured by the sensor, D
is the nozzle diameter, kis the thermal conductivity of the air jet
evaluated at film temperature, Tjis the temperature of the hot air jet
and Twis the temperature of the plate at the stagnation point. The
average Nusselt number based on the local temperature difference
was calculated by numerically integrating the heat flux measurement
over the impingement area. The uncertainty in the heat flux
measurement is estimated at 5%.
The tests were conducted(59) in the pulse frequency range of 10 to
80Hz and Reynolds number based on the jet diameter in the range
16,000 to 32,000. The nozzle to plate spacing, x/Dwas fixed at 4.
The duty cycle of the pulse airflow, that is the fraction of time in a
pulse cycle during which there was jet flow, was 0⋅33.
Figure 11 shows the velocity profile at pulsation frequencies of 10
and 20Hz for a jet nozzle diameter of 20mm. The ON part of the
pulsation duty cycle has a velocity variation similar to a half-cycle of
a sinusoid and the velocity at the nozzle exit is close to zero during
()
jw
qD
Nu TTk
′′
=−
and several harmonics. The boundary layer results show that the
disturbances propagate through at the plate excitation frequency and
the second harmonic only. These plots highlight the fact that the
boundary layer exhibits a very narrowband response to the vibration,
located solely around the input frequency and its associated second
harmonic.
The results also showed that the displacement thickness exhibits
some phase shift, suggesting that the boundary layer does not react
instantaneously to the sudden change in wall position; instead there is
a delay. When probability density distributions are considered no
significant change between the vibrating cases and the stationary one
was observed. Finally it can be deduced that within the range of tests
conducted, although the boundary layer responds to normal vibration,
the effects on the velocity and turbulence profiles are very small. It
could be argued therefore, that the effect of normal vibration on heat
transfer is negligible and normal vibration of an engine is not a key
design parameter of an aircraft anti-icing system. However, further
study is required, since these observations from laboratory experi-
ments may not be consistent with data from aircraft flight testing.
2.4 Effect of pulsing the jets on heat transfer(51-59)
Heat transfer in pulsating jet flows has been the subject of renewed
interest in recent years in an effort to enhance the performance of the
many industrial applications that use hot or cold jets. The purpose of
the investigations at QUB was to assess the current understanding of
the effect of pulsing a hot jet on impingement heat transfer with
reference to anti-icing applications.
Test carried out by Nevins and Ball(51) on heat transfer between a
flat plate and a pulsating jet showed that no significant heat transfer
enhancement was obtained by using a pulsed air jet. The test was
conducted at 1,200 < ReG < 120,000, 10-4 < St < 10-2, and nozzle to
plate spacing from 8 to 32 nozzle diameters. The average Reynolds
number ReGwas calculated based on the average mass flow rate for
each test frequency, and corresponding Strouhal number St deter-
mined using Equation (2):
. . . (2)
where, mis the average mass flow rate calculated from the jet exit
velocity profile, Dis the nozzle diameter, Ais the nozzle area, µis
dynamic viscosity and ρis density of air at supply air temperature,
and ƒis pulsation frequency.
Nevins and Ball(51) did not document the extent of secondary flow
structures in their experiments. Further, the experiment was studied
at a very low Strouhal number which might have affected the ability
to demonstrate pulsed flow heat transfer enhancement.
Kataoka and Suguro(52) show that stagnation point heat transfer for
axisymmetric submerged jets is enhanced by the impingement of
large-scale structures on the boundary layer, such as vortex rings
which occurred in pulse flow. Further tests carried out by Sailor et
al(53) on the effect of duty cycle variation on heat transfer
enhancement for an impinging air jet showed significant heat
transfer enhancement.
Mladin and Zumbrunnen(54) investigated theoretically the influence
of pulse shape, frequency and amplitude on instantaneous and time-
averaged convective heat transfer in a planar stagnation region using
a detailed boundary layer model. They reported that there exists a
threshold Strouhal number, St > 0⋅26 below which no significant
heat transfer enhancement was obtained. Results obtained by
Zumbrunnen and Aziz(54) on the effect of flow intermittency on
convective heat transfer to a planar water jet impinging on a constant
heat flux surface reinforces this finding. This experiment carried out
at St > 0⋅26 found that local Nusselt number increases by up to
100%. However, Sailor et al(53) used Strouhal numbers between
0⋅009 and 0⋅042 and still recorded significant enhancement in
stagnation point heat transfer for pulse flow.
fD A
St m
ρ
=
Re
G
mD
A
=µ
RAGHUNATHAN ET AL KEY AERODYNAMIC TECHNOLOGIES FOR AIRCRAFT ENGINE NACELLES 271
⋅
272 THE AERONAUTICAL JOURNAL MAY 2006
Figure 12. Variation of local Nusselt numbers with radial distance at frequencies of 10 and 20Hz for ReG = 16,000, 23,300 and 32,000.
Figure 13. Variation of stagnation Nusselt numbers with
frequency for Reynolds number of 16,000, 23,300 and 32,000.
Comparison of current results with Azevedo et al.
Figure 14. Variation of Nusselt numbers averaged over
the impingement area with frequency for Reynolds
Number = 16,000, 23,300 and 32,000.
(b) (c) types of thrust reversers.
Figure 15. Thrust reverser.
(b) (c)
(a) Concept of thrust reversal.
transfer at local distances away from the stagnation point resulted in
higher average Nusselt numbers for pulse flow compared to steady
flow. Significant turbulence intensity caused by pulsating the jet
resulted in the increase recorded. The degradation in heat transfer at
the stagnation point is believed to be due to small turbulent inten-
sities of the pulse flow at this position.
When compared with a steady jet system a pulse jet anti-icing
system could produce an enhancement in heat transfer of 10-20% or
reduce the bleed air requirements by 10-20%.
3.0 THRUST REVERSER(60-77)
Typically on modern aircraft the thrust reverser is built into the
engine nacelle (Fig. 15). A thrust reverser uses the power of a jet
engine as a deceleration force by reversing the direction of exit
airflow, which generates forward thrust. Fig. 15(ashows a section
through a jet engine with a typical cold stream thrust reverser. Some
of the thrust reversers employ a cascade to enhance the turning of
the flow(60). There are broadly two types of thrust reversers: (a)
operating on both core and fan flow; and (b) operating on the fan
flow only. Thrust reversers operating on the fan flow could be
cascade or petal type (Figs 15(b) and (c)).
A thrust reverser offers a number of operational advantages(60,61)
shortening of landing runs, less wear and tear of aircraft brakes, safer
landing in adverse weather conditions, additional safety and control
margins during aborted take-offs. Thrust reversers significantly
affect the nacelle design, increasing weight and resulting in higher
manufacturing and operational costs.
3.1 Natural blockage thrust reverser
One of the newer types of thrust reverser that operates on the fan
flow is the natural blockage concept, first implemented on an aircraft
by Bombardier (Fig. 16). This design concept has the potential to be
more reliable and maintainable than other types of thrust reverser,
since it has fewer moving parts, notably no blocker doors or links.
The Bombardier natural blockage thrust reverser design for the
CRJ700/900 has increased reverse thrust efficiency, allowing the
engine to run slower in reverse mode. It also has a novel counter-
balance mechanism, which eliminates the need for powered door
operation. Studies on a natural blockage thrust reverser have found
that the flow entering the fan duct is presented with a rapid
expansion, while the reverser is fully deployed. This occurs just as
the flow is accelerating around the curved surface of the diverter
fairing. The combination of rapid expansion and high rate of turning
results in flow separation. However, this separation point does not
occur at a steady point, and instead moves up and down the diverter
fairing surface as the trans-cowl cavity pressure rises and falls in a
cyclic fashion. The resultant pressure fluctuations could have an
adverse effect on the structure fatigue life, and therefore need to be
taken into account in the design process.
the nominally OFF part of the duty cycle. This shows minimal air
leakage in contrast with the work carried out by Azevedo et al(56)
where the leakage is quite significant. The stable flow structure
created in the experiment is important in order to measure the instan-
taneous heat transfer correctly.
Figure 12 shows typical graphs of local Nusselt number against
radial distance for frequencies of 10Hz and 20Hz for all the three
Reynolds numbers under investigation. The graph shows that the local
Nusselt numbers for pulse flow are higher than for steady flow at
positions from 1-diameter outwards. Higher turbulence intensity at
these positions is believed to contribute to the increase in heat transfer.
Figure 13 shows the dependence of stagnation Nusselt number on
frequency for three Reynolds numbers at . For stagnation point heat
transfer, jet pulsation has the effect of decreasing the heat transfer
for all the frequencies studied. The highest frequency degradations
are between 20 to 50Hz and above 80Hz. This generally is in
agreement with the results obtained by Azevedo et al(56). In their
experiments, where the Reynolds numbers were less than 25,000, the
heat transfer results at all frequencies showed significant degra-
dation. The results show that higher mass flow rate can influence the
heat transfer measurements.
Figure 14 shows the variation of Nusselt number averaged over
the impingement area with pulsed frequency f, for jet Reynolds
numbers at 16,000, 23,300 and 32,000. The flow structure changes
with frequency in a complex manner so the Nusselt number versus
frequency curve shows no clear trend. The average Nusselt numbers
for the pulse jet are equal or higher than the steady jet for all the
frequencies tested except at 50Hz. The maximum average heat
transfer enhancement occurs at a frequency of 70Hz for a Reynolds
number of 32,000. The enhancement of heat transfer at this
frequency is approximately 40%.
It should be noted that the present results were obtained at non-
dimensional distance, x/D equal to four while those obtained by
Azevedo et al(56) were at six. This explains why a higher stagnation
heat transfer value was obtained here.
The present results show that the average heat transfer on the
impingement area is enhanced for almost all of the pulse frequencies
investigated even though the stagnation point heat transfer decreases.
These increases are shown to be available for a system that uses a
single pulse jet impinging on an area with radius up to six times the
nozzle diameter.
The results of the experiments show that there is significant
enhancement in the local heat transfer of the pulse flow at positions
one nozzle diameter or more away from the stagnation point for all
the pulse frequencies. The stagnation point heat transfer does not
show any enhancement for the three Reynolds numbers investigated.
The average Nusselt number for the pulse jet is enhanced for all the
frequencies investigated except at 50Hz. The degree of enhancement
is in the range 0-40% with the greatest benefit at frequency of 70Hz
for Re = 32,000.
Heat transfer in the pulse flow mode is complex and dependent on
the flow structure of the jet. The significant enhancement of the heat
RAGHUNATHAN ET AL KEY AERODYNAMIC TECHNOLOGIES FOR AIRCRAFT ENGINE NACELLES 273
Figure 16. Schematic diagram of natural blockage thrust reverser.
of a porous surface with a plenum chamber: there is no active
suction. It is seen that the basic configuration (Fig. 19(a)) has
regions of supersonic flow with separation and associated insta-
bilities, whereas the configuration with a porous surface (Fig.
19(b)) has no supersonic flow or flow separation on the diverter
fairing. The flow field in the fan duct has been significantly
improved by the inclusion of a porous area in the diverter
fairing, although there is a small reduction in the reverse thrust.
Typical cascade configurations used to investigate the effect of
cascade blades on the performance and weight are shown in Fig. 20.
The corresponding Mach number contours for the flow are shown in
Fig. 21. For design 1 (Fig. 21(a)), with the thrust reverser deployed,
there exists supersonic flow at the diverter fairing. Mach number
contours for design 2 (Fig. 21(b)) show that the flow is subsonic.
This design produced a 10% reduction in weight and an
improvement in the structural performance through reductions in
both maximum vane displacement and mid-vane stress levels.
However there is a penalty in reverse thrust, which was reduced by
9%. Cost and ease of manufacture will also have to be taken into
consideration before any firm conclusions are drawn regarding a
final cascade configuration.
CFD simulations were performed on a two dimensional model
using Fluent 5™ in order to understand the aerodynamic perfor-
mance of the natural blockage thrust reverser, fully deployed.
The test cases included three thrust reverser configurations: a
base line with cascade vanes, a configuration with a porous
diverter fairing and the cascade vanes removed, and a configu-
ration with a reduced number of cascade turning blades. Two
types of grids were used: an unstructured non-uniform grid for
the whole computational domain (Fig. 17 – Grid 1) and a hybrid,
unstructured non-uniform grid with a structured grid inside the
boundary layer (Grid 2). In both cases grid independence studies
were performed. There was no significant difference between the
solutions of those two grid systems (Fig. 18). The CFD predic-
tions for the baseline model were validated against wind tunnel
test data collected at Flow Science Ltd, Manchester, UK, by
Bombardier Belfast on a 40% scale model of the natural
blockage thrust reverser (to be published). Details of the investi-
gations on the natural blockage thrust reverser are given in Refs
62-77.
The Mach number distribution in the natural blockage thrust
reverser with and without flow control in the form of a porous
diverter is shown in Fig. 19. The passive control device consists
274 THE AERONAUTICAL JOURNAL MAY 2006
Figure 17. The unstructured grid. Figure 18. Normalized mass flow rate distributions of
approaching the fan duct flow to the cascade.
Figure 19. (a) No porous surface on diverter fairing, (b) Effect of porous diverter Mach number contours.
constraints on the design of adequate ventilation systems are legion,
including: externally, the position of the engine on the aircraft
(wing or aft fuselage mounted); internally, whether the zone to be
ventilated is the fan cowl casing or core engine casing, structural
design, the locations of major systems (such as the thrust reverser,
gear box, oil cooler, fuel system and lubrication system); and opera-
tional considerations such as the required number of volume
changes per minute and the outside air temperature and Mach
number. With such a range of constraints it is not altogether
surprising that the system design is finalised after the other major
items have been frozen. This can result in over-design of nacelle
components, leading to increased weight, part numbers and cost.
However, before the certificating authorities approve any new
system for operation, it is necessary to demonstrate that it meets the
airworthiness requirements, and the certification tests can be fraught
with uncertainty because of the general ad hoc nature of the venti-
lation system design.
Within the last decade, Bombardier and QUB have collaborated
on four research projects aimed at elucidating the nature and perfor-
mance of ventilation systems. These were: fire propagation and heat
transfer modelling within the BR710 nacelle for certification
purposes(79); performance enhancement of auxiliary air intakes(80,81);
aerodynamic performance of pressure relief doors on engine
nacelles(82); and fire suppressant dispersion(83,90).
4.1 Fire zone modelling(78,79)
An attempt was made to model fire propagation and heat transfer in
the fan compartment, (fire zone one), of the BR710 nacelle. Full-
scale experimental tests are prohibitively expensive, and it is
possible, in principle, to gain valuable and instructive insights
through the application of computational models. However, fire
zones are typically geometrically cluttered environments, and initial
attempts to model the fan compartment zone of the BR710 nacelle
3.2 Natural blockage thrust reverser – cascade-less
Cascade vanes can be costly to manufacture and maintain, and have
a weight penalty associated with them. Having no cascade vanes
removes these disadvantages and allows the cavity in the transcowl
to be closed over. However, the reverse thrust performance is
markedly reduced with such a configuration. The inclusion of a slat
close to the diverter fairing improves the reverse thrust performance
by controlling the flow at the diverter fairing. CFD analyses of the
cascade-less configurations, with and without a slat, have been
carried out and typical Mach number contours are presented in Fig.
22. These CFD predictions have been validated against wind tunnel
test data collected at Flow Science Ltd, Manchester, UK on a 40%
scale model. Although the tests were not fully exhaustive the
cascade-less configuration gave a reverse thrust performance that
was 75% of the configuration with cascade vanes. Consideration of
the test results indicate that with careful choice of diverter faring
and transcowl geometry, slat profile and slat position a reverse
thrust performance that is 90% of the cascade vanes configuration is
achievable.
4.0 FIRE ZONE VENTILATION(78-90)
The safe and efficient operation of aircraft turbofan engines requires
effective ventilation of the nacelle zones. Ventilation of the nacelle
is required to avoid the build up of hazardous vapours, which may
contribute to the initiation or continuance of an engine fire, and in
the event of a fire the ventilation must ensure correct dispersal of
the fire-extinguishing agent. In addition, cooling of critical compo-
nents within nacelle zones is also achieved through the provision of
ventilating flows. This may include forced cooling of the engine
casing in order to improve specific fuel consumption by
maintaining minimum compressor and turbine tip clearances. The
RAGHUNATHAN ET AL KEY AERODYNAMIC TECHNOLOGIES FOR AIRCRAFT ENGINE NACELLES 275
Figure 21. Cascade design for weight reduction. Mach number contours. (a) Design 1, Original, (b) Design 2, 10% Weight reduction.
(a) Cascade Design 1, Original. (b) Cascade Design 3, 10% Weight Reduction.
276 THE AERONAUTICAL JOURNAL MAY 2006
Figure 22. Cascadeless thrust reverser Mach number contours. (a). Without Slat; (b). With Slat.
Figure 25. Side view of predicted temperature distribution on a surface
midway between the inner and outer walls of the annulus
Figure 23. Measured and predicted temperatures inside fire zone.
Figure 24. Side view of inner cylinder black line added to
show limit of carbon deposits after sustained combustion.
(a) Tunnel setup.
(b) T injection nozzle. The flow direction is into the plane of the page.
Figure 26. Schematic of the experimental setup(88).
The current design procedure for the fire-suppression system
relies on engineering judgment and empiricism. It has been
reasonably successful in the past, mainly due to the effectiveness of
Halon 1301, the fire-suppression agent traditionally used in aircraft
engine nacelles. However, Halon 1301 has a high ozone-depleting
potential, and its production was banned as a result of the 1994
amendment to the 1987 Montreal Protocol(84). Despite significant
research effort(85-87), a ‘drop-in’ replacement for Halon 1301, with
acceptable levels of suppression effectiveness, ozone-depleting
potential, global warming potential, and toxicity has not yet been
found. It seems likely, however, that Halon 1301 will be replaced
with a less effective agent, and so it will be important to improve the
performance of the fire-suppression systems, to make the most
efficient use of the agent. It is envisaged that CFD could provide a
better understanding of the agent dispersion, and ultimately deliver a
validated engineering tool to aid the design and certification of the
fire-suppression system. With the aid of CFD, future fire suppression
systems will be lighter, cheaper and safer than today’s overly conser-
vative designs.
The dispersion of fire suppression agents in aircraft nacelles has
not received a great deal of attention in the literature. Hamins et al(88)
performed both experiments and simulations of suppressant transport
in a smooth annular geometry. They studied the dispersion of N2,
which was injected into the annular domain through either a T-
junction or a round nozzle. In some cases clutter was placed on the
inner wall of the duct, in the form of a small transverse rib. Hamins
et al. only had a limited degree of success with their CFD simula-
tions; in most cases the correct trends were predicted, but the overall
correlation with their experimental data was relatively poor. They
suggested that deviations were probably due to the use of the k-
turbulence model. Lopez et al(89) studied the flow of a suppressant in
an uncluttered F-18 engine nacelle; they claimed that this was the
first attempt to model agent dispersion in a realistic geometry. They
obtained a good agreement between their CFD results and earlier
experimental results. The oxygen concentration was slightly over-
predicted, and this was attributed to the fact that they did not model
the fire that was present in the experiments. Later, the same group of
workers performed simulations using the VULCAN fire field model
and reported that the velocity field was relatively unaltered by the
presence of a fire(90).
The aim of the work at Queen’s University of Belfast was to use
CFD to replicate the experiments of Hamins et al(88), to begin
validating the use of CFD for predicting agent dispersion in engine
nacelles. The experimental geometry is shown in Fig. 26. The N2
was injected through either a T-junction (Fig. 26(b)) or a round
nozzle. The commercial CFD code Fluent 6™ was used to study
three main configurations: T injection nozzle, T injection nozzle
with a rib on the inner wall of the duct, and round injection nozzle.
The N2 was injected at a rate of 0⋅0137kgs–1 and the tunnel velocity
was 3ms–1. The species transport model was used in conjunction with
the main flow and turbulence equations. Various k-εand k-ωturbu-
lence models were tested. Unstructured wall-function meshes were
used, since the ultimate aim is to simulate the flow inside a cluttered
nacelle. To obtain a good correlation with the experimental results, it
was essential to use adaptive mesh refinement in the N2plume.
As the agent issues from the T injection nozzle, it impinges on the
wall of the duct, and spreads in the circumferential direction as it
travels downstream (Fig. 27). The volumetric concentration profile
for the T injection nozzle case is plotted in Fig. 28. The measure-
ments were taken at various circumferential positions midway
between the inner and outer duct walls, 96cm downstream of the
injection nozzle. There is a definite asymmetry in the experimental
concentration profiles, which could not be explained by Hamins et
al(88). For the clean duct (Fig. 28(a)), the correlation between the
CFD results and the experimental results is excellent. When the rib
was introduced (Fig. 28(b)), the N2concentration was lower for 135°
< θ< 225°, since the rib slowed the spread of the agent in the
circumferential direction. The CFD results were qualitatively
faithfully were abandoned as it was recognised that there was too
much uncertainty about achieving a reliable model of the
convective flow in the zone, not to mention the difficulties
associated with modelling the progress of a combustion process.
Instead, the fire zone was modelled as an annular cylinder, approxi-
mately 1:5 scale of the fire zone of which it was an idealisation.
Forward and aft Y-shaped vents placed at 12 o’clock on the outer
surface of the annulus, and a rectangular outlet normal to the axis of
the cylinder and placed at 6 o’clock on the outer surface provided
ventilation. Both experimental and computational flows, non-
combusting and combusting, were investigated.
The fire test simulation was that of a ‘ventilated limited flame’,
i.e. the fuel source enters the zone at a strategic point corresponding
to the likelihood of a fuel pipe leak with an ignition point placed
directly opposite the point of fuel entry. The only source of oxygen
to sustain the flame is that of the ventilation system. Five sets of
conditions, typical of take-off, climb and cruise, were investigated.
Two fire scenarios were considered, namely, explosive potential
and sustained combustion. Fluent 5™ was used for the calculations.
Generalised finite rate models were used for the explosive scenario,
and the mixture-fraction/probability density function (pdf) formu-
lation for the sustained combustion cases. The tests employed
propane as the fuel, rather than kerosene, because it is clean,
gaseous at the operating conditions, more easily modelled than
kerosene, and yet exhibits similar flammability properties, such as
heat of combustion.
The explosive scenario proved to be very difficult to model.
Thermal boundary conditions modelling the radiation of heat away
from the annulus could not be applied because of computer memory
constraints. The explosion was very rapid, and very small time steps
were required to capture the first 0⋅5 seconds, after which the
temperature of the outlet gases dropped very rapidly to a steady
state that indicated the end of combustion. The response times of
the temperature probes used to measure the temperature in the
annulus were too long, by an order of magnitude, to capture the
transient temperature rise, and consequently, no tie-up between the
calculations and the measurements was possible.
The sustained combustion case was more yielding to investi-
gation. Temperature was measured at several locations around the
annulus, and a comparison of the measurements and predictions
showed similar trends (Fig. 23). Moreover, the deposition of soot on
the inside of the experimental annulus was taken as an indication of
the regions, within which combustion was concentrated, and
therefore also indicative of flow patterns (Fig. 24). Similar patterns
were observed in the predictions (Fig. 25). Both experiments and
calculations revealed consistent trends, indicating hot-spots in
which the air temperature rose to between 800°C and 1,100°C,
while the air temperature in the rest of the zone remained in the
range 40°C to 100°C. This suggests that whole zone heating is not
as prevalent as was initially assumed.
4.2 Fire suppressant dispersion
The complex ventilation flow in the fire zone makes it difficult to
predict the dispersion of the fire-suppression agent, and hence the
performance of the fire-suppression system. At present, the perfor-
mance of the fire-suppression system is evaluated by ground tests or
flight tests. Typically, the injection nozzle type, number, location
and orientation and the layout of the delivery lines are modified
until the system passes the certification test outlined by the FAA(83).
The certification test procedure involves releasing the agent from
the bottle, and measuring the agent concentration at twelve different
locations inside the nacelle. To pass the test, a minimum agent
concentration (6% by volume for Halon 1301) must be observed
simultaneously at all twelve probe locations for a minimum of 0⋅5s.
Certification is a time consuming and expensive process, which
comes late in the overall design cycle, and so can lead to
programme delays and cost overruns.
RAGHUNATHAN ET AL KEY AERODYNAMIC TECHNOLOGIES FOR AIRCRAFT ENGINE NACELLES 277
278 THE AERONAUTICAL JOURNAL MAY 2006
Figure 27. Volumetric concentration of N2(%) at various
axial planes downstream of the T injection nozzle.
(a). Clean duct.
(b). With a rib on the inner duct wall.
Figure 28. N2concentration versus circumferential
position for T injection nozzle case.
Figure. 29. N2concentration versus circumferential position,
for round injection nozzle case.
(a). Location of pressure relief door.
(b). close-up of door.
Figure 30. Typical engine nacelle and pressure relief door.
was 37⋅8mm. A flat, rectangular flap 25⋅4mm wide by 25⋅4mm long
was attached to the upstream edge of the duct orifice. The orifice
leading edge was placed 203mm downstream of the inflow
boundary. The computational domain being symmetric, only one
half, measuring 432mm long by 79⋅4mm wide by 114⋅3mm tall was
modelled.
Flap angles of 15° to 45°, in 5° increments, were studied. The free
stream Mach number was varied from 0⋅4 to 0⋅85 in increments of
00⋅15. As a result, the ratio of boundary-layer thickness to orifice
length varied between 0⋅095 and 0⋅110. The pressure ratio was
varied between 00⋅64 and 00⋅97 in order to obtain the range of
discharge flow ratio coefficients (DFR) required. The realisable k-ε
turbulence model was used because it is of known accuracy when
dealing with flows involving jets, separations and secondary flows.
A mesh dependence study was performed to ensure that converged
solutions were mesh independent.
Typical results are shown in Figs 31 and 32. Generally good
agreement was obtained between measurement and prediction of
integral quantities such as DFR and thrust coefficient. Predictions
under-estimated the discharge by between 5% and 20%, depending
on pressure ratio; the thrust coefficient was slightly over-predicted.
Calculations were performed for a wider range of flap angles than
was considered in the experiments, and it was shown that the DFR
increases with flap angle up to an optimum value, after which
increasing the flap angle decreases the DFR. The value of optimum
angle falls with increasing pressure ratio, but is insensitive to free
stream Mach number. These effects may be explained, in part, by the
strength and orientation of the vortex system generated by the
inclined flap, which sets a boundary condition on the flow emerging
from the exhaust duct.
It was shown that the angle for which the hinge moment on the
flap was zero lay in the range of 10 to 15 degrees for all cases. A
freely hinged, weightless flap would, therefore, achieve a trimmed
balance in that range of angles. Increasing Mach number decreases
this angle, while increasing pressure ratio increases it.
4.4 Auxiliary air intakes
Auxiliary air intakes perform a variety of functions on aircraft, from
cabin air supply, to engine component cooling and fire zone purging.
The intakes come in a number of forms, including pitot designs,
correct, but the concentration levels were slightly high for 135° < θ
< 225°. For the round injection nozzle (Fig. 29), there is a significant
change in the concentration profile compared to the T injection
nozzle. The concentration levels are extremely low for 90° < θ<
270°, and very high elsewhere. The agent dispersion is much better
for the T nozzle. The CFD results for the round nozzle show a fair
correlation with the experimetal values. To conclude, CFD can
model the dispersion of a model fire suppressant with a reasonable
degree of accuracy, provided adaptive grid refinement is used in the
agent plume region. Further validation is underway to study the
effect of introducing more clutter into the domain.
4.3 Pressure relief doors
An important set of auxiliary engine outlets is related to pressure
relief in the case of duct burst in the nacelle compartment. Typically
pressure relief doors (PRDs) are rectangular in geometry with a door
hinged on either the forward or side edge (Fig. 30). In the event of a
duct burst, the pressure that builds within the compartment would
compromise the nacelle structure were it not for the pressure relief
doors that are installed to regulate the pressure to an acceptable
value.
The designer of PRD installations must have reliable data on force
and discharge characteristics. Very little research has been done on
this subject, especially in recent years. Current designs have been
based on experimental data presented in NACA TN400791
regarding the discharge characteristics of flapped, curved duct
outlets in transonic flows. Consequently a conservative approach to
analysis has been adopted by the aerospace industry. More detailed
information is now required to achieve a design that satisfies
customers and certificating authorities but is not overly conservative.
It is proposed to achieve this through systematic computational
investigations, supported by experimental studies. The database
should help directly the designer to improve the overall perfor-
mance, both aerodynamic and structural, of the device.
The commercial CFD package, Fluent 6™, was used to model82
the experimental work described in NACA TN4007. The computa-
tional domain was a rectangular duct 25⋅4mm wide by 46⋅7mm long,
which turned the exhaust flow through 90° about a radius of
curvature of 50⋅8mm into the stream-wise direction. The upstream
edge of the orifice was extended 9⋅5mm so that the orifice length
RAGHUNATHAN ET AL KEY AERODYNAMIC TECHNOLOGIES FOR AIRCRAFT ENGINE NACELLES 279
Figure 31. Measured and predicted discharge coefficients as
functions of pressure ratio for flap angle = 25°. Figure 32. Measured and predicted thrust coefficients as
functions of discharge coefficient for flap angle = 25°.
fared surface mounted blisters and flush intakes. Intake performance
is characterised by the amount of air that can be swallowed for a
given pressure drop (Fig. 33). Flush intakes are generally either
parallel walled or of the NACA contoured converging wall design.
The former design is simpler, and less expensive to manufacture and
maintain, but has inferior performance, so that a larger, heavier
intake, with a greater drag penalty is necessary to match the capacity
of the equivalent NACA intake. The volume flow rate versus ram
pressure recovery (RPR) characteristic of the flush intake is a
function of the ratio of the boundary layer thickness (δ) to the intake
scale, usually represented by depth (d). The performance of the
intake is degraded as this ratio increases.
Vortex generators (VGs) have been used for decades to improve the
high lift capability of aircraft wings. Vane type generators are the most
common form, but all have much the same mode of operation: a longi-
tudinal vortex is generated with a diameter of the same order of
magnitude as the boundary layer thickness. It is embedded within the
turbulent boundary layer so that it draws high momentum flow from
the external mainstream down into the lower third of the boundary
layer. The boundary layer is thinned in the downwash region, and
thickened in the upwash region. VGs are often used in arrays: co-
rotating (all the VGs aligned the same way) and counter-rotating (VGs
alternately at positive and negative incidence) arrays are possible. In
counter-rotating arrays, the region of downwash between VGs is
referred to as common-flow-down, while the region of upwash
between VGs is referred to as common-flow-up. It is hypothesised that
a pair of vortex generators placed upstream of a parallel walled intake,
such that the intake lies in the common-flow-down region should
reduce the local value of δ/dand improve the intake performance.
Experiments were conducted in a closed loop wind tunnel with a
test section 575mm × 375mm. An idealised model of an auxiliary
intake was mounted in a flat plate, 870mm downstream of the plate
leading edge. The intake had a ramp angle of 10°, a width to depth
ratio of 4 and a cross sectional area of 6450⋅16mm2. A plenum
chamber 405mm × 265mm × 550mm was placed beneath the intake
and a flow meter and vacuum pump placed downstream of it. The
test speed was 45ms–1, and at the intake Reθ= 6,000 and δ/d= 1⋅5.
There was no external stream-wise pressure gradient. A pair of
vortex generators, height 20mm, chord 40mm, apex separation
50⋅8mm, angle-of-attack 22°, in common-flow-down configuration
(i.e. the region between the VGs is common-flow-down) was placed
in a turbulent boundary layer of δ= 10mm, 400mm upstream of the
intake. The total pressure inside the intake was measured using a
pitot rake of five probes, which could be traversed normal to the
intake wall and laterally. Single wire hot wire anemometry was used
to measure the boundary layer in the free stream 150mm upstream.
The computational fluid dynamics package Fluent 6™ was used
to test more configurations. The domain measured 1,210mm ×
1,050mm × 500mm and included the test plate and intake, with the
intake placed 82mm downstream of the inflow boundary so that
experimental data could be used as the inflow boundary condition. A
fully block structured mesh of 350,000 cells was created. The
software was run in its implicit segregated mode, with several
versions of the k-ωturbulence model tested.
Typical results are shown in Figs 33 to 35. The application of the
vortex generators typically gave ram pressure recovery improve-
ments of between 35% and 40%. Although ingestion of the vortex
pair must be avoided, this does not appear to be a problem as the
vortices naturally migrate away from the centreline of the intake.
The treatment gives the intakes the potential for a peak performance
similar to that of the more complex NACA intake. However, the
vortices generated by the NACA intake are ingested at high flow
rates and the NACA intake’s performance drops markedly. Hence,
the treated parallel walled intake performs better over a wider range
of flow conditions. Designers/manufacturers may be able to use
either smaller examples of the treated intake or smaller numbers of
them. This would have benefits for aircraft weight, part count and
maintenance.
280 THE AERONAUTICAL JOURNAL MAY 2006
(a) Effect of flow control, measured and predicted.
(b) Predicted effect of spacing of vortex generators.
(c) Predicted effect of height of vortex generator.
Figure 33. RAM pressure recovery vs velocity ratio.
Godard et al(97) studied the effect of changing the position of the
nacelle both vertically and horizontally. The most sensitive change
in position of the nacelle was found to be the horizontal position.
The best result for minimum drag was found to be furthest upstream
from the wing. The change in vertical distance had much less effect
on drag. The reason for this was that when the nacelle was closest to
the wing the area between the top of the nacelle, the pylon and the
bottom of the wing was contracting. This resulted in an acceleration
of the local flow and a shock wave. This caused the drag of the
aircraft to increase and the lift to decrease. When the nacelle was at
the most upstream position, the area in this region was greater,
which reduced the local velocity and the adverse interference effects.
However, the further the nacelle is away from the wing the greater
the structural weight penalty is.
The effect of larger BPR nacelles has received the most attention
in previous projects. This is because larger BPR engines have better
fuel efficiency(98), but, of course, larger nacelle diameters. The
largest BPR engine today is on the Boeing 777, which has an engine
BPR of nine. These large diameter nacelles lead to greater inter-
ference effects(99). However it is also reported that the higher BPR
engines could have the same impact on aircraft drag as smaller
engines 100.
The CFD methods used to simulate the nacelle, pylon and wing
interference effects have been developed over the past 15 years. The
earlier calculations were performed using inviscid Euler codes(101).
The next step was the used of Euler codes with viscous correc-
tions(102). In recent years Reynolds Averaged Navier Stokes (RANS)
calculations have been performed(103). The capability and computa-
tional resource required of each in predicting the interference effect
is different. Euler calculations can predict the change in lift when the
nacelle and pylon are installed and is computationally inexpensive.
However the accuracy of the lift values is inadequate and drag impli-
cations cannot be predicted. Euler calculations with viscous effects
can predict the lift more accurately and also the changes in lift.
5.0 NACELLE, PYLON AND WING
AERODYNAMIC INTEGRATION(92-102)
With today’s ever increasing demand for air travel, the environ-
mental impact of aircraft is becoming a major concern. As well as
the obvious problem of noise pollution near airports, there is also the
issue of CO2emissions. As previously stated, ACARE have targeted
a 50% reduction in aircraft CO2emissions by the year 2020(92). A
way of achieving these ambitions targets is to minimise drag for a
given lift. In particular when a nacelle and pylon are installed onto
an aircraft there is an increase in the aircraft total drag. A part of this
extra drag can be reduced through better analysis of the complex
interactions between aircraft components. With modern nacelle
designs the use of laminar flow technology can also have a signif-
icant saving of aircraft fuel consumption. This was demonstrated in
flights tests carried out by Rolls-Royce, MTU and DLR(93) where nett
sfc was decreased by 2%.
There has been a number of investigations carried out in the past
on the aerodynamic interactions of the nacelle, pylon and wing.
These can be broken down into four areas: interference effects of the
nacelle, pylon and wing; the effect of different nacelle position; the
effect of high bypass ratio (BPR) and ultra high BPR nacelles; and
the ability of CFD to predict the interference effects.
Once the nacelle and pylon are installed onto an aircraft wing, the
aerodynamic performance of the wing changes significantly. There
is an increase in total drag, a loss of total lift(94) and the shock wave
position on the upper surface of the wing changes. Also affected is
the wing spanwise loading, which will in turn cause an change in
induced drag. There is no agreement as to the source of these degra-
dations, with some authors concluding the pylon shape is to blame(95)
and others concluding the channel effect of the nacelle, pylon and
wing is to blame(96).
Another area of investigation is the effect different nacelle and
pylon locations have on the aerodynamic performance of the aircraft.
RAGHUNATHAN ET AL KEY AERODYNAMIC TECHNOLOGIES FOR AIRCRAFT ENGINE NACELLES 281
Figure 34(a). VR = 0⋅5.
Figure 34(b). VR = 0⋅7.
Figure 34(c). VR = 0⋅9.
Figure 35(a). VR = 0⋅5.
Figure 35(b). VR = 0⋅7.
Figure 35(c). VR = 0⋅9.
282 THE AERONAUTICAL JOURNAL MAY 2006
Figure 36. Prismatic layers from fuselage surface. Figure 37. Predicted and experimental lift coefficient vs drag coefficient.
Figure 38. Compression pylon shape. Figure 39. Conventional pylon shape (DLR F6).
Figure 40. Predicted Lift vs drag of baseline
case to compression pylon case.
Figure 41. Typical percentage DOC
distribution for a commercial aircraft.
Figure 42. Illustration of typical nacelle design. Figure 43. Results from DFMA implementation
on the redesign of a nacelle torque box.
Before
After
Percentage
Reduction
Number of Parts
110
86
22
No. of Fasteners
1090
916
16
Assembly time (Hrs)
116
96
17
Weight (lbs)
86
71
17
Recurring manufacturing Cost
500
A typical Direct Operating Cost (DOC) breakdown of the Airbus
A320 class of aircraft (150 passengers and 2,800 nautical mile range,
Fig. 41) shows that the aircraft cost contribution to DOC is more
than four times that of the fuel contribution. This is even more
extreme for smaller aircraft with a lower payload range. It is under-
stood in the aerospace industry that any reduction in DOC requires
not only a better understanding of aerospace disciplines but also an
understanding of the sensitivity of each discipline to the other. QUB
in partnership with Bombardier Belfast took the initiative in this area
of research, in 1996; through the EPSRC IMI managed programme
(SOAMATAS), where a novel concept of trade-off study between
costs associated with two design drivers, aerodynamics (fuel) and
manufacturing were conducted. The findings from this programme
led to the just completed EPSRC funded programme DEMARDOC
where the concept was extended to trade-off studies between wider
ranges of design drivers; including aerodynamics, structural configu-
ration, manufacturing and assembly. Details of these investigations
are given in refrences (105-111).
A particular component of an aircraft was chosen for a detailed
programme of research. This was the nose cowl of two generic
nacelles described in the next Section. The road map for DOC
estimation consisted of the development of a rapid cost model based
on the baseline design and validated against the later design. Cost
reduction measures for the structural part fabrication and their
assembly was investigated. The savings were extrapolated for the
aircraft and corresponding DOC savings for the mission profile were
estimated.
6.1 Development of an optimised design to reduce
nacelle manufacturing cost
The major objectives involved: (a) development of an accurate cost
model and (b) development of a methodology for evaluation of the
cost benefits of design for manufacturing and assembly (DFMA) in
addition to tradeoffs between aerodynamics and manufacturing.
A rapid cost-modelling methodology was developed that was
specifically aimed at catering for the industrial needs of DFMA
during the conceptual design phase. The rapid cost model was based
on parametric methods taking into consideration the effects of
engineering decisions on cost and the link between design statistics
and manufacturing cost. In the process of cost modelling, method-
ologies were also established for cost data collection, definition and
accounting. From a baseline cost, the method demonstrated a fast
and relatively accurate prediction for the later design.
The task concentrated on cost modelling of the structural elements
of generic nacelles. The task was conducted in two steps. First, a
detailed cost model was developed for the ‘Nose Cowl’ and then the
model was extended to the complete nacelle. The generic nacelle is
However the drag implications cannot be predicted and the computa-
tional cost is three times that of the basic Euler calculations(34).
RANS calculations can predict the lift accurately and the change in
lift. In addition, the drag implications can be predicted, which means
a full investigation can be performed. However the computational
cost can be as much as hundred times greater than the basic Euler
calculations(34).
This research was completed using an unstructured mesh and
solver. The computational fluid dynamics package Fluent 6™ was
used and the mesh was generated using ICEMCFD™. The reason
for choosing unstructured meshing techniques was the relative
automation of mesh generation. To meet the mesh requirements for
an accurate simulation using unstructured meshes is difficult,
especially in the near wall regions. A solution to this problem is to
create a hybrid mesh, which allows for sufficient gird resolution in
the near wall region. To create a hybrid mesh using ICEMCFD,
prismatic layers were grown from a surface mesh. The first cell
height was set to 0⋅003mm and 30 prismatic layers were grown at a
growth rate 1⋅2 (Fig. 36), resulting in a mesh y+ of approximately 1.
To validate the computational methodology the flow over a variant
of the DLR-F6 was simulated. This configuration was subjected to
many wind-tunnel experiments in the 1990s. The CFD results
obtained in this work were compared to published experimental data
on the DLR-F6. The predicted and experimental drag polar can be
seen in Fig. 37 for the wing body configuration and the wing body
nacelle pylon configuration. From this figure it can be seen the total
drag for both cases is over predicted for every design point, by approx-
imately 12%. However, with closer analysis of the results the
predicted drag rise due to the installation of the nacelle and pylon was
predicted to with in 1% of the experimental result. Therefore using
these computational methods one could compare different configura-
tions and accurately predict the drag change of each within sensible
limits. Using this validated technique a new pylon design was investi-
gated to determine the influence of the pylon. The conventional pylon
shape was redesigned to a compression pylon shape (Figs 38 and 39).
The new design had an impact on the drag of the aircraft. Studying the
result in Fig. 40 it is clear the drag has reduced for each design point
by as much as five drag counts.
6.0 AERODYNAMIC DESIGN TRADE-OFF
MODEL FOR MANUFACTURING AND
COST(104-111)
In recent years, customer needs in the commercial Aerospace sector
have been governed by reduced lead-time and cost, while also
looking for enhanced aircraft performance and safety measures(104,105).
RAGHUNATHAN ET AL KEY AERODYNAMIC TECHNOLOGIES FOR AIRCRAFT ENGINE NACELLES 283
Figure 44. Part counts of nacelle B normalised with respect to nacelle A. Figure 45. Trade off between aerodynamic
and manufacturing tolerances for cost.
this methodology several aspects of manufacturing and assembly
were taken into consideration. These include, part count for the
assembly, man-hours involved in fabrication and assembly,
rework/concessions and quality of the surface finish as typified in
Figs 43 and 44. The analysis of certain components such as that
shown in Fig. 43 was for re-design change only, where the material
type and manufacturing process were not affected. In that case, the
frames were manufactured in three sections and joined with splices
and rib clip attachments. These were re-designed into one-piece
frames that did not require the use of the clips. The results of the
approach adopted on Nacelle B (higher specifications and 50%
increase in thrust) althogh, led to an increase in part count for the
door by a factor of ten (Fig. 44) resulted in an overall part count
reduction of 24% compared to Nacelle A. The increase in structural
requirements from the much higher thrust is highlighted by the fact
that fastener count actually increased, e.g. by 8% on the nose cowl in
order to provide the required stiffness. However, there was a saving
of 12% in assembly costs due to the reduction in part count, and the
increased number of complex components, which were not manufac-
tured in-house, influenced the material cost increase of 16%. The
total reduction in cost was found to be 2% per kilo of structure and
therefore the evaluation methodology has been applied to show that
DFMA has been able to improve the manufacturability of the design
to maintain, and even reduce cost slightly, regardless of a 50%
increase in technical design specifications.
6.2 Prediction of aerodynamic performance and cost
arising from surface tolerance
The prediction of aerodynamic performance arising out of surface
tolerances involved two tasks: (a) trade-off studies between aerody-
namics and manufacturing to optimise allocation of the tolerances
and (b) prediction of performance and DOC.
Aircraft surface smoothness requirements were aerodynamically
driven with tighter manufacturing tolerance to minimise drag, where
the tighter the tolerance, the higher is the assembly cost in the
process of manufacture. A typical DOC distribution shows that fuel
cost is only a fraction of manufacturing cost (Fig. 41). A trade-off
study between fuel cost and manufacturing cost was performed on
an isolated nacelle and the analysis extended to a complete aircraft.
Manufacturing tolerance relaxation at eleven key manufacturing
features on the surface assembly of an isolated nacelle was studied
without unduly penalising parasite drag, using the cost model
developed. The drag estimation was based on ESDU data corrected
for pressure gradients determined by CFD and experimental studies.
The drag data and aircraft operation were used in evaluating the fuel
cost. A large amount of manufacturing cost data were obtained from
Bombardier Belfast and were analysed in detail for the effect of
manufacturing tolerances on cost. Fig. 45 shows DOC and drag
variation with surface tolerance relaxation.
The direct operating cost of aircraft was estimated using
Association of European Airlines (1989) ground rules. The payload-
range for the mission profile was kept constant throughout the
aircraft performance analysis.
With conservative estimation, given below is the typical aircraft
cost reduction through such a DFMA approach to tolerance
allocation, with fuel price taken at US$0⋅75 per gallon. The study
resulted in approximately 1⋅28% DOC saving for a 2% saving in
aircraft cost involving no drag rise, and additionally, approximately
0⋅42% DOC saving for a further 1% saving in aircraft cost for
tolerance relaxation that did involve drag rise. The total of 1⋅7%
DOC savings translates into savings of $530 per sortie for the Airbus
320 class of aircraft. With a typical annual utilisation of 500 sorties
that totals to $265K per aircraft. For smaller aircraft the percentage
savings could be higher.
shown in Fig. 33. Two nacelles, A and B, were considered in the
modelling. Nacelle A was an existing product and that was taken as
the baseline design, whereas nacelle B was a newer design with
higher specification standards and 50% more thrust than the baseline
turbofan of that family of nacelles. The aerodynamic mould-lines of
both the nacelles were similar, but their structural design philosophy,
hence the sub-assembly (tooling concept) differed. The manufac-
turing cost of the finished product consisted of: cost of material (raw
and finished product); cost of parts manufacture; cost of parts
assembly to finish the product; cost of support (to ensure quality);
amortisation of non-recurring costs; and additional miscellaneous
cost (contingencies, etc.). (The actual cost data is classified
‘commercial in confidence’, so relative results are shown.)
Structural components and Engine Built Unit (EBU), e.g. anti-
icing units and valves, were considered in the cost modelling, but the
EBU costs were separated and are not included in the presented
research. Eleven cost drivers in two groups were identified for the
analysis. Group 1 related to in-house data within the organisation
and consisted of eight cost drivers, namely: size, material, geometry,
technical specification, structural design concept, manufacturing
philosophy, functionality, man-hour rates (overheads etc.). Group 2
cost drivers consisted of role (e.g. military or commercial), scope
and condition of supply, and programme schedule. As they were not
concerned with the in-house capability issues they were not
considered in this project.
The methodology applied was based on the factors/indices of
nacelle B, in relation to the baseline cost of nacelle A, and the
associated cost drivers mentioned earlier. The total manufacturing
cost of the nacelle was the sum of the individual costs of each of the
four nacelle components, as given below for the nose cowl cost
CNC, being the sum of the following six items:
. . . (4)
Where subscript ‘Mat’ stands for Material, ‘Fab’ for Fabrication ‘Asm’
for Assembly, ‘Sup’ for Support ‘Amr’ for Amortisation and ‘Misc’ for
Miscellaneous Cost.
Methodologies were developed for each of the six cost compo-
nents. For example, nose cowl manufacturing cost C'Man consisted of
two items: parts fabrication and parts assembly to finish. Man-hours
required for the fabrication of each part and assembly were a combi-
nation of operations; machining, forming, fitting and mounting into
jigs. Manufacturing cost was expressed as:
Manufacturing cost = rates × man-hours × learning curve
factor × size factor × manuf. philosophy.
. . . (5)
This analysis led to a simple tool for the calculation of nacelle B
nose cowl cost, as given by:
. . . (6)
Analysis of industrial cost data showed that the nose cowl A cost
fractions were as follows:
. . . (7)
Hence the relative cost of nacelle B was expressed as:
. . . (8)
Analysis of components such as the fan cowl, thrust reverser and tail
cone followed the same procedure as the nose cowl. In developing
284 THE AERONAUTICAL JOURNAL MAY 2006
CNC = ∑16 C′ i = C′Mat + C′ Fab + C′ Asm + C′ Sup + C′ Amr + C′Misc
or C
′
Man = C
′
Fab + CAsm = [(Ksize )0.5 ∑1m [∑1n F1 F2…. F n]m × (man hour × rates ×
learning factor)] Fab + [(Ksize )0.25 ∑1m [∑1n Fi
’]× (man hour × rates × learning factor)]Asm.
CNC_B = 0⋅8306Mat_NacA + 1⋅0878 C
′
Fab_NacA + 0⋅759 C
′
Asm_NacA + 0⋅05 ×
(0⋅828 C
′
Mat_NacA + 1⋅0878 C
′
Fab_NacA + 0⋅759 C
′
Asm_NacA)
C
′
Mat_NacA/CNoseCowl_A = 0⋅408, C
′
Fab_NacA /CNoseCowl_A = 0⋅349
C
′
Asm_NacA /CNoseCowl_A = 0⋅149
C
NoseCowl_B
/C
NoseCowl_A
= 0
⋅
92.
.
.
An auxiliary parallel wall air intake with a pair of vortex generators
placed upstream was studied with a view to enhance the intake perfor-
mance. The application of the vortex generators typically gave ram
pressure recovery improvements of between 35% and 40%. The
treatment gives the intakes the potential for a peak performance
similar to that of the more complex NACA intake but the treated
parallel walled intake performs better over a wider range of flow
conditions. Designers/manufacturers may be able to use either smaller
examples of the treated intake or smaller numbers of them. This would
have benefits for aircraft weight, part count and maintenance.
Validated CFD analysis showed that the drag rise due to the
installation of the nacelle and pylon on a wing can be predicted to
within 1% of the experimental result. Using this validated technique
a new pylon design was investigated to determine the influence the
pylon can have. The new design reduced the drag for each design
point by as much as five drag counts.
Design for manufacturing studies for engine nacelles and cost
analysis were performed for engine nacelles in conjunction with
trade off between manufacturing tolerances and aerodynamic toler-
ances and results were extrapolated to a complete aircraft. With a
fuel price taken at US$0⋅75 per gallon, the study resulted in approxi-
mately 1⋅3% DOC saving for a 2% saving in aircraft cost involving
no drag rise, and additionally, approximately 0⋅4% DOC saving for a
further 1% saving in aircraft cost for tolerance relaxation that did
involve drag rise. The total of 1⋅7% DOC savings translates into
savings of $530 per sortie for the Airbus 320 class of aircraft. With a
typical annual utilisation of 500 sorties that totals to $265K per
aircraft. For smaller aircraft the percentage savings could be higher.
ACKNOWLEDGEMENTS
This paper is based on a collaborative research programme on
Aircraft engine nacelle aerodynamics between the School of
Mechanical and Aerospace Engineering at QUB and Bombardier,
sustained over a ten year period and funded by Bombardier, the
Department for Education and Learning, InvestNI, EPSRC and the
Royal Academy of Engineering. The authors wish to acknowledge
the contribution made during this period by researchers Dr Jeffrey
Brown, Dr Kevin Donaghy, Dr Darren O’Neill, Dr Paul Humphries,
Mr Peter Pratt, Dr Manuel Sanchez, Mr Ciaran Regan, Mr Rozli
Zulkifli, Mr Stephen Crossby and Dr Ajoy Kundu.
REFERENCES
1. Federal Aviation Administration, Proceeding of the FAA International
Conference on Aircraft In-Flight Icing, Iand II, Springfield, Virginia,
USA, 6-8 May 1996. Final Report.
2. United States Department of Transportation, Federal Aviation
Administration. FAA In-Flight Icing Plan, April 1997.
3. THOMAS, S.K., CASSONI, R.P. and MACARTHUR, C.D. Aircraft anti-Icing
and deicing techniques and modeling, AIAA J Aircr, 33, (5), September-
October 1996, pp 841-854.
4. BROWN, J.M., RAGHUNATHAN, S., WATTERSON, J.K., LINTON, A.J. and
RIORDON, D. Heat transfer correlation for anti-icing systems, AIAA J
Aircr, 39, (1), January-February 2002, pp 65-70.
5. JAMBUNATHAN, K., LAI, E., MOSS, M.A. and BUTTON, B.L. A review of
heat transfer data for singular jet impingement, Int J Heat and Fluid
Flow, 1992, 13, pp 106-115.
6. MARTIN, H. Heat and mass Transfer between Impinging Gas Jets and Solid
Surfaces, Advances in Heat Transfer, 13, Academic Press, 1977, pp 1-60.
7. CORNARO, C., FLEISCHER, A.S. and GOLDSTEIN, R.J. Flow visualisation
of a round jet impinging on cylindrical surfaces, Experimental Thermal
and Fluid Science, 1999, 20, pp 66-78.
8. METZGER, D.E., YAMASHITA, T. and JENKINS, C.W. Impingement
cooling of concave surfaces with high velocity impinging air jets, J Eng
for Power-Transactions of the ASME, 1969, 91, pp 149-158.
9. DYBAN, E.P. and MAZUR, A.I. Heat transfer for a planar jet striking a
concave surface, translated from Inzhenerno-Fizicheskii Zhurnal, 17,
(5), November 1969, pp 785-790.
7.0 CONCLUSIONS
The School of Mechanical and Aerospace Engineering at the QUB
and Bombardier have, in recent years, been conducting research into
some of the key aerodynamic safety technologies for the next gener-
ation of aircraft engine nacelles. Investigations have been performed
into anti-icing technology, efficient thrust reversal, engine fire zone
safety, noise attenuation and trade off between aerodynamic perfor-
mance and manufacturing cost.
Studies on the usefulness of the empirical relations developed for
hot-air jet impingement on a flat surface for a highly curved two-
dimensional surface showed that the flat-plate empirical relations are
inadequate for predicting heat transfer on curved surfaces. Extensive
experimental work was performed on a piccolo type of anti-icing
system. The results led to a unique correlation between average
Nusselt number and Reynolds number for such a system. The corre-
lation is independent of the diameter, number and the distance
between the holes and the impinging surface provided these are
within the known optimal range, depending mainly on mass flow per
unit area and weakly on jet spacing. The effect of normal vibration
on heat transfer is negligible and normal vibration of an engine is not
a key design parameter of an aircraft anti-icing system. When
compared with a steady jet piccolo system a pulse jet anti-icing
system could produce an enhancement in heat transfer of 10-20% or
reduce the bleed air requirements by 10-20%.
CFD analysis in conjunction with Finite Element Analysis was
performed on the Natural Blockage Thrust reverser. The results
showed that the quality of the flow field in the fan duct was signifi-
cantly improved and the flow instabilities eliminated, with the intro-
duction of a porous fairing in the duct. A new cascade design with
reduced number of blades eliminated the shock waves, produced a
10% reduction in weight and an improvement in the structural
performance in both maximum vane displacement and mid-vane
stress levels. However, there is a penalty of thrust reversal, which
was reduced by 9%. Cost and ease of manufacture will also have to
be taken into consideration before any firm conclusions are drawn
regarding the final cascade configuration. A cascade-less configu-
ration gave a reverse thrust performance that is 75% of the cascade
configuration. Consideration of the test results indicate that with
careful choice of diverter faring and transcowl geometry, slat profile
and slat position, a reverse thrust performance that is 90% of the
cascade vanes configuration is achievable.
An attempt was made to model fire propagation and heat transfer
in the fan compartment (fire zone one) of the BR710 nacelle. Both
experiments and calculations revealed consistent trends, indicating
hot-spots in which the air temperature rose to between 800°C and
1,100°C, while the air temperature in the rest of the zone remained
in the range 40°C to 100°C. Whole zone heating was not as
prevalent as was initially assumed.
Simulations were performed to begin validating the use of CFD
for predicting the dispersion of a fire suppressant inside the fire zone
of an engine nacelle. Accurate results were obtained for a model fire
suppressant dispersing in an idealised nacelle, provided adaptive
mesh refinement was used in the agent plume region. It was found
that using a T injection nozzle, rather than a circular nozzle,
improved the dispersion of the agent. The introduction of clutter
hinders the dispersion of the agent, and further studies are underway
to investigate the flow in a more realistic, cluttered geometry.
Flap type pressure relief doors with flap angles of 15° to 45°, in
5° increments and in the Mach mumber range 0⋅4 to 0⋅85 were
studied. Generally good agreement was obtained between
measurement and prediction of integral quantities such as DFR
and thrust coefficient. Predictions under estimated discharge by
between 5% and 20%, depending on pressure ratio; the thrust
coefficient was slightly over-predicted. A freely hinged,
weightless flap would achieve a trimmed balance in the range of
angles considered. Increasing Mach number decreases this angle,
while increasing pressure ratio increases it.
RAGHUNATHAN ET AL KEY AERODYNAMIC TECHNOLOGIES FOR AIRCRAFT ENGINE NACELLES 285
32. MAVRIPLIS, D.J. NSU3D (Version 3.4) A CFD package for external
aerodynamics using an unstructured navier-stokes multigrid solver,
Users Manual, Revision 2, September 2000.
33. MAVRIPLIS, D.J. and LEVY, D.W. Transonic drag prediction using an
unstructured multigrid solver, 40th Aerospace Sciences Meeting &
Exhibit, AIAA Paper 2002-0838, Reno, NV, USA, January 2002.
34. MAVRIPLIS, D.J. Aerodynamic drag prediction using unstructured
solvers, VKI Lecture Notes for short course on CFD-Based Drag
Prediction and Reduction, 3-7 February 2003, von Karman Institute for
Fluid Dynamics, Rhode St Genese, Belgium.
35. IZZO, A.J. An experimental investigation of the turbulent characteristics
of a boundary layer.
36. Flow over a vibrating plate – General Dynamics, Electric Boat Division
– Contract NONR-2512(00) May 1969.
37. ROBINSON, S.K. Coherent motions in the turbulent boundary layer – Ann
Rev Fluid Mech, 1991, 23, pp 601-639.
38. JUNG W.J., MANGIAVACCHI N. and AKHAVAN R. Suppression of turbu-
lence in wall bounded flows by high-frequency spanwise oscillations –
Phys. Fluids A, 1992, (4), pp 1605-1607.
39. LAADHAR, F, SKANDAJI, L. and MOREL, R. Turbulence reduction in a
boundary layer by a local spanwise oscillating surface – Phys Fluids,
October 1994, 6, (10).
40. TRUJILLO, S.M., BOGARD, D.G. and BALL, K.S. Turbulent boundary
layer drag reduction using an oscillating wall – AIAA Paper, 971870,
1997.
41. KENDALL, J.M. The turbulent boundary layer over a wall with
progressive surface waves – J. Fluid Mech, 1970, 41, (2), pp 259-281.
42. KRAMER, M.O. Boundary layer stabilisation by distributed damping – J.
A. Soc Naval Engrs, 1960, 72, pp 25-33.
43. GAD-EL-HAK, M. Introduction to flow control – in flow control – funda-
mentals and practices, GAD-EL-HAK, M., POLLARD, A., BONNET and
J.P., (Eds), Springer Verlag, 1998, pp 1-107, 1998.
44. WILBY, J.F. and GLOYNA, F.L. Vibration measurements of an airplane
fuselage structure – II. Jet Noise Excitation – J Sound Vib, 1972, 23,
(4), pp 467-486.
45. MAESTRELLO, L. Design criterion of panel structure excited by turbulent
boundary layer – J Aircr, July 1968, 5, (4), pp 321-328.
46. HUSSAIN A.K.M.F. and REYNOLDS W.C. The mechanics of an organised
wave in turbulent shear flow – J Fluid Mech., 41, (2), pp 241-258.
47. KLEBANOFF, P.S. Characteristics of turbulence in a boundary layer with
zero pressure gradient – NACA Report 1247.
48. HAMA, R. Turbulent boundary layer along a flat plate I & II. – Rep Inst
Science and Tech., Tokyo University, January 1947.
49. SCHULTZ-GRUNOW F. New frictional resistance law for smooth plates –
NACA Report TM 986, 1941.
50. REGAN, C., BENARD, E., RAGHUNATHAN, S., RIORDAN, D. and LINTON,
A., The effect of a wall normal vibration on a turbulent boundary layer,
AIAA Paper 2002-0946, 40th AIAA Aerospace Sciences Meeting and
Exhibit, 14-17 January 2002, Reno, NV, USA.
51. NEVINS, R.G. and BALL H.D. Heat transfer between a flat plate and a
pulsating impinging jet, National Heat Transfer Conference Boulder,
1961, CO, ASME.
52. KATAOKA, K. and SUGURO, M. The effect of surface renewal due to
large scale eddies on jet impingement heat transfer, Int J Heat and Mass
Transfer, 1987, 30, pp 559-567.
53. SAILOR, D.J., DANIEL, J.R. and QIANLI, F. Effect of variable duty cycle
flow pulsations on heat transfer enhancement for an impinging air jet,
Int J Heat and Fluid Flow, 1999, 20, pp 574-580.
54. MLADIN, E.C. and ZUMBRUNNEN, D.A. dependence of heat transfer to a
pulsating stagnation flow on pulse characteristics, J Thermophysics and
Heat Transfer, 1995, 9, (1), pp 181-192.
55. ZUMBRUNNEN, D.A. and AZIZ, M. Convective heat transfer enhancement
Due to intermittency in an impinging jet, J Heat Transfer, 1993, 115,
pp 91-98.
56. AZEVEDO, L.F.A., WEBB, B.W. and QUEIROZ, M., Pulsed air jet
impingement Heat Transfer, Experimental Thermal and Fluid Science,
1994, 8, pp 206-213.
57. FARRINGTON, R.B. and CLAUNCH, S.D. Infrared imaging of large
amplitude, low frequency disturbances on a planar jet, AIAA J, 1994,
32, pp 317-323.
58. LIU, T. and SULLIVAN, J.P. Heat transfer and flow structures in an
excited circular impinging jet, Int J of Heat and Mass Transfer, 39,
(17), pp 3695-3706.
59. ZULKIFILI, R., BENARD.E., RAGHUNATHAN, S. and LINTON, A. Effect of
pulse jet frequency on impingement heat transfer., AIAA 42th
Aerospace Sceinces Meeting, Reno. NV, USA. Paper 2004-1343.
10. YEOMAN, E.K. Efficiency of a bleed air powered inlet icing protective
system, 32nd Aerospace Sciences Meeting & Exhibit, AIAA Paper 94-
0717, Reno, NV, USA, January 1994.
11. CROCE, G., HABASHI, W.G., GUÉVREMONT, G. and TEZOK, F. 3D thermal
analysis of an anti-Icing device using FENSAP-ICE, 36th Aerospace
Sciences Meeting & Exhibit, AIAA Paper 98-0198, Reno, NV, USA,
January 1998.
12. DEMATTOS, B. and S., OLIVEIRA, G.L. Three-dimensional thermal
coupled analysis of a wing slice slat with a piccolo tube, 18th AIAA
Applied Aerodynamics Conference, AIAA Paper 2000-3921, Denver,
CO, USA, 14-17 August 2000.
13. MORENCY, F., TEZOK, F. and PARASCHIVOIU, I. Anti-icing system
simulation using CANICE, AIAA J Aircr, 36, (6), November-December
1999, pp 999-1006.
14. MORENCY, F., TEZOK, F. and PARASCHIVOIU, I. Heat and mass transfer in
the case of an anti-icing system modelisation, AIAA J Aircr, 37, (2),
March-April 2000, pp 245-252. Also AIAA Paper 99-0623, January
1999.
15. MORENCY, F., TESSIER, P., SAEED, F. and PARASCHIVOIU, I. Anti-icing
system simulation on multielement airfoil, CASI 46th Annual
Conference/DNP-ASIP, Montréal, Canada, May 1999, pp 463-470.
16. TRAN, P., BRAHIMI, M.T., SANKAR, L.N. and PARASCHIVOIU, I. Ice
accretion prediction on single and multi-element airfoils and the
resulting performance degradation, 35th Aerospace Sciences
Meeting & Exhibit, AIAA Paper 97-0178, Reno, NV, USA 6-9
January 1997.
17. TRAN, P., BRAHIMI, M.T., PARASCHIVOIU, I., PUEYO, A. and TEZOK, F.,
ice accretion on aircraft wings with thermodynamic effects, AIAA J
Aircr, 1995, 32, (2), pp 444-446.
18. PARASCHIVOIU, I., TRAN, P. and BRAHIMI, M.T. Prediction of the ice
accretion with viscous effects on aircraft wings, AIAA J Aircr, 31, (4),
July-August 1994, pp 855-861.
19. TRAN, P., BRAHIMI, M.T. and PARASCHIVOIU, I. Ice accretion on aircraft
wings, Canadian Aeronautics and Space J, September 1994, 40, (3), pp
185-192.
20. RUFF, G.A. and BERKOWITZ, B.M. User manual for the NASA Lewis ice
accretion code prediction code LEWICE, NASA Contractor Report
185129, May 1990.
21. WRIGHT, W.B. User manual for the improved NASA Lewis ice
accretion code LEWICE 1.6, NASA Contractor Report 198355, June
1995.
22. AL-KHALIL, K.M. Numerical simulation of an aircraft anti-icing system
incorporating a rivulet model for the runback water, Ph.D. Thesis,
University of Toledo, Ohio, USA, June 1991.
23. AL-KHALIL, K.M. and Potapczuk, M.G. Numerical modeling of anti-
icing systems and numerical comparison to test results on a NACA
0012 Airfoil, 31st Aerospace Sciences Meeting & Exhibit, AIAA Paper
93-0170, Reno, NV, January 1993.
24. AL-KHALIL, K.M., FERGUSON, T.W. and PHILLIPS, D.M. A hybrid anti-
icing ice protection system, 35th Aerospace Science Meeting &
Exhibit, AIAA Paper 97-0302, Reno, NV, January 1997.
25. MEOLA, C., CARLOMAGNO, G.M., RIEGEL, E. and SALVATO, F. An exper-
imental study of an anti-icing hot air spray-Tube System, 19th Congress
ICAS, Anaheim, CA, September 1994.
26. Fregeau, Study and Simulation of the Ice accretion on aircraft and
modeling of thermal anti-icing systems, M.S. Thesis, École
Polytechnique de Montréal, Quebec, Canada, May 2004.
27. SAEED, F., MORENCY, F. and PARASCHIVOIU, I. Numerical simulation of
a Hot-Air Anti-Icing Simulation, 38th Aerospace Sciences Meeting &
Exhibit, AIAA Paper 2000-0630, Reno, NV, USA, January 2000.
28. SAEED, F. and PARASCHIVOIU, I. Numerical correlation for local Nusselt
number distribution for hot-air jet impingement on concave surfaces,
Proceedings of the 8th Annual Conference of the CFD Society of
Canada, CFD2K, Montréal, Québec, Canada, 11-13 June 2000, 2, pp
897-904.
29. CROCE, G., BEAUGENDRE, H. and HABASHI, W.G. CHT3D: FENSAP-
ICE Conjugate heat transfer computations with droplet impingement
and runback water, 40th Aerospace Sciences Meeting & Exhibit, AIAA
Paper 2002-0386, Reno, NV, USA, January 2002.
30. PUEYO, A., CHOCRON, D., MOKHTARIAN, F. and KAFYEKE, F. CHT2D: A
2D hot air anti-icing analysis tool, Proceedings of the 50th Annual
General Meeting and Conference of CASI, Montreal, Canada, 28-30
April 2003.
31. MAVRIPLIS, D.J. and PIRZADEH, S. Large-scale parallel unstructured
mesh computations for 3D high-lift analysis, ICASE Report No. 99-9,
February 1999.
286 THE AERONAUTICAL JOURNAL MAY 2006
80. DEVINE, R.J. WATTERSON, J.K. and COOPER, R.K. Performance
improvement of flush, parallel walled auxiliary intakes by means of
vortex generators, 24th International Congress of the Aeronautical
Sciences, Paper ICAS2004-4.2.3, August 2004, Japan.
81. DEVINE, R.D. The Performance of Nacelle Ventilation Intakes at Low
Speed, PhD thesis, Queens University Belfast, 2003.
82. PRATT, P, WATTERSON, J.K. and BENARD, E., Performance of a flapped
duct exhausting into a compressible external flow, 24th International
Congress of the Aeronautical Sciences, Paper ICAS2004-04.2.R,
August 2004, Japan. Augustine, N. Autustine Laes, 6th Edition, AIAA,
Reston, VA, 1997.
83. Federal Aviation Administration, General guidelines for measuring fire-
extinguishing agent concentration in powerplant compartments,
Advisory Circular, AC No 20-100, 1977.
84. United Nations, Report of the Halon Fire Extinguishing Agents
Technical Options Committee, United Nations Environment
Programme (UNEP), 1994.
85. PITTS, W., NYDEN, M., GANN, R., MALLARD, W. and TSANG, W.
Construction of an exploratory list of chemicals to initiate the search for
Halon alternatives, NIST TN-1279, National Institute of Standards and
Technology, 1990.
86. GROSSHANDLER, W., GANN, R. and PITTS, W. (Eds) Evaluations of alter-
native in-flight fire suppressants for full-scale simulated aircraft engine
nacelles and dry bays, SP-861, National Institute of Standards and
Technology, 1994.
87. GANN, R. (Ed), Fire suppression system performance of alternative
agents in aircraft engine and dry bay laboratory simulations, SP-890,
National Institute of Standards and Technology, 1995.
88. HAMINS, A., CLEARY, T., BORTHWICK, P., GORCHKOV, N., MCGRATTAN,
K., FORNEY, G., GROSSHANDLER, W., PRESSER, C. and MELTON,L.
Suppression of engine nacelle fires, fire suppression system perfor-
mance of alternative agents in aircraft engine and dry bay laboratory
simulations, edited by GANN, R., Vol 2 of SP-890, chap. 9, National
Institute of Standards and Technology, 1995, pp 1-199.
89. LOPEZ, A., GRITZO, L. and HASSAN, B. Computational fluid dynamics
simulation of the air/suppressant flow in an uncluttered F18 engine
nacelle, Proc. Halon Options Tech. Working Conference, Albuquerque,
New Mexico, USA, 1997, pp 281-297.
90. NICOLETTE, V., LOPEZ, A. and GRITZO, L. F18 Nacelle fire simulation,
Sandia National Laboratories Technical Memorandum to Leo Budd,
1997.
91. VICK, A.R. An investigation of discharge and thrust characteristics of
flapped outlets for stream Mach numbers from 0⋅4 to 1⋅3, NACA TN-
4007, 1957.
92. ACARE. European Aeronautics: A vision for 2020 – meeting societys
needs and winning global leadership, Report of the group of personal-
ities, January 2001.
93. BARRY, B., PARKE, S.J., BOWN, N.W., RIEDEL, H. and SITZMANN, M.
Flight testing of natural and hybrid laminar flow nacelles, ASME
Proceedings of the International Gas Turbine and Aeroengine Congress
and Exposition, 13-16 June 1994, Hague, Neth, ASME 94 -GT-408.
94. BRODERSEN, O. Drag prediction of engine-airframe interference effects
using unstructured Navier-Stokes calculations, J Aircr, 2002, 39, (6),
2002, pp 927-935.
95. ROSSOW, C.C., GODARD, J.–L., HOHEISEL, H. and SCHMITT, V.
Investigation of ropulsion integration interference effects on a transport
aircraft configuration, J Aircr, 1994, 31, (5), pp 1022-1030.
96. HENDERSON, W.P., Airframe/Propulsion Integration at Transonic
Speeds, J Engineering for Gas Turbines and Power, Transactions of the
ASME, v 113, n 1, January 1991, pp 51-59.
97. GODARD, J.L., HOHEISEL, H., ROSSOW, C.C. and SCHMITT, V.
Investigation of Interference Effects for Different Engine Positions on a
Transport Aircraft Configuration, Proceedings of DLR Workshop on
Aspects of Engine Airframe Integration For Transport Aircraft, edited
by H. HOHEISEL, German Aerospace Research Centre, DLR, TR 96-01,
Brunswick, Germany, 1996, pp 11.1-11.22.
98. BERRY, D. The Boeing 777 engine/airframe integration aerodynamics
design process, international council of the aeronautical sciences, 94-
6.4.4, September 1994.
99. RUDNIK, R. and ROSSOW, C.C. Numerical simulations of
engine/airframe integration for high-bypass engines, ECCOMAS –
2000 Paper, 2000.
100. INGRALDI, A.M., KARIYA, T.T., RE, R.J. and PENDERGRAFT, O.C.
Interference effects of very high bypass ratio nacelle installations on a
low-wing transport, J Engineering for Gas Turbines and Power,
Transactions of the ASME, v 114, 4, October 1992, pp 809-815.
60. LORD, M.W.K., MACMARTIN, P.G. and TILLMAN, T.G. Flow control
opportunities in gas turbine engines, AIAA, 2000-2234, 2000.
61. POLAND, D.T. The aerodynamics of thrust reversers for high bypass
Turbofans, AIAA, 67-418, July 1967.
62. YAO, H., BENARD, E., COOPER, R.K., RAGHUNATHAN, S., TWEEDIE, J.
and RIORDAN, D. Aerodynamics of natural blockage thrust reverser,
proceedings of CASI 50th AGM and Conference, 28-30 April,
Montreal, QC, Canada, Paper 329, 2003.
63. KRAL, L.D. Recent experience with different turbulence models applied
to the calculation of flow over aircraft components, Progress in
Aerospace Sciences, 1998, 34, pp 481-541.
64. ROMINE, Jr B.M. and JOHNSON, W.A. Performance investigation of a fan
thrust reverser for a high bypass turbofan engine, AIAA-84-1178,
AIAA/SAE/ASME 20th Joint Propulsion Conference, Cincinnati, USA,
June 1984.
65. YAO, H., BUTTERFIELD, J., RAGHUNATHAN, S., COOPER, R. and BENARD,
E. Optimization design for thrust reverser cascade with a view to
Reduce Noise, Accepted for publication in Int J Multidiscipline
Modeling in Materials and Structures.
66. YAO, H., BUTTERFIELD, J. and RAGHUNATHAN, S. The aerodynamic
performance of thrust reverser cascade, Proceeding of the 24th
International Congress of Aeronautical Sciences (ICAS 2004), ICAS
2004-4.2.R, Yokohama, Japan, August 2004.
67. YAO, H., BENARD, E., CURRAN, R., PRICE, M., ARMSTRONG, C.G. and
RAGHUNATHAN, S. Integration of aerodynamic, structural, cost and
manufacturing considerations during the conceptual design of a thrust
reverser cascade, AIAA Paper 2004-1239, 42nd AIAA Aerospace
Sciences Meeting and Exhibit, Reno, Nevada, USA, January 2004.
68. HALL, S., YAO, H., COOPER, R.K., BENARD, E. and RAGHUANTHAN, S.
Experimental investigation of a cascade thrust reverser, 9th
Aerodynamics Symposium, 239b, Montreal, Canada, 2003.
69. HALL, S., BENARD, E. and RAGHUNATHAN, S. Progress in developing
innovative flow control in a cascade thrust reverser, 24th International
Congress of the Aeronautical Sciences, 2004.
70. YAO, H., BENARD, E., COOPER, R. K., RAGHUNATHAN, S., TWEEDIE, J.
and RIORDAN, D. Aerodynamics of natural blockage thrust reverser. 9th
Aerodynamics Symposium, Montreal, Canada, 28-30 April 2003.
71. BUTTERFIELD, J., YAO, H., PRICE. M., RAGHUNATHAN, S. and
ARMSTRONG, C. Methodologies for structural optimisation of a thrust
reverser cascade, AIAA Paper 2003-0107, 41st AIAA Aerospace
Sciences Meeting and Exhibit, 6-9 January 2003, Reno. NV, USA.
72. BUTTERFIELD, J., YAO, H., BENARD, E., PRICE, M., COOPER, R.,
MONAGHAN, D., ARMSTRONG, C.G. and RAGHUNATHAN. S. Optimisation
of thrust reverser cascade performance using aerodynamic and structural
integration. Proceedings of CEAS Aerospace Aerodynamics Research
Conference, 10-12 June, 2003, Royal Aeronautical Society, London.
73. BUTTERFIELD, J., YAO, H., BENARD, E., PRICE, M., COOPER, R.,
MONAGHAN, D., ARMSTRONG and C.G., RAGHUNATHAN. S. Investigation
of weight reduction in a thrust reverser Cascade Using Aerodynamic
and Structural Integration. AIAA Paper 2003-1576. 44th
AIAA/ASME/AHS Structures, Structural Dynamics and Materials
Conference, 7–10 April, 2003, Norfolk, Va, USA.
74. BUTTERFIELD, J., YAO, H., BENARD, E., PRICE, M., COOPER, R.,
MONAGHAN, D., ARMSTRONG, C.G. and RAGHUNATHAN. S. Optimisation
of a thrust reverser cascade: An assessment of dynamic response with a
view to reducing weight. AIAA Paper 2003-6748. 3rd Annual
Aviation Technology, Integration and Operations (ATIO) Technical
Forum, 17–19 November 2003, Denver, Colorado, USA.
75. BUTTERFIELD, J., YAO, H., CURRAN, R., PRICE, M., ARMSTRONG, C.G.
and RAGHUNATHAN. S. Integration of aerodynamic, structural, cost and
manufacturing considerations during the conceptual design of a thrust
reverser cascade. AIAA Paper 2003. 42nd AIAA Aerospace Sciences
Meeting and Exhibit, 5-8 January 2004.
76. YAO. H., RAGHUNATHAN, BENARD, E. and COOPER, R. Numerical
simulation of natural blockage thrust reverser. AIAA 43rd Aerospace
Sciences meeting. Reno Nevada, USA, January 2005. Paper 2005-0631.
77. BUTTERFIELD, J., YAO, H., PRICE, M., BENARD, E., COOPER, R.,
MONAGHAN, D., ARMSTRONG, C. and RAGHUNATHAN, S. Enhancement
of thrust reverser cascade performance using aerodynamic and struc-
tural integration, Aeronaut J, 2004, 108, (1090), pp 621-628.
78. DONAGHY, K. Fire Propagation and Heat transfer Modelling Within The
BR710 Nacelle for Certification Purposes, PhD thesis, Queens
University Belfast, 2000.
79. DONAGHY, K., RAGHUNATHAN, S. and RIORDON, D. Fire zone modelling
of aircraft engines, 37th AIAA Aerospace Sciences Meeting, 1999,
Paper 99- 0236, Reno, NV, USA.
RAGHUNATHAN ET AL KEY AERODYNAMIC TECHNOLOGIES FOR AIRCRAFT ENGINE NACELLES 287
288 THE AERONAUTICAL JOURNAL MAY 2006
101. ROSSOW, C.C. and HOHEISEL, H. Numerical study of interference effects
of wing-mounted advanced engine Concepts, International Council of
the Aeronautical Sciences, 94-6.4.1, September 1994.
102. JIE, LI., FENGWEI, LI. and QIN, E. Numerical simulation of transonic
flow over wing-mounted twin-engine transport aircraft, J Aircraft,
2000, 37, (3), pp 469-478.
103. RUDNIK, R., ROSSOW, C.C. and V. GEYR, H.F. Numerical simulations of
engine/airframe integration for high-bypass engines, Aerospace Science
and Technology, 2002, 6, pp 31-42.
104. MURMAN, E.M., WALTON, M. and REBENTISCH, E. Challenges in better,
faster, cheaper era of aeronautical design, engineering and manufac-
turing, Aeronaut J, Oct 2001, pp 481-489.
105. KUNDU, A., RAGHUNATHAN, S. and MORRIS, W. Effect of manufacturing
Tolerances on aircraft aerodynamics and cost world manufacturing
congress. New Zealand, November 1997. Paper 711-003.
106. KUNDU, A., RAGHUNATHAN, S. and COOPER, R.K. Effect of aircraft
smoothness requirements on cost, Aeronaut J, December 2000, 104,
(1039), Paper 2389, pp 415-420.
107. CURRAN, R., KUNDU, A., RAGHUNATHAN, S. and MCSPADDEN, R. Impact
of the aerodynamic surface tolerance on aircraft cost driver, J
aerospace engineering, Proc of the IMechE, Part C, 215, 2001, pp 29-
39.
108. CURRAN, R., KUNDU, A., RAGHUNATHAN, S. and EAKIN, D. Costing tools
for decision making within integrated aerospace design, concurrent
engineering: research and applications, 2001, 9, (4), pp 327-338.
109. KUNDU, A., WATTERSON, J.K., RAGHUNATHAN, S. and MCFADDEN, R.
Parametric optimisation of manufacturing tolerances at the aircraft
surface, J Aircr,AIAA, 2002, 39, (2), pp 271-279.
110. CURRAN, R., KUNDU, A., RAGHUNATHAN, S. and MCFADDEN, R.
Influence of manufacturing tolerance on aircraft direct operating cost
(DOC), J Materials Processing Technology, Elsevier Science B.V.,
ISSN: 0924-0136, 138, 2002, pp 208-213.
111. SANCHEZ, M., KUNDU, A.K., HINDS, B.K. and RAGHUNATHAN, S. A
methodology for assessing manufacturing cost due to tolerance on
aerodynamic surface features on turbo fan nacelles, international J
advanced manufacturing Technology, 1999, 14, pp 894-900.
