Available via license: CC BY 4.0
Content may be subject to copyright.
Copyright © 2020 e Author(s). Published by VGTU Press
is is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unre-
stricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
*Corresponding author. E-mail: vilius.portapas@uwe.ac.uk
SIMULATED PILOT-IN-THE-LOOP TESTING OF HANDLING QUALITIES
OF THE FLEXIBLE WING AIRCRAFT
Vilius PORTAPAS 1, *, Alastair COOKE 2
1Department of Engineering Design and Mathematics, University of the West of England, Bristol, United Kingdom
2Centre for Aeronautics, Craneld University, Craneld, United Kingdom
Received 20 November 2019; accepted 09 February 2020
Abstract. is article aims to indicate the dierences between rigid and exible wing aircra ying (FQ) and handling
(HQ) qualities. e Simulation Framework for Flexible Aircra was used to provide a generic cockpit environment and a
piloted mathematical model of a bare airframe generic high aspect ratio wing aircra (GA) model. ree highly qualied
test pilots participated in the piloted simulation trials campaign and ew the GA model with both rigid and exible wing
congurations. e results showed a negligible dierence for the longitudinal HQs between rigid and exible wing aircra.
However, signicant changes were indicated for the lateral/directional HQs of the exible wing aircra. A wing ratcheting
phenomenon manifested itself during the roll mode tests, the spiral mode exhibited neutral stability and the Dutch roll
mode shape changed from a horizontal to a vertical ellipse. e slalom task ight tests, performed to assess the FQs of the
aircra, revealed the degradation of both the longitudinal and lateral/directional FQs.
Keywords: aeroelasticity, exible aircra, ight dynamics, handling qualities, piloted simulation trials.
Introduction
Over the last few decades society has witnessed signicant
achievements in aviation. Air travel has become so aord-
able that the number of passengers is expected to double
in the next 20 years and airlines keep introducing longer
revenue ights, which range for about 15000 km and last
for more than 17 hours, to connect the most distant places
of our planet. Economic and social benets that the avia-
tion industry induces globally are greatly appreciated and
understood (Anonymous, 2017; ATAG, 2016). However,
they come at a cost of a negative environmental impact,
for which the aviation sector is highly criticised. To reduce
this impact, international aviation organizations such as
the International Civil Aviation Organization (ICAO), the
International Air Transport Association (IATA) and the
European Civil Aviation Conference (ECAC) set environ-
mental targets to signicantly reduce emissions (IATA,
2013; Tollefson, 2016) that can be achieved only by im-
proving aircra eciency. One of the potential improve-
ments is to increase the wing aspect ratio (AR) to reduce
the induced drag. High AR unswept wings are usually
seen in sailplane designs to provide very high li-to-drag
ratios, but could not be used for large transport aircra
in the past because of arising structural issues due to the
much higher wing loading when compared to sailplanes.
With the advent of composite materials and novel manu-
facturing technologies (Kellari, Crawley, & Cameron,
2018), previous structural issues have been solved. As
AR increases, larger wing bending deformations develop.
us, understanding their eect on ying (FQ) and han-
dling (HQ) qualities is vital, as they will aect design con-
siderations of airframes and ight control systems (FCS).
Previous approaches to assess HQs of exible air-
cra were mainly based on the longitudinal dynamics
(Andrews, 2011; Damveld, 2009; Field & Rossitto, 1999;
Waszak & Schmidt, 1988). As pilots, placed in front of
an aircra, experienced dierent dynamics than the ones
at the centre of gravity (CG) due to the bending of long
slender fuselages, it was important to investigate the way
those deformations altered pilots’ perception of HQs. One
of such assessments was done by Waszak, Davidson, and
Schmidt (1987) at the NASA Langley Visual/Motion Sim-
ulator facility for the Rockwell B-1 aircra. Based on this
research Damveld (2009) developed a new method to in-
vestigate and quantify the longitudinal HQs – experimen-
tal behaviour measurement method. Andrews (2011), on
the other hand, investigated the impact of both the fuse-
lage and the wing exibility on HQs of a large commercial
AVIATION
ISSN: 1648-7788 / eISSN: 1822-4180
2020 Volume 24 Issue 1: 1–9
https://doi.org/10.3846/aviation.2020.12175
2V. Portapas, A. Cooke. Simulated pilot-in-the-loop testing of handling qualities of the exible wing aircra
transport aircra. He also developed the AX-1 model –
the foundational part of the Craneld Accelerated Aircra
Loads Model (CA2LM) framework used for the modelling
and simulation activities described in this paper. However,
the commercial aircra that Andrews investigated pos-
sessed a relatively rigid wing – a feature of large commer-
cial aircra until the Boeing 787 (Dodt, 2011) and Airbus
A350 were introduced. As large aircra were tted with
relatively rigid wings, the unmanned aerial vehicles (UAV)
served as test beds for high AR wing (HARW) technolo-
gies. Much eort was made in the USA to develop meth-
odologies and technologies to both manufacture and in-
vestigate HARW UAVs. e Helios Prototype (Noll etal.,
2004) is the most famous example of HARW UAV. Helios
is also a well-known example of the in-ight structural
disintegration due to high wing deformations. As the man-
ufacturing technologies advanced, the researchers started
to look into HARW applicability to large commercial air-
cra. e Subsonic Ultra Green Aircra Research, initi-
ated by Boeing and NASA, is a well-known investigation
into the applicability of high AR low swept wings for com-
mercial aircra (Bradley, Droney, & Allen, 2015). Airbus
along with its academic partners Craneld University and
the University of Bristol proposed the High Aspect Ratio
Technology ENabler (Cooper etal., 2014). Both concepts
consider HARW as well as many other novel technologies
designed to enable improvements in aircra eciency.
ese new HARW concepts not only increase aircra
eciency, but also introduce new aircra dynamics. As
the structural issues of HARW were analysed in the Helios
report (Noll etal., 2004), the FQs and HQs of such aircra
have never been assessed before. Hence, piloted simula-
tion trials were carried out during the research activities
described in this article to compare the dierences of FQs
and HQs between relatively rigid and highly exible wing
aircra through a set of standard ight test manoeuvres
and a recently developed slalom task. e Simulation
Framework for Flexible Aircra (SFFA) was developed to
perform the trials. Moreover, the opinions collected from
professional test pilots, who participated in the simulation
campaign, allowed an assessment of both FQs and HQs of
a generic HARW aircra model (GA).
is article reveals signicant changes in the lateral/
directional dynamics between rigid and exible wing air-
cra congurations. It also proves the suitability of the
SFFA to be used as a pilot-in-the-loop testing tool of new
aircra congurations. Lastly, it proves the suitability of
the slalom task as a rapid ight testing technique that is
capable to reveal the deciencies in FQs and HQs of new
aircra designs.
is article consists of two sections. Section 1 briey
reviews SFFA, the experience of test pilots, who partici-
pated in the simulation campaign, and denes the GA
model along with the ight test manoeuvres used for the
piloted simulation trials campaign. Section 2 reviews the
results for both longitudinal and lateral/directional HQs as
well as the results from the slalom task simulations used
to assess FQs.
1. Methodology
e simulation campaign was carried out using the newly
developed SFFA at the Aerospace Integration Research
Centre at Craneld University. ree certied test pilots
ew exible and rigid wing congurations of the GA
model through a series of ight test manoeuvres to induce
longitudinal and lateral/directional modes. ese piloted
simulation trials allowed to assess HQs of aircra congu-
rations. Pilots also assessed FQs during the slalom task
with respect to the Cooper-Harper Rating (CHR) scale
Figure 1. e architecture of SFFA
Aviation, 2020, 24(1): 1–9 3
(Cooper & Harper, 1969). Comparison of the dynamics
of exible and rigid wing GA allowed identication of the
dierences between the two congurations.
1.1. Simulation framework for exible aircra
SFFA consists of three main components, which are the
CA2LM framework, the engineering flight simulator
EFS500 and the Speedgoat real-time target machine that
connects the previous two components. Figure 1 shows
the overall architecture of SFFA.
e CA2LM framework (Dussart et al., 2018; Por-
tapas, Cooke, & Lone, 2016) provides an environment
for the aeroservoelastic modelling of exible aircra.
e Leishman (1988, 1993, 1994) unsteady aerodynam-
ics model that makes use of eodorsen (1949) and
Wagner (1925) functions along with the Modied Strip
eory (DeLaurier, 1993) denes aerodynamics of the
wing, horizontal (HTP) and vertical (VTP) tailplanes.
Such a modelling approach combines the capability of
real-time pilot-in-the-loop simulations, whilst capturing
the aerodynamic eects necessary for the aeroservoe-
lastic analysis of an aircra (Kim etal., 2008). For the
current test case NASA SC(2)-0610 aerofoil was used to
model aerodynamics of the wing and NASA SC(2)-0010
aerofoil to model aerodynamics of HTP and VTP. Fu-
selage exibility eects are divided into two parts and
its forebody is modelled as a slender axisymmetric fore-
body-cylinder combination (ESDU, 1990, 2004), while
the abody is modelled as a conical boat-tail (ESDU,
1992). Engine nacelles are modelled as annular aerofoils
(ESDU, 2013). Once calculated the aerodynamic load-
ing is then transferred to the structural frame to get the
airframe deformations. e structural dynamics in the
CA2LM framework use the modal form, which means
that the overall deformation of the structure depends
on the sum of the modeshapes, and is the implemen-
tation of the linear structural dynamics model. Widely
accepted opinion states that wingtip deections of less
than 10% of semi-span can be treated as linear deforma-
tions. However, large wing deformations are of interest
in this research and a nonlinear structural model should
be used instead. Nonetheless, Patil and Hodges (2004)
showed that the linear structural model predicts defor-
mations with similar accuracy as the nonlinear structural
model for the static wingtip deections of around 25%
semi-span. However, for the dynamic loading case the
linear structural model could not capture same dynam-
ics as the nonlinear one (Patil & Hodges, 2004). Due to
this limitation current version of the CA2LM framework
could only be used for investigations of low frequency
phenomena. Hence, the assumption of low frequency
wing deformations was made.
e EFS500 ight simulator provides a cockpit en-
vironment and an image generation capability enabling
pilot-in-the-loop simulations. Although EFS500 was de-
scribed previously by Lopez Matos etal. (2018), the struc-
ture of the simulator is briey covered here. e EFS500
ight simulator consists of the following components (Al-
lerton, 2016):
1. Input/output system is based on the Raspberry Pi
platform and performs an analogue-digital signal
conversion and broadcasts between the computer
system and the inceptors.
2. Flight dynamics computer calculates the state and
position parameters of a simulated aircra. How-
ever, for the current case, the CA2LM framework
overwrote its calculations to allow exible wing air-
cra simulations, but the computer still provided
the primary ight display.
3. Cockpit environment provides a generic single seat
cockpit layout with a sidestick, rudder pedals and
throttle lever. Basic pilot instruments, i.e. the pri-
mary ight display and the electronic horizontal
situation indicator, are also provided.
4. Articial loading system provides force feedback to
the rudder pedals.
5. Computer system of six PC stations provides cockpit
displays, avionics, instructor operating system and
image generation capabilities.
6. Sound generation system generates environment and
warning sounds.
7. Image generation system consists of three projectors
and a spherical 140° by 40° eld-of-view screen.
All computers are connected into a local network by
a 16-channel Ethernet switch and run at the frequency of
50 Hz.
e Speedgoat real-time target machine – another es-
sential component of the system – speeds up the CA2LM
framework and provides an interface between the frame-
work and the EFS500 ight simulator. e CA2LM frame-
work is set to run at the frequency of 1000 Hz, hence the
information is sent and received to/from the EFS500 ight
simulator at every 20th time step.
1.2. Generic high aspect ratio wing aircra model
Figure 2 shows the generic HARW aircra model used
during the piloted simulation trials. It represents a medi-
um-size T-tail conguration aircra model with two en-
gines mounted at the back of the fuselage.
Figure 2. Generic HARW aircra model
4V. Portapas, A. Cooke. Simulated pilot-in-the-loop testing of handling qualities of the exible wing aircra
e model was used as a test case of the highly deform-
able wing aircra, which is referred to as a exible aircra in
this article. e same model, but with an increased stiness
of the wing, was used as a test case of the relatively rigid
wing aircra, which is referred to as a rigid aircra in this
article. GA was simulated as a bare airframe, i.e. without
FCS. Its mass was set to 79 t and distributed in a way that the
CG would be at 20% of its mean aerodynamic chord for the
piloted simulation trials. However, the dierence from other
similar size aircra lies within the wing parameters, i.e. its
quarter chord sweep angle (
λ
c/4) is 0.46° and its AR is 17.7.
us, it qualies as low-swept HARW aircra. e exibil-
ity of the fuselage, HTP and VTP was increased by setting
the Young’s modulus to innity. Hence, these components
of GA are considered as non-deformable throughout this
article. e Young’s modulus of the rigid wing was chosen to
be 80 GPa, representing the usual values of aluminium alloys
(Ashby, 2017), and 8 GPa for the exible wing to consider-
ably increase deformations for the purpose of this research.
Figure 3 shows resulting trim state wing deections, when
the aircra was trimmed for the cruise ight at 10,000 and
204 kt. e same trim ight conditions were used as initial
conditions to start each simulation.
1.3. Flight test manoeuvres
Several ight test manoeuvres were performed to identify
the dierences between rigid and exible aircra congu-
rations. Firstly, the wing exibility eect on the longitudi-
nal HQs was assessed through the short period pitching
oscillation (SPPO) mode and the phugoid mode tests. e
typical period of the SPPO mode is 1–2 s (Stinton, 1996),
which translates into the frequency of 0.5–1 Hz. Figure 4
shows that this frequency range overlaps with the lowest
modeshape frequencies of the exible wing aircra as well
as human natural frequencies. Hence, it is essential for the
SPPO mode be well damped. On the other hand, the phu-
goid mode is of much longer period, i.e. 40–100 s (Stin-
ton, 1996). en, the wing exibility eect on the lateral/
directional HQs was assessed through the roll mode, the
spiral mode and the Dutch roll mode tests. e roll mode
is a non-oscillatory mode dened by the time constant τr,
which typically should be less than a second. e spiral
mode is also a non-oscillatory mode dened by the time
constant T2, which equals the time needed to double the
amplitude of the bank angle
Φ
and is usually as large as 40
s. Being a long period mode it is of a little interest for HQs
research as long as T2 is more than 20 s or the period of
the Dutch roll mode (Stinton, 1996). e Dutch roll mode
is the only oscillatory lateral/directional mode (Ward &
Strganac, 2001) dened by the frequency, damping and
the ratio of bank angle to sideslip angle Δ
Φ
/Δβ. Low ratio,
when sideslip dominates the mode, is preferred by pilots.
e aircra response was also compared against the
MIL-F-8785C requirements (Moorhouse & Woodcock,
1981), for which the GA model qualied as class II (medi-
um transport) aircra and the initial ight test conditions
corresponded to the category A ight phase (V ≥ 1.2VS
and 10,000 altitude).
Pilots used slalom task (Dussart etal., 2018) to test
FQs of GA. e task was developed to assess an aircra-
pilot couple’s capability to perform real-life high-gain
tasks. It aims to replicate an oset landing manoeuvre at
a higher altitude to mitigate the deciencies of SFFA, i.e.
the absence of the ground eect and landing gear. Hence,
the pilots had to y through the spheres, as shown in Fig-
ure 5, which allowed assessment of whether a aircra-pilot
couple achieved desired (inner blue sphere) or adequate
(outer orange sphere) performance.
e size of the spheres was dened according to usual
runway dimensions, i.e. the diameter of 66 for the small-
er sphere represented the width of a usual touchdown
zone, while the diameter of 262 for the larger sphere
represented the width of the widest runway at large aero-
dromes. For the distance considerations, it was assumed
that the pilot’s initial visual contact with the runway was
at 200 altitude, which represents the CAT I ILS mini-
mum, and an aircra was o the runway centreline by
2.5°, which is the full scale deection of the ILS indicator.
is led to the lateral oset of 430 and the longitudinal
separation of 7300 , also considering the fact that the
aircra was own in a clean conguration at a higher ve-
locity than it would experience when landing.
Figure 3. Trim state wing deformations for rigid (–) and
exible (- - -) wing aircra. e shape of Boeing 787 aircra in
trimmed ight (Dodt, 2011) is given for illustration purpose.
Here: Δz – wingtip deection in vertical direction under the
load; b – wing semi-span;
η
– relative position along the wing
Figure 4. Frequencies of natural aircra and aeroelastic modes
for rigid and exible wing aircra
Aviation, 2020, 24(1): 1–9 5
2. Results and discussion
e SPPO mode tests are indicated as SP in the follow-
ing tables, the phugoid mode – PH, the roll mode – RM,
the spiral mode – SM, the Dutch roll mode – DR and
the slalom task – SL. e rigid aircra conguration is
indicated as (R) and the exible aircra conguration is
indicated as (F).
2.1. Longitudinal HQs
Two sets of tests inducing the short period pitching oscil-
lation and the phugoid modes were carried out to identify
the dierences between the longitudinal HQs of rigid and
exible wing aircra. Although the phugoid mode mani-
fests itself as a trimming problem and usually is of a little
interest for HQs assessment, pilots continuously reported
it as the dominant and intrusive longitudinal mode during
piloted simulation trials presented in this article.
Both longitudinal modes were induced by the elevator
deection. e SPPO mode was induced by the doublet
input (see Figure 6) and the phugoid mode was induced
by the pulse input (see Figure 7). Response parameters of
both modes are summarised in Table1.
Figure 5. Denition of the slalom task
Table1. SPPO and phugoid modes’ parameters. Here: T – period;
ω
d – damped frequency;
ω
n – natural frequency;
ζ
– damping
ratio; n
α
– acceleration sensitivity; CAP – Control Anticipation
Parameter
Test SP(R) SP(F) PH(R) PH(F)
T [s] 3.9 3.6 52.1 55.5
ω
d [Hz] 0.26 0.28 0.02 0.02
ω
n[Hz] 0.28 0.29 0.02 0.02
[rad/s] 1.76 1.82
ζ
[—] 0.38 0.32 0.02 0.03
n
α
[1/rad] 4.58 8.59 – –
CAP 0.68 0.39 – –
e results show that the rigid aircra exhibited higher
damping of the SPPO mode. Although pilots assessed the
SPPO mode of both rigid and exible aircra as heav-
ily damped, the damping factor of the rigid aircra met
Level 1 requirement (0.35 ≤
ζ
sp ≤ 1.30), while the damp-
ing factor of the exible aircra met Level 2 requirement
(0.25 ≤
ζ
sp ≤ 2.00). Both rigid and exible aircra met
Level 1 requirements with respect to the natural frequency
ω
n and the acceleration sensitivity nα of the SPPO mode.
e Control Anticipation Parameter (CAP), which com-
bines both
ω
n and nα and is expressed as (Cook, 2013):
2
n
CAP n
α
ω
=
, (1)
showed a reduced value for the exible aircra. Lower
CAP value means that the exible aircra’s response to
a pilot’s input is more sluggish and could potentially lead
to an overshoot when a pilot makes corrective action to
achieve targeted attitude. e values of
ω
n, nα and CAP
along with the limits for Levels 1, 2 and 3 are graphically
represented in Figure 8.
e exhibited period of the SPPO mode was about
twice longer when compared to the typical period length.
e mode induction technique, i.e. the elevator doublet
Figure 6. SPPO mode of rigid (–) and exible (- - -) wing
aircra. Here:
δ
e – elevator deection;
α
– angle of attack
Figure 7. Phugoid mode of rigid (–) and exible (- - -) wing
aircra. Here:
δ
e – elevator deection;
θ
– pitch attitude;
EAS – equivalent air speed
Figure 8. Mapping of rigid (x) and exible (o) wing aircra
congurations CAP parameter against SPPO mode frequency
requirements for ight phase category A (Moorhouse &
Woodcock, 1981)
6V. Portapas, A. Cooke. Simulated pilot-in-the-loop testing of handling qualities of the exible wing aircra
input, could be the reason for such dierence. A brief trial
to apply the frequency sweep technique, which covers a
broad spectral range of frequencies, to measure the SPPO
mode parameters resulted in the period of Tsp= 1.6–1.8
s. However, the pilots assessed both aircra congurations
as “qualitatively similar” when using the frequency sweep
input.
CAP values were 0.68 for the rigid aircra and 0.39 for
the exible aircra. Figure 8 indicates both values within
Level 1 requirements.
e phugoid mode, on the other hand, showed a dif-
ferent change of the damping factor, i.e. an increase for
the exible aircra compared to the rigid aircra. e
damping factor of the mode was very low and met Level
2 requirement (0 ≤
ζ
ph ≤ 0.04). Low damping of the mode
also reected in the comments of the test pilots, where
they noted many times and also in other tests that the
exible aircra “started phugoiding without even making
an input”.
e tests of the longitudinal HQs showed negligible
dierences between rigid and exible aircra. Although
the longitudinal HQs degraded for the exible aircra, no
signicant dierences were recorded within the main air-
cra response parameters. Hence, it was concluded that
the two aircra congurations are qualitatively similar in
terms of longitudinal dynamics.
2.2. Lateral/directional handling qualities
e usual mode assessment techniques for all three lat-
eral/directional modes are dierent. Hence, the roll mode
was assessed by banking the aircra between Φ= ±30°.
e spiral mode was induced by banking the aircra to
Φ
= ±15°, stabilising it at the set bank angle and then re-
moving all control inputs to see further development of
the bank angle. e Dutch roll mode was induced by a
rudder doublet input.
e roll mode parameters of interest are presented in
Table2. Although the calculated roll mode time constant
was 0.5–0.6 s and met Level 1 requirement (
τ
r ≤ 1.4 s) for
all test cases, the roll was not smooth as it could be antici-
pated from the
τ
r parameter. An investigation showed that
another roll performance parameter
τ
±30°, which denes
the amount of time needed to roll an aircra between two
bank angles with opposite signs, was more than two times
greater for the exible aircra. It was noticed by pilots that
the reason for this was wing ratcheting induced by the in-
creased exibility and, thus, the dihedral of the wing. Fig-
ure 9 clearly indicates the wing ratcheting phenomenon.
Pilots expect that as long as the ailerons are deected the
aircra should continue rolling. However, the exible air-
cra stops rolling aer 3–3.5 s as indicated by roll ratep
in Figure 9. At this time frame the roll rate goes back to
zero from its previous maximum value. Such a poor roll
performance is an outcome of lost ailerons eciency due
to highly deformed wing. is phenomenon is highly un-
desirable in roll dynamics as it makes it a challenging task
for pilots to anticipate the behaviour of an aircra.
e spiral mode parameters are presented in Table 3.
e rigid aircra exhibited divergent spiral mode, which
would be typically expected for an aircra with a dihedral
low-wing conguration. According to the T2 parameter,
the rigid aircra met Level 1 requirement (T2 ≥ 12 s).
However, the exible aircra exhibited neutral spiral mode
stability, which is also shown in Figure 10. is stabilis-
ing eect is a clear evidence of the exibility eect on the
Table2. Roll mode parameters. Here: Φ – bank angle; δa –
aileron deection; p – roll rate; τr – roll mode time constant;
τ±30° – time to roll between ±30° of bank angle
Test RM(R) RM(F) RM(R) RM(F)
Direction LR LR RL RL
Φdatum –30.8° –34.5° 32.2° 27.9°
δa–25.0° –25.0° 25.0° 25.0°
pmax 30.8°/s 23.4°/s –32.8°/s –29.8°/s
τr0.5 s 0.6 s 0.5 s 0.5 s
τ±30° 2.4 s 5.0 s 2.0 s 5.5 s
Figure 9. Roll performance of rigid (–) and exible (- - -) wing
aircra. Here:
δ
a – aileron deection;
Φ
– bank angle;
p – roll rate
Figure 10. Spiral mode parameters of rigid (–) and exible
(- - -) wing aircra. Here:
δ
a – aileron deection;
δ
r – rudder
deection; Φ – bank angle; p – roll rate
Table3. Spiral mode parameters. Here: T2 – time to double
bank angle
Test SM(R) SM(F)
T2le 32.8 s ∞
T2right 34.4 s ∞
Aviation, 2020, 24(1): 1–9 7
dihedral, i.e. highly deformed wing transformed itself
into laterally stabilising surface. Figure11 shows the roll-
sideslip coupling oscillation, which indicates the exible
aircra tendency for the Dutch roll. Figure11 also shows
greater increment of the roll rate rather than the sideslip
angle oscillation magnitude for the exible aircra. is
means a decrease of Clβ parameter. Nicolai and Carich-
ner (2010) have shown that Clβ parameter reduces with a
reducing wing AR. For the current test case the eective
wing AR of the exible aircra reduced due to a decrease
of the eective wingspan due to an increased wing dihe-
dral.
Table4 presents the Dutch roll mode parameters. e
mode was reported as an intrusive dynamic mode in the
previous lateral/directional tests. As dened by the MIL-
F-8785C requirements, the Dutch roll damping, its natural
frequency and their product are the parameters to be as-
sessed while testing aircra characteristics. According to
the mode’s damping factor, both aircra met Level 2 re-
quirement (0.02 ≤ ζDr ≤ 0.19). Pilots’ comments also noted
the poorly damped mode. Both aircra congurations met
the natural frequency Level 1 requirement (
ω
nDr ≥ 0.06 Hz).
Considering the product of both the damping factor and the
natural frequency, the rigid aircra met Level 2 requirement
(0.008 Hz ≤
ζ
DrωnDr ≤ 0.056 Hz) and the exible aircra
met Level 3 requirement (
ζ
Dr
ω
nDr ≥ 0 Hz). e ΔΦ/Δβ pa-
rameter in Table4 reveals a signicant change in the Dutch
roll mode dynamics for the exible aircra. e horizontal
ellipse motion of the mode for the rigid aircra changed to
the vertical ellipse motion of the mode for the exible wing
aircra, as shown in Figure 12. As mentioned in the dis-
cussion of the spiral mode, this change happened due to
an increased wing dihedral, which in turn decreased Clβ
parameter. is was a signicant change and was reported
by the test pilots as a serious deciency of the exible wing
aircra.
2.3. Flying qualities
e slalom task allowed comparison of FQs between the
rigid and exible aircra. e pilots were asked to assess
the longitudinal and lateral/directional FQs separately.
e resulting CHR scores are provided in Table5.
Overall, the longitudinal FQs and the rigid aircra
were evaluated better than the lateral/directional FQs and
the exible aircra. However, as the previous tests showed
negligible changes between the rigid and exible aircra
longitudinal dynamics, the pilot’s perception during the
slalom task showed that these changes in dynamics were
more signicant. e rigid aircra longitudinal FQs were
assigned CHR-2, hence meeting desired tolerances, while
the exible aircra longitudinal FQs were assigned CHR-4
meeting only adequate. is evaluation means that the
exible aircra FQs are unsatisfactory without improve-
ments, while the rigid wing aircra FQs are satisfactory.
e lateral/directional FQs, on the other hand, were as-
signed CHR-5 for the rigid aircra assessing it as unsat-
isfactory. e exible aircra lateral/directional FQs were
assigned CHR-7 not meeting even adequate tolerances.
However, the aircra was still controllable. Such an as-
sessment of FQs does not surprise as: 1) the bare airframe
without FCS was simulated, and 2) signicant changes in
the lateral/directional dynamics, indicated in the previous
Figure 11. Coupling between the roll rate p and the sideslip β
during the spiral mode test; rigid (–) and exible (- - -) wing
aircra. Here: p – roll rate;
β
– sideslip angle
Table4. Dutch roll mode parameters. Here: T – period;
ω
d –
damped frequency;
ω
n – natural frequency;
ζ
– damping ratio;
Δ
Φ
– bank angle change; Δβ – sideslip angle change
Test DR(R) DR(F)
TDr 4.37 s 4.22 s
ω
dDr 0.23 Hz 0.24 Hz
ω
nDr 0.23 Hz 0.24 Hz
ζ
Dr 0.06 0.02
Δ
Φ
/Δβ 2/3 9/4
Figure 12. Dutch roll mode shape change from the horizontal
ellipse of rigid (–) to the vertical ellipse of exible (- - -) wing
aircra. Here: Δ
Φ
– bank angle change; Δ
β
– sideslip angle
change
Table5. Cooper-Harper Rating scores during the slalom task
Test SL(R) SL(F)
Longitudinal 2 4
Lateral/directional 5 7
8V. Portapas, A. Cooke. Simulated pilot-in-the-loop testing of handling qualities of the exible wing aircra
tests and discussed above, may be undesirable for most
pilots.
One of the pilots, who is currently operating commer-
cial airliners, was asked to perform only the slalom task
and no other tests. In this case the pilot had less structured
preparation for the task and less exposure to the decien-
cies of the aircra. is acted as a surprise factor in the
test. Firstly, he was asked to y the rigid aircra and then
the exible one. e pilot commented that such a signi-
cant change in aircra dynamics due to an increased wing
exibility might cause signicant issues for airline pilots.
e most important nding was that the slalom task
successfully revealed most of the deciencies of the air-
cra. Adverse yaw, leading to the Dutch roll, and wing
ratcheting, leading to the unpredictability of the roll dy-
namics were commented by test pilots. Figure 13 shows
approximations of the roll rate p and the yaw rate r re-
sponse to the aileron input during the slalom tests, as Dur-
ham (2013) states that for the adverse yaw the positive
aileron deection should result in a negative roll and a
positive yaw. p(δa) slopes are negative for both rigid and
exible cases. However, r(δa) slopes are positive for the
rigid cases, which evidence the adverse yaw phenomenon.
According to one of the pilots the “combination of adverse
yaw and Dutch roll” made the “bank angle control most
challenging” in the rigid case. e other pilot concluded
that the “nose was pitching up and down and yawing le
and right – it was hard to know what input to make”. e
exible aircra exhibited less adverse yaw, but much of
the wing rocking, which was experienced and reported by
one of the pilots as “a lot of roll oscillations for any lateral
control input”.
Conclusions
Exposure of the dierences between the rigid and ex-
ible aircra FQs and HQs was achieved through the pi-
loted simulation trials of GA. SFFA was used to provide
the exible aircra mathematical model and the cockpit
environment.
e piloted simulation trials of the longitudinal modes
indicated negligible dierences between rigid and exible
aircra. e main dierences between two congurations
were within the damping factor, i.e. the SPPO’s damping
reduced for the exible aircra, while the phugoid’s damp-
ing increased for the exible aircra. Although changes in
the damping factor of the two congurations were negli-
gible, the HQs degraded from Level 1 to Level 2 with an
increased wing exibility.
Trials of the lateral/directional modes indicated sig-
nicant dierences between the rigid and exible aircra.
Wing ratcheting was present during roll performance tests
and doubled the time needed for the exible wing aircra
to perform a ±30° roll reversal. e roll tests also indicated
a problem of over reliance on the roll mode time constant
τr to dene the roll dynamics as it did not properly re-
veal the lag in roll. e roll performance parameter τ±30°
was suggested for the further assessments of the exible
aircra. e spiral mode tests revealed neutral stability
of the exible aircra compared to negative stability of
the rigid aircra. e Dutch roll mode tests revealed the
reduction of the mode’s damping for the exible aircra.
It also showed signicant change in the motion shape of
the mode, i.e. the horizontal ellipse shape, which is typi-
cal for the most aeroplanes, changed to the vertical ellipse
shape for the exible aircra. ese changes in the lateral/
directional dynamics were mostly attributed to decreasing
lateral static stability parameter Clβ due to increasing wing
exibility and dihedral angle under load.
e slalom task showed degraded longitudinal and
lateral/directional FQs for the exible aircra when com-
pared to the rigid aircra. Although the longitudinal FQs
of the rigid aircra were assessed as desired, the lateral/
directional FQs were assessed as unsatisfactory. e ex-
ible aircra exhibited a degradation of both the longitudi-
nal FQs and the lateral/directional FQs. is degradation
is well explained by the signicant changes in its lateral/
directional dynamics and, thus, an increased workload for
the pilots and the absence of a FCS.
Although the above mentioned piloted simulation tri-
als were performed only by three test pilots they revealed
many qualitative changes between the rigid and exible
wing aircra dynamics. For further research it is recom-
mended to increase the sample of pilots, to include both
commercial and test pilots. Using a motion based ight
simulator should also be considered for further research
as it would allow higher delity assessment of ying and
handling qualities.
Acknowledgements
e authors wish to thank Dr. Mohammad M. Lone from
Craneld University for the project management activi-
ties.
Funding
is work was supported by Airbus, Aerospace Technol-
ogy Institute and Innovate UK through the Agile Wing
Integration (AWI) project.
Figure 13. Adverse yaw during slalom tests; rigid (–) and
exible (- - -) wing aircra. Here:
δ
a – aileron deection; p –
roll rate; r – yaw rate
Aviation, 2020, 24(1): 1–9 9
Disclosure statement
e Authors declare that there is no conict of interest.
References
Allerton,D.J. (2016). EFS500 – user guide. Aeroso.
Andrews,S. P. (2011). Modelling and simulation of exible air-
cra: Handling qualities with active load control. Craneld
University.
Anonymous. (2017). Aviation benets. Industry High Level Group.
Ashby, M. F. (2017). Materials selection in mechanical design
(5th ed.). Butterworth-Heinemann.
ATAG. (2016). Aviation: Benets beyond borders. Air Transport
Action Group.
Bradley,M.K., Droney,C.K., & Allen,T.J. (2015). Subsonic Ultra
Green Aircra Research phase II: volume I – truss braced wing
design exploration (NASA/CR-2015-218704/Volume I). NASA.
Cook,M.V. (2013). Flight dynamics principles: A linear systems
approach to aircra stability and control (3rd ed.). Butter-
worth-Heinemann.
Cooper,G.E., & Harper,R.P. (1969). e use of pilot rating in the
evaluation of aircra handling qualities (NASA-TN-D-5153).
NASA Ames Research Center.
Cooper, J. E., Lowenberg, M. H., Lone, M. M., Garry, K.,
Cooke,A. K., & Coetzee, E. (2014). High Aspect Ratio Tech-
nology Enablers – HARTEn [Research proposal]. Airbus, Uni-
versity of Bristol, Craneld University.
Damveld,H.J. (2009). A cybernetic approach to assess the longitu-
dinal handling qualities of aeroelastic aircra. Del University
of Technology.
DeLaurier,J. D. (1993). An aerodynamic model for apping-
wing ight. e Aeronautical Journal, 97(964), 125–130.
https://doi.org/10.1017/S0001924000026002
Dodt, T. (2011, September 15). Introducing the 787. ISASI 2011,
Salt Lake City, UT.
Durham, W. (2013). Aircra ight dynamics and control. Wiley.
Dussart,G.X., Portapas, V., Pontillo, A., & Lone,M.M. (2018).
Flight dynamic modelling and simulation of large exible air-
cra. In K. Volkov (Ed.), Flight physics – models, techniques
and technologies (pp. 49–72). InTech.
https://doi.org/10.5772/intechopen.71050
Dussart,G. X., Yusuf,S.Y., Portapas, V., Lopez Matos,G.E., &
Lone,M. M. (2018, January 8). Method to assess lateral han-
dling qualities of aircra with wingtip morphing. AIAA At-
mospheric Flight Mechanics Conference, Kissimmee, FL.
https://doi.org/10.2514/6.2018-1015
ESDU. (1990). Normal-force-curve and pitching-moment-curve
slopes of forebody-cylinder combinations at zero angle of attack
for Mach numbers up to 5 (ESDU 89008). ESDU International.
ESDU. (1992). Normal force and pitching moment of conical boat-
tails (ESDU 87033). ESDU International.
ESDU. (2004). Normal force, pitching moment and side force of fore-
body-cylinder combinations for angles of attack up to 90 degrees
and Mach numbers up to 5 (ESDU 89014). ESDU International.
ESDU. (2013). Aerodynamic centre of wing-fuselage-nacelle
combinations: Eect of wing-pylon mounted nacelles (ESDU
77012). ESDU International.
Field,E. J., & Rossitto,K. F. (1999, August 9). Approach and
landing longitudinal ying qualities for large transports based
on in-ight results. 24th Atmospheric Flight Mechanics Confer-
ence. Portland, OR. https://doi.org/10.2514/6.1999-4095
IATA. (2013). IATA technology roadmap. International Air Trans-
port Association.
Kellari, D., Crawley,E.F., & Cameron,B. G. (2018). Architec-
tural decisions in commercial aircra from the DC-3 to the
787. Journal of Aircra, 55(2), 792–804.
https://doi.org/10.2514/1.C034130
Kim, D.-K., Lee, J.-S., Lee, J.-Y., & Han, J.-H. (2008). An aeroe-
lastic analysis of a exible apping wing using modied strip
th eor y. Active and Passive Smart Structures and Integrated Sys-
tems 2008, 6928. https://doi.org/10.1117/12.776137
Leishman,J.G. (1988). Validation of approximate indicial aero-
dynamic functions for two-dimensional subsonic ow. Journal
of Aircra, 25(10), 914–922. https://doi.org/10.2514/3.45680
Leishman,J. G. (1993). Indicial li approximations for two-di-
mensional subsonic ow as obtained from oscillatory meas-
urements. Journal of Aircra, 30(3), 340–351.
https://doi.org/10.2514/3.46340
Leishman,J.G. (1994). Unsteady li of a apped airfoil by indi-
cial concepts. Journal of Aircra, 31(2), 288–297.
https://doi.org/10.2514/3.46486
Lopez Matos,G.E., Portapas, V., Dussart,G.X., Lone,M.M., &
Coetzee, E. (2018, January 8). Pilot-in-the-loop ight simula-
tion of exible aircra in Matlab/Simulink: Implementation
and coding peculiarities. 2018 AIAA Modeling and Simulation
Technologies Conference. Kissimmee, FL.
https://doi.org/10.2514/6.2018-0426
Moorhouse,D.J., & Woodcock,R.J. (1981). Background informa-
tion and user guide for MIL-F-8785C, military specication–
Flying qualities of piloted airplanes (AFWAL-TR-81-3109).
USAF Wright Aeronautical Laboratories.
Nicolai,L. M., & Carichner, G. (2010). Static stability and con-
trol. In Fundamentals of aircra and airship design (pp. 575–
600). American Institute of Aeronautics and Astronautics.
https://doi.org/10.2514/5.9781600867538.0575.0600
Noll,T.E., Brown,J.M., Perez-Davis,M.E., Ishmael,S.D., Tif-
fany,G. C., & Gaier, M. (2004). Helios mishap investigation
report. NASA.
Patil,M.J., & Hodges,D.H. (2004). On the importance of aero-
dynamic and structural geometrical nonlinearities in aeroe-
lastic behavior of high-aspect-ratio wings. Journal of Fluids
and Structures, 19(7), 905–915.
https://doi.org/10.1016/j.juidstructs.2004.04.012
Portapas, V., Cooke,A. K., & Lone, M. M. (2016). Modelling
framework for ight dynamics of exible aircra. Aviation,
20(4), 173–182.
https://doi.org/10.3846/16487788.2016.1264719
Stinton, D. (1996). Flying qualities and ight testing of the aero-
plane. Blackwell Science.
eodorsen, T. (1949). General theory of aerodynamic instability
and the mechanism of utter (NACA-TR-496). NACA Langley
Aeronautical Laboratory.
Tollefson, J. (2016). UN agency proposes greenhouse-gas stand-
ard for aircra. Nature, 530(7590), 266–266.
https://doi.org/10.1038/nature.2016.19336
Wagner, H. (1925). Uber die entstehung des dynamischen
auriebs von tragun. Zeitschri Fur Angewandte Mathematic
and Mechanic, 5(1), 17–35.
https://doi.org/10.1002/zamm.19250050103
Ward,D.T., & Strganac,T.W. (2001). Introduction to ight test
engineering (2nd ed.). Kendall/Hunt Publishing Company.
Waszak,M.R., Davidson,J.B., & Schmidt,D.K. (1987). A simu-
lation study of the ight dynamics of elastic aircra (NASA
CR-4102). NASA Langley Research Center.
Waszak,M.R., & Schmidt,D.K. (1988). Flight dynamics of aer-
oelastic vehicles. Journal of Aircra, 25(6), 563–571.
https://doi.org/10.2514/3.45623
