Aerospace Science and Technology 99 (2020) 105764

Contents lists available at ScienceDirect

Aerospace Science and Technology

www.elsevier.com/locate/aescte

Multi-disciplinary design and performance assessment of effective,

agile NATO air vehicles
Carsten M. Liersch a,∗,1 , Russell M. Cummings b,2 , Andreas Schütte a,1 , Jan Vormweg a,1 ,
Ryan G. Maye c,3 , Tiger L. Jeans c,4
a
DLR – German Aerospace Center, Braunschweig, Germany
b
United States Air Force Academy, Colorado Springs, USA
c
University of New Brunswick, Fredericton, Canada

a r t i c l e i n f o a b s t r a c t

Article history: This article belongs to a series of publications about the activities performed within the NATO STO
Received 27 February 2019 Research Task Group AVT-251 on “Multi-Disciplinary design and performance assessment of effective,
Received in revised form 17 January 2020 agile NATO Air Vehicles”. The article concentrates on the development and investigation of the MULDICON
Accepted 30 January 2020
UCAV configuration, as well as on the organization and assessment of the AVT-251 task group itself. After
Available online 5 February 2020
a brief introduction to the preceding task groups and the research questions that lead to AVT-251, the
Communicated by Scott Morton
selection of design requirements is discussed and the chosen way for developing MULDICON out of its
predecessor, the SACCON concept, is sketched. A special focus is placed on the various aerodynamic
investigations with the aim to control the vortex flow topology at medium to high angles of attack.
Thereafter, the overall aircraft design work on the re-designed outer shape is presented and the resulting
MULDICON configuration is investigated and assessed. Finally, a concluding summary of the MULDICON
concept and the AVT-251 task group is presented.
© 2020 Elsevier Masson SAS. All rights reserved.

1. Introduction formed within the NATO6 STO7 Research Task Group AVT8 -251 on
“Multi-Disciplinary design and performance assessment of effec-
tive, agile NATO Air Vehicles” and its predecessor groups. While
The ability to accurately predict both static and dynamic stabil-
the focus of the previous groups was placed on the assessment
ity characteristics of air vehicles using CFD5 methods could revolu-
and validation of the CFD capabilities to predict complex vortical
tionize the air vehicle design process, especially for military air ve-
flows accurately, AVT-251 was dedicated to the application of CFD
hicles [1]. A validated CFD capability would significantly reduce the
in the early phase of aircraft design.
number of ground tests required to verify vehicle concepts and, in
general, could eliminate costly vehicle ‘repair’ campaigns required
1.1. Background
to fix performance anomalies that were not adequately predicted
prior to full-scale vehicle development [2–5]. This article outlines In order to evaluate and improve the prediction of S&C9 charac-
the extended integrated experimental and numerical approach per- teristics of highly swept wings at medium to high angles of attack,
a number of NATO RTO10 and STO task groups have been formed
in AVT during the past decades. AVT-080 focused on determining
*
Corresponding author. the ability of CFD to predict vortical flow structures on delta wings
E-mail addresses: [email protected] (C.M. Liersch), [6]. In AVT-113 [7,8] the focus was on experimental and numerical
[email protected] (R.M. Cummings), [email protected] (A. Schütte),
investigations on delta wing configurations with various leading
[email protected] (T.L. Jeans).
1
Institute of Aerodynamics and Flow Technology, Lilienthalplatz 7, 38108 Braun-
edges from sharp to different round radii. AVT-113 started from
schweig, Germany.
2
Hypersonic Vehicle Simulation Institute, USAFA, Colorado Springs, CO, 80840,
6
USA. North Atlantic Treaty Organization.
3 7
University of New Brunswick, Fredericton, E3B 5A3, NB, Canada. NATO Science and Technology Organization.
4 8
Department of Mechanical Engineering, University of New Brunswick, Frederic- Applied Vehicle Technology Panel.
9
ton, E3B 5A3, NB, Canada. Stability & Control.
5 10
Computational Fluid Dynamics. NATO Research and Technology Organization.

https://doi.org/10.1016/j.ast.2020.105764
1270-9638/© 2020 Elsevier Masson SAS. All rights reserved.
2 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764

Nomenclature

A , B , C Locations of airfoil sections εT E Twist angle (around trailing edge) . . . . . . . . . . . . . . . . . . . ◦

A ref Reference area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . m2 ϕL E Leading edge sweep angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . ◦
CD Drag coefficient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - ϕT E Trailing edge sweep angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . ◦
CL Lift coefficient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - c ref Reference length (MAC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . m
C m , C M Pitching moment coefficient . . . . . . . . . . . . . . . . . . . . . . . . . . - deg Degree of arc (also written as ◦ )
CT Thrust coefficient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - kn Knots (nautical miles per hour)
I xx , I y y , I zz Mass moments of inertia around X , Y , Z . . . kg m2 nz Vertical load factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -
R Leading edge curvature radius . . . . . . . . . . . . . . . . . . . . mm r Relative leading edge curvature radius . . . . . . . . . mm/m
X , Y , Z Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . m s Halfspan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . m
α Angle of Attack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ◦ y+ Non-dimensional value to assess resolution of
εL E Twist angle (around leading edge) . . . . . . . . . . . . . . . . . . . ◦ boundary-layer wall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -

given fundamental wind tunnel data by NASA followed by several (between 45◦ and 57◦ sweep angle), and with leading edge nose
pre-test CFD results which supported the wind tunnel investiga- radii varying from sharp to medium and large roundness. The ap-
tions with advanced experimental methods. proach was to provide most (if not all) flow features common to
typical UCAV and fighter aircraft configurations, and to investigate
1.2. AVT-161 the aerodynamic challenges which have to be captured by com-
putational methods. An overview of the AVT-161 Task Group is
Following the abovementioned task groups, the NATO RTO provided by Cummings et al. [9,10].
AVT-161 Task Group was established as a next step to determine
the ability of computational methods to accurately predict both 1.3. AVT-183
static and dynamic stability of air and sea vehicles. Whereas the
group concentrated on the air vehicle application, the overall ap- During AVT-161, the SACCON configuration exhibited a num-
proach was to identify major synergy in terms of physical mod- ber of very uncommon flow features, related to its variable lead-
eling, fluid structures, or transition effects. The Task Group joined ing edge roundness along the wingspan. As a consequence, the
three major avenues of interest: the experimental part to provide AVT-183 Task Group on “Reliable Prediction of Separated Flow On-
highly accurate static and dynamic validation data, the CFD com- set and Progression for Air and Sea Vehicles” was founded to gain
munity trying to predict the steady state and dynamic behavior a deeper understanding of the separation onset and progression of
of the target configurations, and the S&C group which was ana- the flow at round leading edges. Experimental and numerical re-
lyzing the experimental and numerical data. The objective of the sults of AVT-183 are, beside others, published by Hövelmann and
group was to provide best practice procedures to predict the static Breitsamter [11], as well as by Frink [12]. All of these investiga-
and dynamic behavior especially for configurations with vortex- tions resulted in improved understanding of the flow physics and
dominated flow fields where non-linear effects have a significant new best practice methods for computational simulation of vorti-
impact. These non-linear regimes are the areas where typical lin- cal flows.
ear S&C methods fail, or where wind tunnel data are only available
for non-full-scale flight flow regimes. Currently these deficiencies
1.4. AVT-201
can only be addressed through costly flight testing. Because of this
the main focus was the prediction with CFD methods rather than
The NATO STO AVT-201 Task Group was established as an ex-
enhancing existing S&C system identification methods.
tension to the AVT-161 Task Group. While the purpose of AVT-161
In an attempt to insure that the computational requirements
was to determine the ability of computational methods to accu-
for the experimental data were included in the planning as the
rately predict both static and dynamic stability of air and sea ve-
Task Group progressed, the CFD participants were asked early in
hicles, AVT-201 took on the additional tasks of including control
the program to identify a “wish list” of experimental results. The
surface deflections in the aerodynamic evaluation, as well as to
over-arching theme of the responses can best be summarized as:
investigate ways to create full flight simulations using CFD. Again,
understand the developing flow structures. In other words, the CFD
the group concentrated on the air vehicle application, but the over-
community not only wanted to know the gross aerodynamics of
all approach still was to identify major synergy in terms of physical
the vehicle, but also the causes of any interesting/unusual flow
modeling, fluid structures, or transition effects [13,14].
phenomena. This request was nearly unanimous and quite strongly
The topics covered by AVT-201 included the following:
stated by the CFD members, and the experimental researchers kept
the requests in mind as they designed the wind tunnel tests.
Since the overall goal of AVT-161 was to determine the abil-
• Perform additional in-depth correlation studies
ity of modern CFD tools to adequately predict static and dynamic
• Evaluate the ability of CFD to accurately predict S&C for
dynamic maneuvers using experimental data obtained by
S&C parameters for modern aircraft, two candidate configurations
were chosen: the X-31 and a generic UCAV11 configuration called AVT-161
SACCON.12 The latter one was designed especially for the AVT-161 • Use the detailed flow field measurements (such as PIV data)
task group, with the aim to exhibit a highly complex aerodynamic obtained by AVT-161 to enhance understanding of discrep-
behavior, serving as a challenge for numerical flow prediction us- ancies between predicted and experimental dynamic deriva-
ing CFD methods. Both AVT-161 Task Group target configurations tives
possess a delta wing planform with medium sweep leading edges • Further analyses of AVT-161 data for cases with flow asym-
metry and highly unsteady flow to extend understanding of
vehicle dynamics (air activity)
11
Unmanned Combat Aerial Vehicle. • Perform additional wind and/or water tunnel/channel test-
12
Stability And Control CONfiguration. ing to extend the dynamic data set to include multiple
 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764 3

frequency and amplitude maneuvers to improve the deter- are very interesting to study, but which are highly undesirable for
mination of realistic dynamic derivatives a realistic aircraft concept.
• Obtain, where possible, full-scale test data for a maneuver- AVT-251 was established in order to accept that challenge:
ing vehicle that can be used for validation of the methods Within a three-year-period of time, a multi-disciplinary re-design
and capabilities that are developed of the SACCON configuration towards a realistic aircraft concept
• Investigate control surface effects on dynamic S&C named MULDICON13 should be performed. Therefore, the group
• Design, build, and test modified SACCON wind tunnel model would have to deal with non-linear aerodynamic flow physics, con-
with trailing-edge control surfaces (static deflections) trol device strategies for the medium to high angle of attack flight
• Evaluate the ability to predict control effectiveness, stability regimes and vortical flow fields, as well as with the design aspects
characteristics, and other flight mechanic characteristics of regarding propulsion systems and signature constraints – every-
the configuration with controls deflected thing relying purely on CFD and other numerical methods. From
• Investigate techniques for creating flight simulation models the beginning on it was clear that a comprehensive investigation
from CFD predictions covering all relevant aspects of the design would be beyond the
• Build S&C data bases from experimental and CFD predictions scope of the group. Instead, it was tried to focus the available re-
to compare impact on flight simulation accuracy sources and partners to some of the most critical aspects and link
• Determine level of accuracy and sensitivity of flight simula- everything together using conceptual aircraft design methods [21].
tion using CFD when compared with the experimental data Specifically, the objectives of AVT-251 were:
model
• Explore range of strategies for creating CFD-derived simu- • Re-design an effective and agile UAV
lation models across the flight envelope (such as reduced- • Highest possible contribution of multiple disciplines
order modeling or combined low-fidelity/high-fidelity ap- • Development of a flight mechanics model by use of CFD,
proaches) ROM,14 etc. (no wind tunnel testing)
• International collaboration • Assess the performance at specific points of a defined flight
• The concept of a virtual laboratory, as pioneered by AVT-113, envelope
and used to great effect in AVT-161, shall be employed by • Development of control laws for specific points of the defined
AVT-201 in order to make the data being measured and flight envelope
computed available to the participants on a timely basis • Development of a design process

There were 16 different organizations making contributions In addition, AVT-251 was to answer the following questions:
from 5 different NATO nations, as well as Sweden and Australia. A
wide variety of contributions were included, such as wind tunnel • How do the tools contribute to the design process?
model development, wind tunnel testing, CFD predictions, engi- • How do the tools accelerate the design process?
neering method analysis, and development of S&C models of vari- • How do we arrange the tools in sequence or in parallel during
ous types. This represents a wide variety of participation that made the design process?
the AVT-201 a highly productive and successful task group. • To what degree can CFD methods provide sufficient data for a
flight mechanics model?
1.5. AVT-251 • To what degree can other disciplines provide inputs to the pro-
cess?
After completing AVT-161 and AVT-201, a comprehensive
knowledgebase about the flow physics of SACCON-like configu- It was important to remember that AVT-251 was not designing
rations had been gathered. These are experimental (see among a competitive configuration, rather about finding ways to improve
others [15,16]) and numerical (see among others [17–20]) data at the design process while designing a realistic vehicle. The basic
a Mach number range from 0.12 to 0.9. This data includes sym- approach for AVT-251 was to:
metrical and asymmetrical cases for steady and unsteady flow
conditions. Furthermore, a large amount of expertise was available, • Use the SACCON as a starting planform
covering the knowledge of how to simulate such sort of configu- • Use the existing experiments from AVT-161 and AVT-201 for
ration using CFD methods and the reliability of the corresponding validation
results. The next logical step was to use all this knowledge and • Use CFD experience to provide the aerodynamic data set (for
experience to re-design the SACCON configuration into a more re- parts of the mission or mission points)
alistic aircraft concept. • Integrate disciplines in parallel or in sequence depending on
the group contributions:
1.5.1. Objectives • Aerodynamics
As mentioned above, the initial design of the SACCON outer • Control device strategy
shape was developed in 2007 and 2008 by partners from AVT-161 • Propulsion system: intake, engine, nozzle
in order to have a common, generic UCAV configuration for re- • Signature requirements
search purposes. One of the ideas behind SACCON was to have • Structure and/or Aeroelasticity, etc.
a geometry which could exactly be reproduced in a wind tunnel • Formulate a design strategy
model as well as in a CFD mesh. The conceptual design task per-
formed in AVT-201 was to design an internal layout into the origi- 1.5.2. Approach
nal SACCON outer shape. This means, that it was only permitted to Before the group officially started, some of the partners con-
scale the entire SACCON geometry to a suitable size and to cut out tributed to a preparatory team, called “Design Specification Group”
parts for integrating components like control surfaces or engine in- (DSG). The aim of the team was to prepare a set of requirements,
lets and nozzles. This work resulted in a 25% increase compared to
the initially planned SACCON size, in order to fit the required fuel
and all the main components into the aircraft [13]. However, the fi- 13
MULti-DIsciplinary CONfiguration.
14
nal configuration still exposed the complex flow phenomena which Reduced Order Model.
4 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764

serving as a basis for the planned SACCON re-design work. Out of

this team the “Design Specification and Assessment Group” (DSAG)
was then formed later on. During AVT-251, the DSAG was–on the
one hand–responsible for extending and updating the design re-
quirements document. On the other hand, the DSAG had to per-
form overall design assessment studies using conceptual aircraft
design methods.
Recent low observable military flight vehicles from different na-
tions can be assumed to be flying within the low angle of attack
range. The aerodynamic behavior is typically linear and critical
aerodynamic states are usually avoided. The requirements of some Fig. 1. Outer shape of the SACCON configuration.
recent designs might demand for being able to fly at higher an-
gles of attack and having more agility compared with the earlier
designs. These advanced requirements lead to the necessity to be
able to handle the resulting non-linear aerodynamic effects, vor-
tical flow aerodynamics and to be able to adequately predict to-
tally different S&C behavior, as it appears for configurations like
SACCON. Thus, one of the main activities in AVT-251 was the re-
design of the outer shape of SACCON, in order to improve its aero-
dynamic performance and remove the undesirable characteristics
which were revealed in the AVT predecessor groups. This task was
given to the second (and biggest) design team within AVT-251, the
“Aerodynamic Shaping Group (ASG).
The third design team of AVT-251 was the “Control Concept
Group” (CCG). Its task was two-fold: On the one hand, the team
had to investigate and specify a suitable set of control surfaces,
sufficient to achieve the demanded performance requirements; on
the other hand, they had to investigate the performance charac- Fig. 2. SACCON UCAV concept. (For interpretation of the colors in the figure(s), the
teristics of the overall aircraft as it came out of the overall aircraft reader is referred to the web version of this article.)
design process. Especially the second task could only be performed
Fig. 1). SACCON is a common research concept which was initially
in a very limited way, due to a lack of contributing partners in that
defined by the NATO STO Task Group AVT-161 on “Assessment
field.
of Stability and Control Prediction Methods for NATO Air & Sea
For such a highly integrated aircraft configuration with the gen-
Vehicles” as a benchmark for CFD methods and wind tunnel ex-
eral demand for low observability, the integration of the engine
periments [9,10]. Due to its leading edge shape defining radius,
(including intake and nozzle) is another crucial aspect. While the
varying continuously from sharp to round and back to sharp, it ex-
design of the engine itself was not a direct part of AVT-251, but
poses complex, vortex-dominated flow structures which are highly
kindly provided by DLR, the integration of that engine into the air-
challenging to predict.
craft was investigated by the “Engine Integration Group” (EIG).
The SACCON concept, as it was specified in AVT-161, was a
The last design team in AVT-251 was dedicated to the investiga-
pure outer shape. Within the successor Task Group AVT-201 on
tion of structural and aeroelastic aspects. The “Structural Concept
“Extended Assessment of Stability and Control Prediction Methods
Group” (SCG) provided a basic concept for the structural topology
for NATO Air Vehicles”, an attempt was made to turn the SACCON
and a corresponding primary structure mass to the overall aircraft
shape into a reasonable aircraft concept. Therefore, a set of design
design process. Furthermore, static and dynamic aeroelastic effects
requirements were defined for SACCON, including a design mission
were investigated for the new configuration.
to be flown, a payload to be carried, and some further parameters
Besides the task to provide a general introduction and an
like fuel reserve and stability margin. As it was decided that the
overview of the activities performed within AVT-251, this article
outer shape of SACCON should be preserved in order to stay con-
focuses on three major topics. The first part (see Chapter 2) is ded-
sistent with existing CFD and wind tunnel data, the aircraft had to
icated to the selection and specification of the design requirements
be scaled until the internal volume and dimensions became suf-
for the new aircraft concept, which was performed by the DSAG.
ficiently large. Convergence was finally reached at a wingspan of
The second topic (see Chapter 3) presents some basic aerodynamic
15.375 m and a MTOM15 of approximately 15 metric tons. The
studies which were performed by the ASG in an attempt to con-
whole concept is sketched in Fig. 2.
trol the vortex development with increasing angle of attack. The
During these studies it turned out that an internal arrange-
third part of this article (see Chapter 4) provides an overview of
ment with a single, central engine (shown in green) and two
the overall aircraft design activities of the DSAG and the resulting
payload/weapon bays aside (shown in yellow) offers the best op-
aircraft configuration. A final section (see Chapter 5) summarizes
portunity for efficient propulsion with a low SFC.16 Even though
the work of AVT-251 with respect to the design task and the task
this concept limits the possible size of a single, large payload, it is
group itself.
an essential prerequisite for reaching the specified mission range
of 3 000 km plus an additional fuel reserve of 45 min. In order
2. Design task
to limit the CG17 movement due to fuel consumption, a concept
with two fuel tanks (shown in red) on each side having a com-
2.1. Background
mon CG within the specified CG range was chosen. The complete

The starting point for the aircraft design work within AVT-251
was the abovementioned SACCON configuration, a tailless, lambda- 15
Maximum Take Off Mass.
shaped flying wing UCAV concept, characterized by a 53◦ swept 16
Specific Fuel Consumption.
17
wing with parallel edges for low radar signature purposes (see Center of Gravity.
 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764 5

Fig. 3. MULDICON design mission.

design process carried out during AVT-201 is described in detail in

Refs. [13,14,22].
Even though a basic mission capability could be reached for the
SACCON concept, several unsatisfactory aspects that demanded sig-
nificant changes in the outer shape were identified. In AVT-201’s
successor group, AVT-251, the main focus was to re-design the
outer shape using high-fidelity aerodynamic methods (proven for
this kind of flow topology during the SACCON related AVT Task
Groups) in order to overcome the identified shortcomings of the
concept. The subsequent chapters of this article describe the iden-
tification and specification of design requirements, as well as the Fig. 4. SACCON control surfaces (taken from [16]).
design process and the resulting MULDICON concept.
back. An additional fuel reserve of approximately 45 min of flight
2.2. MULDICON design requirements time is desired.
In addition to the basic mission, as it was specified for SACCON,
The rationale behind the development of the MULDICON con- maneuver agility and other requirements for each section of the
cept is to overcome the known deficiencies of the SACCON concept mission were defined, mostly based on the United States Military
and to evolve it into controllable and agile UCAV configuration Standard MIL-STD-1797A18 [24].
which is consistent from a conceptual aircraft design point of view.
In order to stay as close to the SACCON concept as possible, it was 2.2.2. Payload
agreed that most of the requirements from SACCON should remain As for SACCON, it was decided to also use the configuration lay-
the same and that the changes of the geometry should be limited out one central engine with a payload/weapon bay on either side
to a minimum. The following sections discuss the various design for MULDICON. In order to determine the required size of the bays
requirements in detail. and the maximum payload mass to be carried, a weapon systems
study was performed. With respect to different possible weapons,
2.2.1. Design mission a payload/weapon bay length of 4.2 m and a width of 1.0 m was
The design mission for MULDICON was taken from SACCON chosen. The height of the bay depends on the outer shape and
without any change. Altitude and Mach number are sketched over the structural concept of the aircraft. The maximum payload mass
distance to the starting point (from left to right and back to the for each weapon bay has been agreed to be 1 250 kg. The pay-
left) in Fig. 3. load/weapon bay shall be located within the permitted CG range
It is a rather classical long range transport mission with a ra- in order to minimize CG movement due to weapon release.
dius of 1 500 km and no aerial refueling, closely related to “Bomber
– low-level penetration” from the United States Military Standard
2.2.3. Agility and control concept
MIL-STD-3013 [23]. It consists of two main parts, a high altitude
With respect to the demand of evolving SACCON into a control-
cruise segment, followed by a low altitude dash approaching the
lable, agile UCAV, there are several new requirements coming from
target. After passing the target, it continues with a turn and re-
the field of S&C. First of all, SACCON is lacking a reasonable control
turns to base over the same flight profile. Nearly all the mission is
concept. It is equipped with two trailing edge control surfaces on
flown with a Mach number of 0.8 – except for the initial climb/fi-
each side of the wing for roll and pitch control, as can be seen in
nal descent and the turn after passing the target. During the last
Fig. 4. During the investigations of AVT-201 it turned out that–due
75 km before reaching the target area, the speed could be fur-
to the high trailing edge sweep of the wing–these control surfaces
ther increased up to a Mach number of 0.9 or below – depending
exhibit very poor performance. Such a performance assessment
on the available thrust of the engine (and maybe other limitations
can be done, e.g. by comparing data from high-fidelity RANS19
like permitted dynamic pressure). However, this acceleration is just
CFD or wind tunnel tests to results from low-fidelity CFD based
an option to exploit existing reserves of the aircraft and not a
design requirement itself. The high level cruise flight segment is
performed at an altitude of 11 km, the low level dash when ap- 18
Even though MULDICON is not a piloted aircraft, many of the requirements form
proaching and leaving the target is started at an altitude of 1 000 this standard seem to be appropriate here, as well.
19
m, descending to 300 m and climbing back to 1 000 m on the way Reynolds-Averaged Navier-Stokes Equations.
6 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764

Table 1
MULDICON design points.

Cruise Takeoff Combat Low Approach Combat High

Altitude Altitude
Flight conditions Altitude 11 000 m 0m 0m 0m 11 000 m
Mach number 0.8 0.2 0.8 0.4 0.8
Velocity 236 m/s 68 m/s 272 m/s 136 m/s 236 m/s
Mass 15 000 kg 15 000 kg 13 000 kg 13 000 kg 13 000 kg
Vertical load factor 1.0 1.5 4.5 2.5 4.5
Approximated turn radius – 422 m 1 722 m 824 m 1 314 m
Assumed lift-to-drag ratio 20.0 7.0 20.0 16.1 7.0

Design requirements Roll performance 90◦ / 1.7 s 30◦ / 1.1 s 130◦ / 1.0 s 90◦ / 1.7 s 130◦ / 1.0 s
Pitch rate 20◦ /s 20◦ /s 20◦ /s 20◦ /s 20◦ /s
Yaw rate 10-15◦ /s 10-15◦ /s 10-15◦ /s 10-15◦ /s 10-15◦ /s
Permitted crosswind – 30 kn (15.43 m/s) – – –
Lift coefficient 0.184 1 0.162 0.361 0.717
Maximum lift coefficient 0.284 1.1 0.262 0.461 0.817
(Max. lift coefficient SACCON) (≈ 0.65) (≈ 0.90) (≈ 0.65) (-) (≈ 0.65)
(corresponding angle of attack) (≈ 12–14◦ ) (≈ 20–21◦ ) (≈ 12–14◦ ) (-) (≈ 12–14◦ )
Assumed drag coefficient 0.0092 0.1429 0.0081 0.0224 –
(for Engine design)
Sustained turn thrust demand – 32.15 kN 29.23 kN 20.13 kN –
(for Engine design)

sources available in AVT-251 it was agreed to focus on these five

points only. Table 1 lists the flight conditions for the design points
and the corresponding requirements for MULDICON.
Each design point consists of altitude, velocity, mass, and ver-
tical load factor. There are two mass cases: 15 000 kg (Design
MTOM) and 13 000 kg (Design MTOM, after use of 1/3 of maxi-
mum fuel). The vertical load factors can be seen as pull-up or turn
requirements. In case of performing a sustained turn20 with the
defined load factor, an approximation for the corresponding turn
radius is provided, as well. The last entry under “Flight conditions”
contains an assumption for the lift-to-drag ratio in that flight point
Fig. 5. Impression of flow around SACCON with deflected control surfaces (taken
from [27]). (including the load factor), guessed by the DSAG. The “Design re-
quirements” part of the table contains first the specified roll, pitch
and yaw requirements. These are either given for a steady rotation
on the linearized potential flow equations. As the latter meth-
rate in degrees per second, or as a performance requirement to
ods always assume attached flow, their results can be seen as an
reach a certain angle within a specified time, starting from steady,
assumption for control surfaces with an optimal performance–at
least as long as vortex-induced additional lift effects can be ne- horizontal flight. Aside from the specified yaw rate, there are two
glected. Such comparisons are performed in [22,25] and they show further main aspects which are crucial for sizing the yaw control
strong discrepancies between low- and high fidelity results. The concept: the ability to apply yawing moments rapidly, in order to
overall roll performance of SACCON based on low-fidelity aerody- compensate the expected directional instability of MULDICON and
namics, as well as on experimental data from wind tunnel tests is the so-called “de-crab” maneuver for landing under maximum per-
discussed by Ehlers et al. [26] (Fig. 17, p. 658). In that article, the mitted crosswind conditions. The requirement on permitted cross-
roll performance based on the low-fidelity aerodynamics dataset wind for takeoff and landing is again taken from MIL-STD-1797A
is shown to be sufficient for this type of aircraft, but when ap- [24] and provided in the next line of the table. The lift coeffi-
plying the aerodynamic results from wind tunnel tests instead, the cients in the following line are necessary to fly in that point and
roll performance becomes very poor. An impression of the complex with the given load factor; the maximum lift coefficients (which
flow topology of the SACCON configuration with deflected control are the lift coefficients, increased by a safety margin of 0.1) are
surfaces which leads to these effects is provided in Fig. 5. the requirements for the design work of the Aerodynamic Shaping
Considering yaw control, several different flap and spoiler con- Group: MULDICON shall be able to reach that maximum lift co-
cepts were studied for SACCON [28], but none of these performed efficient under the specified flight conditions. In order to provide
well. Again, the main reason for the limited control effectiveness an idea of the required changes compared to SACCON, the subse-
seems to have its roots in the high trailing edge sweep of the quent line gives an assumption of the maximum lift coefficients of
wing. As a consequence of these control problems, it was agreed SACCON for the same cases. Using lift coefficient and lift-to-drag
in AVT-251 that the trailing edge sweep of MULDICON should be ratio from above, the assumed drag coefficient and the required
reduced from 53◦ (as it was for SACCON) to 30◦ . Even though this thrust for a sustained turn are calculated and placed in the last
certainly adds additional peaks to its radar signature, it was as- two lines. These are essential requirements for the engine design
sumed that the performance of the trailing edge devices would be work. It has to be mentioned that for the “Combat High Altitude”
enhanced and that the modified flow would also enable an appro-
case a sustained turn requirement was not specified. Under such
priate yaw control concept. In order to define agility requirements
flight conditions, a sustained turn with a load factor of 4.5 would
for MULDICON, five dedicated design points covering various flight
conditions were selected and specified. Even though agility re-
quirements for each segment of the design mission were already
20
defined (as described in section 2.2.1), with respect to the re- “Sustained turn” means a turn without changing altitude or speed.
 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764 7

Table 2
Main design parameters of MULDICON.

Parameter Value
Outer shape Based on SACCON, ±30◦ trailing edge sweep
Propulsion Single turbofan engine without afterburner
Propulsion integration Internal (due to signature reasons)
Static dry thrust Thrust-to-weight ratio = 0.4 (≈60 kN)
Payload storage Internal (due to signature reasons)
Payload bay size Length: 4.2 m, Width: 1.0 m
Payload mass 2 × 1 250 kg
Design range 3 000 km (without aerial refueling)
Fuel reserve ≈45 min
Cruise altitude 11 km
Cruise Mach number 0.8 (all altitudes)
Stability margin 0–3% MAC (stable)
CG range 5.82 m–6.00 m

shall not exceed 1 m. On top of these requirements, which are

rather similar to those of SACCON, the MULDICON engine needs to
deliver sufficient thrust for performing the sustained turn maneu-
vers that are described above in section 2.2.3.
Fig. 6. SACCON pitching moment curve (taken from [29]).
2.2.6. General aspects
not be reasonable with respect to the selected engine concept 21 With respect to the amount of time and resources being avail-
(see section 2.2.5). able for the design of MULDICON, the disciplinary design teams
had to work in parallel with a very small number of interdisci-
2.2.4. CG range and pitching moment characteristic plinary iterations between them. Hence, it could happen that the
From SACCON it is already known that development, movement results of the different teams are not fully consistent anymore. It
and interaction of vortices with increasing angle of attack severely was the task of the Design Specification and Assessment Group to
influence the pitching moment characteristics. These effects were harmonize the parallel work of the disciplines as far as possible
investigated in detail by Schütte et al. [29] and lead to a diagram and to provide assumptions for parameters that were not available
which is shown in Fig. 6. Looking at the pitching moment curve at the time when they were required. Concluding the specification
of SACCON, it becomes apparent that such a nonlinear character- of design requirements for MULDICON, an overview of the main
istic is not acceptable for a combat aircraft operating at angles of parameters is collected and listed in Table 2.
attack up to 20◦ or even beyond. Thus, one major design task for
the Aerodynamic Shaping Group was to modify the outer shape in 3. Numerical investigations
such a way that the pitching moment characteristic becomes much
smoother.
The understanding and prediction of the flow physics of delta
In order to increase the agility of MULDICON around the pitch
wing configurations with round or variable leading edges is still
axis (compared to SACCON), the permitted CG range was reduced
challenging and a key issue for the design process of aircraft con-
down from 2–8% MAC22 (SACCON) to 0–3% MAC (MULDICON) on
figurations with a vortex dominated flow field. Accepting that chal-
the stable side of the neutral point.23 After performing some first
lenge, the SACCON concept has been investigated experimentally
computations, the neutral point was approximated to lie 6 m be-
and numerically by several NATO AVT research task groups for
hind the nose of the aircraft. This neutral point is also taken as
more than a decade now (see Chapter 1). As a result of these in-
MRP24 for the aerodynamic investigations. Together with a MAC
vestigations, a deeper understanding of the separation onset and
length of 6 m (which is taken as reference length for normalizing
progression of the flow at round leading edges was gained. The de-
all three moment coefficients, as well), this leads to a permitted
scription of the separation mechanism process described by Frink
CG range between 5.82 m and 6 m, counted from the nose of the
[12] has been systematically extended by Schütte [30] to show
aircraft.
how the flow physics develops by changing the leading edge con-
2.2.5. Propulsion tour and onflow conditions. It was documented at which onflow
For efficiency reasons, the MULDICON configuration shall be de- conditions for a particular leading edge geometry the vortex occurs
signed with a single, central engine without afterburner (see sec- and how it progresses depending on the angle of attack, leading
tion 2.1). Based on experience from SACCON, the static dry thrust edge sweep and onflow Mach number. The sensitivity and studies
of the engine shall correspond to a thrust-to-weight ratio of 0.4; of the flow topology on the aerodynamic behavior lead to design
for an estimated MTOM of 15 metric tons, this leads a thrust value guidelines for configurations with swept wings and round leading
of 60 kN. The engine shall be optimized for cruise performance edges.
(see design point “Cruise” in Table 1); however, its fan diameter Based on this extended knowledge about flow physics, de-
sign guidelines and the use of validated computational methods
for such an aircraft shape, several design studies were conducted
21
For that combination of velocity, altitude and load factor, the required thrust within the Aerodynamic Shaping Group of AVT-251. A part of these
would be around 83 kN. studies, performed by the University of New Brunswick in Canada
22
Mean Aerodynamic Chord.
23
and the DLR is presented in this chapter. The aim of these spe-
Due to the vortex-dominated flow topology, there is not one single neutral
point. However, with respect to the design requirement of smoothing the pitching
cific studies was to provide a solid, vortical flow dependent design
moment curve, it is still reasonable to assume an average position for it. based on SACCON, which meets the requirements specified in the
24
Moment Reference Point. previous chapter.
8 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764

Fig. 7. Planform and reference data SACCON and MULDICON.

3.1. MULDICON planform and design parameters advantages offered by prismatic grid structures used for the reso-
lution of viscous shear layers near walls, and the flexibility in grid
The MULDICON shape is based on the SACCON configuration. generation offered by unstructured grids. By sub-dividing the grid,
Both planforms share the same span and leading edge sweep angle TAU permits highly parallelized computations on large computer
of 53◦ . In order to enhance the effectiveness of trailing edge con- clusters.
trol devices (which turned out to be rather inefficient for SACCON The current simulations have been performed using the steady
[16]), the trailing edge sweep was reduced to 30◦ . The differences state and unsteady dual time-stepping approach [32]. The dual
in planform, as well as the main geometric properties are shown in time stepping approach was used to achieve a steady state re-
Fig. 7. With respect to this new shape, the planform area (used as sults by averaging over a certain time period. The latter was al-
reference area A ref ) for MULDICON increases to 77.8 m2 (compared ways applied in cases where no steady state solution could be
to 77.2 m2 planform area and 77 m2 reference area for SACCON), obtained. For the numerical simulations a version of the one equa-
while the halfspan s stays the same (7.696 m). So, both configu- tion Spalart-Allmaras turbulence model [33] called “SA-neg” has
rations share an aspect ratio of 3. As reference length (MAC) for been applied. The SA-neg version allows particularly negative val-
the three moment coefficients and for the Reynolds number, the ues of the transport turbulence quantities [34]. This modification
chord length of the inner kink c ref was selected, which is 6 m should lead to a more efficient solution of the equation without
for MULDICON (compared to 4.79 m for SACCON). All forces and changing the final aerodynamic solution.
moments are related to the MRP (which is identical the most rear- The hybrid unstructured grids used for the simulations with
ward permitted CG location for MULDICON), located 6 m behind TAU have been created with the hybrid grid generator Centaur,
the nose of the aircraft for both configurations. For the design developed by CentaurSoft [35]. The grid topology and local grid
studies presented below, the airfoils used in the definition sections refinement are based on best practice approaches from previous
A, B and C are modified in several different ways. Furthermore, numerical validation of the DLR RANS method TAU, e.g. published
the inner part of the configuration is changed in order to integrate by Schütte et al. in [29]. Fig. 8 shows an example of the grid
a generic intake and nozzle for the engine. topology. The prismatic layer is colored in yellow. Furthermore, the
refinement of the tetrahedral grid in the field can be seen colored
3.2. Numerical approach in green. This refinement is based on an approximation of the ele-
ment size compared to vortex size and is done using a field source
For the aerodynamic design approach two different computa- and increases the grid resolution in the area above the upper wing
tional RANS methods have been applied. Both have in common where the vortices appear. The grid topology and refinement are
that they use a hybrid computational grid approach, but the DLR done in the same way for all computational grids. Fig. 9 shows the
TAU-Code is based on a cell-vertex scheme, while the USAFA25 CFD surface grid refinement at the leading edge. For all hybrid grids the
solver Cobalt uses a cell-centered scheme. In order to make the spacing of the first prismatic layer normal to the wall is 0.005 mm,
grid sizes of these different concepts comparable, they are pro- resulting in a typical y + value of approximately one. The bound-
vided in grid points for TAU, while they are given in cells for ary layer is fully resolved by 30 prismatic layers. Over the entire
Cobalt. surface of the configuration the full 30 prismatic layers can be
achieved therefore no chopping of the prismatic layer occurs. The
3.2.1. DLR flow solver TAU overall size of the different computational grids is approximately
The DLR RANS solver TAU is developed by the DLR Institute of 21 million grid points for a half model configuration.
Aerodynamics and Flow Technology [31]. It solves the compress-
ible, three-dimensional, time-accurate Reynolds-Averaged Navier- 3.2.2. USAFA flow solver Cobalt
Stokes equations using a finite volume formulation. The code is The flow solver Cobalt is developed and provided by the US
based on a hybrid unstructured-grid approach to be able to handle AFRL.26 It solves the unsteady, three-dimensional, compressible
structured and hybrid computational grids, which makes use of the Navier-Stokes equations in an inertial reference frame. The RANS

25 26
United States Air Force Academy. United States Air Force Research Laboratory.
 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764 9

Fig. 10. Cobalt surface grid topology.

Fig. 8. TAU full grid domai.

Fig. 11. Cobalt near body grid distribution representations with leading edge topol-
ogy zoom.

performance and scalability can be achieved even on ten thousands

of processors [39].
All Cobalt simulations were run using the SARC28 turbulence
model. The choice of a SARC turbulence model was due to the
relative success of this model in previous AVT Task Groups, see
Lofthouse et al. [40]. A second order spatial accuracy was chosen,
as it allows for gradients within a given cell. Simulations were run
with first order temporal accuracy, which is often used for steady-
state calculations [41]. The structured grid topology, resolution and
size is related to best practice approaches from previous research
Fig. 9. TAU surface grid topology. publications regarding the SACCON geometry within AVT-201 by
Kennet et al. [42] and Roy et al. [43]. The structured MULDICON
grid is of O-grid style with an initial cell height of 0.012 mm and
solver Cobalt uses hybrid computational grids which mean ar- a growth rate of 1.2 normal to the body for approximately 15 lay-
bitrary cell types in two or three dimensions are applied [36]. ers. The cell count of the entire half span grid is approximately 29
In Cobalt, the Navier-Stokes equations are discretized on arbi- million. The details of the inner grid and the surface mesh on the
trary grid topologies using a cell-centered finite volume method. MULDICON surface are visible in Fig. 10 and Fig. 11. The final grid
A second-order accuracy in space is achieved using the exact Rie- was chosen based on a grid sensitivity study evaluating the lift,
mann solver of Gottlieb and Groth [37], and a least squares gradi- drag and pitching moment versus grid size.
ent calculations using QR factorization. To accelerate the solution
of the discretized system, a point-implicit method using analytic 3.3. Aerodynamic design studies
first-order inviscid and viscous Jacobians is used. A Newtonian
sub-iteration method is used to improve the time accuracy of the The following section presents the design study approaches and
point-implicit method. An approach from Tomaro et al. [38] con- the effects on the flow topology of the MULDICON configuration,
verted the code from explicit to implicit, enabling CFL27 numbers as well as on the aerodynamic performance and S&C behavior. For
as high as 106 . In Cobalt, the computational grid can be divided the latter one only the pitching moment characteristic is discussed.
into group of cells, or zones, for parallel processing, where high For an overall design to achieve an optimum in aerodynamic per-

27 28
Courant-Friedrichs-Lewy. Spalart-Allmaras with Rotational Correction.
10 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764

Fig. 12. Lift, drag, and pitching moment coefficient versus AoA for MULDICON with different airfoils; TAU.

formance for the entire flight envelope an inverse design is appro-

priate, as long as attached flow is considered. The objectives of the
current investigations are the influence of the vortical flow which
could occur at medium to high angles of attack and the influence
of design changes with respect to flow topology, aerodynamic per-
formance and pitching moment characteristic. Thus, by taking the
design constrains into account, the results and knowledge gained
could later be applied to an inverse design to extend the range
of the flight envelope of the aircraft. Even though the most crit-
ical flight points for MULDICON might be at a Mach number of
0.8, the studies presented in this section have all been performed
at a Mach number of 0.4. This flow speed was selected as a com-
promise in order to avoid transonic effects in these fundamental
studies, while still being compressible (making the process more
stable). Following the conclusions from these first studies, a sec-
ond step would then have to incorporate the transonic effects at a
Mach number of 0.8, as well.

3.3.1. Airfoil studies

Fig. 13. Overview of the assessed design rules: Effect of angle of attack, sweep angle,
This section discusses the influence of a variation of airfoils
leading edge contour radius and Mach number (taken from [30]).
applied at location A, B and C (see Fig. 7). The results of the nu-
merical simulations using TAU is shown in Fig. 12. Depicted are
the lift (CL ), drag (CD ) and pitching moment (Cm ) coefficient ver- occurs. For the Baseline the pitching moment coefficient is around
sus AoA29 for four different airfoils. The Baseline configuration uses 0, whereas the Super Critical and NACA-64A-410 airfoils provide
the SACCON airfoils (having varying radii over the wingspan) ap- negative values and the symmetric NACA airfoil provides positive
plied to the MULDICON planform. The other three are a typical ones. Beyond α = 10◦ an increase towards a more front loading
supercritical airfoil, a symmetric NACA30 -64A-010 and a cambered pitching moment can be noted for all configurations.
NACA-65A-410 airfoil. Besides the Baseline, the airfoils are adapted The reason for this is the development of a leading edge vor-
with the CST approach to apply a constant leading edge curvature tex. While the flow is fully attached over the entire upper side
radius of r = 1 mm related to a 1 m chord length (approx. 10 of the wing for an angle of attack of α = 8◦ , the flow topology
mm at full scale). The aim of the current design investigations is has completely changed for α = 12◦ . For all configurations besides
to provide a less complex topology in comparison to the SACCON the Baseline, a strong leading edge vortex has developed. These
leading edge contour distribution and to get rid of the discontinu- leading edge vortices provide a higher suction in front of the mo-
ity in the pitching moment which occurs for the SACCON at higher ment reference point in comparison to the suction behind it. For
AoA, as described amongst others by Schütte et al. in [29]. AoA higher than α = 12◦ this trend does not change because the
The lift distribution versus AoA shows a quite similar linear be- starting point of the vortex at the leading edge moves towards the
havior for all airfoils. The differences in lift versus AoA are based apex with increasing AoA. This vortex development behavior for
on a different cambering. The non-cambered NACA-65A-010 airfoil swept wings with round leading edges has been already described
provides the lowest lift coefficient followed by the Baseline con- by Schütte in [30] (see Fig. 13) and should be one of the elements
figuration, the Super Critical and the NACA-65A-410 airfoil. The for the further design studies in the present article.
same statement applies to the drag distribution. The pitching mo- The results for the Baseline configuration are presented to show
ment coefficient shows a neutral stability behavior (C m ≈ const.) the effect of the design changes in comparison to the SACCON con-
between 0◦ and 8◦ AoA for all applied airfoils. Due to the dif- figuration. In contrast to the first three configurations the leading
ferent cambering a significant difference in rear loading behavior edge contour distribution for SACCON is not constant along the
span. The SACCON configuration has a sharp leading edge at the
apex and wing tip which leads to a two-vortex-topology at an AoA
29
Angle of Attack. of α = 12◦ . Thus, the gradient of the pitching moment is less than
30
National Advisory Committee for Aeronautics. for the configurations with only one vortex close to the apex. In
 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764 11

on the results achieved in the constant leading edge radii profiles.

Upon completion of these studies it should be possible to deter-
mine the relationship between leading edge radii and the overall
aerodynamic characteristics of the MULDICON geometry. The lead-
ing edge contour radius distribution applied to the MULDICON is
depicted in Fig. 14.
Six different airfoils were provided for each airfoil section ( A,
B and C ) along the MULDICON body by manipulating the leading
edge contour radius. The predicted lift, drag and pitching moment
coefficients for the various constant leading edge radius simula-
tions are plotted against angle of attack in Fig. 15. The Baseline
MULDICON results are also plotted, providing a reference for eval-
uating the force characteristics of the modified geometries. A shift
in lift coefficient occurs for all constant radius simulations, as seen
in Fig. 15. This increase in overall lift at a given AoA is due to
the change in wing profiles as the NACA-65A-410 has more cam-
ber than the baseline MULDICON profiles. An increase in camber
moves the angle of zero lift to more negative values, resulting in
an upward shift in overall lift force for a given AoA.
All drag force coefficient data follows the same trend. However,
with decreasing leading edge radius the drag force at medium to
Fig. 14. Leading edge contour radius distribution applied to MULDICON; Cobalt. high angles of attack (α > 10◦ ) increases. For most streamline
bodies a large portion of drag is related to skin friction, caused by
comparison to the constant 10 mm radius for the other configu- the friction of the fluid passing over the body. The reason why drag
rations the apex vortex is smaller due to a higher leading edge often increases with angle of attack (aside from increasing induced
contour radius at section B. In addition to this the tip vortex com- drag) is an effect of flow separation creating larger amounts of
pensates the nose up loading trend slightly. pressure drag. The small radius cases tend to have higher amounts
More important is to note, that the pitching moment behav- of drag, possibly related to earlier separation at low angles of at-
ior is primarily defined by the leading edge contour. The reason tack. With larger radii profiles (R ≥ 10.382 mm) the drag force is
for this effect is that the vortex occurrence, its progression and reduced at higher AoA. For R = 30 mm drag from α = 16◦ to α =
strength depend on the leading edge contour and not on the airfoil 18◦ is consistently lower than for the Baseline case. The mid-range
geometry. According to Fig. 12, the increase in pitching moment radii sizes, R = 10–20 mm, perform almost identical to that of the
occurs approximately at the same AoA. baseline data, with slight variation at α > 12◦ .
The predicted pitching moment coefficients for all constant ra-
3.3.2. Leading edge contour studies dius cases have been shifted downward into the negative pitch
Referring to the results in the previous section the question regime at low angles of attack. All cases predict the same ini-
arises how to influence the flow topology over the wing in a way tial pitching moment coefficient of C M ≈ −0.03, significantly lower
such that the vortex development leads to a stable (or at least than the Baseline result, which predicts a C M = 0.05. The positive
less unstable) pitching moment characteristic. This question leads camber of the airfoils gives rise to this shift, as a large camber
to the second parametric study on the MULDICON planform deal- value causes initial negative pitching moment behavior. For all
ing with the effect of the leading edge radius regarding the vortex cases the pitching moment tends to increase with increasing an-
development and related effects on the pitching moment charac- gle of attack, resulting in a positive slope and unstable pitching
teristic. The first part of these investigations has been performed moment behavior. Furthermore, it can be seen that the increase
using the RANS solver Cobalt and the second part with the RANS in pitching moment slope corresponds with the break from linear
solver DLR-TAU. Both, constant and linearly varying radii cases in the lift coefficient-vs-alpha data. For the small radius case the
were simulated. Based on the design studies by Schütte [30], five pitch slope increases rather quickly, while for larger radii profiles a
different constant leading edge radii profiles were chosen, varying more neutral behavior is seen over a wider AoA range, visible with
from 2.5 to 20 mm. Two varying leading edge profiles, linearly de- the almost zero slope region from α = 2◦ to α = 12◦ . Further
creasing from apex to tip, were also chosen for investigation based improvement is visible when increasing the leading edge radius

Fig. 15. Lift, drag and pitching moment coefficient versus AoA for the various constant leading edge radius simulations; Cobalt.
12 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764

Fig. 16. Effects of flow topology on the pitching moment for small constant leading Fig. 17. Effects of flow topology on the pitching moment for large constant leading
edge radius geometry; Cobalt. edge radius geometries; Cobalt.

to R = 30 mm, as the slope appears to be almost constant up to Above α = 16◦ a significant amount of the low pressure related to
α = 18◦ . the vortex core has moved ahead of the MRP, resulting in a signif-
This type of UCAV aircraft is inherently unstable, requiring icant increase in pitching moment between α = 16◦ to α = 17◦ .
aileron deflection to provide proper pitch correction. The key is Above α = 17◦ the pitching moment behavior stays constant as
to avoid sudden changes in pitching moment over short angle of the low pressure region ahead of the MRP does not strengthen by
attack ranges, as it is almost impossible for the ailerons to com- a substantial amount.
pensate for such dramatic changes. Taking this into consideration Deviation from this pitching moment behavior is significant
the use of a large constant leading edge profile provides quite with further increase in leading edge radii, as represented in
good pitch behavior for this aircraft type, as does the baseline Fig. 17. At incidence angles below α = 10◦ the flow is fully at-
MULDICON. For cases with a leading edge radius less than R = tached, resulting in a neutral pitching moment characteristics (i.e.
20 mm, the stability characteristics are quite poor, providing lit- zero slope relative to AoA). At approximately α = 10◦ flow sepa-
tle advantage when compared to the other leading edge designs. ration commences at the tip of the MULDICON planform, with a
Thus, force results suggest that the use of a large leading edge large region of low pressure across the entire leading edge. This
radius would significantly improve the overall pitching moment low pressure region continues to strengthen with increasing angle
characteristics, with secondary improvement in lift coefficient re- of attack, while continually moving upstream towards the apex.
sults. This finding goes well in line with the classical way of using Gradual increases in pitching moment occur as the region of low
round leading edges for conventional subsonic aircraft in order to pressure behind the MRP continues to decrease in size, resulting
delay separation. in a pitch up characteristic as the distribution of pressure relative
Based on the findings discussed above, two diagrams were cre- to the MRP is unbalanced. At an AoA just above α = 16◦ the vor-
ated in order to explain the effects of flow topology on the pitching tex related to separation moves ahead of the MRP, resulting in a
moment characteristics for constant leading edge radius geome- more significant rise in pitching moment. Above α = 17◦ the vor-
tries. Fig. 16 represents the lower limit of radii investigated (R = tex continues to move upstream towards the apex, resulting in a
2.5 mm), where unfavorable pitch behavior occurs. The second di- steady increase in pitching moment. This rise in pitching moment
agram, Fig. 17, represents the upper limit of radii investigated (R is considerably more gradual in comparison to the smaller radius
= 20–30 mm), where a significant improvement in pitching mo- case. Variation is visible for the pitching moment characteristics
ment performance over a wide angle of attack range is visible. of the largest radius case, R = 30 mm. As the upstream vortex
Each of them will be discussed individually, in order to describe movement is delayed with increase in angle of attack, also the rise
the effects of the size of constant leading edge radius chosen. At in pitching moment is delayed to higher angles of incidence. The
low incidence angles, the case with R = 2.5 mm (Fig. 16) exhibits sudden increase in pitching moment is not visible below α = 18◦ ,
fully attached flow with well distributed surface pressure over the a significant improvement in comparison to all other radii inves-
upper surface of the MULDICON configuration. At α ≈ 6◦ flow tigated. This is the reason, why the R = 30 mm case, together
separation commences directly at the apex, spanning the entire with a downward-twist of the wingtip to delay outboard separa-
leading edge, creating regions of low pressure which are well bal- tion, (“Case 8”) was selected for the engine integration studies in
anced relative to the MRP. Above α = 10◦ , the separation behavior the next section.
and resulting vortex roll-up strengthens, creating larger regions of
low pressure that are not evenly distributed forward and aft of the 3.3.3. Engine integration studies
MRP, which lead to an increase in pitching moment. From α = 10◦ The purpose of this section is to provide the influence of the
to α = 12◦ these regions of low pressure across the entire leading aerodynamic performance by integrating an engine intake and out-
edge re-balances, stalling any increase in pitching moment. This let and applying the related engine conditions. For these investiga-
step like behavior in pressure distribution resulting in the desta- tions, the configuration “Case 8” was selected from the parametric
bilization of pitching moment repeats from α = 12◦ to α = 16◦ . variations described above. It incorporates the NACA-65A-410 air-
 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764 13

Fig. 19. Lift and pitching moment versus AoA for Case 8 with and without engine
integration; TAU.

Fig. 18. MULDICON design Case 8 (R = 30 mm) in clean configuration and with
integrated engine inlet and outlet.

foils with R = 30 mm and a twist distribution of 0◦ for section A,

−2◦ (around trailing edge) for section B and −3◦ (around leading
edge) for section C . Fig. 18 shows a comparison between the Case
8 clean configuration and the same configuration with a generic
integrated engine intake and nozzle. It can be seen that the volume
of the inner wing body section has not been increased but reduced
between the engine inlet duct and the apex. The boundary condi-
tions of the inlet and outlet surfaces as well as the engine condi-
tions have been evaluated within a separate investigation from the Fig. 20. Lift, drag and thrust coefficient and load factor versus AoA for Case 8 with
Engine Integration Group. As described above, a Mach number of integrated engine; TAU.
0.4 has been chosen for the calculations with engine. The engine
flight conditions chosen are the approach condition with a thrust engine meets the required regular onflow condition for the engine
level to provide a load factor of n z = 2.5 estimating a mass of the inlet. The inlet geometry does not cause flow disturbances ahead
aircraft of m = 13 000 kg. for the engine intake.
Fig. 19 shows a comparison of lift and pitching moment coeffi-
cients between the clean configuration and the configuration with 4. MULDICON design
integrated engine. The plots show that for the present investiga-
tion the lift and pitching moment are similar for the clean and the The overall aircraft design work for the MULDICON configura-
engine integrated configuration for the considered AoA range. The tion was performed at DLR. It was based on the DLR conceptual
engine integration is neither affecting the longitudinal stability, nor design system, a flexible concept design toolset being developed
the overall flow topology on the upper wing surface. Fig. 20 shows since 2005 [44,45]. This system consists of three components:
the lift, drag, and thrust coefficients for the configuration with en-
gine integration. In addition, the load factor is evaluated for AoA • A set of analysis tools representing the different disciplines of
of 2◦ to 10◦ . For the present considered design point the load fac- aircraft design.
tor of n z = 2.5 is provided at an AoA of α = 6◦ . The comparison • A flexible and extensible data exchange file format called
between the drag versus thrust coefficient provides a slight lack of CPACS,31 serving as common language for the analysis tools
thrust of ≈ 80 drag counts. This is caused by the installation ef- [46].
fects in comparison to the isolated engine evaluation and needs to • The DLR integration framework software RCE,32 used for con-
be adjusted to receive a trimmed flight condition. Nevertheless, the necting the tools to build up task-specific analysis workflows
present design matches the requirements quite well for the flight [47].
condition considered here.
Fig. 21 shows a comparison of the surface pressure distributions Using this system, a process chain for investigating and as-
and flow topology on the upper surface of the clean MULDICON sessing the MULDICON configuration was set up and applied. It
configuration at α = 6◦ , as well as for the engine integration case contained the generation of aerodynamic and engine performance
for α = 6◦ and 10◦ . These plots show that the flow topology is
not changed for the current flight condition (M = 0.4, Re = 55.9
million) with and without engine integration. Furthermore, it can 31
Common Parametric Aircraft Configuration Schema.
32
be seen that the adjusted apex geometry for the configuration with Remote Component Environment.
14 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764

Fig. 21. Flow topology on the upper wing with and without engine integration: clean configuration at α = 6◦ (left), with engine at α = 6◦ (middle) and α = 10◦ (right);
TAU.

Table 3
Maximum lift coefficients of MULDICON compared to requirements.

Flight case Load factor Required lift coefficient Target maximum lift coefficient MULDICON
Baseline Final design
Takeoff 1.5 1.0 1.1 0.84–0.96 1.11–1.14
Combat High Altitude 4.5 0.717 0.817 0.61 0.72

maps, a convergence loop for sizing important components of the

aircraft, and a post-analysis for investigating performance and fly-
ing qualities of MULDICON. The whole process and the DLR design
system itself are described in more detail by Liersch et al. in [48].

4.1. Results of the AVT-251 design teams

The design of the MULDICON configuration was a collabora-

tive effort within the five design teams of AVT-251. This article
focuses on the overall aircraft design aspects performed in the
Design Specification and Assessment Group. However, the results
Fig. 22. Outer shape of the MULDICON final configuration.
from the other four teams are essential inputs to this work and
are thus presented and discussed.
CPACS data format and investigated in the conceptual design work-
4.1.1. Aerodynamic shaping
flow. At this point it turned out that the first design is not yet
With respect to the outer shape, the Aerodynamic Shaping
usable as its generic airfoils incorporate too much camber (causing
group provided a reference configuration first, which is identical
a strong zero-lift pitch-down moment) for a flying wing aircraft.
to SACCON (same airfoils, same twist distribution), but with the
Since there were no resources available to apply the leading edge
modified trailing edge sweep of 30◦ (see section 3.1). In order to
design to a more suitable set of airfoils, this concept was not in-
satisfy the requirements for maximum lift coefficient and pitching
vestigated further. Thus, the second design concept, which is the
moment characteristics, two different design philosophies were ap-
official final design from the Aerodynamic Shaping group, is taken
plied to the reference configuration. The first one was focused on
as main concept for the conceptual design studies (see Fig. 22).
the understanding of the physical principles behind the complex
The work of the Aerodynamic shaping Group as a whole is docu-
vortex phenomena. Therefore, based on generic airfoils, paramet-
mented by van Rooij and Cummings in [51].
ric studies on varying leading edge radii and twist distributions
were performed. Using these physical principles, the leading edge
was shaped in a way that the movement of the vortices was min- 4.1.2. Control concept
imized and the pitching moment curve became much smoother. The second main task in the development of MULDICON was
An excerpt from these studies was presented in the previous chap- the design of a suitable control concept, which is discussed by
ter. More details on this approach are discussed by Schütte et al. Löchert et al. in [28]. After confirming that the conventional trail-
in [49]. The second design approach aimed at minimizing vor- ing edge devices for roll and pitch worked much better than they
tex effects by designing for attached flow conditions. Therefore, a did for SACCON, the focus of this work was placed upon finding
complete redesign of airfoil shapes and twist distribution was per- a solution for yaw control. As it was not clear at this point, how
formed. Finally, the discontinuities in the pitching moment could much yawing moment coefficient would be required, a target max-
be reduced and the maximum lift coefficient was increased. Details imum value of 0.015 was chosen based on experience. In order to
about this second, inverse design approach are given by Nangia et validate this value, an investigation with varying yaw control effi-
al. in [50]. Table 3 shows the achievements with respect to maxi- ciency was performed by Hasan et al. [52] (Fig. 25, p. 27). Using
mum lift coefficient. It can be seen that the design targets for these the conceptual design workflow for simulating a landing maneuver
three points are nearly met. For the Combat High Altitude case the with maximum permitted crosswind, it turned out that a yawing
maximum lift coefficient is slightly lower than the design target, moment requirement of 0.015 seems to be reasonable for handling
however, the lift demand for the 4.5 g maneuver is still reached. the permitted crosswind of 30 kn. The final control concept coming
For the conceptual design studies, all three versions (reference out of the Control Concept Group was applied to the MULDICON
configuration and the two new designs) were modeled in the CPACS datasets.
 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764 15

Fig. 23. MULDICON planform with inner arrangement and CG locations.

Table 4 estimate for the structural mass of MULDICON was given. A spe-
Parameters of MULDICON engine “UCAV_G”. cial focus had to be placed on the big cutouts due to engine and
Parameter Condition Unit Value payload/weapon bays and on aeroelastic effects like body-freedom-
Static thrust (dry) Takeoff kN 60 flutter. Further details about the structural and aeroelastic design
Bypass ratio Cruise – 1.7 work for MULDICON are presented in [57–60].
Overall pressure ratio Takeoff – 30.5
Mass flow Takeoff kg/s 114
Turbine entry temperature Takeoff K 1740 4.2. Overall aircraft design
Specific fuel consumption Cruise g/(kNs) 23.8
Fan diameter all m 0.908
Length all m 2.2 Based on the results coming from the different design teams,
Mass all kg 1040 the overall aircraft design work was performed at DLR. One of the
central elements of the MULDICON workflow is a spreadsheet con-
taining the main components of the aircraft and a two-dimensional
4.1.3. Aerodynamic shaping
planform view including the CG limits (see Fig. 23). Using this
Another important task was to provide an engine model which
spreadsheet, the main internal components were arranged. As can
satisfies the engine design requirements, as specified in Chapter 2.
be seen in the diagram, large components e.g. engine were placed
This work was an additional contribution dedicated to AVT-251, in
directly. Smaller components such as avionics boxes, for which the
order to close a gap in the design capabilities of the group. Start-
geometric properties are not known at this stage of the design
ing from a permitted fan diameter of 1 m, some engine design
cycle, were placed in free areas, assuming that they will have suffi-
studies were performed. As it became clear, the fan diameter is
cient space there. The filled circles within the components shown
still critical with respect to the integration of the engine and a
represent the CG of that component. It was a difficult, iterative
corresponding intake and nozzle concept, a variation study for the
procedure to arrange all the components such that they have suf-
fan diameter was performed. As a final result, a slightly smaller
ficient space, while the CG positions for all 11 weight & balance
engine was selected and its performance tables were provided to
cases under investigation were kept within the specified limits be-
the AVT-251 group. The engine design work and the sizing study
are explained by Zenkner and Becker in [53]. The main engine tween the red dashed lines. The CG range from the center of the
parameters are provided in Table 4. With respect to engine in- aircraft is further displayed in a magnified detail sketch on the
tegration into MULDICON, several studies were performed by the right side. The most forward location corresponds to the OEM33
Engine Integration Group. Due to limitations in time and resources, (no payload, no fuel) with landing gear up (GU), while the most
their final results could not be incorporated directly into the over- rearward one represents the ferry flight case (no payload, max-
all aircraft concept. However, their demands were considered as imum fuel) with landing gear down (GD). As an impression of
boundary conditions where possible. More Details on engine inte- the very tight limitations here, it shall be mentioned that for the
gration work for MULDICON can be found in References [54–56]. operational empty mass case (yellow and green squares), the CG
movement due to retracting the landing gear already uses around
4.1.4. Structural concept one third of the permitted CG range.
The structural concept of MULDICON was defined by the Struc-
tural Concept Group. Based on experience and Finite Element anal-
33
yses, the main structural elements were placed and sized, and an Operational Empty Mass case: aircraft ready to fly, but without fuel or payload.
16 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764

Table 5
MULDICON component masses, CG locations, and mass moments of inertia around (0,0,0).

CG coordinates Mass moments of inertia (0, 0, 0)

Component Mass [kg] X [m] Y [m] Z [m] I xx [kg m2 ] I y y [kg m2 ] I zz [kg m2 ]
Structures 2 638 6.84 0.00 0.00 25 914 124 619 150 533
Landing gear (down) 321 5.29 0.00 0.00 920 9 796 10 716
Landing gear (up) 321 4.24 0.00 0.00 920 6 561 7 481
Propulsion 1 459 5.95 0.00 0.07 10 51 644 51 633
Systems 1 790 4.49 0.00 0.00 0 48 939 48 939
Other 559 5.93 0.00 0.00 595 19 673 20 268

OEM (GD) 6 767 5.88 0.00 0.02 27 440 254 672 282 091
OEM (GU) 6 767 5.83 0.00 0.02 27 440 251 437 278 856

Maximum payload 2500 5.90 0.00 0.00 3306 87025 90331

10% fuel (Landing) 367 6.11 0.00 0.00 8 288 26 081 34 368
66.7% fuel (Design Maneuver) 4 249 6.11 0.00 0.00 55 251 173 871 229 122
83.8% fuel (Design Mission) 5 341 6.11 0.00 0.00 69 441 218 524 287 964
100% fuel (MTOM, Ferry Flight) 6 374 6.11 0.00 0.00 82 877 260 806 343 683

Table 6
MULDICON weight and balance cases, including mass moments of inertia around CG.

CG coordinates Mass moments of inertia (CG)

Weight and balance case Mass [kg] X [m] Y [m] Z [m] I xx [kg m2 ] I y y [kg m2 ] I zz [kg m2 ]
Gear down OEM 6 767 5.88 0.00 0.02 27 438 20 928 48 349
ZFMa 9 267 5.88 0.00 0.01 30 745 20 929 51 656
Landingb 7 404 5.90 0.00 0.01 35 726 23 250 58 959
Design missionc 14 607 5.97 0.00 0.01 100 186 40 299 140 467
Ferry flightd 13 141 5.99 0.00 0.01 110 316 44 017 154 314
MTOMe 15 641 5.98 0.00 0.01 113 623 44 034 157 637

Gear up OEM 6 767 5.83 0.00 0.02 27 438 21 647 49 068

Design maneuverf 13 516 5.93 0.00 0.01 85 997 37 133 123 110
Design mission 14 607 5.94 0.00 0.01 100 186 41 088 141 255
Ferry flight 13 141 5.96 0.00 0.01 110 316 44 820 155 118
MTOM 15 641 5.95 0.00 0.01 113 623 44 829 158 432
a
Zero Fuel Mass case: OEM + maximum payload.
b
Landing case: OEM + landing fuel.
c
Design Mission case: OEM + mission fuel + maximum payload.
d
Ferry Flight case: OEM + maximum fuel.
e
MTOM case: OEM + maximum fuel + maximum payload.
f
Design Maneuver case: OEM + maneuver fuel + maximum payload.

A mass breakdown of MULDICON, as calculated with the hours and 18 minutes.34 Compared to the maximum fuel capac-
spreadsheet, is provided in Table 5. It contains the masses of the ity of 6374 kg, the fuel reserve is 1033 kg (or 16.2% of maximum
main components, their center of gravity locations and the mass fuel). With respect to an averaged mission fuel burn of 0.45 kg/s,
moments of inertia for the main axis’ around (0, 0, 0). The devi- this reserve would last for another 38 minutes of flight. Consider-
ation moments are currently neglected, as well as the center of ing a fuel flow of around 0.17 kg/s, as it is present at the end of
gravity locations in Z-direction (set to zero). Table 6 lists the se- the cruise segment right before the final descent for landing, even
lected weight and balance cases for MULDICON, together with the a duration of 1 hour and 41 minutes is achieved. So, the design
mass moments of inertia for the main axis’ around the correspond- requirement of providing ≈ 45 minutes of fuel reserve after flying
ing CG. the design mission is satisfied. Compared to the MULDICON base-
One drawback of the spreadsheet is that it only contains a 2D line configuration, the final design of MULDICON requires 304 kg
model of the inner geometry, whereas the thickness of MULDICON (≈5.4%) less fuel for the design mission. In Fig. 25, the main pa-
varies continuously over the chord. As a consequence, from this rameters of the aircraft are plotted over the flight time.
model it is not possible to sufficiently determine, whether a com- After the end of the conceptual design workflow, a rather com-
ponent really fits into the outer shape. As a solution to this prob- prehensive CPACS dataset of the MULDICON configuration was
lem, the spreadsheet was extended by a so-called “Design Table” made available, permitting further, more detailed investigations of
for Dassault’s CATIA CAD software [61]. Combined with an existing the aircraft concept. One such investigation, which has not been
CAD model of the MULDICON outer shape which also incorporates published yet, was dedicated to flight performance and flying qual-
intake, nozzle and the control surfaces of the final control con- ities evaluations. It shows that with respect to roll performance,
cept, the CATIA software uses the construction table to generate the requirements for the “Takeoff” and “Combat High Altitude”
the inner components as specified in the spreadsheet. The CATIA cases could not be met sufficiently, while the other cases are
within the specified limits. Furthermore, it turns out that the avail-
3D model of the UCAV configuration with its main components is
able thrust for the “Combat Low Altitude” case is not sufficient
shown in Fig. 24.
In the center of the convergence loop of the conceptual design
process chain, the simulation of the design mission is located. Af- 34
It hast to be mentioned here that the mission simulation currently neglects the
ter reaching convergence, the results for MULDICON final design trim drag. However, with respect to the very small stability margin and the low
show a required fuel mass of 5341 kg and a flight duration of 3 zero-lift moment, the trim drag over the mission flight is assumed to be rather low.
 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764 17

Fig. 24. 3D model of MULDICON with internal arrangement.

Fig. 25. Trajectory of MULDICON, flying the design mission.

with respect to the sustained turn requirement – a consequence of ing investigations. There were two distinct approaches that were
a lift-to-drag assumption which turned out to be too optimistic for followed by the Aerodynamic Shaping Group as they proceeded
that case. through the re-design: 1) design a new wing which was free of
vortices during the mission. And 2) design a new wing which
5. Summary minimized the impact of the vortices on the aerodynamics of the
vehicle. The enhancement of the SACCON concept had a number
The foreground task of AVT-251 was to specify design require- of specific goals, while still desiring to meet the mission require-
ments for an effective, agile UCAV, and then use these require- ments that had been applied to the original SACCON:
ments to conduct a re-design of the SACCON configuration into a
more realistic aircraft. In parallel, the background-task was to per- • Remove undesirable pitching moment characteristics
form an assessment of how such a re-design could be performed • Increase maximum lift coefficient
within an AVT task group and how CFD could be effectively applied • Develop a control concept for sufficient roll, pitch, and yaw
in such an early phase of the design process. Section 5.1 addresses control
the first question, while section 5.2 is dedicated to the second one. • Integrate an engine (intake & nozzle)
Finally, section 5.3 provides some conclusions drawn of the work • Develop and size a structural concept, suitable for rigid and
being performed in AVT-251 aeroelastic effects

5.1. Design task After completing the various trade studies (including aerody-
namic shaping and flow topology, structural layout and aeroe-
As a first step, a set of design requirements was put together lastics, as well as control concepts and flying qualities), a new
and agreed on with the contributing partners of AVT-251. These configuration named MULDICON was found. The biggest change
requirements were selected to be typical for such sort of UCAV in the planform was the new trailing-edge sweep angle, which
concepts, but highly ambitious with respect to the rather challeng- was greatly reduced compared to SACCON. It increased the in-
ing SACCON configuration as a starting point. Due to the limited ternal volume, changed the CG locations, and made the control
timeframe and resources it was not possible to work with a re- concepts more effective. A detailed engine installation study was
ally comprehensive set of aircraft requirements. Instead, the de- also included as a last design detail, and while the engine in-
manded design targets were reduced to the most critical aspects lets and outlets are still being improved, an acceptable design and
and handed over to the various design teams. Next, a detailed internal layout was achieved for mission requirements and other
design study took place with a number of aerodynamic shap- constraints. Reviewing the results from conceptual design and the
18 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764

different design teams, it has been demonstrated that the mission was definitely very positive, but it was challenging to keep ev-
and payload requirements could be met. Regarding the agility re- eryone mindful of the full purpose of the task group. These are
quirements for the specified design points it can be summarized important considerations to keep in mind as the assessment re-
that there were not enough resources to investigate all five design sults are looked at, but also if another task group with a similar
points to the necessary extent and that one of the selected aerody- purpose is proposed in the future.
namic design paths could not be followed up to its end. However, First it is important to have a look at the time and compu-
for the addressed points the requirements could be satisfied or at tational resources expended to perform the overall design study.
least be nearly satisfied. The pitching moment characteristics for For the 22 participants who have filled out their questionnaire, a
both MULDICON aerodynamic design paths have been smoothed total of approximately 20000 person-hours were used for partic-
– at least for the required range of lift coefficients. The new con- ipation in AVT-251. This includes attending task group meetings,
trol concept fulfills most of its requirements, even though the roll participating in telecons, using the group sharepoint site, and other
performance for The “Takeoff” and “Combat High Altitude” cases communication for the group, but also includes the total time
is still insufficient; just as the thrust for doing a sustained 4.5 spent performing the studies and analysis required for the de-
g turn under “Combat Low Altitude” conditions. With respect to sign of MULDICON. Of the total hours spent, approximately 25%
this, coupling the results from the assessment back to the design was spent within the DSAG (Design Specification and Assessment
teams and performing another design iteration would have been Group), 60% within the ASG (Aerodynamic Shaping Group), and
useful in order to fulfill the requirements completely. With regard 15% for all other groups (CCG, EIG, SCG). Within the ASG a total of
to structures and aeroelasticity, a suitable solution for the struc- 28.3 million CPU hours was used to perform the various CFD stud-
tural concept has been found and investigated. ies and analyses. Even though only half of the members answered
All of these design trade studies were carried out within the 3 the questionnaire, it can be stated that most of the key members
year time period of AVT-251. While done without additional wind of each design group (especially ASG where the vast majority of
tunnel testing, the studies were performed with a high degree of the CFD predictions were made) did contribute to the assessment
confidence based on the large amount of wind tunnel data that survey, so the actual amount of CPU usage, for example, was prob-
was available for SACCON, making the CFD studies trustworthy ably not significantly higher than the 28.3 million hours reported
within regular aircraft design accuracy levels. above.
One aspect of the twice-yearly meetings was allowing signif-
5.2. Assessment task icant time for each group to meet individually, and a few ques-
tions were asked about how those group meetings (and the groups
A detailed questionnaire was sent out to all participants of themselves) functioned. On a scale of 1 to 10 (1 being extremely
AVT-251 to obtain information about how well the task group op- poor and 10 being excellent), the average response to how the
erated, as well as basic information on time and computer hours group meeting format functioned yielded 7.33/10 with a standard
that had been used to perform the work discussed in this paper. deviation of 1.74. Considering the challenges of the logistics and
Finally the questionnaire has been filled out by 22 participants of space issues that took place at our in-person meetings (where only
the task group, which represents slightly over 50% of the members one room was allotted for the total task group, hence breaking into
who participated in design groups. An effort was made to insure 5 or 6 sub-groups required some groups meeting in hotel lobbies,
that key members of the design teams participated in the survey. etc.), this result was quite satisfactory. As for how well the groups
In order to put the assessment results in context, it is helpful to functioned overall, the average was 7.50/10 with a standard devia-
understand how the group operated. First of all, AVT-251 only met tion of 1.62. Again, this was a very satisfactory result and showed
in person two times per year, which means the in-person meetings that, overall, the members of AVT-251 found the methods of meet-
had to contain a great deal of time for the design teams to meet ing and communicating fairly successful.
(approximately half of the meeting time was reserved for discipline The final questions of importance dealt with how well the over-
team meetings). In order to make progress, many of the groups all design of MULDICON took place. The first question was “What
had to conduct tele-conferences at regular intervals between the is your impression of the overall collaboration effectiveness on the
in-person meetings in order to coordinate and update each other design of MULDICON (entire design across the group borders).” The
on group progress. All other communication for the design groups response to this question was 6.40/10 with a standard deviation of
was conducted via email and phone calls among group members, 1.85. The final question was “Do you think the final MULDICON is a
as well as by using the group sharepoint site for exchanging docu- reasonable design for the requirements?” which had a response of
ments and files. This situation represents an unusual situation for 6.53/10 with a standard deviation of 1.50. While these results were
a typical aircraft design team, but in general the team managed to slightly lower than the previous questions regarding how the sub-
work together well and make reasonable progress. groups operated, they still represent reasonably good results for a
Another important aspect which significantly contributes to the task group that was only meeting twice a year, using a sharepoint
performance of the task group is the continuous recall of the aims site to share results and decisions, and occasional telecons. Dur-
and scientific problems to be answered by AVT-251, specifically: ing the final meeting of AVT-251, several members mentioned that
the typical constraints of an aircraft design process led to a great
• How do the tools contribute to the design process? deal of work taking place in the final year of the task group, which
• How do the tools accelerate the design process? did not allow for a re-evaluation and re-design to take place (as
• How do we arrange the tools in sequence or in parallel during would normally be done in industry). Overall, AVT-251 operated
the design process? quite well and the results were interesting and satisfactory for a
• To what degree CFD methods can provide sufficient data for a team that met for a total of three years with the constraints dis-
flight mechanics model? cussed previously.
• To what degree other disciplines can provide inputs to the pro-
cess? 5.3. Conclusions

While focusing on the technical details and challenges of the AVT-251 was a natural follow-on task group to AVT-161 and
MULDICON design work, these questions tended to be moved AVT-201. While these two groups had concentrated on the aero-
slightly into the background. Finally, the overall result of AVT-251 dynamics of the SACCON configuration, AVT-251 had taken on the
 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764 19

challenge of making the vehicle able to achieve specific mission AVT-113, AVT-161, AVT-183, and AVT-201. The hard work and ded-
requirements that were typical for an advanced, agile UCAV con- ication of the members of those task groups have made the work
figuration. The design trade studies were conducted within the of AVT-251 possible.
framework of multiple teams, including design, aerodynamics, con- The authors from UNB would like to thank the USAFA (Andrew
trols, structures, and engine integration. These teams were able Lofthouse and Matthew Satchell) for providing access to HPC clus-
to re-design SACCON with respect to certain constraints and re- ters necessary to conduct the research.
quirements and came up with an enhanced configuration named
MULDICON, which already satisfies most of these requirements. References
All of these studies and design aspects were conducted within a
[1] D. McDaniel, R. Cummings, K. Bergeron, S. Morton, J. Dean, Comparisons of CFD
group that lasted for three years, while only meeting in person
solutions of static and maneuvering fighter aircraft with flight test data, in: 3rd
twice a year. The details of the design studies were included in International Symposium on Integrating CFD and Experiments in Aerodynam-
four special sessions at the AIAA Aviation 2018 conference held ics, 2007, p. 39.
in June 2018, followed by a special issue of the journal Aerospace [2] L.A. Meyn, K.D. James, Full-scale wind-tunnel studies of F/A-18 tail buffet, J.
Science and Technology, to which this article belongs. They are fur- Aircr. 33 (3) (1996) 589–595, https://doi.org/10.2514/3.46986.
[3] D.A. Lovell, Military vortices, in: NATO RTO/AVT Symposium on Advanced Flow
ther published in the final report of AVT-251, which is currently in Management, vol. RTO-MP-069-I, 2003.
the publication process. The time and resource requirements of the [4] S. Jacobson, R. Britt, D. Freim, P. Kelly, Residual pitch oscillation (RPO) flight test
study were recorded, as well as results of how well the task group and analysis on the B-2 bomber, AIAA Paper 1998-1805, in: Structures, Struc-
worked and how effective the resulting design was able to achieve tural Dynamics, and Materials and Co-located Conferences, American Institute
of Aeronautics and Astronautics, 1998.
the requirements and constraints of the mission.
[5] R. Hall, R. Biedron, D. Ball, D. Bogue, J. Chung, B. Green, M. Grismer, G. Brooks,
The major novelty of the AVT-251 design process is the fact J. Chambers, Computational Methods for Stability and Control (COMSAC): the
that all the design work being performed was solely based on Time Has Come, AIAA Paper 2005-6121, in: Guidance, Navigation, and Control
CFD simulations. During the extensive studies of the predecessor and Co-located Conferences, American Institute of Aeronautics and Astronau-
tics, 2005.
task groups, a great amount of experience on the shape of the
[6] Vortex Breakdown over Slender Delta Wings, Final Report of the AVT-080 Task
developing flow structures, as well as on the correct application Group RTO-TR-AVT-080 AC/323(AVT-080)TP/253, NATO RTO/AVT, October 2009,
of modern CFD methods for such type of flow had been devel- https://doi.org/10.14339/RTO-TR-AVT-080.
oped. Relying on this expertise and the corresponding confidence [7] D. Hummel, et al., Understanding and Modeling Vortical Flows to Improve the
Technology Readiness Level for Military Aircraft, Final Report of the AVT-113
in the numerical results, the team members could perform a huge
Task Group RTO-TR-AVT-113 AC/323(AVT-080)TP/253, NATO RTO/AVT, October
amount of different parameter studies in parallel – without the 2009, https://doi.org/10.14339/RTO-TR-AVT-113.
necessity to validate each single step by wind tunnel experiments. [8] D. Hummel, Review of the second international vortex flow experiment
Even though this advantage cannot be quantified due to a lack of (VFE-2), AIAA Paper 2008-377, in: Aerospace Sciences Meetings, American In-
data for a comparable reference effort, it becomes obvious that a stitute of Aeronautics and Astronautics, 2008.
[9] R.M. Cummings, A. Schütte, Integrated computational/experimental approach
similar process incorporating extensive low- and high speed wind to unmanned combat air vehicle stability and control estimation, J. Aircr. 49 (6)
tunnel campaigns for a step-by-step evolving aircraft configuration (2012) 1542–1557.
(including control surface design and intake optimization) would [10] R.M. Cummings, A. Schütte, Assessment of Stability and Control Prediction
not have been possible within a similar three-year task group. Methods for NATO Air and Sea Vehicles, Final Report of the AVT-161 Task
Group RTO-TR-AVT-161 AC/323(AVT-161)TP/440, NATO RTO/AVT, September
So, this study represents a good example of how modern, well
2012, https://doi.org/10.14339/RTO-TR-AVT-161.
validated design and analysis tools can streamline the design pro- [11] A. Hövelmann, C. Breitsamter, Leading-edge geometry effects on the vortex for-
cess, as well as being able to come up with an enhanced config- mation of a diamond-wing configuration, J. Aircr. 52 (5) (2015) 1596–1610,
uration within a reasonable short period of time. The MULDICON https://doi.org/10.2514/1.C033014.
[12] N.T. Frink, Numerical analysis of incipient separation on 53-deg swept diamond
configuration has similarities to a number of other modern UCAVs,
wing, AIAA Paper 2015-88, in: AIAA SciTech Forum, American Institute of Aero-
and represents an already mostly satisfactory design that would nautics and Astronautics, 2015.
have controllable flight characteristics at angles of attack that will [13] R.M. Cummings, C.M. Liersch, A. Schütte, K.C. Huber, Aerodynamics and con-
make the configuration agile and capable of fulfilling more chal- ceptual design studies on an unmanned combat aerial vehicle configuration, J.
Aircr. 55 (2) (2018) 454–474, https://doi.org/10.2514/1.C033808.
lenging missions.
[14] R.M. Cummings, A. Schütte, Extended Assessment of Stability and Control Pre-
diction Methods for NATO Air Vehicles, Final Report of the AVT-201 Task
Declaration of competing interest Group STO-TR-AVT-201, NATO STO/AVT, November 2016, https://doi.org/10.
14339/STO-TR-AVT-201.
[15] D.D. Vicroy, K.C. Huber, A. Schütte, M. Rein, J.P. Irving, G. Rigby, T. Löser, A.-R.
The authors declare that they have no known competing finan- Hübner, T.J. Birch, Experimental investigations of a generic swept unmanned
cial interests or personal relationships that could have appeared to combat air vehicle with controls, J. Aircr. 55 (2) (2016) 475–501, https://doi.
influence the work reported in this paper. org/10.2514/1.C033782.
[16] K.C. Huber, D.D. Vicroy, A. Schütte, A.R. Hübner, UCAV model design and static
experimental investigations to estimate control device effectiveness and S&C
Acknowledgements capabilities, AIAA-Paper 2014-2002, in: AIAA AVIATION Forum, American Insti-
tute of Aeronautics and Astronautics, 2014.
[17] D. Zimper, D. Hummel, Analysis of the transonic flow around a unmanned
The authors would like to acknowledge all of the nations, or-
combat aerial vehicle configuration, J. Aircr. 55 (2) (2016) 571–586, https://
ganizations, and individuals that participated in the various phases doi.org/10.2514/1.C033697.
of the task group. Without the hard work and dedication of all [18] J. Coppin, T. Birch, D. Kennett, G. Hoholis, K. Badcock, Prediction of control
of these contributors the work of the group would not have been effectiveness for a highly swept unmanned air vehicle configuration, J. Aircr.
possible. 55 (2) (2016) 534–548, https://doi.org/10.2514/1.C033988.
[19] M. Ghoreyshi, M.E. Young, A.J. Lofthouse, A. Jirasek, R.M. Cummings, Numerical
The authors from DLR would like to thank the German MoD simulation and reduced-order aerodynamic modeling of a lambda wing config-
and The Federal Office of Bundeswehr Equipment, Information uration, J. Aircr. 55 (2) (2016) 549–570, https://doi.org/10.2514/1.C033776.
Technology and In-Service Support (BAAINBw) for their support for [20] A. Schütte, K.C. Huber, N.T. Frink, O.J. Boelens, Stability and control investiga-
the military research at DLR and the support to attend the NATO tions of generic 53 degree swept wing with control surfaces, J. Aircr. 55 (2)
(2016) 502–533, https://doi.org/10.2514/1.C033700.
STO/AVT Task Group meetings. [21] R.M. Cummings, C.M. Liersch, A. Schütte, Multi-disciplinary design and per-
Russell Cummings would like to thank all of the members of formance assessment of effective, agile nato air vehicles, in: AIAA AVIATION
the Task Groups that preceded and informed AVT-251, including Forum, American Institute of Aeronautics and Astronautics, 2018.
20 C.M. Liersch et al. / Aerospace Science and Technology 99 (2020) 105764

[22] C.M. Liersch, K.C. Huber, A. Schütte, D. Zimper, M. Siggel, Multidisciplinary de- [43] J.-F. Le Roy, S. Morgand, D. Farcy, Static and dynamic derivatives on generic
sign and aerodynamic assessment of an agile and highly swept aircraft configu- UCAV without and with leading edge control, AIAA Paper 2014-2391, in: AIAA
ration, CEAS Aeronaut. J. 7 (4) (2016) 677–694, https://doi.org/10.1007/s13272- AVIATION Forum, American Institute of Aeronautics and Astronautics, 2014.
016-0213-4. [44] C.M. Liersch, M. Hepperle, A distributed toolbox for multidisciplinary pre-
[23] Department of Defense, USA, “Glossary of Definitions, Ground Rules, and Mis- liminary aircraft design, CEAS Aeronaut. J. 2 (1–4) (2011) 57–68, https://
sion Profiles to Define Air Vehicle Performance Capability,” Military Standard, doi.org/10.1007/s13272-011-0024-6.
February 2003. MIL-STD-3013. [45] B. Nagel, T. Zill, E. Moerland, D. Böhnke, Virtual aircraft multidisciplinary analy-
[24] Department of Defense, USA, “Flying Qualities of Piloted Aircraft,” Military sis and design processes - lessons learned from the collaborative design project
Standard, July 2001. MIL-STD-1797A. VAMP, in: CEAS 2013 European Air and Space Conference, 2013.
[25] C.M. Liersch, K.C. Huber, Conceptual design and aerodynamic analyses of a
[46] B. Nagel, D. Böhnke, V. Gollnick, P. Schmollgruber, A. Rizzi, G.L. Rocca, J.J.
generic UCAV configuration, in: AIAA AVIATION Forum, American Institute of
Alonso, Communication in aircraft design: can we establish a common lan-
Aeronautics and Astronautics, 2014.
guage?, in: 28th International Congress of the Aeronautical Sciences (ICAS),
[26] J. Schwithal, D. Rohlf, G. Looye, C.M. Liersch, An innovative route from wind
2012.
tunnel experiments to flight dynamics analysis for a highly swept flying wing,
[47] D. Seider, P. Fischer, M. Litz, A. Schreiber, A. Gerndt, Open source software
CEAS Aeronaut. J. 7 (4) (2016) 645–662, https://doi.org/10.1007/s13272-016-
framework for applications in aeronautics and space, IEEE Aerospace Confer-
0214-3.
ence, 2012.
[27] A. Schütte, K.C. Huber, D. Zimper, Numerische aerodynamische analyse und Be-
wertung einer agilen und hoch gepfeilten Flugzeugkonfiguration, in: Deutscher [48] C.M. Liersch, G. Bishop, Conceptual design of a 53 deg swept flying wing UCAV
Luft- und Raumfahrtkongress, 2015. configuration, in: AIAA AVIATION Forum, American Institute of Aeronautics and
[28] P. Löchert, K.C. Huber, C.M. Liersch, A. Schütte, Control device studies for yaw Astronautics, 2018.
control without vertical tail plane on a 53◦ swept flying wing configuration, [49] A. Schütte, J. Vormweg, R.G. Maye, T.L. Jeans, Aerodynamic shaping design
in: AIAA AVIATION Forum, American Institute of Aeronautics and Astronautics, and vortical flow design aspects of a 53deg swept flying wing configuration,
2018. in: AIAA AVIATION Forum, American Institute of Aeronautics and Astronautics,
[29] A. Schütte, D. Hummel, S.M. Hitzel, Flow physics analyses of a generic un- 2018.
manned combat aerial vehicle configuration, J. Aircr. 49 (6) (2012) 1638–1651, [50] R.K. Nangia, J. Coppin, M. Ghoreyshi, UCAV wing design, assessment and com-
https://doi.org/10.2514/1.C031386. parisons, in: AIAA AVIATION Forum, American Institute of Aeronautics and
[30] A. Schütte, Numerical investigations of vortical flow on swept wings with Astronautics, 2018.
round leading edges, J. Aircr. 54 (2) (2017) 572–601, https://doi.org/10.2514/ [51] M.P.C. van Rooij, R.M. Cummings, Aerodynamic design of an unmanned combat
1.C034057. air vehicle in a collaborative framework, in: AIAA AVIATION Forum, American
[31] N. Kroll, S. Langer, A. Schwöppe, The DLR flow solver TAU - status and re-
Institute of Aeronautics and Astronautics, 2018.
cent algorithmic developments, AIAA Paper 2014-080, in: AIAA SciTech Forum,
[52] Y.J. Hasan, J. Flink, S. Freund, T. Klimmek, R. Kuchar, C.M. Liersch, G. Looye,
American Institute of Aeronautics and Astronautics, 2014.
E. Moerland, T. Pfeiffer, M. Schrader, S. Zenkner, Stability and control investi-
[32] A. Jameson, Application of dual time stepping to fully implicit Runge Kutta
gations in early stages of aircraft design, in: AIAA AVIATION Forum, American
schemes for unsteady flow calculations, AIAA Paper 2015-2753, in: AIAA AVIA-
Institute of Aeronautics and Astronautics, 2018.
TION Forum, American Institute of Aeronautics and Astronautics, 2015.
[33] P.R. Spalart, S.R. Allmaras, A one-equation turbulence model for aerodynamic [53] S. Zenkner, R.-G. Becker, Preliminary engine design for the MULDICON con-
flows, Rech. Aérosp. 1 (1994) 5–21. figuration, in: AIAA AVIATION Forum, American Institute of Aeronautics and
[34] S.R. Allmaras, F.T. Johnson, P.R. Spalart, Modifications and clarifications for the Astronautics, 2018.
implementation of the Spalart-Allmaras turbulence model, in: 7th International [54] C. Voß, M. Trost, R.-G. Becker, Automated optimization of the MULDICON in-
Conference on Computational Fluid Dynamics, Big Island, Hawaii, 2012, IC- let with minimum losses and reduced sight onto the compressor front face,
CFD7–ICCFD1902. in: AIAA AVIATION Forum, American Institute of Aeronautics and Astronautics,
[35] CENTAUR Homepage, http://www.centaursoft.com. (Accessed 15 February 2018.
2019). [55] H. Edefur, M. Tormalm, L. Tysell, M. Quas, Design and integration of a low ob-
[36] W. Strang, R. Tomaro, M. Grismer, The defining methods of Cobalt-60 - a par- servable intake for the MULDICON platform, in: AIAA AVIATION Forum, Ameri-
allel, implicit, unstructured Euler/Navier-Stokes flow solver, in: Aerospace Sci- can Institute of Aeronautics and Astronautics, 2018.
ences Meetings, American Institute of Aeronautics and Astronautics, 1999. [56] P. Aref, A. Jirasek, M. Ghoreyshi, R.M. Cummings, M.J. Satchell, Computational
[37] J. Gottlieb, C. Groth, Assessment of Riemann solvers for unsteady one- design of S-duct intakes for the NATO AVT-251 multi-disciplinary configuration,
dimensional inviscid flows of perfect gases, J. Comput. Phys. 78 (1988) in: AIAA AVIATION Forum, American Institute of Aeronautics and Astronautics,
437–458, https://doi.org/10.1016/0021-9991(88)90059-9. 2018.
[38] R. Tomaro, W. Strang, L. Sankar, An implicit algorithm for solving time depen- [57] J.M. Schweiger, A.M. Cunningham Jr., M. Dalenbring, A. Voß, E. Sakarya, Struc-
dent flows on unstructured grids, in: Aerospace Sciences Meetings, American tural design efforts for the MULDICON configuration, in: AIAA AVIATION Forum,
Institute of Aeronautics and Astronautics, 1997. American Institute of Aeronautics and Astronautics, 2018.
[39] R.F. Tomaro, W.Z. Strang, K.E. Wurtzler, Can legacy codes scale on tens of thou-
[58] A. Voß, Gust loads calculation for a flying wing configuration, in: AIAA AVIA-
sands of PEs or do we need to reinvent the wheel?, in: High Performance
TION Forum, American Institute of Aeronautics and Astronautics, 2018.
Computing Modernization Program Contributions to DoD Mission Success, DoD
[59] G. Voß, D. Schaefer, C. Vidy, Investigations on flutter stability of the DLR-
High Performance Computing Modernization Office, 2012, pp. 231–236.
F19/SACCON configuration, in: AIAA AVIATION Forum, American Institute of
[40] A.J. Lofthouse, M. Ghoreyshi, A. Jirasek, R.M. Cummings, Static and dynamic
Aeronautics and Astronautics, 2018.
simulations of a generic UCAV geometry using the Kestrel flow solver, in: AIAA
[60] E. Sakarya, C. Kocan, B. Okumus, A study on evaluation of aeroelastic charac-
AVIATION Forum, American Institute of Aeronautics and Astronautics, 2014.
[41] Cobalt User’s Manual - Cobalt Version 6.0, Cobalt Solutions, LLC, February 2013. teristics of a UCAV configuration, in: AIAA AVIATION Forum, American Institute
[42] D.J. Kennett, G. Hoholis, K.J. Badcock, Numerical simulation of control surface of Aeronautics and Astronautics, 2018.
deflections over a generic UCAV configuration at off-design flow conditions, [61] CATIA Homepage, URL: http://www.3ds.com/de/produkte-und-services/catia/.
AIAA Paper 2014-2134, in: AIAA AVIATION Forum, American Institute of Aero- (Accessed 15 February 2019).
nautics and Astronautics, 2014.