Modelling & Analysis of a Tilt Wing Aircraft

Felix Rubin∗
KTH, Royal Institute of Technology Stockholm, Sweden

The aim of this report is the investigation of a hybrid vertical take-off and landing (VTOL)
tilt wing aircraft which is in development at the company Dufour Aerospace. Using a model,
programmed in MATLAB® different stages of flight can be simulated and investigated. Main
investigation area of this report is the transition between cruise and hover conditions of the
aircraft.
The simulation is based on the six-degree-of-freedom nonlinear equations of motion for
aircraft modified for tilt wing operation. The model characteristics have been determined
using various CFD programs, wind tunnel data, as well as numerical and handbook methods.
The main focus of modeling lies on the static longitudinal aerodynamic coefficients, the propeller
and engine coefficients as well as a propeller slipstream model. Furthermore lateral directional
aerodynamic coefficients and dynamic effects and a tail blower (Notar) system are modeled.
As results, aerodynamic characteristics of the wing in the propeller slipstream are shown
and discussed and the transition conditions are investigated by ’trimming’ the model at equi-
librium points over its speed range and analyzing the resulting power requirements.

I. Introduction
As is common knowledge in the field of aeronautics, it is extremely hard to design aircraft to be both efficiently
operating in low and high speed conditions. Good low speed performance is wished for in aircraft that have the ability
to use small distances for take off and landing and even sustain hovering flight. These conditions favor rotorcraft
and aeroplanes with low wing loading and high static thrust. High speed performance however, favors fixed wing
aircraft with less wing surface area (and thus higher wing loading), and propellers designed to efficiently operate in said
conditions.
In the last decades improvements in performance of fixed wing aircraft has stagnated while a surge in design and
performance improvements of small rotorcraft manifested. Development and demand of small electrically powered
unmanned aircraft is rising due to decreasing cost of components and advancements in technology. Increased power
density in recent batteries allows for their weight to be reduced so as to be used successfully in such aircraft.

Fig. 1 Dufour Aerospace aeRo 2 in cruise (left) and hover (right) mode [1]

Considering the statements made above, a great challenge is found in attempting the development of an aircraft
that is both able to display vertical take-off and landing (VTOL) capabilities as well as sustain good high speed cruise
performance while being electrically powered. While there is an abundance of small unmanned electrically driven
vehicles as well as quite a few manned VTOL aircraft, the combination of attempts to design an electrically powered
∗ MSc. Student, Aerospace Engineering

1
manned aircraft with VTOL capabilities is sparse to say the least. However, exactly this is the goal of the Swiss company
Dufour Aerospace [1] which is in development phase of the aEro 2, a hybrid tilt wing aircraft.
Tilt wing aeroplanes are defined, as the name says, by their capability to rotate their main wing. Using this possibility,
the wing mounted propeller engines are able to change their thrust force vectors upwards and thus sustain lower flight
speeds than conventional fixed wing aircraft and operate as rotorcraft in the hover regime.
While both, high speed and hover conditions need to be researched, the design of such an aircraft calls for a
throughout investigation of the transition regime between those two phases. This report takes aim at an initial attempt to
do so.

II. Methods

A. Definitions and Assumptions

1. Assumptions and Simplifications

The aircraft is assumed to be:
• ...symmetric around xz plane
• ...modeled as a rigid body with 2 spinning rotors
• ...having constant mass over time
• ...using flat earth as inertial reference
• ...flying under ISA (International Standard Atmosphere) conditions with no additional wind added

2. Axis and Definitions

Figure 2 displays the axis and variable definitions that will be used in the following sections.

Fig. 2 Definitions

B. Equations of Motion
This section intends to lay out the governing equations of motion during the cruise, hover and transition flight
regimes. The implemented nonlinear mathematical model has 6 degrees of freedom. The equations, derived from
Newtons second law of motion (Eq. (1) left), sum up all external forces acting on the body. The resulting force then
equals the time derivative of the momentum of the body (e.g. the aircraft) in respect to inertial space. Applying the
same approach, one can define the total moment of a body as the time derivative of its own moment of momentum
around the center of gravity (Eq. (1) right).
®
F= M=
Í® d ® Í ® mH
dt (mV) and dt
(1)
Using figure 2 as reference for definition of axes and angles and Eq. (1) as guideline, one is now able to derive the
equations which define the movement of a body in all 6 degrees of freedom by summing external forces to the left hand
side and internal momentum to the right hand side in Eq. (2) and Eq. (3) [2].

2
 FG x + FA x + FEx + FB x = m(uÛB + qw B − rv B )
FGy + FAy + FEy + FBy = m(Ûv B + ruB − pw B ) (2)
FGz + FAz + FEz + FBz = m(wÛ B + pv B − quB )

L A + LE + LB = pI
Û xx + qr(Izz − Iyy ) − (rÛ + pq)Ixz + hg L
M A + ME + MB = qI Û yy − pr(Izz − Ixx ) + (p2 − r 2 )Ixz + hg M (3)
N A + NE + NB = rÛ Izz + pq(Iyy − Ixx ) + (qr − p)I
Û xz + hg N
Where u, v and w are linear p, q and r are angular velocities and the superscript B denotes a body fixed coordinate
system of the aircraft. The subscribt G stands for gravitational, E for engine, A for aerodynamic and, B for blower.
Inertial moments are denoted by I with their respective x x,yy,zz,xz planes as subscripts. Additionally, the effects of the
spinning rotors of the engines are added as angluar momentum hg .
Changing references, the angular (Eq. (4)) as well as linear (Eq. (5)) velocities are transformed into an earth fixed
coordinate system as follows:

pE 1 0 −sinθ φÛ
®=­ 0
© ª © ª© ª
­ qE cosφ sinφcosθ ® ­ θÛ ® (4)
­ ® ­ ®­ ®
« r
E
¬ « 0 −sinφ cosφcosθ ¬« ψÛ ¬
xÛ E cosθcosψ sinφsinθcosψ − cosφsinψ cosφsinθcosψ + sinφsinψ uB
® = ­ cosθsinψ
© ª © ª© ª
­ yÛ E
­ ® ­ sinφsinθsinψ + cosφcosψ cosφsinθsinψ − sinφcosψ ® ­ vB
®­
®
® (5)
E B
« zÛ ¬ « −sinθ sinφcosθ cosφcosθ ¬« v
¬
Where φ, θ and ψ are euler angles determining the aircrafts pitch, roll and yaw attitude in reference to earth.
Combining equations (2), (3), (4) and (5), a system of 12 independent differential equations with 12 implicit dependent
variables is found, representing the aircraft in its 6 degrees of freedom in an earth fixed coordinate system. The state
vector containing said variables is defined as:

x = (p, q, r, φ, θ, ψ, u, v, w, x E, y E, zE ) (6)

C. Wing to Body Coordinate System Transformation

Compared to conventional fixed wing aircraft, the effect of a tilted wing allows forces and moments originating at the
wing to change in direction relative to the crafts xz plane in its body axis. Therefore, a wing fixed coordinate reference
system is introduced (fig. 2 - right). The wing fixed coordinate system moves with respect to the body fixed system
when the wing is tilted. Wing forces are calculated in the wing based coordinate system first and then transformed into
the body fixed system via the wing tilt angle δW Eq. (7).

F B cos(δW ) ∗ FWX W
© WX
®=­
ª © ª
­ FW B FWY W ® (7)
­ Y ® ­ ®
B W
« FW Z ¬ « sin(δW ) ∗ FWz ¬
Where the wing force vector FW is split in its x, y and z components and the superscripts B and W signify the body
and wing fixed coordinate systems.
As the wing coordinate system originates at a different reference point (25% of mean aerodynamic chord) than the
body fixed system (origin at center of gravity) (see fig. 2), this has to be accounted for as well when transforming the
moments originating at the wing (Eq. (8).

L B L W + FWZ B ∗ ydist + FWY B ∗ zdist

© W ª © W
® = ­ MW W + FW B ∗ zdist + FW B ∗ xdist
ª
­ MW B ® (8)
­ ® ­ X Z ®
¬ « NW + FYY ∗ xdist + FWX ∗ ydist
B W B B
« NW ¬
Where xdist , ydist and ydist signify the respective distances from wing reference point to center of gravity.

3
D. Trimming
In order to be able to make viable statements about the aircrafts performance, one must be able to ’trim’ the model
to reference flight conditions. This means finding equilibrium points at which the dynamic system described above is in
steady state. In order to find the required equilibrium state, it is useful to introduce the control vector containing the
control variables:

u = (δE , δ A, δR , δF , δW , δ PL , δ PR , δ PI , δY A) (9)
Where the subscripts stand for: elevator, aileron, rudder, f laps, wing tilt, le f t engine power setting, right
engine power setting, tail blower pitch control setting and tail blower yaw control setting.
In order to find an equilibrium around a specific flight condition, the state variables and control inputs which define
said condition are set and ’frozen’ to their specific values. Finding the trim point is then achieved by assigning the rest
of the ’non-frozen’ control and state vector inputs so that (Eq. (10)) is fullfilled. This approach is iterative in nature and
is solved by using a multivariate Newton-Raphson algorithm [3] in order to minimize the state vector derivative.
dx f rozen
=0 (10)
dt

E. Weight and Balance

1. Mass, Center of Gravity & Intertial Moments

With an estimation of the mass of each component as well as their placement to a reference point, the aircrafts total
mass m and center of gravity (C.G.) position are found. Since the aircrafts main power source is electric power, it can be
assumed, that the craft does not vary in mass during flight. However, aircraft mass and center of gravity position are
depending on the number of pilots (1 or 2) as well as the placement and mass of baggage which needs to be accounted
for. Maximum take-off mass is set to be 715 Kg.
Modelling all mayor components of the craft as point masses with their respective x, y,and z lever arms to the center
of gravity, the inertial moments Ixx , Iyy , Izz and Ixz of the complete system are computed. The subscripts express the
moments of inertia about each of its three axis as well as the product of inertia about its xz axis. Assuming symmetry
about the xz plane the Ixy and Iyz products of inertia are found to be zero and can thus be omitted.

2. Tilt Wing Considerations

As the lever arms of the wing internal parts to the C.G will change when changing the tilt angle δW , there will
be a variation in C.G position as well as the moments of inertia. To account for this, the model interpolates these
values between the maximum and minimum wing tilt angle settings, thus I and C.G. position are defined as functions
dependent on δW .

F. Power System
The main propulsive force of the model stems from two wing mounted variable pitch propeller engines driven
by electrical motors with battery storage in the wings. Additionally, a tail blower or no-tail-rotor (NOTAR) system,
which is driven by a small turbine inside the fuselage is used to assist in hover and low speed flight conditions as the
aerodynamic force on the tail surface will not be sufficient to control the craft. Table 1 gives an overview on some
values of the system.

Propeller diameter D pr op 2.4 m

Number of blades Nprop 3
Shaft power main engines Pmax 100 kW
Power of turbine Notar system Pmax B 50 kW
Total power of system PT OT 250 kW
Maximum thrust main engines FTma x 3700 N
Maximum thrust notar system FBma x 500 N

Fig. 3 aEro2 side view with Notar system in tail Table 1 Power System Overview

4
1. Main Engines - Engine Data and Controller
Since the propeller design for the aircraft in question has not been finished yet, the necessary data for this simulation
had to be assumed. In order to build a database for power usage, thrust, propeller speed and propeller blade angle setting,
data from wind tunnel tests of wing mounted propellers optimized for cruise flight [4] was interpolated with theoretic
rotor efficiency of helicopter rotor blades in hover [5] scaled to the correct propeller geometry data resulting in figure 4.

Fig. 4 Propeller efficiency and blade angle

Figure 4 displays the propeller efficiency η plotted over the engine power coefficient CP and propeller advance ratio
J. Additionally the propeller blade angle (at 75% blade radius) can be found for each value of η. Given these data,
engine thrust can be calculated using the relations in Eq. (11) where CT is the propeller thrust coefficient, n and D are
the propeller rotational speed and diameter, P is engine shaft power and T is the thrust force generated by the system.

η= CT
Cp J , J= V∞
nD , CP = P
ρn3 D 5
, CT = T
ρn2 D 4
, (11)
The engine shaft power is known and given by Eq. (12), where Pmax signifies the maximum available shaft power
and δ P is the power setting control variable (ranging form 0 to 1). Also known are the air density ρ at the respective
altitude as well as the propeller diameter D.

P = Pmax ∗ δ P (12)
To find maximum thrust an engine controller is implemented which maximizes the below function (Eq. (13)) in
dependency of propeller rotations for each given power setting and altitude. In a final step, the respective propeller
blade angle β is found by interpolation in figure 4.

CP (n) ∗ η(n)
Tmax = max ( ∗ ρ ∗ n2 ∗ D 4 ) (13)
0≤n≤nma x J(n)

2. Main Engine - Other Forces & Moments

Next to the main thrust force, further engine generated forces and moments cannot be omitted in the simulation.
Therefore, a selection of moments and forces which are implemented is given below [6]:
Engine torque (Eq. (14)), where P stands for engine shaft power and Ω signifies the propeller angular speed.
P
Ltorque = (14)
Ω

5
 Yawing moment due to pitch rate and pitching moment due to yaw rate (Eq. (15)) where I Prop is propeller moment
of inertia modeled as solid thin disk of equal mass (Eq. (16)) with m prop and r prop being propeller mass and radius.

Ngyr o = I Pr op ∗ Ω ∗ q , Mgyro = I Prop ∗ Ω ∗ r (15)

1
I Prop = 2
∗ m prop ∗ r prop (16)
2
Pitch and yaw damping moment produced by the propeller when rotated about the yaw and pitch axes (Eq. (17)),
where k d is a factor dependent on propeller solidity. The damping arises from a difference in lift at opposite sides of the
propeller disk which can be traced back to an in- or decrease in local angle of attack.
ρ q
Mdamping = k d ∗ 2 5
2 ∗ Ω ∗ r prop ∗ arctan r pr o p
ρ (17)
Ndamping = k d ∗ 2 5 r
2 ∗ Ω ∗ r prop ∗ arctan r pr o p

3. Notar System
Implementation of the tail blower system is straight forward, given the maximum thrust produced by the turbine the
blower force is found by Eq. (18):

FB pi t c h = δ pi ∗ FBma x
FBy aw = δya ∗ FBma x
(18)
where δ pi & δya = {−1., .1}
and (0 ≤ |δ pi | + |δya | ≤ 1)
The sign of the controls δ pi and δya then indicates the direction of the force (up, down, left or right).

G. Aerodynamics

1. Aerodynamic Coefficients & Geometry

A table of all the aerodynamic coefficients used in the simulation and the respective methods of modeling can be
found in the attachments to this report next to an overview of the geometry and airfoil profiles of the main aerodynamic
surfaces.
The translation from aerodynamic coefficients to forces and moments is achieved via Eq. (19).

L= 1
2 · ρ · V · S · CL , D = 21 · ρ · V · S · CD , F y = 12 · ρ · V · S · CY
(19)
R= 1
2 · ρ · V · b · S · CR , M = 12 · ρ · V · c · S · CM , N = 12 · ρ · V · b · S · CN
Where S stands for wetted surface area, V for airspeed, ρ for air density, c and b for mean chord and mean wing span.

2. Slipstream Model
As can be expected, the flow conditions which meet the wing behind the propeller will be different from the free
stream conditions. This is illustrated in figure 5. In order to account for this, a slipstream model is implemented with
the goal of defining the correct slipstream velocity, separating wing area inside and outside of the slipstream, finding
the correct angle of attack for all wing areas and correcting the wing lift coefficient due to slipstream effects. Using
propeller momentum theory [7], the slipstream velocity and diameter are defined as ((Eq. (20)):
q
Vs = 1 ρπrT2 + (V∞2 )
pr o p q
2
(20)
Ds = 2 ∗ r prop ∗ 1+µ 2 where µ = V∞
Vs

Where r pr op is the propeller diameter,T is thrust force, ρ is air density and V∞ is the freestream velocity.
Calculation of the slipstream and freestream angles of attack as well as the lift coefficient correction are omitted
here for the sake of brevity and can be found in [8], [8] and [8].

6
Fig. 5 Wing in slip stream (blue) and free stream (red) displayed on right wing and wing airfoil profiles shown
on left wing

3. Wing and Longitudinal Aerodynamic Forces and Moments

As figure 5 displays, the wing is parted into two airfoil profiles (N AC A 2412 &N AC A 2409) which change at the
engine mount. Also, the airflow conditions vary along the wing as is seen in the previos subsection. Thus each side of
the main wing can be split into four sections:
• S1 - Inner profile (CL2412 , CD2412 , Cm2412 ), freestream conditions (α∞ , V∞ )
• S2 - Inner profile (CL2412 , CD2412 , Cm2412 ), slipstream conditions (αS , VS )
• S3 - Outer profile (CL2409 , CD2409 , Cm2409 ), slipstream conditions (αS , VS )
• S4 Outer profile (CL2409 , CD2409 , Cm2409 ), freestream conditions (α∞ , V∞ )
As was found in Eq. (20), the slipstream diameter varies in size depending the propeller thrust and thus the four wing
areas vary in size as well. As the slipstream also changes the airspeed VS and angle of attack αS behind the propeller
all four sections will produce a different amount of aerodynamic forces and moments. Furthermore, because of the
different angles of the airspeed vectors, the lift and drag force vectors of the slipstream and freestream areas do not
act in the same direction as shown in figure 6, where the subscript S denotes slipstream conditions (αS & VS ) and the
resulting lift and drag forces, which act in a different direction when compared to the forces resulting from free stream
conditions (α∞ & V∞ )

Fig. 6 Wing longitudinal aerodynamic forces

7
 This means each aerodynamic force vector must be translated into the wing fixed coordinate system via its respective
angle of attack and can then summed up with the rest of the vectors (Eq 22). Since the reference point of both the airflow
based system as well as the wing fixed system coincides, the pitching moment need not be translated.

FX W = sin(α∞ ) ∗ (L1 + L4 ) + sin(αS ) ∗ (L2 + L3 ) − cos(α∞ ) ∗ (D1 + D4 ) − cos(αS ) ∗ (D2 + D3 )

(21)
FZ W = −cos(α∞ ) ∗ (L1 + L4 ) − cos(αS ) ∗ (L2 + L3 ) − sin(α∞ ) ∗ (D1 + D4 ) − sin(αS ) ∗ (D2 + D3 )

4. Wing Downwash Model

As the wing produce lift a downwash of the airflow behind the wing changes the tail angle of attack, which can be
represented as (Eq. (22)):
κ ∗κ
 = vκb p CARL ≈ π163 CAR
Ll
(22)
αt ail = αwing + 
The values for κv , κ p and κb are graph interpolations depending on wing geometry and distance from wing to tail and
are found in [9]. The tail longitudinal aerodynamic forces are then implemented in similar manner to the wing forces.

5. Lateral Directional Aerodynamic Forces and Moments

The lateral directional aerodynamic force and moment coefficients stem from a full aircraft vortex lattice model
in reference to the aircraft body coordinate system. Thus, the coefficients are derived directly for the full aircraft in
dependence of the control variables δr and δa as well as angle of attack α and sideslip angle β. All coefficients used are
found in the attachment. Additionally, differences in lift, drag and pitching moment between the two wings will have an
impact on the lateral force and moment balance and are thus included as well.

III. Results
The resulting figures and plots mainly depict longitudinal characteristics since the biggest tilt wing effects are seen
in the xz−plane. However, all calculations are carried out in the full 6-degrees- of-freedom simulation model.

A. Power System

Fig. 7 Thrust plots

8
 The plots in figure 7 show engine trust at different airspeeds and power level settings at 2000m above mean sea level
(left hand side) and the maximum thrust at different altitude and airspeeds (right hand side). At the aircraft maximum
weight of 715Kg in order to purely hover without any excess power one propeller/engine need to be able to produce
(W ∗ g/2) ≈ 3507 N of thrust. As can be seen in the right hand side plot in figure 7, this is possible up to an altitude of
roughly 2150 m. Since the propeller model is only a preliminary estimation, these values might change with a different
propeller design.
Figure 7 shows engine maximum thrust diminishing rapidly with increasing airspeeds. Since, however at high
speeds, the aircraft functions purely in ’fixed wing’ mode, the thrust, produced by the engines easily suffices in order to
trim the craft at a goal cruise speed of 400 km/h (111 m/s). The propeller will however have to be carefully designed in
order to diminish mach effects. At high rotational as well as high free stream speeds propeller blade speeds are found to
be Eq. (23):
q
V
VBma x = V∞2 + (π ∗ nmax ∗ d Prop )2 and MBma x = B ma a
x (23)

When inserting the values n = 2400 RPM for maximum propeller rotational speed and Vcruise = 111 m/s for aircraft
cruise speed and using a sea level Mach 1 airspeed of a = 340 m/s and the given propeller radius of 2.4 m the propeller
blade speed is calculated to be VB = 321 m/s or MB = 0.95 which is high enough to produce a significant propeller
efficiency decrease due to mach shock effects on the propeller blades.

B. Aerodynamics

Fig. 8 Slipstream effects example - α = 7, height = 2000m and Airspeed = 25m/s

Figure 8 exemplifies the effects of the air in the propeller slipstream meeting the wing surface. It is visible that, with
increasing power of the engine, the slipstream area diminishes as the airspeed behind the propeller is accelerated. This
effect can be seen in the left and middle plots and is in accordance with the propeller momentum theory formulation in
Eq. (20). The rightmost plot displays diminishing angle of attack with increasing engine power. This has a favorable
effect on ’power on’ stall behavior of the wing as is discussed further below.
The left hand side plot in figure 9 shows the lift curve slope of all wing areas. The inner part of the wing uses a
different airfoil profile (N AC A 2412), than the outer part (N AC A 2409), and the middle section of the wing is subject
to the propeller slipstream and thus a different airspeed and angle of attack. Therefore, four different coefficient curves
exist. The blue graph shows a mean value for the lift coefficient. As is eminent, the coefficients in the slipstream area
(and thus also the mean coefficients) are depending on engine thrust. This is further exemplified in the middle and right
hand side plots of figure 9 which show lift curve slope and drag polar dependency on engine thrust. As thrust increases,
lift and drag coefficient decrease and the stall angle of attack is delayed.
Using Eq. 24 the ’power off’ stall speed can be calculated (using a mass of 715 Kg, a mean wing CLma x of 1.4 as
well as ISA sea level conditions):
s
W
VSi dl e = 1 ≈ 28m/s (24)
2 ρSC L ma x

9
 Fig. 9 Aerodynamic coefficients example

As is discussed above ’power on’ stall will depend on engine thrust and will be less pronounced and at lower
airspeeds as thrust increases. However, since the propeller blades will be affected by high free stream angles of attack,
there might be flow separation on the blades which will cause a decrease in power and might excite wing buffeting and
flow irregularities on the wing part affected by the slipstream. Furthermore, it needs to be considered that only the
middle wing part, which is subject to the propeller slipstream will remain unstalled at higher angles of attack. This
could lead to a loss of aileron effectiveness, since the ailerons are (for the most part) outboard of the area with attached
flow. In general however, flying a tilt wing aircraft at excessively high angles of attack would be unreasonable, since a
tilting wing mechanism is build to be able to avoid exactly such maneuvers.

C. Performance - Cruise

Fig. 10 Cruise Performance at 2000 m

Figure 10 exemplifies the values fo power setting (left plot), tail horizontal stabilizer deflection (middle plot) and
trim angle of attack and trim pitch attitude angle for different airspeeds at 2000 meters above sea level. It can be seen
in the left side plot, that at an airspeed of 50 m/s the aircraft requires minimal power. It can thus be said, that since
the design cruise speed is set to be around double this value, one might consider further optimizing the aerodynamic
design of the aircraft. I terms of stability, when simulated in any cruise condition, the trimmed model shows very stable
longitudinal as well as lateral-directional characteristics.

D. Performance - Transition
Figure 11 depicts power setting in dependency of airspeed at different wing tilt angles on the left side as well as a
transition corridor plot on the right. Again, the plots are representing the model results at an exemplary altitude of 2000
m. Both graphs in the figure stem from trim calculations of the model at different forward airspeeds as well as different
tilt wing angles. During the transition, trim conditions which show pitch attitude angles θ bigger than 15 are omitted,

10
 Fig. 11 Transition Performance at 2000 m

which can be seen as the left side limit (dotted line) of the transition corridor. The right side limit (continuous line) is
given by a maximal main engine power setting of 90%. This narrows the the graph greatly but is done so there is still
reserve power for mitigation of any disturbances (e.g. wind gusts). During the whole transition, the flap deflection angle
is allowed to float (20° up to 40° down). This effect helps greatly at high wing tilt angles. As the slipstream flow ’forces’
the wing to produce undesired lift and drag forces which act backwards (−x direction) and downwards (z direction). An
upward flap deflection can thus help ’destroy’ these forces and reduce wing pitching moment. A further mitigation of
these effects in hover is found in not tilting the wing fully upward to the maximum angle (δw < 90°). This leads to a
slightly forward leaning thrust vector, which results in a cancellation in backwards acting lift force. Finally it was found
that the transition corridor is less narrow at lower altitudes, since the engine thrust will increase.

IV. Discussion
As was found the wing profile(s) does not show unwanted behavior during the transition phase as the aircraft is
trimmable during the whole transition range. Furthermore, the propeller slipstream effect keeps a large part of the wing
at low angles of attack and attached to the profile since the transition maneuver requires high engine power settings. As
the mission design for the aircraft only requires short transition and hover phases, further wing design should be mainly
focused on optimizing for cruise conditions.
One could argue, that the profile can be designed for high speed since the airplane power on stall will delayed greatly.
Further consideration should be given to the profile thickness, since the forces and moments which are produced by
a wing mounted engine are in this case considerably high. Thus, the size of the wing support beam should not be
underestimated.
The transition analysis shows, that the aircraft is trimmable over the whole airspeed range and displays stable
behavior when simulating at the trimmed conditions.With increasing altitudes the transition corridor is narrowing which
results in less excess power. It should be possible to design a fairly straight forward control system for augmenting
the pilot in the transition phase. The challenge for development of such a system will however be reactions to fast
disturbances at high wing tilt settings. Since the propellers are big in radius for an aircraft of such size, the inertia of the
system cannot be underestimated.
A mayor design challenge next to the control system design will clearly be the development of efficient propeller
blades as the range of efficiency needs to be vastly greater than in conventional propellers. One can however if necessary,
sacrifice efficiency in the low speed region as long as hover conditions are still met.
The model was programmed in a way so that geometry as well as engine and aerodynamic data can be easily updated
when necessary. Further, the model does also lend itself to control system design and testing.

11
 References
[1] “Dufour Aerospace aEro 2 Project,” https://dufour.aero/vision/, 2018. Accessed: 2018-10-28.
[2] Bernard Etkin, L. D. R., Dynamics of Flight. Stability and Control, Third Edition, Wiley, 1996.
[3] MIT, “Multidimensional-Newton,” http://web.mit.edu/18.06/www/Spring17/Multidimensional-Newton.pdf,
2018. Accessed: 2018-10-28.
[4] Gray W. H., M. N., “Repersentative operating charts of propellers tested in the NACA 20-foot propeller research tunnel,” Tech.
rep., NACA, Langley Field, Va., 1943.
[5] Theodore, C. R., “Helicopter flight dynamics simulation with refined modeling,” Ph.D. thesis, University of Maryland, College
Park, Maryland, 2000.
[6] S., S. M., “Modeling Propeller Aerodynamics and Slipstream Effects on Small UAVs in Realtime,” AIAA Atmospheric Flight
Mechanics 2010 Conference, 2010.
[7] Yuksek B, e. a., Vuruskan A, “Transition Flight Modeling of a Fixed-Wing VTOL UAV,” J Intell Robot Syst (2016), 2010, p.
83–105.
[8] Murat, B., and Antoine, D., “Preliminary Design Estimation of the V/STOL Airplane Performance,” Conference: IMAV 2015:
International Micro Air Vehicles Conference and Flight Competition, Aachen, Vol. 7, No. 4, 2015, pp. 449–462.
[9] Phillips, W. F., Mechanics of Flight, 2nd edition, Wiley, 2009.
[10] “ AVL Documentation,” http://web.mit.edu/drela/Public/web/avl/, 2018. Accessed: 2018-10-28.
[11] “XFOIL Documentation,” http://web.mit.edu/drela/Public/web/xfoil/, 2018. Accessed: 2018-10-28.
[12] Carter E.C., e. a., “Lift coefficient increment at low speeds due to full-span split flaps,” Tech. rep., ESDU Aerodynamic Library,
1997.
[13] Garner H.C., e. a., “Maximum lift of wings with trailing-edge flaps at low speeds,” Tech. rep., ESDU Aerodynamic Library,
1995.
[14] Garner H.C., e. a., “Lift-curve slope of wing-body combinations,” Tech. rep., ESDU Aerodynamic Library, 1995.
[15] Egbert, T., Synthesis of Subsonic Airplane Design, Kluwer Academic Publishers, 1982.
[16] Jan, R., Airplane Flight Dynamics and Automatic Flight Controls, First Edition, DARcorporation, 1998.
[17] Eschenlohr P., F. A., “VTOL Flight Control Concept,” ZHAW Winterthur, 2018.

Acknowledgments
The author would like to thank the team of Dufour Aerospace for the opportunity to work on their project and for
their support. Furthermore the author would like to thank Mr. Martin Winet for his wise counsel in need.

12
 Appendix

Geometry Overview

Wing Horizontal Tail Vertcal Tail

Wingspan b 7.15 m 3.1 m 2.15 m
Wing mean aerodynamic chord c 1.6 m 0.79 0.8
Wing area S 10.45 m2 2.32 m2 1.65 m2
Wing twist iT 3◦
Airfoil (wing root to engine mount) NACA 2412
Airfoil (engine mount to wingtip) NACA 2409
Airfoil (tail) NACA 0009 NACA 0009
Taper ratio TR 1.07 2.33 2.13
Aspect ratio AR 4.91 4.14 2.81

Table 2 Geometry Table

Aerodynamic Coefficients

A/C Part Coefficient Method

Wing CLα , CDα , Cmα Wind tunnel data with Prantl 2D to 3D correction. Validated with AVL
[10] & XFOIL [11] simulations
Flaps & Ailerons CL δ f , CD δ f , Cmδ f ESDU method [12],[13]
Tail CLα , CDα , Cmα Wind tunnel data with Prantl 2D to 3D correction. Validated with AVL
[10] & XFOIL [11] simulations
Body CLα , Cmα Incorporated in wing coefficients, adapted with ESDU wing fuselage
interaction [14]
Body CDα Estimated with fuselage drag estimation method drag method [15],
extrapolated for high and low angles of attack with flat plate wind tunnel
data.
Full A/C CLq , CDq , Cmq Estimated by using examplery twin propeller aircraft data [16]
Full A/C CLαÛ , CDαÛ , CmαÛ Estimated by using examplery twin propeller aircraft data [16]
Full A/C CYαβ , Clαβ , Cnαβ AVL data Validated with values in [17]
Full A/C CYδr , Clδ r , Cnδr AVL data Validated with values in [17]
Full A/C CYδ a , Clδ a , Cnδ a AVL data Validated with values in [17]
Full A/C CL p Estimated by using examplery twin propeller aircraft data [16]

Table 3 Aerodynamic coefficients

13
Simulation Block Diagram

Fig. 12 Block diagram of simulation.

14
Simulation Flow Chart

Fig. 13 Simulation functions.

15