Lilium Analysis
Lilium Analysis
Lilium Analysis
Germany
d Whittle Laboratory, University of Cambridge, 1 JJ Thomson Avenue, Cambridge, UK
Acknowledgment
I would like to thank Prof. Dr. Bardenhagen, Prof. Dr. Strohmayer, Prof.
Miller, Dr. Grimshaw and Dr. Taylor for the diligent proof read, the technical
assessment of the proposed e-VTOL aircraft on an aircraft architecture level
and the valuable comments on the technical consistency of the content in the
following work. No public or private funds were involved in this collaboration.
Abstract
Urban and Regional Air Mobility (UAM & RAM) seek to extend passenger
transportation to the third dimension and a variety of electric vertical take-off
and landing (e-VTOL) aircraft concepts have been proposed to unlock this new
mode of transport. The UAM market requires an operating range of 5−40km at
a speed of 70 − 120km/h, and the RAM market, involving connection between
metropolitan areas, requires an operating range of 100 − 300km, and a speed of
beyond 200km/h.
This paper analyses the performance and range of a Ducted Vectored Thrust
March 6, 2021
Concept (DVTC) e-VTOL, which is specifically optimised to meet the low-noise
vertical take-off, long range and high payload requirements of the RAM market.
The specific aircraft analysed is the Lilium jet. The aircraft is 7-seater with a
cruise speed of 300km/h and maximum take-off mass of 3, 175kg. The paper
shows that assuming a battery energy density of 320kW h/kg, the maximum
range of the aircraft is 261km. At the same time, the ducted fan architecture
allows the noise emissions in take-off and landing hover to be below 60dB(A)
at a distance of 100m distance. This performance makes the aircraft ideal for
operation in the RAM market.
2
1. Architectural design space for e-VTOL aircraft
The design of e-VTOL aircraft for passenger transport is a new field, see
Thipphavong et al. (2018) and Whittle (2018). Five degrees-of-freedom (DOF)
drive the design: payload, speed, range, noise and simplicity. These degrees-of-
freedom are strongly interlinked and the improvement of one parameter often
disadvantages one or more of the others. Additional degrees-of-freedom, which
are not considered in this paper, include safety and costs. They are essential
for inner city operations and to enable scaling of mobility services. For sim-
plicity, however, this paper only considers the impact of the first five degrees-
of-freedom on the aerodynamic performance of the Lilium jet. In Figure 1,
Figure 1: Comparison of different e-VTOL concepts. All data is based on estimates from
company data in the public domain.
the four most common types of e-VTOL aircraft architectures are shown and
compared against their predicted application in UAM and RAM markets. The
choice of architecture determines the trade between the different DOFs and ul-
timately determines the cruise and hover efficiency and the noise emissions of
an aircraft. As e-VTOL aircraft are designed to operate close to and within
city centers, special emphasis should be given to safety and noise during the
conceptual design phase. The conceptual design of an e-VTOL aircraft results
from the combination of different DOFs and determines the technological com-
plexity while varying a single DOF. Discussion of these trade-offs are provided
by Bacchini and Cestino (2019); Polaczyk (2019); Clarke et al. (2019).
3
lack cruise efficiency, which limits their application to UAM markets only. A
more detailed review of multicopter architecture can be found in Pradeep and
Wei (2019).
The second category of e-VTOL is the lift & cruise architecture. These merge the
multicopter, for vertical take-off and landing, with a standard aircraft for cruise
flight. Doing so allows the aircraft to achieve both efficient vertical take-off and
landing, and efficient cruise. In order to maximize range for these concepts, the
propellers needed for VTOL are designed with fewer blades and shorter chords,
in order to reduce drag during cruise flight. The small size of the propellers for
VTOL operation create a significant challenge in terms of noise emissions, due
to increased blade tip speeds.
The third category of e-VTOL is the tilt rotor architecture which either in-
volves the wing and propellers, or the propellers alone, tilting. This allows the
propeller axis to rotate through 90 degrees as the aircraft transitions from hover
to forward flight. In general, this architecture allows a propeller to be designed
which is more optimal than would be possible with a lift & cruise architecture
of aircraft. However, this comes at the price of higher technical complexity and
larger overall size and weight, due to the tilt and variable pitch mechanisms.
The hover requirement requires the propellers to be large, with low tip speed
and low disc loading. This means that either the motors need to be large and
heavy, to produce the low speed torque, or a gearbox is required, eventually
interfering with the structure. Further challenges can occur in the design of the
flight dynamics during the flight transition process, although significant progress
has been reported in Chauhan and Martins (2020). A design space evaluation
of tilt rotors is given in Hendricks et al. (2019).
The three categories of e-VTOL discussed above all rely on a propulsion system
based on propellers. The fourth category of e-VTOL are known as ducted fan
architectures. A major advantage of ducted fans over unducted propellers is that
the duct acts to significantly mitigate noise, both blade passing and broadband.
This is achieved both by the presence of the duct and by acoustic liners mounted
within them. This is of particular importance when the payload of aircraft is
increased. If the design requirement is to operate from existing helipads then
the maximum footprint of the aircraft is limited, i.e. the largest dimension must
be less than 14m. When the payload is raised on a propeller aircraft the only
way of holding the level of noise constant is to hold the disc loading constant
and to therefore increase the size of the propellers. When the payload is raised
on a ducted fan the designer has an extra degree of freedom: they can let the
disc loading rise and use the duct and acoustic treatment to limit the increase
in noise. For a fixed size of footprint this results in a ducted fan aircraft having
a payload which is approximately 40% higher than a propeller aircraft.
Ducted fans have efficiency advantages which arise from the presence of the
duct. Blade tip losses are reduced and the presence of a stator row removes exit
4
swirl. Another advantage of ducted fans is that the duct can contain a blade-off
in the case of blade loss or bird strike making them safer than open propellers.
The main disadvantage of ducted fans is that because of the higher disc loading,
as discussed above, the hover power is higher than for propeller aircraft.
When employing ducted fans, two aircraft types are feasible. In the first type
the ducted fan is located away from the airframe so that aerodynamic interac-
tions are minimised; this type of aircraft is similar to conventional large civil
airliners. The second type integrates the ducted fan closely with the fuselage
and wings meaning there is close aerodynamic coupling. This architecture is
similar to that used in future aircraft concepts where boundary layer ingestion
is applied, see Hall and Crichton (2006).
The Lilium jet, Lilium (2019b), is an example of the second type of ducted
fan concept, where the ducted fan and wing are closely coupled. 36 ducted fans
are integrated into the aircraft’s canard and main wings. We refer to this as the
Ducted Vectored Thrust Concept (DVTC) with distributed electric propulsion
(DEP).
In the Lilium jet the ducted fans are compactly embedded close to the trail-
ing edge of the wing. This has three main advantages: first, the ducted fan
helps to keep the flow over the wing attached, enabling controlled flow over the
wings throughout the aircraft’s flight envelope, which provides improved flight
performance. Second, it allows the ducted fans to be used for thrust vectoring,
making aerodynamic control surfaces such as rudders, ailerons or tails obsolete
resulting in weight savings and complexity reduction. Third, the compact pack-
aging allows for improved cruise flight performance. By integrating the smaller
ducted fans in the form of DEP at the trailing edge of the wing, around a 30%
reduction in aircraft wetted duct surface area is achieved, thus reducing drag
and increasing range. Finally, it is worth noting that the higher rotational speed
of the fans, relative to propellers, means that they are less likely to excite low
frequency structural modes in the airframe providing improved safety and user
experience.
The ducted fan architecture of the Lilium jet has a disc loading which is up
to 10 times higher than some open propeller architectures making it less effi-
cient in hover. This results in a higher power consumption during the hover
phase when compared to open propellers. However, the paper will show that
the proposed DVTC aircraft architecture only spends a small fraction of the
flight mission in hover flight, and so the contribution of the increased hover
power has a relatively small effect on the overall energy of the flight mission.
Lilium has therefore concluded that the economic benefit of a heavier aircraft
with higher payload and lower noise emission outweighs the higher hover flight
power consumption.
5
VTOL aircraft which can address the market needs of UAM and RAM without
facing the same noise challenges as propeller concepts. This paper is structured
as follows: Section 2 details the aircraft requirements and specification. Sec-
tion 3 describes the approach to the analysis. Section 4, analyses the aircraft
performance. Section 5, provides an overview of the battery, powertrain and
electric motor technologies, and Section 6, analyses the sensitivity to key input
parameters.
To ensure a profitable and scalable business case, it is assumed that the aircraft
can transport 6 passengers with luggage with a total mass of mpax = 6 · 100kg.
The pilot’s mass is assumed to be mp = 100kg. This leads to a total payload
mass of mpl = 700kg. The ratio of payload mass to maximum take-off mass
mpl
is therefore µpl = M T OM = 0.22 and the ratio of empty mass to maximum
take-off mass is µs = M TmOM s
= 0.48. These values are typical of the range
reported in the literature, see Kim et al. (2018). Thus, the ratio of battery mass
to maximum take-off mass is µb = M TmOM b
= 0.30, again typical of the range
seen in the literature for e-VTOL aircraft, see Bacchini and Cestino (2019) and
Polaczyk (2019).
Es = µb · M T OM · eb = 305kW h. (1)
For all of our assessments we take the cell based energy density as an input
parameter, as Lilium assumes the structural modules of the battery cells will
be part of the empty aircraft structure.
The dimensions of the Lilium jet are shown in Figure 2. To make use of exist-
ing helipads and Final Approach and Take-Off areas (FATOs), the span of the
aircraft is sw = 13.9m, the length, width, and height of the cabin are lb , wc ,
and hc and the chord of the main wing is lc .
6
Figure 2: The Lilium jet architecture. Picture taken from Lilium (2019a).
The aim of the analysis is to calculate the aircraft’s range while retaining
enough simplicity to allow the reader to gain a physical understanding of the key
trades. The methodology introduced here, is applicable to any DVTC e-VTOL
concept, captures the most important physics and gives a realistic estimate of
7
range. To begin the assessment, a flight mission is defined, shown in Figure 3.
The mission can be divided into seven phases. The seven phases are made up
of the four modes of flight listed below.
• Hover: This is the period of the flight where the aircraft takes off and lands
vertically. During hover the power demand of the battery Ph is constant.
• Transition and re-transition: This is the period of the flight where the
aircraft is transitioning from hover into forward flight, and forward flight
into hover. During transition, the power demand of the battery Ptr drops
quadratically with time as the source of the aircraft lift changes from lift
generated by the propulsors to lift from the wings.
• Climb and descent: This is the period of flight where the aircraft climbs to
cruise altitude and accelerates to cruise speed. The climb power demand
Pcl during this phase is assumed to be constant.
• Cruise: This is the period of flight where the aircraft is at constant altitude
and speed. During cruise the power demand of the battery Pcr is constant.
A nominal cruise altitude of 3000m will be used in this study. This is the
altitude up to which a DVTC e-VTOL aircraft can operate without the
need of a pressurized passenger cabin.
The methodology for calculating the power demand of the battery during the
hover, transition, climb, and cruise periods of the flight are given in Section 3.1
-3.4 respectively. The methodology for calculating range is given in Section 3.5.
8
where vj,h is the jet velocity at nozzle exit during hover and ṁh is the total
mass flow through the fans during hover. The overall hover efficiency can be
written as
power to jet power to electronics power to motors
η0,h = = ·
power f rom battery power f rom battery power f rom electronics
| {z } | {z }
ηB,h ηP E,h
power to shaf t power to f low power to jet
· · · ,
power f rom motors power f rom shaf t power f rom f low
| {z } | {z } | {z }
ηM,h ηF,h ηD,h
(3)
where ηD,h is the duct efficiency, ηF,h the fan efficiency, ηM,h the motor effi-
ciency, ηP E,h the power electronics efficiency and ηB,h the battery efficiency.
The power use of the battery during hover can then be calculated by rearranging
Equations (2) and (3) and by rewriting the equation for the kinetic energy flux
of the jet in terms of the jet thrust (derivation given in Appendix A) to give
3/2
1 Th
2 √
0.5 · ṁh (vj,h ) 2 ρh ·n·Aj,h
Ph = = , (4)
η0,h ηF,h · ηD,h · ηP E,h · ηM,h · ηB,h
see Spurk and Aksel (2019) or Trancossi and Madonia (2012). In Equation (4),
Th is the thrust of the propulsion system during hover, ρh is the air density
during hover flight calculated for an international standard atmosphere, Aj,h is
the jet area of a single ducted fan in hover, and n is the number of ducted fans.
During hover the thrust requirement is fixed by the weight of the aircraft.
It should be noted that the relationship between the kinetic energy flux of
the jet and the thrust of the jet, shown in the numerator of (4) is only valid
for the case of a ducted fan. For the case of a propeller with no outer duct the
relationship between the kinetic energy flux and the thrust is
3/2
Th
Phopen = 0.5 · ṁh · vj,h
2
=√ , (5)
2ρh nAd
see Bittner (2005). In Equation (5) Ad is the disc area of the open propeller.
9
where Dcr is the drag of the aircraft, and vcr is the cruise velocity of the aircraft.
The efficiency in cruise can be written out as a product of efficiencies as
power to aircraf t power to electronics power to motors
η0,cr = = ·
power f rom battery power f rom battery power f rom electronics
| {z } | {z }
ηB,cr ηP E,cr
power to shaf t power to f low power to jet power to aircraf t
· · · · ,
power f rom motors power f rom shaf t power f rom f low power f rom jet
| {z } | {z } | {z } | {z }
ηM,cr ηF,cr ηD,cr ηP,cr
(7)
where ηP,cr is the propulsive efficiency and the other efficiencies are the same
as in Equation (3).
The power extracted from the battery during cruise can then be calculated
by rearranging Equations (6) and (7) to give
P
i Di,cr · vcr
Pcr = , (8)
ηP,cr · ηD,cr · ηF,cr · ηP E,cr · ηM,cr · ηB,cr
P
where i Di indicates the summation of all individual drag components of the
aircraft.
where Ptr,ef f = Ph /κ. Because in reality the power drops quadratically Equa-
tion (9) overestimates the power consumption and is therefore conservative. For
simplicity the same power consumption relationship will be used for both tran-
sition and re-transition. During transition the aircraft accelerates at 0.2g from
0km/h to vtr = 150km/h, resulting in each transition taking ttr = 21.2s.
10
crafts drag and a component of its weight in the direction of flight. This can be
written as
P
( i Di,climb + M T OM · g · sin(Φ)) · vcl
Pcl = . (10)
ηP,cl · ηD,cl · ηF,cl · ηP E,cl · ηM,cl · ηB,cl
During climb the altitude increases to 3000m and the speed increases from
150km/h to the cruise velocity vcr = 300km/h. This means that climb power
Pcl changes through the climb phase. In practice the component of the air-
craft’s weight term in the numerator of equation (10) is around 2 times larger
than the drag term, depending on the constant climb angle Φ. This results
in a higher climb power Pcl demand compared to cruise flight of approximately
Pcl ≈ 2.25·Pcr . For simplicity the climb power is assumed to be constant and cal-
culated at a mean altitude of 1500m with a mean climb velocity vcl = 275km/h.
During the descent the component of the aircraft weight, in the direction of
flight, changes direction. This reduces the power requirement. In this analysis
we will, for simplicity, set the engine power to 20% of the cruise power. This
is a deliberately conservative choice and in practice the engine power will often
be lower, set in practice by the particular architecture of the DVTC.
The maximum energy use during a mission can be written in terms of power
extracted from the battery during each phase of the flight.
Es (1 − SOCmin ) = t h · Ph + 2·t ·P
| {z } | tr{z tr}
Hover consumption T ransition consumption
(13)
+ tcl · (Pcl + Pdes ) + Pcr tcr ,
| {z } | {z }
Climb consumption Cruise consumption
11
where th is the total time in hover, take-off and landing, ttr is the total time in
transition, take-off and landing, tcl is the time in climb, and tcr is the time in
cruise. This can be rearranged to give an expression for the time in cruise, and
multiplying by the cruise velocity vcr gives a relationship for the cruise range of
the aircraft. The distance travelled during the climb and descent phases of the
flight are added to this to give the overall range:
vcr
R= · (Es (1 − SOCmin ) − th · Ph − 2 · ttr · Ptr
Pcr (14)
−tcl · (Pcl + Pdes )) + 2 · tcl vcl .
Equation (14) will be used for the range assessment of the aircraft.
The analysis starts by calculating the aircraft’s cruise drag, the efficiencies of
the propulsion system and the power use in each phase of flight. The values for
the aircraft’s drag and efficiencies are summarised in Table 2. These values are
then be used to calculate the aircraft’s range. Finally a noise assessment of the
propulsion system will be undertaken.
Table 2: Calculated efficiencies from Section 4 being used for the performance assessment.
The total drag of the aircraft in cruise Dcr is the sum of the drags of each
part of the aircraft, see Dole et al. (2016) and Kuchemann (2012):
Dcr = Dc + Dw + Df + Di , (15)
where Dc is the cabin drag, Dw is the wing drag, Df is the flap drag, and Di is
the induced drag.
12
4.1.1. Cabin drag
The ideal drag of the aircraft cabin, see Nicolosi et al. (2015) among others,
can be calculated as
2
Dc,ideal = CD,c · 0.5 · ρcr · Sc,∅ · vcr = 248N, (16)
where the cabin drag coefficient is CD,c = 0.039 and Sc,∅ is the frontal area of
the cabin. The frontal area of the cabin is
2
π wc + hc
Sc,∅ = · = 2.01m2 , (17)
4 2
where CD,w is the wing drag coefficient and Sw is the wing area. For a wing
thickness to chord ratio of t/lc = 0.12 and a fully turbulent boundary layer
at a Reynolds number in the order of O(107 ), we find a reasonable value of
CD,w = 0.014, see Schlichting and Truckenbrodt (2001a,b).
Sw = lc · (sw − wc ) − nw · ls · ds , (20)
where in the first term on the right hand side represents the area of the whole
wing, and the second term represents the area of the wing that has been removed
and replaced with the ducted fans, see green wing surfaces in Figure 4. In
Equation (20) lc is the wing chord, sw is the wing span, wc is the cabin width,
nw is number of ducted fans on the wing. ls is the length of the fan stage and
ds is the diameter of the fan shroud. Using the wing and fan dimensions from
Section 2, Equation (20) gives a wing area of
Sw = 8.464m2 , (21)
Dw = 374N, (22)
13
Figure 4: Wing and flap configuration. The scales, sizes, proportions, details and measure-
ments are not implemented in the real aircraft and shall serve only for illustrative purposes.
Green areas mark the surfaces being used to calculate wing drag Dw and flap drag Df . The
fan stage efficiency ηF is effective over the length lstage . The duct efficiency ηD is effective
over the remaining red length ls − lstage .
where CD,f the drag coefficient of the flaps, and Af lap,o is the projected surface
area of the flaps
Df = 399N. (25)
14
4.1.4. Induced drag
The induced drag Di can be calculated, see Schlichting and Truckenbrodt
(2001a,b), as
2 · (M T OM · g)2
Di = 2 · s2 · e
, (26)
ρcr · π · vcr w
where for a canard configuration the Oswald factor is e = 0.83, see Kroo (2001,
1982). Substitution of the Oswald factor into Equation (26) gives an induced
drag of
Di = 610N. (27)
This allows a lift to drag ratio to be calculated for the aircraft. The lift is set
by the weight of the aircraft L = M T OM · g giving
M T OM · g
Ω = L/D = = 18.26. (29)
D
The simple analysis employed in this paper does not account for the close aero-
dynamic coupling between the wing and ducted fan. To quantify the interaction
between the wings and the ducted fan intake, for example, would require higher
fidelity tools such as CFD. The boundary layer on the top of the wing is directly
ingested into the ducted fan. This means that the low velocity air in the bound-
ary layer does not have a chance to mix out downstream in the wing’s wake,
and instead is re-energized in the ducted fan. This effect will act to raise the
overall cruise efficiency. At the same time the non-homogeneous flow entering
the fan will act to lower fan efficiency. If correctly designed it is likely that
this boundary layer ingestion mechanism can be exploited to reduce the power
consumption of the aircraft in cruise. However, this is beyond the scope of this
paper and in this analysis its beneficial effects will be ignored.
4.2. Efficiencies
Each of the efficiencies, described in Section 3, will be calculated in turn.
15
the jet is defined by σ, the ratio of the diameter of the exit nozzle Aj and the
area of the duct in which the fan is located Ad
Aj = σAd . (30)
For hover we set σh = 1.3 and for cruise σcr = 0.9 resulting in Aj,h = 0.0742m2
and Aj,cr = 0.0513m2 . At hover the blade tip Mach number is M ab,tip,h = 0.45
and the chord based Reynolds number of the rotor is Rer,h ≈ 610·103 . At cruise
the blade tip Mach number is M ab,tip,h = 0.26 and the chord based Reynolds
number of the rotor is Rer,cr ≈ 280 · 103 . See Appendix B for the detailed
calculation procedure.
Figure 5: Illustrative representation of the variable nozzle concept. Aj,h and Aj,cr represent
the effective jet area in hover and cruise flight respectively.
The effect of changing the area of the variable area nozzle on the fan oper-
ating point is shown in Figure 6, where the fan stage loading coefficient ψ and
flow coefficient φ are defined as
∆h0 vin
ψ= ; φ= . (31)
u2b,mean ub,mean
∆h0 is the stagnation enthalpy across the fan, ub,mean is the blade velocity
at mass averaged radius rmean (see Appendix B), and vin is the axial veloc-
ity at fan inlet. The shape of the characteristic and the particular values of
the flow and stage loading coefficients are calculated in Appendix B. Figure
6 shows an illustrative example of how stage loading and fan efficiency vary
with flow coefficient. The black dot represents the hover operating point and
the two white points represent the cruise operating point with the variable exit
nozzle set at two different areas. For the concept described in this paper, the
fan is designed to have peak performance during hover flight with an estimated
isentropic efficiency ηF,h = 0.88 at σh = 1.3. This leads to a flow and stage
loading coefficient of φh = 1.09 and ψh = 0.40 for hover flight. At cruise flight
conditions, the same fan, with the same variable exit nozzle setting, σ = 1.3
would have flow and stage loading coefficients of φ = 1.47 and ψ = 0.10. At this
high flow coefficient, a fan could drop in isentropic efficiency by ∆ηaero ≈ −20%.
16
hF
88%
83%
0.4
0.40
0.27
y=1+f tan b2
At the cruise operating point, closing the exit nozzle area to σ = 0.67 would
move the operating fan point so that it was identical to that of the hover op-
erating point. This would be ideal in terms of fan performance providing an
efficiency of ηF,cr = 0.88. However, during cruise flight operations, the Lilium
et is controlled via thrust vectoring and RPM adjustments of single or multiple
fans, so a safe stall margin is required for the nominal cruise operating point to
allow for off-design operation during manoeuvres. Closing the variable area noz-
zle also increases the jet velocity which reduces the overall system’s propulsive
efficiency. Taking these effects into account, the variable area nozzle area ratio
for cruise can be optimised at a flow coefficient which is 0.3 higher by setting
σcr = 0.9, which leads to φreal = 1.31, ψcr,real = 0.27 and a fan efficiency of
ηF,cr = 0.83.
The design flow coefficient of the fan φh = 1.09 is higher than those typically
reported in the literature where the optimal flow coefficient is approximately
φ = 0.6. The choice of a high design flow coefficient is deliberate. It allows both
the fan diameter to be kept low, and the blade tip Mach number to be below
0.45. This is especially important for reducing fan noise. High fidelity analysis,
such as CFD and other advanced in-house analysis tools, give Lilium confidence
17
that ducted fans with high flow coefficients φcr = 1.31 can be designed during
cruise flight at efficiencies of ηF,cr = 0.83, while achieving peak performance in
hover flight with a fan efficiency of ηF,h = 0.88.
where vj,cr is the jet velocity during cruise. Equation (32) is commonly known
as the Froude equation.
The jet velocity in cruise vj,cr is calculated by applying the steady flow mo-
mentum equation to a control volume surrounding the propulsor, see Farokhi
(2014); Eckert and Schnell (2013), and rearranging gives
q
2
Aj,cr,tot · ρcr · vcr + (Aj,cr,tot · ρcr · vcr ) + 4 (Aj,cr,tot · ρcr · Dcr )
vj,cr =
2Aj,cr,tot · ρcr
= 93.96m/s,
(33)
18
The duct efficiency is defined as
power to jet
ηD = =
power f rom f low
(35)
T Ṡd,irrev
1− 2
,
0.5 · vcr ṁcr · vj,cr − vcr2 + T Ṡ
d,irrev
where T Ṡd,irrev are the boundary layer losses in the duct due to viscous dis-
sipation which can be calculated for the hover and cruise fan operating points
as
3
T Ṡd,irrev = n · As · CD · ρ · vd3 = n · (π(ls − lstage ) · (dh + ds )) · CD · ρ (vj · σ) . (36)
The jet velocity vj,cr was calculated in the previous section for the cruise case
as vj,cr = 94.11m/s and substituting this result into Equation (36) gives us
For the hover case the procedure is repeated using the jet velocity vj,h for hover.
This is derived from Equation (33) noting that the flight speed is zero and the
thrust is the same as the weight of the aircraft. This yields a nozzle exit velocity
of:
s
M T OM · g
vj,h = = 97.59m/s, (39)
ρh Aj,h
Substituting the results from Equation (39) into Equation (36), the duct loss at
hover is calculated as:
19
giving a duct hover efficiency of
20
where the overall aircraft efficiency during hover is η0,h = ηF,h · ηD,h · ηP E,h ·
ηM,h · ηB,h ≈ 59% and Pboard = 8kW through the whole flight. Substituting
into equation (46) gives
Ph = 2,570kW. (47)
It is an interesting academic question to calculate maximum hover time which
the power in Equation (46) and the total battery energy in Equation (1) would
allow.
Es · (1 − SOCmin )
tmax
h = = 384s. (48)
Ph
It should, however, be noted that this hover time would in practice be impossible
because it would lead to overheating of the battery packs.
(Ph + Ptr,ef f )
Ptr,avg = + Pboard . (51)
2
where Ptr,ef f = P h/10. Substituting into Equation (51) gives
During climb, we set the climb angle to Φ = 5◦ . This means that the thrust
of the aircraft has to overcome both the aircraft’s drag and a component of its
weight. This results in the power in climb being around twice the power in
cruise. As discussed earlier, during climb the altitude increases to 3000m and
21
the speed raises to vcr = 300km/h. This means that the power changes through
the climb phase. However, in practice this change is not large and for simplicity
a mean altitude of 1500m with a mean climb velocity vcl = 275km/h is set.
Taking the drag value from Table 2, the climb drag is
The jet velocity in climb is higher than in cruise because the thrust of the
propulsion system must overcome both the drag and a component of the weight.
This means that the propulsive efficiency of the aircraft and duct efficiency must
be recalculated. The final values of the drag and efficiencies used in the climb
calculation are shown in Table 2. Substituting in (53) gives
The comparison between Equation (55) and Equation (50) reveals that the
climbing power is around a fifth of the hover power. This demonstrates the
benefit of the DVTC architecture on RAM missions.
As discussed earlier, during the descent, the component of the aircraft weight
switches direction. To ensure a powered descent the engines are operated during
the descent phase of the flight at 20% of the cruise power.
22
tuting into Equation (14) gives a range
vcr
R= (·Es (1 − SOCmin ) − th · Ph − 2 · ttr · Ptr
Pcr (57)
−tcl · (Pcl + Pdes )) + 2 · tcl · vcl = 261km.
The focus of this paper has been on providing a simplified method of calculating
the realistic maximum range of a DVTC e-VTOL aircraft. The analysis shows
that for the DVTC e-VTOL aircraft considered this is R = 261km. The result
demonstrates that the aircraft is suitable for RAM markets.
The range obtained assumes that all of the available battery energy is con-
sumed. For an operational aircraft, pre-planned operational manoeuvres e.g.,
diverts or emergency procedures, will also need to be accounted for. Discus-
sion with authorities on these pre-planned operational manoeuvres are ongoing.
However, the operational mission planning framework for e-VTOL aircraft is
not expected to be based on a percentage of total onboard energy since the di-
versity of e-VTOL concepts would lead to inconsistent reserve planning. Instead
missions are likely to include absolute reserve planning, meaning that a range
above 200km is realistic for aircraft similar to the one presented in this paper.
In order to approximate Leng for any axial ducted fan, the NASA source di-
agnostic test (STD) described among others by the Regenscheit analogy in
VDI (1982) is used. Using the engine geometry described in Section 2, we
calculate Leng = 94dB(A) at Reng = 1m. Substituting into Equation (59)
the aircraft overall sound pressure level Lac at a distance of Rac = 100m
gives Lac = 65dB(A). The Regenscheit analogy has a standard deviation of
±2.7dB(A). This low noise emission is the result of the fan being located in
23
a duct and to the tip speed of the blade being limited to ub,tip = 154.38m/s,
which is equivalent to a Mach number M a = 0.45, see also Appendix B.
In addition, acoustic liners are included in the duct. These are designed to
reduce tonal noise through Helmholtz and porous liners. Sutliff (2019) and
Aurégan (2018) show that in small ducted fans these reduce noise by ≈ 1dB(A)
for every centimeter of liner length. This has the effect of significantly reducing
tone noise from the noise spectrum, and thus, reduces the overall engine noise.
With the acoustic liners in place we calculate the overall sound power will be
reduced to Leng = 89dB(A) at Reng = 1m. Substituting into Equation (59) the
aircraft overall sound pressure level Lac at a distance of Rac = 100m drops to
Lac = 60.6dB(A).
Finally, the ducted fan configuration allows noise attenuation techniques through
blade design, e.g. by the used of swept rotors and stators, Envia and Nallasamy
(1999). In this assessment it is assumed that this can be used to further reduce
the aircraft overall sound pressure level below
at Rac = 100m for the real aircraft. This is equivalent to the sound level created
by normal conversation. Finally, for reference, in cruise, due to lower thrust,
the aircraft’s overall sound pressure level Lac at a distance of Rac = 100m is
Lac,cr = 54.4dB(A).
Based on first order assumptions and applying standard industry tools during
the architecture design phase, Lilium believe that the inner city operation of a
7-seat, 3,175kg e-VTOL aircraft can be below Ltarget < 60dB(A) at a distance
of 100m.
5. Technology assessment
24
in excess of 350W h/kg by using different anode materials.
The specific power required during short periods of hover, reported in Section
4.3, is Ph /mb = 2.7kW/kg. The cell family introduced by Shirk and Chinh
(2020), demonstrates that a constant specific power beyond 2kW/kg can be
achieved with eb ≈ 300W h/kg. Beyond that, they demonstrate the ability to
deliver constant high power and high energy densities, while allowing the batter-
ies to be discharged to SOCmin = 10%. A value of SOCmin = 10% is therefore
used in this paper.
These values are available for devices commercially available from aerospace
suppliers. These assumptions are supported by Seitz et al. (2012), where overall
electric system efficiencies of greater than 90% are reported.
In this section, the effect on range of the batteries energy density, hover time
and the ratio of battery mass to payload are studied. These parameters are cho-
sen as they have the largest impact on range and are most likely to affect how
the aircraft is operated in the near future. Aerodynamic, powertrain, motor
and battery efficiencies are fixed for this analysis as technological development
in these areas is expected to be incremental. Five variations from the concept
described in the previous sections (DVTC-1) are considered and the results are
shown in Table 4.
DVTC-3: The total operating hover time is increased from th = 60s to th = 90s.
With all other assumptions fixed, including eB = 320W h/kg, the range drops
to R = 232km. This highlights an important observation that DVTC aircraft
25
Input Range
Aircraft variation eB [W h/kg] th [s] µb /µpl [−] R [km] (Section 4.4)
DVTC-1 320 60 1.36 261
DVTC-2 250 60 1.36 181
DVTC-3 320 90 1.36 232
DVTC-4 400 60 1.36 353
DVTC-5a 320 60 2.30 337
DVTC-5b 400 60 2.30 448
Table 4: Parameter variation for an investigated e-VTOL aircraft architecture. All range
values and hover times are calculated for a minimum acceptable state-of-charge of SOCmin =
10%.
with relatively highly loaded ducted fans should be designed for as little hover
time as possible to exploit the cruise efficiency.
DVTC-5a: This version examines the trade-off between battery mass, payload
and range by configuring DVTC-5a as a 5-seat aircraft instead of a 7-seater.
The payload, mpl is reduced by 200kg so that battery mass, mb is increased by
200kg and, thus, µb /µpl increases from 1.36 to 2.30. The battery energy density
for this version is set to eB = 320W h/kg. This change in configuration is not
expected to effect the other assumptions presented in this paper and the range
is calculated as R = 337km.
7. Conclusion
26
physical understanding of the key trades.
The analysis shows that if an available battery technology with energy den-
sity of 320W h/kg is assumed then the 7-seater DVTC aircraft with a maximum
take-off mass of 3,175kg can achieve a range of 261km. It is shown that the de-
velopment of solid-state batteries with an energy density in excess of 400W h/kg
would increase the range of the aircraft beyond 330km.
The range was assessed without operational penalties being considered. How-
ever, it is believed that inclusion of all necessary reserves will still allow an
aircraft with a range greater than 200km allowing significant business opportu-
nities in the RAM markets.
The analysis shows that noise levels of less than 60dB(A) at an observe dis-
tance of 100m, are feasible during hover. This is equivalent to the sound level
created by normal conversation.
References
Y. Aurégan. Ultra-thin low frequency perfect sound absorber with high ratio
of active area. Applied Physics Letters, 113:201904, 11 2018. doi: 10.1063/
1.5063504.
A. Bacchini and E. Cestino. Electric vtol configurations comparison. Aerospace,
6, 2019. doi: https://doi.org/10.3390/aerospace6030026.
W. Bittner. Flugmechanik der Hubschrauber. Springer-Verlag, 2nd edition, 2005.
doi: https://doi.org/10.1007/3-540-27541-X.
M. P. Boyce. Preface to the First Edition. Butterworth-Heinemann,
Oxford, fourth edition edition, 2012. ISBN 978-0-12-383842-1. doi:
https://doi.org/10.1016/B978-0-12-383842-1.00043-3. URL https:
//www.sciencedirect.com/science/article/pii/B9780123838421000433.
S. S. Chauhan and J. R. R. A. Martins. Tilt-wing evtol takeoff trajectory op-
timization. Journal of Aircraft, 57(1):93–112, 2020. doi: 10.2514/1.C035476.
URL https://doi.org/10.2514/1.C035476.
27
M. Clarke, J. Smart, E. M. Botero, W. Maier, and J. J. Alonso. Strategies
for Posing a Well-Defined Problem for Urban Air Mobility Vehicles. 2019.
doi: 10.2514/6.2019-0818. URL https://arc.aiaa.org/doi/abs/10.2514/
6.2019-0818.
J. D. Denton. Loss Mechanisms in Turbomachines, volume Volume 2: Combus-
tion and Fuels; Oil and Gas Applications; Cycle Innovations; Heat Transfer;
Electric Power; Industrial and Cogeneration; Ceramics; Structures and Dy-
namics; Controls, Diagnostics and Instrumentation; IGTI Scholar Award of
Turbo Expo: Power for Land, Sea, and Air. 05 1993. doi: 10.1115/93-GT-435.
URL https://doi.org/10.1115/93-GT-435. V002T14A001.
28
H. D. Kim, A. T. Perry, and P. J. Ansell. A Review of Distributed Electric
Propulsion Concepts for Air Vehicle Technology. 2018. doi: 10.2514/6.2018-
4998. URL https://arc.aiaa.org/doi/abs/10.2514/6.2018-4998.
S. Kim, D. Kim, C. Son, K. Kim, M. Kim, and S. Min. A full en-
gine cycle analysis of a turbofan engine for optimum scheduling of vari-
able guide vanes. Aerospace Science and Technology, 47:21 – 30, 2015.
ISSN 1270-9638. doi: https://doi.org/10.1016/j.ast.2015.09.007. URL http:
//www.sciencedirect.com/science/article/pii/S1270963815002655.
I. Kroo. Drag due to lift: Concepts for prediction and reduction. Annual Review
of Fluid Mechanics, 33(1):587–617, 2001. doi: 10.1146/annurev.fluid.33.1.587.
I. M. Kroo. Minimum induced drag of canard configurations. Journal of Aircraft,
19(9):792–794, 1982. doi: 10.2514/3.61557.
D. Kuchemann. The Aerodynamic Design of Aircraft. American Institute of
Aeronautics and Astronautics (AIAA) Education Series, 2012.
Lilium. The Lilium jet technology. https://lilium.com/the-jet, 2019a.
Lilium. Visual representation of the planform area of the Lilium jet. https:
//photos.app.goo.gl/ozik5SJcpsQE8fdQ7, 2019b.
D. Lin, Y. Liu, and Y. Cui. Reviving the lithium metal anode for high-
energy batteries. Nature Nanotechnology, 12(3):194–206, 2017. doi: 10.1038/
nnano.2017.16. URL https://doi.org/10.1038/nnano.2017.16.
F. Nicolosi, P. Vecchia, D. Ciliberti, and V. Cusati. Fuselage aerodynamic drag
prediction method by cfd. 09 2015.
M. T. Pascu. Modern Layout and Design Strategy for Axial Fans. doctoralthesis,
Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), 2009.
C. Pitis. Energy efficient single stage axial fan (enef). pages 280 – 285, 11 2007.
ISBN 978-1-4244-1444-4. doi: 10.1109/EPC.2007.4520343.
N. Polaczyk. A review of current technology and research in ur-
ban on-demand air mobility applications. 2019. URL https:
//repository.tudelft.nl/islandora/object/uuid:10d1c68a-dfb9-
4419-a912-26e6c1742853?collection=research.
P. Pradeep and P. Wei. Energy-efficient arrival with rta constraint for multirotor
evtol in urban air mobility. Journal of Aerospace Information Systems, 16(7):
263–277, 2019. doi: 10.2514/1.I010710. URL https://doi.org/10.2514/
1.I010710.
E. J. Rickley, G. G. Fleming, and C. J. Roof. Simplified procedure for
computing the absorption of sound by the atmosphere. Noise Control
Engineering Journal, 55(6):482–494, 2007. doi: https://doi.org/10.3397/
1.2820987. URL https://www.ingentaconnect.com/content/ince/ncej/
2007/00000055/00000006/art00001.
29
F. Schieber. Procedure for Estimating the Combined Noise Level of Multi-
ple Acoustic Sources. 2021. URL http://apps.usd.edu/coglab/schieber/
psyc707/pdf/AddedNoiseLevels.pdf.
H. Schlichting and E. A. Truckenbrodt. Aerodynamik des Flugzeuges: Erster
Band: Grundlagen aus der Strömungstechnik Aerodynamik des Tragflügels
(Teil I). Springer-Verlag, 3rd edition, 2001a.
H. Schlichting and E. A. Truckenbrodt. Aerodynamik des Flugzeuges: Zweiter
Band: Aerodynamik des Tragflügels (Teil II), des Rumpfes, der Flügel-Rumpf-
Anordnung und der Leitwerke. Springer-Verlag, 3rd edition, 2001b.
30
R. Whittle. Air mobility bonanza beckons electric vtol
developers. https://pdfs.semanticscholar.org/a132/
ba1a11012da261627597f1d074c3f3ea67e9.pdf, 2018. Accessed: 2010-
09-30.
X. Ye, P. Li, and C. Li. Numerical investigation of grooved blade tips effect
on the performance of an axial flow fan. Zhongguo Dianji Gongcheng Xue-
bao/Proceedings of the Chinese Society of Electrical Engineering, 35:652–659,
2015. doi: 10.13334/j.0258-8013.pcsee.2015.03.019.
E. Zenlabs. Idaho national laboratory validates 1000 charge-discharge
cycles from zenlabs’ silicon anode-based lithium-ion cells. 2021.
URL https://www.prnewswire.com/news-releases/idaho-national-
laboratory-validates-1000-charge-discharge-cycles-from-zenlabs-
silicon-anode-based-lithium-ion-cells-301226389.html.
Applying the steady flow momentum equation to the ducted fan, the total thrust
can be shown to be equal to the momentum flux of the exit jet
2
Th = ṁh · vj,h = ρh · Aj,h · vj,h . (A.2)
Substitution Equation (A.2) into Equation (A.1) gives an equation of the power
of the jet in hover
s
3/2
3 T h Th 1 T
Pj = 0.5 · ρh · Aj,h · vj,h = = p h . (A.3)
2 ρh Aj,h 2 ρh Aj,h
Th,tot = M T OM · g. (A.4)
3 1 (M T OM · g)3/2
Pj = 0.5 cot ρh · Aj,h · vj,h = p . (A.5)
2 ρh · n · Aj,h
The jet area of a single engine Aj,h = Ad · σh and total jet area of the n
ducted fans Aj,h,tot during hover flight is calculated from
d2s − d2h
Aj,h = π · σh = 0.0742m2 Aj,h,tot = Ad,h · n = 2.67m2 . (A.6)
4
31
Equation (A.3) differs from the power consumption for open rotor systems
open
Ph,ideal
3/2
T
Phopen = p h , (A.7)
2ρh Ad,tot
see Bittner (2005).
At cruise and climb operating conditions the fan speed is set by the thrust,
flight speed and the requirement that the cruise operating point lies on the fan
characteristic shown in Figure 6. The relationship between stage loading and
flow coefficient is approximated by
ψ = 1 + φ tan(β2 ), (B.2)
where β2 is the rotor exit flow angle in the relative frame. β2 is fixed once φ
and ψ in hover are specified by selecting the blade speed and is always negative.
At any point on the characteristic, Equation (B.2) can be rewritten as
where ηF,α is the fan efficiency at operating condition α (hover ?h or cruise ?cr ).
Equation B.3 can be rearranged to a quadratic equation in ub,mean and this can
be solved to give
q
−uin,α tan(β2 ) (uin,α · tan(β2 ))2 − η2F (vα2 − (uout 2
j,α ) )
ub,mean,α = . (B.4)
2
The values of ub,mean for cruise and climb are shown in the final column of Table
B.5 and this is used to calculate the non-dimensional parameters in Table B.6.
The Reynolds number given in Table B.6 is defined as
ub,mean · cr
Rer = , (B.5)
ν
32
where the rotor chord length is cr = 0.052m. The demonstrated procedure can
be repeated for different radial cuts, but doesn’t change the architectural deci-
sion making process.
φ ψ Rer Mb,tip
Hover 1.085 0.396 610,415 0.45
Cruise 1.306 0.273 280,434 0.26
Climb 1.138 0.367 365,760 0.31
For low pressure ratio fans with similar aerodynamic parameters, isentropic
engine
stage efficiencies exceeding ηaero > 0.9 have been reported and validated in
Boyce (2012); Gourdain (2015); Ye et al. (2015); Kim et al. (2015); Pascu (2009);
Pitis (2007) and, thus, we set ηF,h = 0.88.
Abbreviation Description
DEP Distributed Electric Propulsion
DVTC Ducted Vectored Thrust Concept
e-VTOL Electric vertical take-off and landing
FATO Final approach and take-off areas
MTOM Maximum take-off mass
RAM Regional air mobility
RPM Revolutions per minute
SOC State of charge
UAM Urban air mobility
Symbol Description
Ad Disc area of ducted fans
Aj,α Effective jet area of a ducted fan
Aj,α,tot Effective jet area of all ducted fans
Af lap,o Surface area without shroud surface
C Idealized compressor characteristic
CD Drag coefficient
CD,c Cabin drag coefficient in cruise flight
CD,f Flap drag coefficient in cruise flight
CD,w Wing drag coefficient in cruise flight
cr Rotor chord length
cu,2 Post rotor circumferential velocity
33
Dcr Total drag
in cruise flight
Dc Total cabin drag
in cruise flight
Dc,ideal Ideal cabin drag
in cruise flight
Df Total flap drag
in cruise flight
Di Induced drag
in cruise flight
Dw Total wing drag
in cruise flight
dh Hub diameter
ds Shroud diameter
Eacc Accessible energy in the battery
Es Total stored energy in the battery
e Oswald factor
eb Energy density of batteries
g Gravitational acceleration g = 9.81m/s2
∆h0 Enthalpie increase over the stage
lb Length of cabin
hc Height of cabin
Lac Aircraft overall sound pressure level
Lac,cr Aircraft overall sound pressure level during cruise flight
Latm Atmospheric noise correction
Leng Sound pressure level correlated at a reference distance
Ltarget Target overall sound pressure level in 100m distance
ls Shroud length
lstage Fan stage length
lc Chord length
Ma Mach number
M ab,tip,α Mach number at the blade tip for the respective flight state
ṁα Mass flow through the ducts in the respective flight state
mb Mass of batteries
mpax Passenger mass
mp Pilot mass
mpl Total payload
ms Empty aircraft mass
n Total amount of engines
nc Engines on canard wing
nw Engines on main wing
Pboard Total power required for onboard systems
Pcl Total power required during climb flight
Pcr Total power required during cruise flight
Pdes Total power required during descent flight
Ph Ideal power required during hover flight
Ptr,ef f Approximate power demand after transition phase
Ptr,avg Average power demand during transition phase
Rac Arbitrary distance from aircraft
Rmin Minimal expected range for RAM
Rr Average Reynolds number based on rotor
Rs Average Reynolds number based on stator
Rr,α Reynolds number on the rotor stage for the respective flight state
Sc,∅ Frontal surface area of the fuselage
U RPM of the rotor stage
u1 Circumferential blade velocity
u1,tip Circumferential blade velocity at blade tip
vin,α Average inlet velocity before the fan stage
ub,mean Average blade velocity
rmean Mass averaged radius of the ducted fan
Sw Total main wing reference area
sw Main wing span
sc Canard span
34
Tα Total thrust required during the respective flight state
T Ṡd,irrev,α Total losses due to viscous dissipation
tcl Operational climb time
tcr Operational cruise time
tmax
h Theoretical maximum hover time
th Total hover time
th,l Operational hover time during landing
th,to Operational hover time during take-off
ttr Operational transition time
V̇ Volume flow in ducted fan
vcl Average climb velocity
vcr Cruise velocity
vd,α Flow velocity in fan duct
vj,α Nozzle exit velocity
vcl Total climb path velocity
vv,cl Vertical climb velocity
wc Width of cabin
Greek symbols Description
α α = cr for cruise, α = cl for climb and α = h for hover respectively
η0,α Total efficiency of the aircraft during the respective flight state
ηB,α Battery efficiency of the aircraft during the respective flight state
ηP E,α Power electronics efficiency of the aircraft during the respective flight state
ηM,α Motor efficiency of the aircraft during the respective flight state
ηF,α Fan efficiency of the aircraft during the respective flight state
ηD,α Duct efficiency of the aircraft during the respective flight state
ηP,α Propulsive efficiency of the aircraft during the respective flight state
κ Power ratio between hover and the end of transition phase
Λ Wing aspect ratio
λ Disc loading
µb Mass fraction of batteries
µpl Mass fraction of payload
µs Empty mass fraction of aircraft
ν Air viscosity under atmospheric standard conditions at sea level
Φ Climb angle
φ Flow coefficient
φα Flow coefficient during the respective flight state
φα,real Effective flow coefficient during the respective flight state
ψα Stage loading coefficient during the respective flight state
ψα,real Effective stage loading coefficient during the respective flight state
Ω Glide ratio
ρα Density during the respective flight state
σα Nozzle exit ratio for hover flight
35