Technology behind the Lilium Jet
At the beginning of December 2020, I joined Lilium as Chief Technology Officer. Prior to joining, I was Chief Engineer and Managing Director at Rolls-Royce Germany. In my career I have been responsible for delivering a number of world leading turbofan jet engine programmes including the BR725 for the Gulfstream G650 and the Trent XWB for the Airbus A350. Like many other aerospace engineers who joined Lilium before me, I was curious as to how Lilium’s technology would really perform in practice. At first glance, it certainly appeared tricky for Lilium’s ambitious architecture to achieve the stated mission profile. For example, it appeared to me that the associated fan size and power consumption implications of electric ducted jet engines might make such an architecture impractical, or at least economically non-viable in an eVTOL aircraft given the thrust required for take-off.
I reviewed this technology and architecture in detail as part of my due diligence. I found it to be technically sound and genuinely impressive and now as part of the team, I have seen much of the innovative technology, analysis and, importantly, test data that underwrites the concept. To share some of that insight, this blog lays out the core elements of the Lilium architecture and the principal technical arguments as to why this architecture is powerful when applied to the aircraft mission profile. Where possible, I have also addressed common misconceptions regarding the aircraft architecture.
It is fair to say that the team at Lilium have been keeping their heads down, quietly getting on with things. Going forward, our sincere hope is to be more transparent (sensitive IP notwithstanding) about our technological progress and design thinking. The goal of this blog and updates to come, is to invite a wider audience from the aerospace and high-technology communities into the conversation. Hopefully this is also the beginning of making these exciting technologies more accessible for everyone.
What are we trying to achieve?
To set context, let me quickly explain the ambitions behind the service and its technical requirements.
Lilium envisions directly connecting inner towns and cities across ranges of between 40 and 200km at launch (and up to 500km longer term) at speeds of up to 300km/h, while enabling significant time savings for individual passengers compared to alternatives. We call this Regional Air Mobility (RAM) - not to be confused with Urban Air Mobility (UAM) that typically seeks to connect points within a city over much shorter distances (<20km). In parallel we have aimed for the highest payload of passengers (or freight) in the market since this translates into improved operating economics and, by extension, the flexibility to offer truly competitive ticket prices to customers. The revenue potential of an aircraft can be assessed in terms of passenger-kilometers per day per aircraft, which is ultimately a function of passenger seat-count multiplied by speed.
To deliver the desired range and top speed requires an aircraft that is highly efficient in cruise flight. Equally, the noise emissions during hover flight must be very low in order to guarantee regular market access near populated areas. The physical footprint (its length and wingspan projection) should be compatible with standard helicopter landing pads to make use of existing infrastructure, notwithstanding the inevitable need to build new facilities. It is important that any future higher-payload aircraft has the ability to leverage this same infrastructure. And finally, it goes without saying that it needs to be an environmentally friendly and sustainable solution.
So, the Lilium aircraft requirements can be summarised as follows:
- Zero operating emissions
- Highly efficient cruise phase
- Vertical take-off & landing for inner city accessibility
- Low noise for high frequency inner city flight operations and customer acceptance
- High seat capacity to achieve attractive unit economics and affordable pricing over time
- Scalability whilst maintaining ground footprint and low noise.
Similar to most new product design, the first reaction is that these requirements are contradictory and compete against each other. Traditional thinking drives trade-offs which leads to meeting one objective and precludes meeting some of the others.
Lilium’s greatest breakthrough is not accepting that traditional approach to trade-offs but rather getting to first principle understanding of the physics and then systematically innovating and optimising each subsystem in an integrated and coherent manner such that the resulting aircraft achieves all stated performance objectives. And this starts at the whole aircraft concept level. Let’s be very clear: from a technical perspective this path is the more difficult one, but at Lilium we are now confident that this approach will succeed and will ultimately pay off by offering new means of mobility to everyone.
Given the context of VTOL, before I go any further we need to touch briefly on the attribute of disc loading. For those not so familiar, disc load is the total aircraft weight to be lifted divided by the aggregate area of the rotors used during lift. High disc load is associated with low lift efficiency in hover and a high power requirement. Therefore, conventional eVTOL aircraft designs will attempt to minimise this attribute via larger diameter rotors but then at the expense of other attributes, such as noise, cruise efficiency and ground footprint. This leads to aircraft architectures that are either (i) low-disc load, low range, multicopter concepts or (ii) more complex open-rotor, higher range concepts. Both directions offer low-noise emissions at fixed payload but the payload of such systems cannot be scaled for the same noise profile. When you step back, the end result with this conventional approach is to effectively adapt existing non-electric VTOL concepts to make them electric.
Lilium approached this challenge from the other direction: exploiting a flight envelope more typical of a commercial fixed wing aircraft (rather than rotorcraft) we can meet our stated requirements by combining small electric ducted fan jets, with higher disc load, with a vertical take-off and landing aircraft.
In Figure 1, we can see that a lift fan concept such as Lilium's incurs a 50% penalty in power consumption versus a tilt rotor concept during hover, for the same weight. At first glance this appears like an inefficient design - but only when considering the hover phase.
This obvious weakness of hover phase efficiency (as illustrated in Figure 1) can be compensated for by an extremely efficient cruise flight enabled by designing an optimised fixed wing architecture. Other important mission-critical objectives, such as low noise, small footprint and scalability, can also be met.
What is the Lilium Jet Architecture?
Figure 2 shows our 7-Seater certification architecture. Yes, we are building a 7-Seater! It’s been in development for several years and represents the fifth evolution of our technology over our six year history. The aircraft features forward canards, main wings, and a distributed propulsion system providing vectored thrust. The main wing-span is limited to <14 meters to enable the use of existing helipads (approximately 14,000 possible locations in the US alone). Simple by design, there are no ailerons and there is no need for a vertical stabiliser. The landing gear is fixed and there are no hydraulics. Directional stability is provided by active electronic differential thrust control. The aircraft is controlled through a fly-by-wire avionics system. The main wings generate ~60% of the lift, the canards ~20%, and the remaining ~20% is generated across the fuselage. The canards and wings are positioned as far apart as practicable, to enable the aircraft to be stable in pitch.
The propulsion system consists of 36 individually controllable flaps, which also serve as lifting and control surfaces and each flap contains a ducted electric fan. The 36 ducted fans are embedded in a 1:2 ratio on the canard to main wing. Embedding the ducted fans into the wings eliminates the need for dedicated nacelles, reducing weight and minimising aerodynamic drag loss. The flap is rotated by an integrated servo unit, which can rotate the whole flap unit for controllability during hover and cruise flight. The flaps only receive two signals, fan speed and flap angle, by which the aircraft can be controlled throughout the flight envelope via thrust vectoring.
The fan stage is designed to be most efficient during hover flight. In order to achieve approximately the same aerodynamic stage efficiencies during cruise flight, we make use of a variable nozzle. Insights about this particular technology will be provided later in this blog.
The associated flight envelope
A common misconception of eVTOL aircraft is that they must emulate the mission profiles and performance characteristics of helicopters - that they must be capable of hovering for a long time as a central feature of their mission purpose. The goal of Lilium’s mission profile is to simply connect two geographical points as quickly and efficiently as possible. In other words, like a commercial airliner, to maximise time spent in the efficient cruise flight phase and to minimise takeoff and landing (i.e. hover) time.
Let us first have a look at a simplified visualisation of the Lilium Jet’s flight path to enable RAM. Figure 3b illustrates a simplified version of our flight path consisting of a vertical take-off, transition, climb, cruise, descent, re-transition and vertical landing.
Take-off and landing: during take-off we are only hovering approximately ~10-25 seconds and we also assume approximately ~20 seconds during a standard landing phase, whilst keeping approximately ~ 60 seconds as reserve. This leads to an overall total hover time (on a typical mission) of <60 seconds in the most power-hungry flight phase. The Lilium Jet demands approximately 2x more power during this phase compared to an open rotor concept on the market at equal weight (as indicated in Fig 1). However, since this phase is quite short, it represents approximately 5% of the mission energy budget.
Transition: during transition flight the aircraft accelerates forward and steadily increases its efficiency by creating more lift on the fixed wings as forward speed increases. The transition to forward flight and re-transition back to hover take approximately 20 seconds each. The total hover, transition and re-transition time consumes <15% of the total stored energy on a typical mission; or <20% if including reserves.
Climb, cruise and descent: the major part of the aircraft’s flight envelope consists of the climb, cruise and descent phase in which the efficiency is more than 10x higher than during hover flight (i.e. the power consumption is a tenth of hover in this phase). In this highly efficient phase, the Lilium jet flight physics resembles a commercial fixed wing jet aircraft such as the Boeing 777 or Airbus A350.
This profile underpins, in part, the decision to go with small ducted fans and thus accepting a higher disc-load. We intentionally spend only a very short time in takeoff and landing so we can optimise the aircraft design for the dominant period of a flight, which is cruise. So despite 2x higher power consumption in hover than an equivalent open rotor concept, this increased power demand is compensated by optimising cruise flight performance, which only requires 1/10th of hover power and also makes up to more around 90-95% of the flight time, see also Figure 3b.
A more detailed breakdown of the Lilium architecture, the aircraft’s power demand during the different flight phases and the predicted range is presented in our recently published technology paper, which was recently developed with colleagues from TU Berlin, TU Stuttgart and The Whittle Laboratory at Cambridge University. The data in the paper does not represent the data for the certification aircraft, but introduces our design thinking process and the respective architectural aircraft decisions.
The key technology enablers for the RAM mission
Item 1: Low noise electric ducted fans
At Lilium we quickly understood that at least one essential technology suited to such a flight profile would be electric ducted fans. During hover, ducted fans have roughly 40% increased efficiency compared to an open propeller at same discload since nearly all blade tip losses and swirl losses are removed thanks to the duct and the presence of stator vanes. This efficiency improvement goes some way to compensate for the higher power demand induced by the 10x disc loading while hovering. The power demand for our aircraft during hover and how the ducted fans can be sized accordingly was introduced here.
The low noise signature of ducted fans is a key enabler giving market access to inner city operation. The duct casings and acoustic liners contain the noise, stopping it from radiating in all directions, as it does with open propellers. This effect is visualized in Figure 4.
The team has spent significant time optimising the fan blade and stator vane aerofoil design such that the noise emission is dropped even further. Several in-house tools have been developed to optimize the aerodynamic performance whilst reducing the creation of noise caused by turbulent flow fluctuations. Among other things we have developed a high fidelity Computational Fluid Dynamics (CFD) code that allows for in-depth analysis of turbulent fluctuations by means of large-eddy simulations, see the snapshot in Figure 5. The analysis tool and models have been validated with actual test data measured during the test campaign of the 5-seater technology demonstrator aircraft. Reassuringly, even at this early stage of optimisation, the measured noise levels not only correlated well with pre-test predictions but also gave maximum peak values in the 60-65dB range, at 100m distance, without the addition of acoustic liners.
In a similar manner to commercial jet engines, further noise reduction can be achieved by the addition of acoustic liners to the internal walls of the duct structure. The liners absorb the acoustic energy of the rotor blade passing frequency, and the harmonics, which leads to a substantial reduction (up to 10dB(A) depending on conditions) in the overall perceived noise level. Subsequently, only broadband noise is emitted, which arises from the turbulent aerodynamic fluctuations. These fluctuations are mostly evident in the high frequency bandwidth of the audible spectrum. Nature helps us to perceive the resulting noise even at lower volume levels as a function of distance, which is why we can hear the low frequency impulse of helicopter rotors from very far distances (but not the rotor’s airflow itself).
I am quite familiar with acoustic liner technology from my time developing turbofan engines at Rolls-Royce, thus I am very comfortable introducing these into our baseline certification standard.
Figure 6 shows how our engineering efforts translate into the predicted absolute perceived noise level across the given flight envelope. Except for the initial hover phase with the corresponding 60dBA at 100 meter distance, the aircraft will be virtually inaudible during cruise flight. A good reference: normal conversational speech is circa 65dBA at 1 meter distance whilst a dishwasher is circa 60dBA.
Item 2: System architecture to maximise cruise flight efficiency
The Lilium Jet’s highly efficient cruise flight is enabled through three key architectural aspects:
- The fixed wings and canards create dynamic lift in forward flight
- Distributed propulsion enabled by embedding compact ducted fans in the wing
- Designing the flaps with a variable nozzle to optimise the fan flow during cruise
As the wings and canards create dynamic lift, the power consumption during the efficient cruise flight is only 10% of the hover consumption. As mentioned earlier, the cruise flight phase is substantially longer than the hover phase so the lower power consumption needed for cruise is compatible with the higher energy consumption of the hover phase.
We discussed earlier the concept of purposely designing the aircraft for high disc load during take-off. Instead of having large ducted fans which lead to structural challenges and increased cruise flight drag, this high disc load design leads to the freedom of having 36 smaller ducted fans and embedding them efficiently in the rear wing of our canard aircraft concept. Smaller embedded fans, lead to less wetted nacelle surface area, which increase the range due to less aerodynamic drag. This is visually demonstrated by Figure 7, in which the compact engine integration makes the cruise flight benefits visible.
This concept is known as distributed electric propulsion. Furthermore, in case of the Lilium Jet, we can exploit this even further by controlling the aircraft using thrust vectoring without any need for standard aerodynamic control surfaces such as tails, ailerons and rudders. This essential feature reduces structural weight, aerodynamic drag and structural complexity for the whole aircraft, with positive consequences for mission performance.
The aerodynamics of the fan are designed to be most efficient during hover flight. As the thrust required for cruise only a tenth of that required for hover flight, the flow field around the fan changes significantly. This would lead to the consequence of a significantly reduced aerodynamic efficiency during cruise flight compared to hover flight. However, at the exhaust of the ducted fan we use a variable area nozzle which changes the exhaust cross-sectional area during the flight and thereby guarantees high levels of fan efficiency in all phases of the mission profile. For hover, the nozzle is fully open whilst for cruise the nozzle area is reduced.
Since the fans are installed at the rear of the wing, during the transition flight, we achieve constant attached flow. This leads to an efficient high lift flow field and allows extreme controllability during the critical transition flight phase. The greater the mass flow in the engines then the greater the lift generated. This high lift effect combined with the fixed wing architecture allows us to approximate the reduction of power consumption during transition to be inverse of the path velocity squared.
Flight tests with our technology demonstrator proved this effect during the transition phase. In Figure 9 the decrease in power consumption is illustrated as a function of the flight velocity. It is visible that the required power is only around 40% of the hover power at approximately 25% of the cruise velocity. It is important to note that Lilium does not intend flight velocities below this threshold as part of the intended flight envelope and, thus, during our standard approaches the power consumption is significantly less compared to the hover power consumption. Velocities below 25% of the cruise velocity are used to decelerate only.
By embedding the ducted fans in the rear of the wing, the aircraft also profits from boundary layer ingestion which reduces the total pressure loss. Although the assessment of this effect is heavily coupled to general drag housekeeping, due to the low disc exposure during cruise flight, boundary layer ingestion benefits offset blocking effects from the distributed fans.
Scalability - noise versus footprint
For an equivalent thrust required to lift 1000kg in hover phase, an open propeller requires a 30x larger footprint than a ducted fan with acoustic liners for the same noise level.
The Lilium Jet is designed with a footprint that is approximately 10-15x smaller compared to an open rotor eVTOL competitor aircraft, which gives the Lilium Jet a 6dB(A) lower noise level prediction at a distance of 100m. This holds true for open propeller eVTOL concepts with a disc diameter of 2-3m and a rotating frequency during hover between 50-70Hz.
The low noise profile of a ducted fan comes from 4 important factors introduced earlier: firstly, the duct generates straight inflow on fan blades with low turbulence, which leads to less noise production. Secondly, the stator takes out swirl and thus allows for very low rotor tip speeds. We design the serial aircraft for blade tips speeds below a Mach number of 0,45. Thirdly, duct and acoustic liners shield and dissipate tonal and broadband noise, and finally, ducted fans require less power at same disc load and thrust due to the significant performance advantages.
In the same way as the disc load is 10-15x higher compared to competitor open propeller eVTOL aircraft, the compactness of Lilium’s ducted fan concept leads to 10-15x less footprint. This allows e. g. to build a 16-seater eVTOL with 63dBA @100m that fits into the same 15m helipad which is usually fully covered by a 5-seat propeller eVTOL. So this technology allows almost 4x the passenger throughput from the same infrastructure, which underlines the significant advantages in unit economics for ducted fan concepts.
Item 3: Energy system
A common, and reasonable, concern that is often raised is whether current battery technology can support Lilium’s architecture and its energy needs. The answer is simple: Yes. Batteries for an eVTOL aircraft need to deliver two things. 1) sufficient energy density to deliver long range and 2) sufficient specific power to support the hover phase. The latter is of special interest for us, as we’ve designed an aircraft with specifically high disc load as discussed earlier.
The total range of the aircraft is challenged by two factors of the battery cell: total energy in the cells and the minimum accessible state-of-charge (SOC), which represents the total accessible energy in a battery cell. One of the challenges in designing the battery system for the Lilium Jet is the provision of high power by the battery cells at low SOC.
Battery technology from earlier than 2014, which we used in our demonstrator, allowed for a minimum SOC of approximately 30-40% only, as the discharge capacity of the batteries did not provide sufficient power at lower levels of the SOC. Today, this is mitigated by state-of-the-art cell technology applying more advanced anode material, such as silicon, which increases the discharge capacity of the battery cell and, thus, the power provision at low SOC is significantly improved. For our 7-Seater serial Lilium Jet, a minimum SOC of 10-15% is enabled.
The performance of battery cell development has rapidly improved over the past decades and today cells are available with energy densities of >300Wh/kg and power density of >3kW/kg.
This will lead to a battery system design for the serial aircraft with >300kWh of total stored energy of which we can access 85-90% for our mission profile to reach a physical mission range of minimum 250km which does not account for operational reserves.
As Figure 11 shows, not only the battery density has grown significantly, but also the cells ability to provide sufficient power. While the battery cells being used for our demonstrator were not able to provide sufficient power and energy at the same time to design a 7-Seater aircraft at target range, since two years ago technology development allowed us to pivot the aircraft architecture enabling a payload increase to 7 seats. The energy team at Lilium have been working closely with our battery cell technology partners to deliver cell performance suitable to fulfil our mission profile and in a package that can be produced on a full scale manufacturing line.
By making use of our custom cells and chemistry with our aerospace grade supplier, we will be able to expand the range of the Lilium Jet to approximately 250km at entry into service in 2024. More details on the influence of future battery technology on our range, as well as the current cell technology itself, can be reviewed in this paper as well.
The battery system is always of critical interest when designing an eVTOL aircraft. The Lilium Jet has a very distinct power demand due to the varying disc loading between the hover and cruise flight phase and, therefore, the discussion of the battery system and the respective requirements gets rightfully a lot of attention. We will publish soon an additional blog article, which is fully focused around our battery design and we can't wait to show more.
In this blog I introduced you to the Lilium Jet architecture, a 7-Seater eVTOL aircraft reaching RAM markets initially within a minimum 200km range at speeds up to 300km/h. I hope that I have been able to give you insights into our thinking process, the rationale behind our requirements and why we make use of a unique aircraft architecture and technologies to enable significant time savings and enriched connectivity for customers.
One key aspect is our approach to our intended flight mission for the Lilium Jet (where we draw more comparison to a commercial airliner profile than a helicopter). Our choices recognise both the limited time period spent in the power-hungry hover mode, and the need for speed in the dominant cruise phase. Both these attributes can be addressed using small electric ducted fans arranged in an embedded, distributed propulsion system.
Although conventional thinking indicates that a higher disc load should be avoided, Lilium’s architecture has consciously chosen to increase disc load and power consumption during hover in order to achieve its mission objectives. Since the aircraft’s hover time is <90 seconds in total, the net impact energy usage is manageable.
The low power consumption during cruise is achieved by the combination of a fixed wing layout and compact, integrated ducted fans, which lead to low aerodynamic drag and excellent cruise efficiency. We can comfortably predict an operational range of approximately 200km, with battery technology available today whilst considering operational reserves. Battery technology will only continue to improve with time, which will benefit the range proportionally.
During this blog I showed you our thinking process linking low noise emissions with scalability and the key role ducted fans play. Ducted fans are well known for their noise shielding, asides from their aerodynamic efficiency advantages compared to open rotors, and the ducts themselves allow for the installation of acoustic liners. This gives the Lilium Jet a unique competitive advantage. We can scale our aircraft architecture to higher payloads, while maintaining the noise levels well below regulatory thresholds. The scalability of our technology and the unique architecture enables us to pursue a large scale inter-city shuttle business model in which we transport more passengers on a single flight. And we can leverage existing infrastructure to the full.
With the Lilium Jet we have secured simplicity in our solution which is always a good thing in engineering. The intrinsically stable aerodynamics of the architecture in combination with the vectored thrust propulsion system negates the need for much of the complications associated with traditional aircraft, eg. vertical stabiliser and rudder, ailerons, and hydraulics. All delivered in an eye-catching, attractive aircraft with a base architecture which is scalable for higher payload (be that passengers or freight) and market leading unit economics.
I hope this blog has provided you with a window into the design and engineering thinking at Lilium which is underpinned by a strong desire to solve the real market requirements of Regional Air Mobility. We have chosen an ambitious architecture and to reject the limitations of a traditional approach, while substantiating our technical claims with test evidence. And likewise, I hope you have gained a sense of my own journey in understanding this novel aircraft concept. I have confidence that we will not only become an exciting and innovative aerospace company but also that our aircraft technology will be transformative for regional air mobility. I am looking forward to the exciting months ahead as we reveal more about our technology, serial aircraft design, and commercial launch plans as we build Lilium to serve future customers and industry partners alike.
- Media Release