2.1 NDARC

Primary sizing and performance analysis of the aircraft is performed by NASA Design and Analysis of Rotorcraft (NDARC, [Reference Johnson11–Reference Johnson12]). NDARC has a modular architecture, facilitating its application to new aircraft and propulsion types, including non-traditional propulsion systems. Aircraft are represented as a collection of components, with surrogate or semi-analytic models for performance and weight of each component. NDARC performs fixed-point iteration to find the vehicle size based on specified sizing rules, missions and flight conditions. The most complete description of the concept aircraft is encapsulated in NDARC output files. NDARC is capable of assessing the impact of advanced technology and design choices. NDARC calculates economic measures, such as productivity, flyaway cost and operating cost for the concept aircraft. For a sized aircraft, NDARC calculates the weights and speeds to fly the required operations, such as noise metric conditions, as these are vehicle-specific, and must consider nonlinearities and limits in the propulsion system, not just aerodynamic performance. Advanced technology is incorporated in NDARC models in terms of the surrogate models for rotor and engine/motor performance, battery weight and efficiency and technology factors for component weights. Often useful aircraft technologies (such as active flow control) must be integrated into the design, accounting for weight and power increments as well as performance and efficiency improvements.

2.2 CAMRAD II or CHARM

A comprehensive analysis code simultaneously solves the rotor dynamics and aerodynamics for a trimmed or transient flight condition. CAMRAD II was used in the optimisation of the rotor geometry for the aircraft sizing conditions, and to develop calibrated rotor performance models for NDARC; as well as to calculate rotor and hub loads for structural design, assess rotor and blade stability and whirl flutter, and provide rotor blade airloads for mid-fidelity acoustics calculations. NASA is not actively developing its own comprehensive analysis solution, rather is applying commercial software for this step in the process. Two commercial codes, CAMRAD II [Reference Johnson13–Reference Johnson14] and CHARM [Reference Wachspress, Quackenbush and Boschitsch15], are being integrated interchangeably into the toolchain. By using two codes, a robust application programming interface can be defined that will allow other comprehensive analysis codes to be integrated into the toolchain. Both CAMRAD II and CHARM have well-validated free wake modeling capability. For interacting rotors, a freely convecting wake is necessary to accurately predict performance, as well as blade-vortex-interaction airloads that are responsible for significant noise. For example, lift+cruise, quadrotor, and side-by-side aircraft have rotors that interact to varying degrees in various stages of flight. Similarly, for practical UAM rotors larger than a couple of feet in radius, elastic and kinematic motion should be considered to calculate the rotor trim and blade-vortex interactions. Blade airloads and velocities will in general depend on blade motion, induced velocity, and wake convection. Whether or not the aerodynamic model of a comprehensive analysis code is adequate for evaluating the noise sources, a comprehensive analysis model of the structure is required. Building a comprehensive analysis model is generally a necessary step in the development of a meaningful high-fidelity simulation, as vehicle trim, rotor trim and blade kinematic and elastic motion are usually computed by a comprehensive analysis, which is coupled to CFD airloads.

2.3 AARON/ANOPP2

In order to predict the acoustic metrics, a tool that predicts thickness, loading, and broadband noise, acoustic propagation and observer noise is necessary. The Aircraft NOise Prediction Program 2 (ANOPP2, [Reference Lopes and Burley16–Reference Brooks, Pope and Marcolini18]) and AeroAcoustic ROtor Noise (AARON) tools provide the acoustic calculations in the toolchain. AARON is the user code to perform rotorcraft calculations with ANOPP2. A Python script provides a simplified interface that is geared toward generating the certification noise and other typical rotorcraft calculations with a manageably small number of inputs. AARON/ANOPP2 calculates the loading noise with a compact loading model, using the blade section loading and motion as well as the aircraft operating condition from the comprehensive analysis. Consistently, thickness noise is calculated with a compact monopole model. An aerofoil self-noise model extended to rotating blades is used for broadband noise [Reference Brooks, Pope and Marcolini18]. AARON/ANOPP2 can also calculate the loading and rotational noise from a higher-fidelity solution for the blade surface pressures and geometry, such as from CFD calculations.

2.4 RCOTools

Rotorcraft Optimisation Tools (RCOTools, [Reference Meyn19]) is a set of Python libraries that serve as application wrappers for input/output and program execution of several tools. In this toolchain, RCOTools interfaces for NDARC and CAMRAD II/CHARM are used. The data connections depicted by grey lines in the XDSM diagrams of Fig. 1 are facilitated by RCOTools.

2.5 Additional tools

Other tools are available to provide detailed models and data for the principal codes: engine models (NPSS [Reference Snyder and Tong20]); handling qualities assessment (SIMPLI-FLYD [Reference Lawrence, Theodore, Johnson and Berger21]; and FlightCODE); structural design and analysis (IXGEN [Reference Rohl, Dorman, Cesnik and Kumar22] and M4 Structures Studio [Reference Winter, Márquez and Scheneman23]); geometry and layout (OpenVSP [Reference Hahn24] for initial geometry; Rhino for final geometry); and high-fidelity aerodynamic loading (OVERFLOW [Reference Derlaga, Jackson and Buning25] and FUN3D [Reference Biedron, Carlson, Derlaga, Gnoffo, Hammond, Jacobson, Jones, Kleb, Lee-Rausch, Nielsen, Park, Rumsey, Thomas, Thompson, Walden, Wang and Wood26]).

2.6 Assessment of tools

A goal of the present tool suite development effort within NASA’s RVLT project is to provide robust computational methods that facilitate design space exploration with varied problem definitions and with the ability to concurrently consider several different potential solutions. The development of the NASA concept vehicles has demonstrated that the computational tools available for rotorcraft aeromechanics analysis and design are generally applicable to VTOL AAM aircraft. In particular, the toolchain has been capable of quantitatively trading several relevant noise reduction technologies and design approaches for aircraft that can perform the Urban Air Mobility mission [Reference Silva and Johnson10]. The calculations performed for these trades are quick enough to allow several simultaneous aircraft types to be evaluated by an individual designer or design team, with the calculations performed on workstation computers. The tools are practical, in terms of the amount of input required and the computation times [Reference Silva and Johnson10].

4.1 Battery technology

Light and efficient batteries are crucial to producing good designs for electric aircraft. The electric designs for these aircraft assume an installed, usable battery specific energy of 400 Wh/kg (pack). Internal resistance reduces battery efficiency at high discharge rates, so typical Li-ion battery discharge characteristics are used to calculate the efficiency. Margins for maximum charge and discharge are established to prolong battery life (in terms of discharge-charge cycles): charge to within 5%–10% of full capacity, discharge to 15%–20% capacity. Current delivery limits for cells are specified as a C-rate (capacity/hr). Even with a high maximum burst discharge capability (high maximum power), discharge currents in today’s high-specific-energy Li-ion cells must be limited to 2–3C for good battery life. The installed specific energy is reduced by packaging and conditioning requirements, including thermal management systems. For the electric propulsion variants of the three configurations, design excursions of range and battery technology were generated to quantify the impact on aircraft weight, power, and feasibility. Specific weight, the crucial aspect of battery technology for aircraft design, was considered from 400 Wh/kg down to 150 Wh/kg (values for pack, not cell). Below about 300 Wh/kg, the range capability was reduced while the aircraft weight increased. Having established these trends, a high level of battery technology had to be assumed for the UAM concept vehicles to ensure that the designs would meet the mission requirements.

5.1 Quiet Single Main Rotor Helicopter

The Quiet Single Main Rotor helicopter (QSMR) is representative of a state-of-the-art helicopter designed specifically for the UAM mission [Reference Johnson8]. In addition to shorter range for UAM applications, several design decisions were made to bias the design toward low noise. The tools in the toolchain have been validated with legacy helicopters [Reference Johnson12–Reference Lopes and Burley16], including some helicopters with the noise reduction technologies employed in this demonstration. Therefore, the predicted size, weight, performance and noise of the QSMR is expected to be quite close to that of a real aircraft designed to these criteria and with the assumed level of technology. The relative predictions of the other vehicles can therefore be compared to the QSMR, in order to establish a reasonable expectation of the absolute merits and costs.

The impact of rotor noise on community acceptance of helicopters was recognised early, so there have been five decades of work on reducing noise. From the beginning, much of the focus was on the design variables of tip speed and tail rotor size. Based on demonstrated technology, the QSMR concept vehicle uses a turboshaft engine with a sound-absorbing installation (accounted for in weight of engine installation), and the NOTAR (NO TAil Rotor) anti-torque system to eliminate tail-rotor noise. The NOTAR system is heavier than a traditional tail rotor, but can reduce the anti-torque contribution to noise to such an extent that the main rotor is by far the dominant noise source. For low main rotor noise, the design has a low disk loading and low tip speed, hence large blade area and a large number of blades. The direct and indirect results of low tip speed include high rotor and transmission weight, which increases the aircraft takeoff weight. Design excursions include turboshaft and electric propulsion. The effects of blade tip droop and higher-harmonic control on the aircraft noise were examined [Reference Johnson8].

5.2 Quadrotor

The quadrotor aircraft is intended to represent multirotor-type aircraft, which use collective or rotor speed control for flight control, without cyclic control [Reference Silva, Johnson, Antcliff and Patterson6]. One interesting attribute of the designs for the NASA quadrotor concept aircraft is that the rear rotors are mounted higher on the aircraft than the forward rotors. Initial design work with the free wake comprehensive analysis in CAMRAD II led to this design decision, as a significant reduction in power required for cruise was predicted. Subsequent higher fidelity analysis confirmed the beneficial effect, albeit with a different absolute magnitude. Design excursions include turboshaft, hybrid and electric propulsion. The implications of hingeless blades, rpm control and reciprocating engine propulsion were examined for the single-passenger quadrotor [Reference Johnson, Silva and Solis4].

5.3 Side-by-side helicopter

The side-by-side helicopter [Reference Silva, Johnson and Solis3, Reference Antcliff, Whiteside, Silva and Kohlman7] is representative of a high-performance helicopter, which has main rotors that intentionally interact as they physically overlap and intermesh. The side-by-side aircraft has a support crossbar to support the rotors, but this bar does not generate significant lift. The interaction of the main rotors is intended to reduce induced power in forward flight. Design excursions include turboshaft, electric and hybrid propulsion.

5.4 Lift+Cruise aircraft

The Lift+Cruise aircraft [Reference Silva, Johnson, Antcliff and Patterson6] is a stopping-rotor thrust- and lift-compound helicopter. There are three distinct flight modes for the aircraft: helicopter mode with the lifting rotors turning, compound mode with lifting and thrusting rotors operating and airplane mode with the lifting rotors stopped and forward thrust is provided by a pusher propeller. In airplane mode, the lifting rotors are stopped with the blade axis pointed along the vehicle longitudinal axis, and therefore nominally aligned with the free stream to minimise drag. Unlike the other concept vehicles, the lift+cruise aircraft can only have two blades on the lifting rotors, so as the rotor speed and solidity changes, no change is allowed in number of blades. The vast majority of lift+cruise aircraft being proposed have fixed rotor pitch and use variable rpm for control. However, to facilitate calculations of aerodynamic interference and noise, investigations can model the aircraft using collective control of the lifting rotors with fixed (identical) rotational speed. Similar to the quadrotor, the lift+cruise aircraft has the rear rotors mounted higher than the front rotors. The performance effects are similar for rotor-rotor interference, but there are now two additional considerations: the influence of the boom for an under-mounted rotor and the interference with the wing. By placing the front rotor under the support boom, a simpler and more compact support can be built. Design excursions include turbo-electric and electric propulsion.

5.5 Tiltwing

The tiltwing vehicle [Reference Whiteside, Pollard, Antcliff, Zawodny, Fei, Silva and Medina9] uses a turbo-electric propulsion system to power six proprotors positioned on a tilting main wing and two tilting proprotors positioned on the horizontal tail. All rotors are directly connected to electric motors, with no interconnect shaft. There is a single turboshaft engine connected to a generator, with a relatively small battery just to enable emergency landing after loss of turboshaft power. Collective control is used, with hover tip speed of 550 ft/sec and cruise tip speed of 300 ft/sec. A tiltwing is of interest for UAM due to its potential to increase cruise speed and efficiency while reducing noise in cruise. The other concept vehicles all have rotors that do not tilt and so either operate in edgewise flight (which limits cruise speed and typically generates more noise than a proprotor in axial flight) or are stopped during cruise (generating drag and weight penalties).

All proprotors are located ahead of the wing leading edge by 3/4 the proprotor radius, in order to reduce the noise caused by wing-rotor interaction; however, this introduces additional challenges in structures (long moment arm) and aerodynamics (the contracted proprotor wake does not yield beneficial aero-propulsive effects along the full span of the wing). The main wing is swept so that the proprotors are longitudinally staggered to prevent cascading proprotor failures. Laterally, the main wing proprotors are placed such that the inboard proprotor is close to the fuselage and the proprotor disks are tangential to one another as viewed from the front. Although this will leave sections of the wing unblown as the slipstream aft of the proprotors contracts, the inflow to the proprotors will be cleaner and the reduced rotor-rotor interaction is expected to reduce noise and increase efficiency. The proprotors at wing tips are positioned such that the shaft axis is coincident with the wing tip to assist in counteracting the wing tip vortex.

6.1 Battery technology

The most important factor in the feasibility of electrical propulsion systems is the requirement for light-weight, high-power batteries. The baseline designs here assume an installed, usable specific energy of 400 Wh/kg. Current state-of-the-art batteries have installed specific energy of 100–150 Wh/kg. The weight and power variation with range and battery technology are shown in Fig. 7 for an electric side-by-side aircraft. High specific energy enables a useful range with a reasonable aircraft weight and power. Aircraft size does not change the conclusions from these figures, as similar results are obtained for both single-passenger and 15-passenger side-by-side designs [Reference Johnson, Silva and Solis4].

Figure 7. Electric side-by-side helicopter (six passengers): weight and power variation with range and battery technology; discharge current for battery technology 150–400 Wh/kg (pack).

The power capability of batteries is also important. High power is obtained with high current, and current can be characterised by fraction x of the charge capacity $C$ : $I = xC$ , with units of 1/hr for $x$ . A maximum burst discharge current of 10 $C$ to 30 $C$ (fully discharged in 6 to 2 minutes) is possible for emergency use, but long battery life typically requires currents of 1 $C$ to 3 $C$ . The discharge current variation with range is shown in Fig. 7 for the electric side-by-side aircraft. The cruise current is less than the hover current for these designs, since cruise speed is fallout (not specified) and the power is sized by the hover condition. The battery capacity is the sum of hover, cruise and reserve requirements:

(1) \begin{align} {E_{cap}} = {E_{cruise}} + {E_{hover}} + {E_{reserve}} \end{align}

Writing cruise power in terms of the aircraft effective lift-to-drag ratio ( ${P_c} = WV/\left( {L/{D_e}} \right)$ ), the cruise energy is proportional to range:

(2) \begin{align}{E_{cruise}} = \left( {{P_c}/{\eta _c}} \right) \times {\rm{time}} = WV \times {\rm{time}}/\left( {\left( {L/{D_e}} \right){\eta _c}} \right) = WR/\left( {\left( {L/{D_e}} \right){\eta _c}} \right)\end{align}

where ${\eta _c}$ is the propulsion system efficiency in cruise, and the range is $R = V \times {\rm{time}}$ . From the charge capacity ${C_{cap}} = {E_{cap}}/\nu $ (for voltage $\nu $ ), and hover current $I = \left( {{P_h}/\nu } \right)/{\eta _h})$ ( ${\eta _h}$ is the propulsion system efficiency in hover), the hover discharge current is

(3) \begin{align}{x_{hover}} = I/{C_{cap}} = {P_h}/\left( {{\eta _h}{E_{cap}}} \right) = \left( {W\sqrt {W/2\rho A} /FM} \right)/\left( {{\eta _h}{E_{cap}}} \right)\end{align}

with hover power in terms of figure of merit ( ${P_h} = W\sqrt {W/2\rho A} /FM$ ). Substituting for ${E_{cap}}$ gives

(4) \begin{align}1/{x_{hover}} = R\frac{{{\eta _h}FM}}{{{\eta _c}\left( {L/{D_e}} \right)}}/\sqrt {W/2\rho A} + {\rm{constant}}\end{align}

where the constant comes from the hover and reserve energy capacity. Ignoring the constant gives

(5) \begin{align}{x_{hover}} = \sqrt {W/2\rho A} \frac{{{\eta _c}\left( {L/{D_e}} \right)}}{{{\eta _h}FM}}\frac{1}{R}\end{align}

High hover efficiency (low disk loading and high figure of merit) reduces the current, but short range or high cruise efficiency $\left( {L/{D_e}} \right)$ reduces the battery capacity required, hence increases the hover current ${x_{hover}}$ . As illustrated in Fig. 7, this result is independent of battery technology, except as it impacts the range that is achievable by a design. For the side-by-side aircraft, the hover current ${x_{hover}} \lt 1C$ if the range is greater than about 160 nm, and ${x_{hover}} \lt 2C$ if the range is greater than about 60 nm.

The conclusion from Fig. 7 is that reasonable hover current requires that the aircraft have a large battery, preferably because of good weight efficiency permitting long range, not because of low cruise efficiency.

6.2 Aircraft aerodynamic efficiency

Electric propulsion is enabled by aerodynamic efficiency of the aircraft, in both hover and cruise. For each design, there is a disk loading that minimises aircraft weight, power, and energy. Small aircraft with edgewise rotors optimise with low disk loading. Rotor-rotor interference must be considered for optimum cruise performance. Interactional aerodynamics impact performance and operation: for tiltwings, wing separation or buffet during conversion must be considered; for tiltrotors, hover download and rotor-tail interactions must be considered. Rotor shape optimisation covers blade twist and taper, tip sweep and droop. Drag minimisation includes hub, rotor support and airframe.

For aircraft with two or more main rotors, interactions between the rotors have a significant impact on performance, vibration, noise and handling qualities. The interactions depend on the arrangement of the rotors. Figure 8 illustrates the wake geometry of the quadrotor and side-by-side aircraft in cruise flight. The overlap of the side-by-side rotors significantly improves the efficiency of cruise flight, because the twin rotors act as a single, large-span wing system. Figure 9 (from [Reference Silva, Johnson, Antcliff and Patterson6]) shows the influence of overlap (wing span = 1.0D = rotor diameter means the rotor disks are tangent) on rotor efficiency (rotor $L/{D_e} = WV/\left( {{P_i} + {P_o}} \right)$ ).

Figure 8. Wake geometry of side-by-side and quadrotor aircraft at cruise speed.

Figure 9. Rotor cruise efficiency as function of overlap for side-by-side helicopter.

Elevating the rear rotors above the front rotors on the quadrotor reduces the cruise power, as shown in Fig. 10 (from [Reference Silva, Johnson and Solis3]), by reducing the interference of the wakes of the front rotors on the rear rotors. Elevating the rear rotors is expected to reduce vibration and noise and improve handling qualities as well. Moving the aircraft center of gravity forward of the mid-point between the rotors, so the front and rear rotors trim closer to the same blade loading C_T/ $\sigma$ at cruise speed, further reduces the power (Fig. 10).

Figure 10. Influence of elevation of rear rotors on cruise performance of quadrotor.

The effects of the rotor-rotor interactions may require vibration and load alleviation systems. The present designs have a weight allocation for vibration control.

6.3 Trim of multi-rotor aircraft

For the quadrotor, both collective and rotor speed control were considered. Figure 11 shows the trim operating conditions of the front and rear rotors for the two control methods. Edgewise rotor flight has reduced induced power (relative hover, for the same lift) due to increased flow through the disk, followed by power increasing with speed as the parasite power increases. The thrust of the rear rotors increases with speed relative the front rotors, for aircraft pitch moment trim. With collective control and fixed rotor speed, the rotor blade loading C_T/ $\sigma$ follows the thrust variation, remaining moderate, while the collective control follows the power variation with speed. With rotor speed control and fixed collective, the rotor rpm follows the power variation with speed, while the rotor C_T/ $\sigma$ increases with speed initially and then decreases (Fig. 11). The increase in C_T/ $\sigma$ might be limited by maximum blade loading, perhaps requiring a smaller design C_T/ $\sigma$ at hover (hence larger blade area).

Figure 11. Electric quadrotor trim as a function of flight speed, for collective control and rotor speed control (with collective control, tip speed and V/V_tip are same for front and rear rotors).

For the lift+cruise aircraft, Fig. 12 shows the variation with speed of the rotor blade loading C_T/ $\sigma$ and wing lift coefficient. C_L. This design has rigid rotors for hover and low speed lift, using fixed pitch and rotor speed control. The lift rotors are stopped for cruise, with the wing providing all lift. As the induced power decreases with speed, the rotor blade loading C_T/ $\sigma$ increases. As speed decreases, the wing lift coefficient increases, bringing the wing closer to stall. Figure 12 also shows the rotor loading limits due to stall, according to four measures: maximum thrust, change in thrust derivative with collective, four times the profile power and twice the profile power. These hingeless, fixed-pitch rotors operate with large hub roll moments in forward flight (here advance ratio 0.36 and lift offset 0.34 at 90 knots), which increases the loading limits due to stall (a flapping rotor has a steady stall limit of about C_T/ $\sigma$ = 0.12 at 90 knots). At sufficiently high speeds (here above 90 knots), blade stall encompasses most of the rotor disk, and the C_T/ $\sigma$ limit decreases. The design hover C_T/ $\sigma$ must be low enough and the wing area large enough that transition from rotor-borne to wing-borne flight is possible over a reasonable speed range.

Figure 12. Lift+cruise aircraft rotor blade loading and wing loading variation with flight speed.

6.4 Rotor/propeller design

Design of the rotor or propeller impacts weight, vibration, and handling qualities. For the quadrotor aircraft, both flapping and hingeless rotors were considered. The flapping rotor had 4% hinge offset, with 45° pitch-flap coupling to minimise flapping relative the shaft. The hingeless or rigid rotor generates higher blade and hub loads, which implies higher rotor weight and vibration-control weight, and the resulting aircraft has 25% larger design gross weight than with flapping rotors.

Figure 13 shows the calculated mean hub moments for the four rotors on the quadrotor aircraft, in level flight and 2g turn. The hub moment is given as lift offset = moment divided by thrust times rotor radius. With no flap hinge and no cyclic pitch, the rotor in edgewise flight operates with increasing lift offset as speed increases. For reference, the design hub moments are shown in Fig. 13 for typical hingeless rotors (Bo-105 level flap stiffness with 10 deg tip-path plane tilt) and lift offset rotors (design load for 200–250 knot aircraft).

Figure 13. Quadrotor (fixed pitch, hingeless rotors) mean hub moments in level flight and 2g turn (rotor designation: FR = front right, FL = front left, AR = aft right, AL = aft left); lift offset is M/WR = hub moment divided by thrust times rotor radius.

All UAM designs will likely require active control of rotorcraft vibration. Fixed-frame active vibration alleviation is common for rotorcraft. In addition, up to 90% reduction of loads and vibration using higher-harmonic control (HHC) or individual blade control (IBC) has been demonstrated through analysis, wind tunnel test and flight test [Reference Johnson86]. The UAM designs presented here have technology factors that represent the weight of vibration alleviation systems, excluding HHC or IBC.

6.5 Propulsion system weights

The trend in UAM to explore electric, and in particular distributed electric propulsion (DEP), has introduced the need to model motors, generators, motor controllers, heat exchangers, wiring, and batteries. Since 2014, NDARC has had the capability to model electrical propulsion system components, and as more aircraft have been modeled, further empirical and semi-empirical models for elements in the electrical propulsion system have been added to NDARC’s built-in capabilities.

Researchers at NASA have used the reference aircraft designs to explore electric architectures including DC, DC/AC, turboelectric and parallel hybrid [Reference Hanlon, Thomas, Csank and Sadey35, Reference Thomas, Chapman, Alencar, Hasseeb, Sadey and Csank51], resulting in extensions to the NPSS software tool. Other research has looked at power transmission wiring design for different material systems [Reference Aretskin-Hariton, Schnulo, Hendricks and Chapman36], finding that the minimisation of vehicle weight may favor lower-density higher-resistance wiring for some architectures. Research into thermal management systems for electric VTOL applications have found that the high power, low mass-flow and small temperature-gradient conditions of eVTOL aircraft in hover present a challenge for cooling design that probably demands a dedicated fan-driven heat exchanger, rather than being able to utilise surface cooling as would be possible for an electric fixed-wing aircraft [Reference Chapman, Hasseeb and Schnulo53].

Two particular design trades for electric propulsion systems are indicative of the new trades, which may be considered and modeled with the advent of electric propulsion: direct-driven rotors versus motor-plus-gearbox; and fewer large rotors driven by larger motors versus many smaller rotors driven by smaller motors (DEP). The weight of the motor plus transmission of an electric rotorcraft is shown in Fig. 14, as a function of rotor radius for a disk loading of 4 lb/ft² and tip speed of 550 ft/sec (typical of the present designs). These weights are calculated using parametric equations for motor and drive system weights [Reference Johnson11], with technology factors appropriate for future designs. N is the motor speed (rpm) for cases with a transmission. For direct drive the motor speed equals the rotor speed. These results show that for current and projected future technology, a high-speed (low-torque) motor with an advanced transmission is lighter than direct drive. For direct drive to be the lighter design approach, a light-weight, high-torque motor is required, operating with large mean and oscillatory loads from the rotor.

Figure 14. Motor+transmission weight variation with rotor size and motor speed N (rpm).

Figure 15 shows the motor plus transmission weight variation with number of rotors, for a disk loading of 4 lb/ft², tip speed of 550 ft/sec, and aircraft weight 5,000 lb. Figure 16 adds the rotor weight to the motor and transmission weight, for design C_T/ $\sigma$ = 0.10 and flap frequency of 1.25/rev (typical of hingeless helicopter rotors). The rotor weight is calculated using parametric equations [Reference Johnson11], which are based on data that includes aircraft weights and rotor diameters corresponding to these designs. Considering just the propulsion system (motor, transmission and rotor), the weight decreases significantly as the number of rotors increases (Fig. 16). Similar results are obtained considering turboshaft engines instead of electric motors, and for a large range of aircraft size. This result is therefore not consistent with the observation that for helicopter designs the single-main-rotor configuration (even with a tail rotor) is preferred to tandem or side-by-side aircraft types. Usually, adding the weight and drag of the structure that supports the rotors changes the optimum solution. Figure 6 shows the empty weights for the concept vehicles designed for the UAM mission. With eight lifting rotors, the lift+cruise type has higher structural weight even though it has good cruise efficiency.

Figure 15. Motor+transmission weight variation with number of rotors and motor speed N (rpm).

Figure 16. Motor+transmission+rotor weight variation with number of rotors and number of blades.

6.6 Structural weights

With the introduction of new aircraft types having very different structural arrangements and different loading conditions, the use of empirical models based on existing helicopters, tiltwings, tiltrotors and fixed wing aircraft is of questionable validity. For the initial design work on the NASA UAM concept vehicles, the existing weight relationships in NDARC were used to model the vehicles, with adjustments to capture the likely impact of design differences. For instance, the side-by-side aircraft was modeled with a tiltrotor wing for the rotor support beam, using the jump takeoff sizing and preventing the whirl-flutter constraint from becoming active. For the quadrotor rotor support beams, a fixed-weight increment was added to NDARC based on an external calculation of the support beam weight. For the lift+cruise aircraft, the wing was modeled with the NDARC fixed-wing weight equation, and the rotor supports were modeled using the same methodology as applied for the quadrotor.

To address the structural weight uncertainty in the absence of any certified multi-rotor UAM aircraft to be used in development of new weight models, NASA has undertaken to introduce higher-fidelity weight assessment into conceptual design while trying to maintain the flexibility and lack of detailed information that are key features of the conceptual design stage of development. M4 Structures Studio [Reference Winter, Márquez and Scheneman23] is a tool that interfaces between parametric conceptual geometry in OpenVSP [Reference Hahn24] and NASTRAN to rapidly and robustly build finite-element models and maintain the parametric connection to the conceptual design geometry. Using this tool, the structural weights of the initial reference aircraft are predicted to have been in some instances conservatively heavy (side-by-side in [Reference Winter, Márquez and Scheneman23]) and in other instances optimistically light (lift+cruise in [Reference Winter, Márquez and Scheneman23]).

As UAM vehicles go through certification and enter operational service, NASA will make an effort to gather as much information as possible about actual weights and driving load conditions, and use this information to calibrate weight models and technology factors.

6.7 Handling qualities

The NASA concept vehicles are being used in explorations of the handling qualities and resulting design implications of these new rotorcraft configurations [Reference Malpica and Withrow-Maser40, Reference Schuet, Malpica, Lombaerts, Kaneshige, Withrow, Hardy and Aires47, Reference Withrow-Maser, Malpica and Nagami60, Reference Withrow-Maser, Malpica and Nagami75]. Similar investigations are in progress throughout the industry [Reference Walter, McKay, Niemiec, Gandhi and Ivler87–Reference Walter, McKay, Niemiec, Gandhi and Ivler91]. In [Reference Withrow-Maser, Malpica and Nagami75], control models of three vehicles (the quadrotor, an octocopter and the lift+cruise aircraft) were created and compared to determine the effect of rotor number and disk loading on control margin and design. All of these aircraft had electric propulsion with rpm control. In variable rotor speed-controlled vehicles, the propulsion system is directly in the path of the control system, which means the stability and control of the vehicle must be well understood and considered in the process to determine power requirements and the related overall weight of the vehicle.

NDARC was used to size the three, six-passenger vehicles to the same mission. SIMPLI-FLYD [Reference Lawrence, Theodore, Johnson and Berger21] was then utilised to generate the bare airframe model. Flight control systems were synthesised using CONDUIT [Reference Tischler, Colbourne, Morel, Biezad, Levine and Moldoveanu92] to a common set of handling qualities-oriented guidelines following a multi-objective function optimisation procedure. A model following control system architecture was used, with an integrated electric motor and speed controller. The general optimisation philosophy was to jointly minimise motor usage and crossover frequency, while primarily enforcing closed-loop robust stability margin and performance constraints — disturbance rejection bandwidth (DRB) (and disturbance rejection peak) for the main (heave, roll, pitch and yaw) control loops and minimum steady-state error for the engine speed controller (ESC) loops. Additionally, constraints to ensure minimum eigenvalue damping and crossover frequency were imposed.

All three vehicles met Level 1 handling qualities specifications when the motor current was not constrained. For all three vehicles, the heave and yaw axes demanded more actuator usage (rms current) than the roll and pitch axes. Between heave and yaw, heave was the more demanding because of the dependence of heave on the engine speed controller. The optimised ESC gain and phase margins were large, so the motor control was not a limiting factor. When actuator usage was translated to current margin, torque margin, and power margin, heave was the most demanding axis, followed by yaw, roll, and then pitch for all three vehicles (Fig. 17). the required motor power and torque margins, relative hover, are P_max/P_hover = 1.67 for quadcopter, 2.49 for octocopter and 2.15 for lift+cruise aircraft. The results emphasise the importance of an accurate motor model within the control system architecture.

Figure 17. Power margin (as fraction of hover power) required to achieve level 1 handling qualities [Reference Withrow-Maser, Malpica and Nagami75].

NASA is also conducting a series of human-in-the-loop (HITL) simulation studies in the NASA Ames Vertical Motion Simulator (VMS). A simulation completed in 2021 investigated the connection between engine/motor usage, heave-axis control law design requirements and handling, and ride qualities for all-electric multirotor (quadrotor) architectures using rotor collective or rotor speed (RPM) for control. The studies considered the six-passenger reference quadrotor [Reference Silva, Johnson, Antcliff and Patterson6] and a specially designed RPM-controlled variant [Reference Withrow-Maser, Malpica and Nagami75]. The first aircraft was employed as the test control subject, and theoretically configured to be in deep Level 1 handling qualities according to ADS-33. The RPM-controlled variant was configured with varying heave axis pilot response time constants and feedback loop disturbance rejection bandwidths. This approach permitted an examination of tradeoffs in required engine power and handling qualities performance. Low-speed and hover-handling qualities of these configurations, with an Attitude Command-Attitude Hold response type, were evaluated by rotorcraft experimental test pilots in a combination of ADS-33 and specially tailored, UAM-representative mission task elements that emphasised precision over pilot aggressiveness. Preliminary trends frequently showed performance degradation with increased pilot aggressiveness. However, increased time for manoeuvring did not universally correlate with an ability to be more precise. Also, to gain initial insight into UAM passenger susceptibility to motion sickness during flight under smooth (collective control with high-disturbance rejection bandwidth) and moderately rough (RPM control with medium-disturbance rejection bandwidth) flight conditions, a second, complementary study was conducted in the VMS using passenger test subjects. Each participant took two brief (approximately 10 minute) prerecorded simulated quadrotor flights that took off, cruised at an altitude of about 500 ft and then landed at the same San Francisco vertiport. Preliminary findings suggest statistically significant increases in motion sickness symptoms reported, with positive correlation with individual predisposition to motion sickness, for flight under RPM control (medium-disturbance rejection bandwidth) compared to preflight baselines.

6.8 Safety and airworthiness

Airworthiness approval means a document, issued by the FAA for an aircraft, which certifies that the aircraft conforms to its approved design and is in a condition for safe operation (14 CFR 21.1(b)(2)). While certification requirements and procedures for air taxi aircraft may be debated, negotiated, or even contested, for aeromechanics research the focus is on safe operation. Every innovative aircraft type and non-traditional propulsion system requires an extensive failure mode, effects and criticality analysis (FMECA). Important for air taxi aircraft are crashworthiness and the consequences of propulsion system failure. Crashworthiness requirements affect design of airframe structure, landing gear, and passenger accommodation and restraint. Propulsion system failures must be considered in detail. In particular, single as well as complete engine failure must be considered, with requirements for control and methods for safe landing. In the concept vehicles, degraded weather operation, propulsion system failures and crashworthiness have been assumed to be requirements, and technology factors representing these design considerations have been applied.

6.9 Reliability and safety assessments

NASA commissioned several investigations to contrast major design differences of UAM aircraft and explore the safety and reliability implications from the perspective of vehicle design, propulsion architecture and flight dynamics and control [Reference Darmstadt, Catanese, Beiderman, Dones, Chen, Mistry, Babie, Beckman and Preator32, Reference Beiderman, Darmstadt, Dillard and Silva68–Reference Darmstadt, Pathak and Mistry70, Reference Darmstadt, Pathak, Chen, Mistry, Arkebauer, Beiderman, Dillard, Zierten, Beckman, Monroe, Catanese, Greene and Preator84–Reference Justin, Patel, Bouchard, Gladin, Verberne, Li, Ozcan, Rajaram, Mavris, D’Arpino, Aldemir, Rizzoni, Türkmen, Mayo and Bivens85]. The industry is pursuing the issues of UAM aircraft safety [Reference Woodham, Graydon, Borer, Papathakis, Stoia and Balan93–Reference Bendarkar, Behere, Briceno and Mavris95], particularly since EASA’s issue of special conditions for VTOL certification and the proposed means of compliance [96–97]. EASA SC-VTOL-01 establishes certification criteria for vertical takeoff and landing vehicles with unique propulsion and control architectures, including distributing the flight controls and electric propulsion elements.

In the initial investigation sponsored by NASA [Reference Darmstadt, Catanese, Beiderman, Dones, Chen, Mistry, Babie, Beckman and Preator32, Reference Beiderman, Darmstadt, Dillard and Silva68], the failure modes and hazards associated with four NASA concept vehicles (1-passenger quadcopter, 6-passenger side-by-side and lift+cruise and 15-passenger tiltwing) were identified and a functional hazard analyses (FHA) and failure modes and effects criticality analyses (FMECA) was performed. A Fault Tree Analysis (FTA) was created for each of the concept vehicles. In addition, conceptual designs of a notional power-train configuration for each of the four concept vehicles were developed to support the reliability and safety analysis for the study. Hazards were identified and the severity of each were categorised in the FHA. Guidelines for reliability targets for both the air vehicle and the operation in the UAM mission were provided. This study provided a methodology for evaluating the safety and reliability of a propulsion architecture.

The recent work sponsored by NASA [Reference Darmstadt, Pathak and Mistry70, Reference Darmstadt, Pathak, Chen, Mistry, Arkebauer, Beiderman, Dillard, Zierten, Beckman, Monroe, Catanese, Greene and Preator84, Reference Justin, Patel, Bouchard, Gladin, Verberne, Li, Ozcan, Rajaram, Mavris, D’Arpino, Aldemir, Rizzoni, Türkmen, Mayo and Bivens85] developed for several of the concept vehicles an all-electric distributed propulsion and flight control (DPFC) architecture that will have no more than 10^–9 catastrophic failures per flight hour, based on the requirements of EASA SC-VTOL-01. The DPFC architecture included electrical power and distribution system, drive and power system, thermal management system and flight control system. Specific questions regarding number of rotors, control schemes and propulsion type were examined:

1. A) impact of number of rotors: compare hexacopter and octocopter (electric, rpm control)

2. B) impact of control schemes: compare rpm control and collective control (hexacopter, electric)

3. C) impact of propulsion architectures: compare electric, hybrid and turboshaft (quadcopter, collective control, with cross shaft)

The assessments were in terms of the safety level achieved, and aircraft components and features needed to meet the required safety level. Transient stability and control models were developed to help inform vehicle handling characteristics and for system requirements definition. To help characterise safety in all operating environments, transient power requirements were investigated in the presence of atmospheric disturbances representative of an urban canyon environment.

Sizing criteria for power plant components are not currently available for multirotor aircraft; therefore, early stability and control simulations are required to characterise power transients to feed design sizing for both normal and emergency flight conditions. Such simulations show: (a) removing interconnecting shafts results in higher power transients at each motor; (b) rpm control results in higher power transients at each motor, compared to collective control; (c) increasing the number of rotors has a small increase in power transients, but changes to power transients are not as severe as the change to no interconnecting shafts or to rpm control. Table 4 summarises the results for power transients.

Table 4. Required motor power transient capability, P/P_hover

DPFC architecture excursions show realistic means to UAM aircraft safety using hybrid-electric or turboshaft power sources: the required safety levels can be achieved using state-of-the-art technology. The turboshaft propulsion architecture is the most reliable of the three considered. Battery networks located at each motor system were required to have at most 10^–10 failures per flight hour to comply with SC-VTOL-01, which was accomplished by pairing redundant battery packs at each motor, with an impact on aircraft weight. The aircraft with collective control and rpm control exhibit similar reliability levels. Increasing the number of rotors (from 6 to 8) does not inherently increase reliability, due to common cause failures, particularly in the battery system.

All aircraft evaluated likely have paths to comply with the stringent, probabilistic catastrophic failure criteria of EASA SC-VTOL-01. However, the stability and control models showed large power transients that must be addressed. Preliminary system safety assessment results show that future work is needed in single load path structures, high voltage power storage and distribution, and in motor/rotor overspeed protection.

EASA SC-VTOL-01 extends fail-safe and redundancy management design techniques to leverage the potential to segregate failures through unique, multirotor configurations, with stringent design requirements associated with redundancy management and single or multiple load path designs. SC-VTOL-01 requires that a single failure must not have a catastrophic effect upon the aircraft. Certification of conventional rotorcraft configurations imposes similar single failure criteria, except that some exceptions are allowed if the probability of failure is extremely remote. SC-VTOL-01 does not permit catastrophic effects of single failures, regardless of the probability of failure; thus the safety assessment process must demonstrate that all single failures are not catastrophic. The single failure criterion has a potentially large impact on all aircraft intended to be type certificated under SC-VTOL-01. Single-load path structures are of prime concern, being mechanical systems in which redundant load path structures can be difficult to integrate. For example, blade loss, rotor loss or excessive vibrations leading to structural failure are reasonable and conceivable potentially catastrophic failure modes that must be addressed.

6.10 Cost

Purchase cost of aircraft is roughly (to about 20% accuracy) driven by aircraft empty weight, installed power, and complexity, plus the costs of electronic systems. For electric propulsion, the cost of batteries should be explicitly included in the purchase cost estimate. Data are available for maintenance cost of helicopters flying traditional missions, but not for unconventional aircraft types engaged in air taxi operations. A significant component of operating costs is the cost of fuel or energy. If the mission range is small enough so that electric propulsion is feasible, energy costs are generally smaller for the all-electric propulsion configuration, even though the aircraft weight is larger.

A method to estimate the purchase price of rotary wing aircraft was developed by Harris and Scully [Reference Harris and Scully98], revised and extended by Scott [Reference Scott99–Reference Scott100]. The method gives the price within 20%, as ${\rm{Price}} = K\left( {{\rm{SF}}} \right)$ , with the size factor

(6) \begin{align}{\rm{SF}} = {\rm{W}}{{\rm{E}}^{0.4638}}{P^{0.5945}}\end{align}

Here ${\rm{WE}}$ is the weight empty and $P$ is the installed power; the factor $K$ depends on the aircraft type, and some measures of complexity. The equation ${\rm{Price}} = K\left( {{\rm{SF}}} \right)$ is applicable to helicopters, tiltrotors, general aviation and airlines [Reference Justin, Patel, Bouchard, Gladin, Verberne, Li, Ozcan, Rajaram, Mavris, D’Arpino, Aldemir, Rizzoni, Türkmen, Mayo and Bivens85], and to commercial turboprops and unmanned aircraft [Reference Johnson86]. The size factor ${\rm{SF}}$ appears as a universal scaling parameter for aircraft cost (Fig. 18). The NASA concept vehicles have ${\rm{SF}} = $ 1,500 to 6,000, which would imply a unit flyaway cost in the range $ {\$} $ 3M to $ {\$} $ 10M following the data in Fig. 18. However, the UAM industry is projecting prices an order of magnitude less than existing rotorcraft and turboprops when high-rate manufacturing is employed and larger numbers of UAM vehicles are produced. There is no data yet to evaluate the factor $K$ appropriate for UAM aircraft, but it is reasonable to use this size factor to estimate the relative costs of the concept vehicles.

Figure 18. Unit flyaway price for several classes of rotary wing and propeller aircraft [Reference Scott99].

6.11 Noise and annoyance

The UAM concept vehicles have been designed with low hover tip speed, in anticipation of a significant requirement for noise reduction in the urban environment. Rotor-rotor interactions, such as rear rotors operating in the wake of front rotors, and wake interactions on retreating sides of overlapped side-by-side rotors, will increase blade-vortex interaction noise. Blade shape and spacing can be optimised for low blade-vortex-interaction and high-speed-impulsive noise. Noise metrics and requirements are established by regulation for rotorcraft, but suitability and applicability of these to air taxi operations must be established.

To compare the impact of design variables on the aircraft noise, the noise certification requirements as specified in FAR Part 36, Appendix H, “Noise Requirements for Helicopters,” [101] are used. The aircraft is flown over three microphones, at the centerline and 150 m to starboard and port. The operating conditions are sea level pressure, and a temperature of 25^oC. Three trajectories are flown: takeoff, flyover and approach. The takeoff profile starts 500 m from the center microphone, 20 m above ground level, and is flown at best rate of climb speed, V_y, and 100% maximum rated power (MRP). The flyover profile is 150 m above ground level, for these aircraft flown at 0.9V_H, where V_H is the speed at 100% maximum continuous power (MCP). The approach profile has a 6-degree descent angle, flown at V_y, 120 m above ground level at the center microphone. Certification requirements specify the maximum EPNL for these trajectories. NDARC was used to calculate V_y, V_H, and the takeoff climb angle. CAMRAD II calculated the section loading at each condition, and AARON/ANOPP2 calculated the EPNL levels. For each of the reference vehicles, these noise metrics were calculated for a range of design tip speeds.

The aircraft designed to the UAM mission all have relatively high takeoff performance, due to the 900 ft/min climb rate requirement. This results in a higher installed power, so that for the takeoff noise certification condition, the aircraft is both further from the observer than for existing helicopters and the rotor is operating with a quickly convecting wake. These two effects combine to reduce the noise for the QSMR, quadrotor, and side-by-side by at least 10 EPNdB compared to helicopters of similar weight, as seen in Fig. 19 (takeoff). The colours and markers are consistent in the figures, with QSMR a yellow circle, quadrotor a blue square, lift+cruise a red triangle and side-by-side a purple diamond. The existing helicopter data are shown as grey circles, with grey lines for Stage 2 and Stage 3 limits. The lift+cruise in compound mode trim is the loudest and heaviest of the vehicles, due in large part to the high number of rotors, rotor-rotor interactions and operating at a higher disk loading due to the rotor radius being constrained. Flyover results with varying rotor tip speed are shown in Fig. 19 (flyover). For flyover, the QSMR, quadrotor and side-by-side are also in general quieter than existing helicopters, while the lift+cruise is once again similar to existing helicopters. Approach results for varying rotor tip speed are shown in Fig. 19 (approach). Approach noise is close to existing helicopters if the tip speeds are at the higher end of the range of speeds examined, but noise quickly reduces as tip speed goes to the lower end of the range. The lowest tip speed, 375 ft/s for the quadrotor, does not improve the approach noise or the flyover noise compared to the 400 ft/s case.

Figure 19. Noise and max gross weight for the four types of aircraft with varying tip speed compared to existing helicopters (QSMR yellow circle, quadrotor blue square, lift+cruise red triangle, side-by-side purple diamond).

Vehicle flyaway cost generally increases with rotor tip speed reduction as shown in Fig. 20 (relative costs are shown in this figure). The increase in cost with reducing tip speed is not always accompanied by a reduction in noise, and this trade is likely to be of importance in UAM conceptual design.

Figure 20. Approach noise and relative flyaway cost for the four types of aircraft with varying tip speed (QSMR yellow circle, quadrotor blue square, lift+cruise red triangle, side-by-side purple diamond).