Original Article

Proc IMechE Part G:

J Aerospace Engineering
An integrated approach to vehicle and 0(0) 1–19
! IMechE 2015

subsystem sizing and analysis for novel Reprints and permissions:

sagepub.co.uk/journalsPermissions.nav

subsystem architectures DOI: 10.1177/0954410015594399

uk.sagepub.com/jaero

Imon Chakraborty1, Dimitri N Mavris1, Mathias Emeneth2 and

Alexander Schneegans2

Abstract
The aerospace industry is currently transitioning to More Electric subsystem architectures due to steadily improving
electric technologies and the technology saturation of established conventional architectures. For aircraft with such
unconventional architectures, the lack of historical information and the presence of increased inter-subsystem inter-
actions create a significant design challenge. These necessitate a greater focus on subsystems design earlier in the design
process than typically seen for aircraft with conventional subsystem architectures. At the same time, however, to be
suitable for the early design phases, the subsystem analyses must be computationally inexpensive and not require detailed
aircraft definition. This work presents an integrated, modular, and tool-independent approach to the sizing and per-
formance analysis of the aircraft and its subsystems, in which inter-dependencies are established between relevant
aircraft and subsystem level parameters. The approach allows the assessment of subsystem architectures using vehicle
and mission level metrics for a fixed vehicle design, and also the amplified effect of re-sizing the vehicle in accordance with
a pre-defined rule-set. The proposed approach was demonstrated through a comparative assessment of a predominantly
electric subsystem architecture and a conventional one for a representative single-aisle aircraft. In this assessment, the
impacts of changing the form of secondary power extraction, the change in vehicle empty weight, and re-sizing of the
vehicle were successively identified.

Keywords
All electric aircraft, more electric aircraft, more electric initiative, subsystems, bleedless, sizing

Date received: 1 August 2014; accepted: 11 June 2015

loads (galley loads, in-flight entertainment), etc. The

Introduction and motivation effect of the subsystems is through:
The primary goal of aircraft and propulsion system
sizing in early design is to find a converged triplet of 1. Weight – Subsystem weight is part of aircraft fixed
aircraft takeoff gross weight (WTO), required sea-level equipment weight and affects the empty weight
static thrust (TSL), and required wing planform area fraction WE/WTO.
(Sw) that allows the aircraft to satisfy stated perform- 2. Power consumption – Extracting power to operate
ance requirements. These typically include mission the subsystems increases engine specific fuel con-
performance requirements (e.g. payload, range) and sumption (SFC).
point performance requirements (e.g. takeoff field
length, landing approach speed, etc. which constrain
the design in terms of feasible thrust-to-weight ratio
TSL/WTO and wing loading WTO/Sw). 1
Aerospace Systems Design Laboratory, School of Aerospace
The sizing process is influenced by the characteris- Engineering, Georgia Institute of Technology, Atlanta, GA, USA
2
tics of the aircraft subsystems, also known as aircraft PACE America Inc., Seattle, WA, USA
equipment systems (AES).1 They include the environ-
Corresponding author:
mental control system (ECS), the ice protection sys-
Imon Chakraborty, Aerospace Systems Design Laboratory, School of
tems (IPS), the actuation systems for flight controls, Aerospace Engineering, Georgia Institute of Technology, Atlanta, GA
landing gear, thrust reversers, brakes, and nose-wheel 30332, USA.
steering, in addition to the avionics systems, cabin Email: [email protected]

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

2 Proc IMechE Part G: J Aerospace Engineering 0(0)

Figure 1. Conventional subsystem architecture (CSA) versus electric subsystem architecture (ESA). The latter is a ‘‘bleedless’’
architecture (no bleed air is extracted from the engine) with all hydraulic systems removed.

3. Drag – Subsystems requiring incorporation of ram some significant and unique challenges for designers.
air inlets or external modifications to the aircraft For conventional subsystem architectures, the
increase drag. designer had access to a large historical database,
statistical regression equations for subsystem
The design of commercial aircraft subsystems has weights,12–14 and other heuristic rules that aided in
largely gravitated towards towards a conventional accounting for subsystem effects during early design.
architecture,2,3 in which the nonpropulsive (second- This, and the relatively mild or nonexistent inter-
ary) power for the subsystems is distributed in four actions among subsystems allowed for relatively iso-
forms: mechanical, hydraulic, pneumatic, and electric lated/independent design of the subsystems further
(Figure 1). Pneumatic power in the form of pressur- downstream in the design. No equivalent historical
ized air bled from the engine compressor stages is used database for estimating subsystem characteristics is
for ECS and IPS operation. Mechanical shaft-power available to the designers of MEA, for which a need
is used to drive engine fuel and oil pumps, and to to address the subsystems collectively at the aircraft
generate hydraulic power (using hydraulic pumps) level has been identified.1 Further, the Boeing 787 and
and electric power (using electric generators). The Airbus A380 suggest that the transition to an All
hydraulic power is used for actuation functions, Electric subsystem architecture (Figure 1) will likely
while electric power is used for avionics, lighting, involve intermediate More Electric architectures that
and cabin and galley loads. feature a combination of conventional and electric
This proven conventional secondary (nonpropul- subsystems. This is largely driven by the need to min-
sive) power architecture has nevertheless reached a imize technological uncertainty and risk. This creates
point of technology saturation that makes further a combinatorial architectural design space from
improvements in efficiency progressively more difficult within which the designer may be interested in com-
to attain. This and the rapid improvements in power paring and contrasting subsystem architectures of
electronics and electric drives2,4 have led to interest in interest using vehicle and mission level metrics.
electric subsystem solutions (the so-called more elec-
tric initiative). This has thus far resulted in more elec-
tric aircraft (MEA) such as the Boeing 787 (electrified
Proposed approach
ECS, wing IPS, engine-starting, wheel-braking5) and The investigation of the vehicle and mission level
the Airbus A380 (use of electric actuators for some impact of subsystem architectures may be facilitated
control surfaces, electric thrust reverser actuation6), in through an integrated approach to the sizing and ana-
addition to several research programs and technology lysis of the aircraft and its subsystems which acknow-
demonstrations.7–11 Potential benefits of MEA ledges the following two characteristics of the early
include reduction in weight at the aircraft level and design phases: (i) only limited information is available
an increase in overall efficiency and reconfigurability.3 regarding the design, and (ii) the underlying analyses
Despite the presence of two in-service commercial must be computationally inexpensive to facilitate
MEA, the early design phase of such aircraft poses rapid tradeoffs among multiple candidate solutions.

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

Chakraborty et al. 3

Figure 2. Illustration of proposed integrated and modular approach for sizing and analysis of the aircraft, propulsion system, and
subsystems.

As shown in Figure 2, the integrated approach pro- models (meta-models) of component characteristics.
posed by the authors interfaces the sizing and analysis The first two options were used in the current work.
of subsystem architectures with the aircraft and pro- In each case, the component is sized using the chosen
pulsion system sizing. The latter is first used to estab- component model and the previously derived func-
lish a baseline aircraft configuration (assuming a tional requirements.
conventional subsystem architecture) based on mis- The subsystem architecture is synthesized by link-
sion and point performance requirements. The use ing the sized components with appropriate generators
of such a conventional baseline to serve as a yardstick of secondary power using distribution elements
for assessing alternative architectures has been well- (piping in a hydraulic network, wiring/cabling in an
established through notable prior studies.15,16 electric network, ducting in a pneumatic network).
Using the baseline aircraft mass properties, geo- This requires that the distribution elements be sized
metric definition, and mission profile, the functional as well as power generation and transformation
requirements of each subsystem are then computed. devices. The conclusion of the sizing process yields
Such functional requirements are derived without Wsub, the change in overall subsystems weight
considering specific subsystem solutions and involve relative to a conventional baseline. Evaluation of
only the required causal effect of the subsystems (e.g. subsystems operation over a mission yields time his-
the required actuation load, rate, and range for actu- tories of the nonpropulsive power requirement
ation functions, the pressurization and thermal Psec ðtÞ ¼ ½Pspx ðtÞ, Pbx ðtÞ (subscripts ‘‘spx’’:
regulation required for ECS operation, the required shaft-power extraction, ‘‘bx’’: bleed extraction) and
thermal heat flux to provide IPS function). direct drag increments CD(t). Re-evaluation of the
Subsequent steps explicitly consider the identified sub- mission performance accounting for Wsub , Psec ðtÞ,
system architecture(s) of interest, as the characteristics and CD ðtÞ yields the changes in empty weight WE,
of architecture-specific components must be deter- fuel required WF, and takeoff weight WTO driven
mined based on the derived functional requirements. by subsystem architecture.
The models used for the subsystem components must If analysis of an unconventional subsystem archi-
be determined by a tradeoff between fidelity and com- tecture yields an unfavorable impact on performance
putational expense, and depending on the case at (e.g. WF > 0 and/or WTO 4 0), then the process
hand may include the direct incorporation of math- may be terminated (‘‘Cycle? ¼ No’’ in Figure 2). On
ematical descriptions of the component, the use of the other hand, a favorable impact may be exploited
multi-dimensional data tables, or the use of surrogate in one of two ways: (i) the payload-range capabilities

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

4 Proc IMechE Part G: J Aerospace Engineering 0(0)

Figure 3. Example of dependencies between system and subsystem parameters. A re-sizing rule of the form Sw ¼ f ðWTO Þ allows
actuator weight (for instance) to be influenced by weight changes of other unrelated subsystems through aircraft level parameter WTO.
and thus the utility of the vehicle may be increased, or shown in Figure 1 for completeness, was not con-
(ii) the vehicle may be further down-sized while retain- sidered in this work).
ing the same performance capabilities. The remainder of this paper is organized as fol-
This work assumes that the aircraft is being sized to lows: forthcoming sections describe the analysis
a given (fixed) set of mission and point performance approach for the aircraft subsystems considered.
requirements, and thus follows the second option. By Next the sizing approach for the power generation,
following the ‘‘Cycle? ¼ Yes’’ path in Figure 2, the transformation, and distribution elements is described
aircraft may now be re-sized in accordance with a pre- and a comparative assessment of the two subsystem
defined set of re-sizing rules, one of which may state architectures considered is done. Finally, the last sec-
for instance that the same margin from active con- tion concludes and provides a brief description of
straints in the TSL/WTO versus WTO/Sw design space avenues for future work.
be maintained. Since the re-sizing of the aircraft also
affects the subsystem sizing, an iterative process is
Actuation subsystems
established in Figure 2, which terminates only when
the aircraft and subsystem level parameters simultan- For each actuation subsystem, the actuation require-
eously converge. A salient feature of this approach is ments were first determined in terms of peak actuation
that it allows the ‘‘snowballing’’ effects of subsystem load, maximum actuation rate, and range of move-
architecture changes to be effectively captured, in part ment, following which actuators were sized to meet
through the establishment of inter-subsystem and these requirements. The conventional and electric
system-subsystem dependencies of the form shown actuators were required to be equivalent in terms of
in Figure 3. their peak load, rate, and stroke capabilities, rather
It is well-known that within the aerospace industry, than in terms of their peak power (which is not a valid
airframers and subsystem solution providers each measure for comparison between conventional and
have their own preferred methods, tools, and models electric actuation18). The specific approach taken for
(which are typically proprietary) to aid in the execu- each subsystem is explained in the following sections.
tion of their work packages. Another salient feature
of the proposed approach is that it is not specific to
Actuation of flight control surfaces
any particular set of tools, and may be implemented
with any choice of tools that collectively provide the For determining the control surface actuation loads,
following capabilities: (i) aircraft sizing and mission control surfaces such as ailerons, elevators, and
performance analysis, (ii) propulsion cycle analysis, rudder were defined parametrically with respect to
(iii) subsystem sizing and analysis, and (iv) analysis their corresponding main lifting surfaces using well-
integration framework (for sequencing execution of established early-design sizing guidelines12 for control
functions (i)–(iii), passing relevant information surface span and chord ratios. The layout of control
between modules, and checking for analysis conver- surfaces is shown in Figure 4(a), while the layout of
gence). Further, the same tool may be used to satisfy electric actuators (for the ESA design) is shown in
more than one of the above functions. For example, Figure 4(b). A similar approach was taken for second-
this work uses Pacelab SysArc to perform functions ary flight control surfaces such as flight and ground
(iii) and (iv), Pacelab APD (the underlying framework spoilers. The planform areas of horizontal and verti-
of Pacelab SysArc) for function (i), and the numerical cal stabilizers were linked to the wing planform area
propulsion system simulation (NPSS) tool17 for func- using representative horizontal and vertical tail
tion (ii). volume coefficients applicable for this category of air-
The proposed approach is demonstrated by con- craft.12 A similar parametric representation of lead-
sidering a single-aisle narrow-body commercial ing-edge and trailing-edge high-lift devices was
aircraft whose major characteristics are shown in implemented based on a review of high-lift devices
Table 1. The two subsystem architectures illustrated on existing commercial aircraft.19
in Figure 1 are considered: a conventional subsystem Based on the control surface dimensions and type,
architecture (CSA) and an electric subsystem architec- the actuation loads were determined either by con-
ture (ESA) in which ECS, IPS, and actuation func- sidering hinge moment coefficients20,21 and relevant
tions are electrified (thrust reverser actuation, though Federal Aviation Regulations (FARs),22 by scaling

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

Chakraborty et al. 5

Figure 4. Flight control surface actuation – Flight control surface definition and association of electric actuators with control
surfaces.

The dimensions a and b were linked to landing

gear leg length Llg and early design guidelines13 were
Table 1. Description of baseline single-aisle narrow-body used to set the CG position ‘cg . The mechanism
aircraft (CSA design). kinematics were used to determine the actuator’s
Parameter Symbol Value
stall load F0, stroke s, and maximum ram speed vmax
as follows
Wing reference area Sw 124.8 m2 (1343 ft2)
Wing span bw 34.3 m (112.6 ft)
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

cðÞ ¼ a2 þ b2  2ab cosð þ Þ,
Wing sweep (25% chord) c/4 25.02 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Wing aspect ratio Aw 9.2 Wlg ‘cg cos  a2 þ b2  2ab cosð þ Þ
FðÞ ¼
Wing taper ratio w 0.21 m a b sinð þ Þ
Wing loading WTO/Sw 622.5 kg/m2 (127.5 lb/ft2) F0 ¼ max FðÞ, s ¼ cðdn Þ  cðup Þ,
Sea-level static thrust TSL 242.9 kN (54,600 lbf) 2½up , dn 

Thrust-to-weight ratio TSL/WTO 0.3189 s

vmax ¼ , ð1Þ
Operating empty weight OEW 41,871 kg (92,310 lb) t ð1  Þ
Maximum takeoff weight MTOW 77,655 kg (171,200 lb)
Payload WPL 14,987 kg (33,040 lb) where up and dn are respectively the ‘‘up’’ and
‘‘down’’ leg angles, and t is the retraction time.
force and moment coefficients from published wind- The mechanism was assumed to accelerate/decelerate
tunnel studies23,24 to the appropriate control surface over equal time intervals   t,  2 ð0, 1Þ and move at
dimensions, or based on the specifications of existing the maximum ram speed vmax for the remaining time
actuators. The first approach was used for the ail- t ð1  2Þ. While variations in the weight of the
erons, elevators, rudder, and spoilers, while the brake and steering mechanisms do have an effect on
latter two were used for the high-lift devices, yielding the weight and CG position of the landing gear leg, a
the actuation loads summarized in Table 8 in sensitivity study showed these dependencies to be
Appendix. For further details regarding the assess- small for the aircraft size considered, and thus
ment of control surface actuation requirements, the they were not included. Assumptions regarding the
reader is referred to Chakraborty et al.25 landing gear mechanism are tabulated in Table 9 in
Appendix.
Retraction/extension of landing gear
Since during extension/retraction of the landing
Application of wheel brakes
gear, the gravitational load dominates over the aero- Brake actuation requirements were determined from
dynamic, frictional, and inertial loads,26 the weight the constraining braking cases shown in Table 2. In
of the main and nose landing gears was set as a addition to static cases, these included dynamic cases
function of the aircraft’s MTOW.27 The required such as a rejected takeoff at MTOW (accelerate-stop
actuator force F was computed as a function of the case) and landing deceleration at maximum landing
gear leg’s angular position  using an assumed linkage weight.28 The dynamic cases are described in more
efficiency m and the kinematics shown in Figure 5. detail in FAR §25.735 (f)(1) and (f)(2).

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

6 Proc IMechE Part G: J Aerospace Engineering 0(0)

For each case, the braking force Fb for each of nbrk 2 b

Fa ¼ ð4Þ
¼ 4 braked tires per main gear was found from a force  nf ðro þ ri Þ
balance
where ro and ri are the outer and inner radii of
0 0 W dV the friction material (Figure 6(a)),  the coefficient
Ftot
b ¼T% D  F%f þ W sin   ,
g dt of friction, and nf the number of friction pairs
Ftot formed between rotor and stator elements (in modern
Fb ¼ b ð2Þ aircraft, four to five rotors is common). The speed and
nbrk
stroke were estimated based on the requirement that
where the aerodynamic drag D and rolling friction the running clearance (&10 mm) be closed upon
Ff were conservatively neglected. The required brake application in a specified time (&50 ms).28
braking torque was then computed from a moment bal-
ance, neglecting the effect of wheel inertia I wheel
Steering of nose gear
d For determining nose-wheel steering system (NWS)
I wheel !wheel ¼ Fb Rtire  b ) b  Fb Rtire ð3Þ
dt actuation requirements, the aircraft was assumed to
be at its maximum ramp weight (MRW). The max-
Assuming a case of uniform brake wear, the brak- imum nose gear normal reaction Fnlgz was computed
ing torque is related to the required axial force Fa by assuming a dynamic load factor ¼ 1.33 (FAR

Figure 5. Landing gear retraction kinematics (A: actuator-strut attachment point, B: actuator-airframe attachment point, C: landing
gear pivot point, G: landing gear leg center of gravity).

Table 2. Static and dynamic cases for establishing braking requirements.

Case description Full-throttle run-up Parking on incline Accelerate-stop Landing decel.

Case type Static (S1) Static (S2) Dynamic (D1) Dynamic (D2)

Aircraft weight MTOW MRW MTOW MLW

Thrust setting Max. SLS – Idle Idle
Ground speed – – 72 m/s (140 kts) 75 m/s (145 kts)
Deceleration – – 1:8 m=s2 ð6:0 ft=s2 Þ 3:0 m=s2 ð10 ft=s2 Þ
Opposing gradient 1.1 1.1 1.1 1.1
MTOW: maximum takeoff weight, MRW: maximum ramp weight, MLW: maximum landing weight, SLS: sea-level static.

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

Chakraborty et al. 7

Figure 6. (a) Translating braking torque to required axial force for braked main wheels, (b) translating ground lateral force to
steering moment for steered nose-wheel.

§25.499). The critical horizontal load at the tire-to- displacement pump, which actuates a hydraulic cylin-
ground contact point was set to 0.8 times the der, and (ii) electromechanical actuators (EMAs), in
normal reaction (Fnlg nlg
y =Fz ¼ 0:8). As seen from which a variable-speed bi-directional motor drives a
Figure 6(b), the moment due to this lateral force reduction gearbox which in turn drives a ballscrew.30
depends on the effective moment arm, which is related These two actuator types were sized following the pro-
to the NWS geometry as Leff ¼ xtire cos r , using cess described in Chakraborty et al.,25 in which actu-
which the sizing moment evaluates to ator weight was estimated using a component weight
! build-up method. The predicted weights were validated
nws nlg
Fnlg
y against published actuator weights (with matched spe-
Mmax ¼ Leff Fy,max ¼ Leff nlg
Fnlg
z,max
Fz cifications), and the agreement was found to be reason-
! able with a small calibration factor applied. As shown
Fnlg
y
¼ ðxtire cos r Þ ð s MRWÞ ð5Þ in Figure 4(b), the control surfaces in the ESA were
Fnlg
z assumed to be EHA-driven, with the exception of the
trailing-edge flaps, which were EMA-driven. The land-
The geometric parameters in equation (5) were set ing gear, brake, and nose-wheel steering mechanisms
based on preliminary guidelines for the class of aircraft were considered to be EMA-driven.
considered12 (Table 11 in Appendix). The required
peak steering moment predicted using this approach
showed good agreement with that published from the
Environmental control system
ELGEAR project29 for the Airbus A320 aircraft. The The passenger cabin of a modern commercial aircraft
available moment was required to be independent of comprises multiple zones, with the provision of a sep-
the nose-wheel angle, and allowed to reduce linearly to arate target temperature setting for each. For this
zero above 50% of the maximum angular rate. study, all target temperatures were set to 22 C. The
cabin pressure for a given altitude was determined in
accordance with a pressurization schedule.31,32 Within
Sizing of actuators zone ‘‘i’’, the minimum volume flow rate V_ i
min
was
In the preceding sections, the actuation requirements determined based on the air cycle time (the time
were estimated in terms of required load, rate, and tcyc required to have a complete exchange of the
range of movement (either linear or angular). air volume V zone,i within the zone) and a minimum
These actuation requirements are the starting volume flow rate requirement of V_ per pax ¼ 20 ft3/
point for the sizing of both hydraulic and electric actu- min per occupant31
ators. However, since this work considers a complete  
transition to electric actuation, it was not necessary to _V min ¼ max V zone,i ,
i V_ per pax npax, i ð6Þ
develop sizing codes for the hydraulic actuators and tcyc
power control units (PCUs) of the CSA, since the air-
craft sizing tool (like most other such tools) already The minimum zone mass flow rate m_ min i was com-
min
provided an estimate of hydraulic surface controls puted using V_ i , the cabin pressure, and desired zone
weight based on regression equations. temperature Ti. Applying the first law of thermo-
On the other hand, since very limited historical dynamics to the zone control volume, the required
data exists for electrical actuators used in the ESA, air inlet temperature was computed as
their sizing was considered explicitly. Two types of
electric actuators were considered: (i) electrohydro- Q_ loss,i  Q_ int,i
static actuators (EHAs), in which a variable-speed Tin,i ¼ Ti þ ð7Þ
m_ min
i Cp,air
bi-directional electric motor directly drives a fixed-

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

8 Proc IMechE Part G: J Aerospace Engineering 0(0)

In equation (7), Q_ int,i 4 0 included thermal loads respectively ram drag and weight penalties, and
due to metabolic heat of occupants (&73–100 W per removing bleed extraction penalty but imposing
occupant) and also heat generated by galley loads, in- shaft-power extraction penalty instead). This work
flight entertainment (IFE) systems, etc. (Table 6). The utilized a well-established thermodynamic model of
heat exchange across the thermal resistance offered by the ECS pack33 but did not consider the effect of air
the fuselage wall was computed as humidity for sake of simplicity. Details regarding
   some relevant pack parameters are provided in
1
1 2 Table 12 in the Appendix.
Q_ loss,i ¼ Ti  T1 1 þ Rc M ð8Þ
Rtot 2

where Rtot is the fuselage thermal resistance and Rc is

Ice protection system
the temperature recovery factor (both explained in the The first requirement for sizing and evaluating the IPS
Appendix). If the air inlet temperature computed operation is to define the extent of protection (EOP),34
from equation (7) fell outside a permissible range, which for the aircraft considered included all leading-
i.e. Tin,i 2 = ½Tmin max
in , Tin , equation (7) was solved with edge slats except the most outboard ones and the inlets
Tin,i set to the applicable temperature limit to find of the two engine nacelles. These formed the wing ice
the required mass flow rate m_ i 4 m_ min i . Otherwise, protection system (WIPS) and the cowl ice protection
for Tin,i 2 ½Tmin in , T max
in , _
m i ¼ _
m min
i was set. Since the system (CIPS), respectively. Analysis was limited to
ECS mixing manifold (which combines fresh pack two types of thermal ice protection (other papers
air with a portion of recirculated cabin air) can only describe mechanical/impulse-based or hybrid sys-
supply air at a single temperature Ts, this is set to the tems35–37 and associated preliminary power esti-
minimum of the required zone air inlet temperatures: mates38). For each, it was assumed that the rate of
Ts ¼ mini Ti , i ¼ 1, . . . , n. The required air inlet tem- thermal energy supply could be modulated based on
perature for zones with Tin,i > Ts is achieved by the flight condition (Note: This is not the case for most
mixing hot ‘‘trim’’ air bled from upstream of the conventional pneumatic systems34). For both, the
ECS packs to the supply. Assuming adiabatic WIPS was assumed to be of the running-wet type,37,39
mixing, the trim air requirement for each cabin zone where sufficient heat is supplied to maintain the surface
was computed as skin temperature Tskin above freezing (Tskin 2 4–10 C),
while the CIPS was of the evaporative type, where
Q_ i  m_ i Cp,air ðTs  Ti Þ
m_ trim,i ¼ ð9Þ higher heat flux is supplied (Tskin 2 37–50 C) to com-
Cp,air ðTtrim  Ti Þ
pletely evaporate impinging droplets.15
Analyses using computational fluid dynamics
Under the assumption that the ECS comprised np packs (CFD) and finite element methods40–43 were deemed
and employed a recirculation fraction of 2 ð0, 1Þ, the inapplicable due to computational expense and the
mass flow requirement
P for each pack was obtained as lack of detailed geometric definition in early design.
m_ pack ¼ ð1  Þ i m_ i =np . Considering the aircraft type Instead, the heat flux required to maintain a protected
analyzed, two ECS packs (np ¼ 2) and 50% recirculation surface at a specified temperature Tskin against an
( ¼ 0.5) were used. ambient temperature T1 was computed using a lim-
To assess the effect of operating the ECS with elec- ited set of geometric parameters. This heat flux q_tot is
tric rather than pneumatic power without the con- a function of flight condition and is the combined
founding effect of varying pack designs, the ECS in result of convective, sensible, evaporative, and kinetic
both the CSA and the ESA was assumed to comprise heating39 (see Appendix for details regarding evalu-
conventional bootstrap units. While customized packs ation of each component).
containing motorized-compressor-turbine (MCT)
units may be more optimal for electric ECS, conven- q_tot ¼ q_conv þ q_sens þ q_ evap þ q_ kin ð10Þ
tional bootstrap units are feasible solutions for such
layouts as well.16 In such packs, cooling is provided With the exception of convection, the remaining pro-
by two heat exchangers and expansion through the air cesses depend on the mass of water impinging per unit
cycle machine turbine. Ram air admitted through ram time per unit area. This is called the local water ‘‘catch
air inlets serves as the cooling flow for the heat rate’’ and is defined as
exchangers (imposing ram drag penalty). The conven- v 0:613
1
tional and electric ECS packs differ in the source of m_ local ¼ v1 LWC Em , Em ¼ 0:00324
the pack air. For the conventional system, this is t
extracted (directly) from the engine compressor ð11Þ
stages (which imposes a fuel burn penalty). For the In equation (11), the liquid water content (LWC) of
electric ECS, the pack air is admitted from the ambi- air LWC was obtained from FAR §25, Appendix C22
ent through dedicated inlets and compressed using as a function of air temperature and with an assumed
electrically powered cabin air compressors prior to mean droplet size of 20 mm. In the expression for the
being introduced into the packs (imposing water catch efficiency factor Em, t is the maximum

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

Chakraborty et al. 9

thickness presented by the protected surface perpen- cyc ¼ theat =tcyc . The heat melts a small thickness of
dicular to the freestream. This expression was derived ice ‘ice , i.e. a mass per unit area of mice ¼ ice ‘ice ,
by Meier and Scholz44 and stated to have acceptable allowing ice above the melted zone to be dispatched
error between sea level and 6000 m (20,000 ft). by freestream pressure.
The cyclic heat flux q_ cyc , overall effective heat flux
q_eff , and electrical power requirement Preq may be
Pneumatic ice protection
found as
In conventional pneumatic IPS designs, hot bleed air
extracted from the engine is ducted to the surfaces to mice

q_ cyc ¼ ci ðTref  T1 Þ þ Lf , ) q_ eff ¼ q_ps ps

be protected and discharged through ‘‘piccolo’’ tubes theat
to transfer heat to the protected surfaces, as shown in Aprot q_ eff
þ q_ cyc ð1  ps Þ cyc , ) Preq ¼ ,
Figure 7(a). The heat transfer was modeled as the heat
following processes occurring in series: (i) heat infu- ð13Þ
sion into the heated bay by the incoming stream of
mass flow rate m_ bleed at temperature Tsupply, which is where heat is the overall efficiency of the electrother-
exhausted overboard at the mean bay temperature mal ice protection system. The geometric and oper-
T bay , (ii) convection from the heated bay air at mean ational parameters assumed for this paper are
temperature T bay to the inner bay wall at temperature summarized in Table 13 in Appendix.
Twall, and (iii) conduction through the leading-edge
skin of thickness tskin and thermal conductivity kskin. Power generation and distribution
Under steady-state conditions, the heat flux for the
system architecture
above is identical to the total heat flux of q_ tot (equa-
tion (10)). Using this, the required mass flow rate The focus of the previous sections has been the iden-
m_ bleed was computed as follows tification of the characteristics of the power con-
sumers (end-users of secondary power): mass,
q_ tot tskin electrical power requirement, hydraulic flow require-
Twall ¼ Tskin þ  Tskin ,
kskin ment, or pneumatic mass flow requirement. The
q_ tot required secondary power is supplied to the subsys-
) T bay ¼ Twall þ , tems from the power sources (in this case the engines
hint
q_tot Aprot and the APU) through the power generation and dis-
) m_ bleed ¼ ð12Þ tribution system comprising
Cp,air ðTsupply  T bay Þ
1. Power generation and transformation devices:
The assumptions of the mean bay temperature T bay These include hydraulic pumps and electric gener-
and that the incoming airstream would be exhausted ators, which generate hydraulic and electric power
at that temperature were found justifiable based on respectively from extracted engine shaft-power.
inspection of the flow-field analysis results of Hua Transformation devices include power converters
and Liu.43 The internal convection coefficient hint within the electrical system that can convert
varies significantly with distance from the piccolo jet between power types (e.g. AC to DC) or voltage
streams and jet arrangement, and a mean value from levels
ranges identified in more detailed analyses45 was used. 2. Power distribution system: These are responsible
for the physical transfer of the appropriate form of
secondary power to the power consumers. They
Electrothermal ice protection include hydraulic piping, electrical wiring/cabling,
In these systems, electric heating pads are bonded to and pneumatic ducting
the surface to be protected and may be energized sim-
ultaneously for anti-icing or sequentially for de-icing.5 In order to size these components, it is first neces-
Krammer and Scholz38 showed that reasonable esti- sary to define the association/connectivity of subsystem
mates for the power requirements of such a system power generation devices, distribution elements, and
may be obtained by considering only the main geo- the power consumers. While such associations are
metric features, as shown in Figure 7(b). well-defined for the CSA, the same is not the case for
Spanwise and chordwise heating strips known as the ESA. In this case, a feasible configuration or top-
‘‘parting strips’’ take up ps 2 ð0, 1Þ fraction of the ology46 may be chosen, whose components may then
total protected area Aprot and are heated continuously be sized accordingly. The layout considered for the
with a heat flux q_ ps ¼ q_tot to prevent ice formation. ESA power generation and transformation system in
The remaining ð1  ps Þ fraction of the protected this paper is shown in Figure 8 and was based on
area comprises cyclically heated surfaces. These are trends observed in existing MEA (e.g. Boeing 787) or
energized (heated) for a duration theat within a cycle concept studies: (i) Variable frequency starter gener-
of duration tcyc, yielding an effective activity ratio ators (VFSGs) to eliminate the weight of the constant

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

10 Proc IMechE Part G: J Aerospace Engineering 0(0)

Figure 7. Pneumatic and electrothermal wing ice protection systems (WIPS).

Figure 8. Electrical power generation and transformation system architecture considered for ESA.

speed drive (CSD) required otherwise for constant fre- requirement for multiple redundancy, which contrib-
quency power generation,3 (ii) Two generators driven utes significantly to architecture weight. In lieu of an
off each engine (and also the APU),5 providing greater explicit fault tree analysis (FTA) or failure modes
availability and also facilitating integration into a and effects analysis (FMEA), this work made the
lower-drag nacelle,15 (iii) Each generator connected assumption that the overall reliability of an electric
to its own variable frequency AC bus to avoid paral- system (e.g. one of the DC busses) would have to be
leling of generators, with tie-switches (shown by at least equivalent to that of a conventional hydrau-
dotted black lines) to prevent loss of a bus due to fail- lic system for a CSA !ESA transition to be con-
ure of a single power source,47 (iv) AC bus voltage of sidered feasible. Subject to this assumption, some
230 V (as opposed to 115 V), as higher voltage systems heuristics derived from inspection of subsystem
facilitate lower feeder/cable weights3 and have been architecture information for Airbus and Boeing air-
shown to be feasible,5 (v) Direct 230 VAC supply to craft obtained from public domain sources (e.g.,
large resistive loads (insensitive to frequency of sup- http://www.smartcockpit.com/aircrafts-models.html)
plied power) such as the IPS and galley loads,32 thus are applicable to the ESA as well. Particular exam-
reducing magnitude of downstream power conversion ples include the number of independent power sys-
(and thus conversion losses), and (vi) 230 VAC power tems associated with a power consumer and the
converted to  270 VDC (volts DC) using auto-trans- distribution of multiple loads of the same type (e.g.
former rectifier units (ATRUs), to supply the ECS and a row of spoilers) among available power systems.
actuation loads.5 Based on such heuristics, the association of power
To size components such as generators and systems to power consumers shown in Table 3 was
ATRUs within this template (Figure 8), it is further developed.
necessary to consider the associations of the power Using these associations between power systems
consumers (loads) with the power systems estab- and power consumers and their geometric locations
lished above. These are largely driven by the (which were defined parametrically with respect to

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

Chakraborty et al. 11

Table 3. Allocation of electric loads (power consumers) to energy systems for ESA (see also Figure 8).
Energy system

230 VAC power  270 VDC power

Load AC-1 AC-2 AC-3 AC-4 DC-1 DC-2 DC-3

Ice protection systems

Engine 1 CIPS X
Engine 2 CIPS X
Wing IPS X X
DC power system
 270 VDC Bus 1 (DC-1) X
 270 VDC Bus 2 (DC-2) X X
 270 VDC Bus 3 (DC-3) X
Environmental control system
Left ECS Pack CACs X X
Right ECS Pack CACs X X
Actuation systems
Left aileron X X
Right aileron X X
Left elevator X X
Right elevator X X
Rudder X X X
Spoilers 1-L, 1-R X
Spoilers 2-L, 2-R X
Spoilers 3-L, 3-R X
Spoilers 4-L, 4-R X
Spoilers 5-L, 5-R X
Spoilers 6-L, 6-R X
Kruegers 1-L, 1-R X
Kruegers 2-L, 2-R X
Slats 1-L, 1-R X
Slats 2-L, 2-R X
Slats 3-L, 3-R X
Slats 4-L, 4-R X
Inboard T/E Flaps X X
Outboard T/E Flaps X X
Landing gear actuation X
Main gear brakes X X
Nose-wheel steering X

(i) Only loads unique to ESA are shown. (ii) For spoilers, Krueger flaps, and slats, numbering is from inboard panels to outboard panels, with ‘-L’ and ‘-R’
signifying left and right wing affiliations (see Figure 4).

the aircraft geometry), the lengths of the distribution as the generator sizing condition. Since the ECS
system elements connecting them were computed. and IPS were essential loads, the critical flight condi-
Hereafter, the sizing of such connecting elements tion yielding the greatest joint ECS and IPS power
involves the determination of weight per unit length. demand was determined by evaluating their operation
For electric wiring, this comes from the wire gage over the altitude/Mach number flight envelope of the
determined based on permissible voltage drop per aircraft
unit length.  
The sizing of the power generation and transform- n ðiÞ
Pmax
ecsþips ¼ max PðiÞ
ecs ðh,M,ISA ðiÞ
ÞþPips ðh,M,ISA ðiÞ
Þ ,
ation devices requires the identification of the con- i¼1
straining load-cases where these components handle i ¼ # of ISA deviation conditions tested
the maximum secondary power. For example, for an ð14Þ
architecture with electric ECS and IPS, Liscouet-
Hanke et al.34 established the one engine inoperative Similarly, the flight-critical control surface peak
(OEI) condition (which would neutralize the gener- power demand was estimated by assuming continued
ation capability of two out of the four generators) operability of the ailerons, elevators, rudder, and four

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

12 Proc IMechE Part G: J Aerospace Engineering 0(0)

Table 4. Mass deletions and additions of ESA design relative to CSA design.

Deletions Additions  Mass (kg)

Hydraulics & pneumatics –540

Conventional surface controls –1451
Electric CS & LG actuators þ 996
Conventional brakes (4) –760
EMA brakes (4) þ 582
Electric nose-wheel steering þ 79
Pneumatic WIPS & CIPS –61
Electrothermal WIPS & CIPS þ 95
ECS cabin air compressors (4) þ 80
(50 kW each, 2.5 kW/kg)
2  120 kVA generators –87
(2.76 kVA/kg)
4  135 kVA generators þ 196
(2.76 kVA/kg)
Additional wiring/cabling þ 84
ATRU (x 4) þ 130
Higher capacity APU (from þ 98
100 kW to 280 kW, 1.83 kW/kg)
–2899 þ 2339

Net fixed equipment  Mass: 560 kg (– 1232 lb)

out of eight flight spoilers in the OEI condition.25 The surfaces48 at low pump r/min. Pump weight may be
sum of the corner powers of these surfaces (product of readily determined as a function of rated pressure and
maximum hinge moment and maximum rate) pro- flow based on commonly available manufacturer
vides a very conservative estimate of the peak power data-sheets. For this specific analysis, the above pro-
requirement in a OEI crosswind landing condition.15 cedures were unnecessary since the entire hydraulic
Since generators are designed with overload capacity system was effectively removed during the CSA
for brief periods, 40% of the corner power sum was !ESA transition considered, and since the aircraft
considered to avoid over-conservatism. As a result of sizing tool used already provided an estimated
the same overload capacity, short-term loads from hydraulic system group weight based on regression
operation of high-lift devices, landing gear, brakes, equations.
etc. do not affect generator sizing.15 The ATRUs
were sized by considering the maximum power Comparative assessment of subsystem
demand of downstream connected loads. The sizing
of APU generators was directly driven by the require-
architectures
ment to support ground power requirements of the For evaluating the effect of novel subsystem architec-
electric ECS in addition to previously existing tures, the change in fixed equipment weight of such
ground loads. Finally, the weights of electric gener- architectures relative to a conventional architecture
ators and ATRUs were determined by considering a must be ascertained. This may be done by accounting
power-to-weight ratio as a figure-of-merit for these for relevant mass additions, deletions, and adjust-
components. ments,15,16 as shown in Table 4.
In case of hydraulic systems, the actuation require- The weight of the hydraulic and pneumatic systems
ments (loads and rates) can be converted into flow and conventional surface controls for the CSA were
rate requirements using the well-known fluid power already available from the aircraft sizing tool (most
formula. The pipe wall thickness is driven by the nom- aircraft sizing codes have such group weight break-
inal system pressure, while the bore (diameter) is downs), and were deleted for the ESA. Table 4 shows
determined by the requirement that this necessary that the (added) weight of electric actuators estimated
flow rate be passed without exceeding a permissible by sizing them for the actuation loads (section ‘‘Sizing
pressure drop per unit length. The sizing case for of actuators’’) was less than that of the hydraulic sur-
hydraulic pumps is typically set by the approach face controls (removed), even though the opposite
phase when high-lift devices and landing gear have trend is generally expected.18 This trend reversal
to be actuated in addition to the flight control was driven by the assumed power-to-weight ratio

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

Chakraborty et al. 13

Figure 9. Flight profile for comparative mission performance assessment of subsystem architectures.

(i.e. state-of-the-art) of the actuators’ electric motors design. The second, representing an engine meant to
and power electronics, and has also been seen in pre- provide only shaft-power off-take, was used for the
vious studies.49 For both CSA and ESA, the brake ESA designs. The engine decks were generated using
heat-stack mass was estimated based on the thermal the NPSS tool,17 and were representative of the
capacity required to dissipate the aircraft kinetic CFM56-7B27 engine, but did not contain or utilize
energy as heat without exceeding a permissible gross any proprietary information.
temperature rise: mhs ¼ KE=ðc Tmax Þ. The brake
weights were estimated with the material properties
and conditions described in Table 10 in Appendix,
Subsystem activity during mission segments
and yielded acceptable agreement with published fig- The effect of subsystem architecture on the mission
ures.50,51 Conventional IPS weight was estimated performance of the CSA and ESA designs was
based on previous studies,15 while the weights of the assessed over the course of the flight profile shown
electrothermal IPS heating pads were estimated based in Figure 9. In the absence of flight test data or
on the material properties and thicknesses reported by closed-loop flight dynamics analysis, normalized
Al-Khalil et al.39 for such an arrangement. The deflection amplitudes þ=    and frequencies f were
weights of main engine and APU generators and used to characterize control surface movements,52 as
ATRUs were determined based on their power ratings shown in Table 5. Given their short duration, the
(section ‘‘Power generation and distribution system operation of high-lift devices, landing gear, wheel
architecture’’) using power-to-weight ratios (kVA/kg brakes, and nose-wheel steering had negligible effects
or kW/kg) that were established through a review of on mission fuel burn.
manufacturer product data-sheets available in the As shown in Table 6, both subsystem architectures
public domain. considered have some common loads, whose magni-
tudes per flight segment were derived from previous
Effects of secondary power extraction on engine AEA studies.4,15 The CSA hydraulic system power
demand was computed based on the control surface
performance movements (Table 5) and a hydraulic load profile dia-
Since the CSA !ESA transition involves a change in gram for a narrow-body aircraft.53 For the ECS, the
both the form and magnitude of secondary power required mass flow, cabin inlet temperature, and pres-
extraction, the choice of architecture must be sure were computed as discussed in ‘‘Environmental
accounted for in determining the engine cycle control’’ section. For the IPS, the required heat flux
(through cycle optimization). In fact, as noted by was evaluated as discussed in ‘‘Ice protection system’’
Jones,2 failure to do so by either considering pure section, and the system was assumed to be deactivated
shaft-power extraction from an engine designed for at cruise altitudes.46
‘‘mixed’’ off-take (both shaft-power and bleed), or
conversely ‘‘mixed’’ extraction from an engine opti- Comparison and discussion of architecture
mized to provide only shaft-power may result in
performance
mis-representation of the effect of secondary power
off-take on the engine performance. To avoid such For making a comparative assessment of the ESA
discrepancies, this approach used two separate design relative to the CSA design, the effects of the
engine performance data tables (commonly called change in the form and magnitude of secondary
engine decks). The first, representing an engine power extraction, change in vehicle OEW, and re-
designed for mixed power extraction, was used to sizing of the vehicle were isolated by considering a
evaluate the mission performance of the CSA three-step transition from one architecture to the

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

14 Proc IMechE Part G: J Aerospace Engineering 0(0)

Table 5. Characterization of aileron, elevator, rudder, and flight spoiler movements. Ground activity represents a flight controls
check. Activity in cruise represents turbulence encounters, which were assumed for 20% of cruise segment.

— Aileron — — Elevator — — Rudder — – Flt. Spoiler –

Surface !
Phase #   
þ = f (Hz)   
þ/ f (Hz)   
þ/ f (Hz)  f (Hz)

Ground þ1/1 0.4 þ1/1 0.4 þ1/1 0.4 1 0.4

Takeoff þ0.12/0.2 0.4 þ0.2/0.2 0.4 þ0.2/0.2 0.4 0.12 0.4
Climb þ0.12/0.2 0.2 þ0.2/0.2 0.2 þ0.2/0.2 0.2 0.12 0.2
Cruise þ0.5/0.5 0.4 þ0.3/0.3 0.4 þ0.3/0.3 0.4 0.5 0.4
Descent þ0.12/0.2 0.3 þ0.2/0.2 0.3 þ0.2/0.2 0.3 0.12 0.3
Landing þ0.4/0.53 0.3 þ0.53/0.53 0.3 þ0.53/0.53 0.3 0.4 0.3

Table 6. Subsystem loads for CSA and ESA designs per flight phase.

CSA and ESA Unit Ground Takeoff Climb Cruise Descent Landing OEI

Instruments kW 8 8 8 8 8 8 8
In-flight entertainment kW 10 10 10 10 10 10 0
Galley loads kW 0 0 40 40 0 0 0
Essential lighting kW 10 10 10 10 10 10 10
Engine-driven fuel pump kW 8 8 8 8 8 8 4
Electric fuel pump kW 8 8 8 8 8 8 4
ECS recirculation kW 5 5 5 5 5 5 5
E/E cooling kW 6 6 6 6 6 6 6
Miscellaneous loads kW 5 5 5 5 5 5 5

þ Only CSA
Hydraulic system kW 9.8 17.1 13.1 12.5 15.1 23.5
ECS kg/s 1.05 1.05 1.05 1.05 1.05 1.05
Wing IPS (running-wet) kg/s 0.97 0.97 0.2 0 0.31 0.97
Nacelle IPS (evaporative) kg/s 1.13 1.13 2.58 0 1.09 1.13

Figure 10. CSA !ESA transition with intermediate, uncycled, and cycled designs.

other, as shown in Figure 10: CSA !ESA-I !ESA- 2.88% reduction in cruise fuel consumption and
U !ESA-C. The corresponding changes in relevant 3.50% reduction in block fuel. While a hypothetical
vehicle and mission parameters are shown in Table 7, intermediate design, the ESA-I allows two important
where the percentage changes are relative to the per- observations to be made. First, with engine cycle
formance of the baseline CSA design. changes accounted for, there is a significant fuel
The first step, CSA !ESA-I, was aimed at isolat- burn reduction directly attributable to the change to
ing the effect of changing the form and magnitude of a bleedless subsystem architecture with only shaft-
secondary power extraction by transitioning from power off-take. Second, this fuel burn reduction
mixed off-take to pure shaft-power off-take. In this would allow even a certain amount of OEW increase
step, the predicted OEW (Table 4) was not applied to be tolerated without an overall MTOW or mission
to the mission performance evaluation. The resulting fuel penalty. Such an observation is significant given
intermediate ESA-I design (I: intermediate) showed the considerable uncertainty regarding subsystem

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

Chakraborty et al. 15

Table 7. Comparison of CSA (baseline) design with ESA design.

Parameters CSA ESA-I ESA-U ESA-C

Wing
Ref. area 124.8 m2 (1343 ft2) ¼ ¼ – 2.92%
Loading 622.5 kg/m2 (127.5 lb/ft2) – 0.88% – 1.98% ¼

Propulsion
Sea-level static thrust 242.9 kN (54,600 lbf) ¼ ¼ – 2.92%
Thrust/weight ratio 0.3189 þ 0.88% þ 2.02% ¼

Mass
OEW 41,871 kg (92,310 lb) ¼ – 1.69% – 2.97%
MTOW 77,655 kg (171,200 lb) – 0.88% – 1.98% – 2.92%

Fuel
Block fuel 17,556 kg (38,704 lb) – 3.50% – 4.20% – 5.12%
Cruise fuel 14,869 kg (32,780 lb) – 2.88% – 3.52% – 4.43%
I: Intermediate, U: Uncycled, C: Cycled. Percentage changes are shown relative to the baseline CSA design. Equality sign (¼) indicates the original
baseline value.

component weights that exists in early design phases, established in this approach (an example of which is
especially for unconventional architectures. seen in Figure 3), the application of this re-sizing rule-
The second step, ESA-I !ESA-U, considered add- set resulted in the simultaneous re-sizing of the vehicle
itionally the effect of the predicted change in fixed and its subsystems (‘‘Cycle? ¼ Yes’’ in Figure 2), yield-
equipment weight (Table 4) while maintaining con- ing a converged solution where MTOW, wing area,
stant vehicle geometry and rated thrust, resulting in and required thrust were reduced by 2.92%, and mis-
the uncycled ESA-U design (U: uncycled). For this sion fuel burn by 5.12% relative to the baseline.
case, the fuel-burn reduction was driven by the com- The predicted fuel burn reduction is higher than
bined effects of the OEW reduction and the transition that noted by Jones2 for an Airbus A320 AEA
to electric secondary power architecture. The OEW ( 4%). However, it must be noted that the current
and mission fuel burn reduction together resulted in analysis did not account for drag increments due to (i)
a reduction in MTOW of the ESA-U relative to the the two additional ram air inlets required to supply air
CSA. This indicates that an increase in payload cap- to the electric ECS compressors of the ESA and (ii)
ability (maintaining fixed range), or range capability the possible build-up of ice on the protected wing sur-
(maintaining fixed payload), or a combination thereof faces during the ‘‘OFF’’ period of the cyclic heater
may be possible. Thus the ESA-U, which is more rep- cycle. Had they been accounted for, these two effects
resentative of a retro-fit design than an entirely new (the first continuous and the second present in mission
design, shows the potential for increased utility that is segments with icing risk) would likely reduce the pre-
driven by the subsystem architecture changes. dicted fuel burn savings.
However, if the payload-range and operational
requirements are fixed, then the ESA-U is over-designed
as it has a higher thrust-to-weight ratio (TSL/WTO) and
Conclusions and future work
a lower wing loading (WTO/Sw) than required to satisfy This work presented an approach for integrating the
point performance requirements. The logical step is to sizing and analysis of novel subsystem architectures
‘‘cycle’’ or re-size this design, which was done in the with aircraft and propulsion system sizing. The
third and final step, resulting in the cycled ESA-C approach involves the derivation of subsystem func-
design (C: cycled). For re-sizing the aircraft, the base- tional requirements, the subsequent sizing of architec-
line payload-range capabilities were maintained. The ture components by computationally inexpensive
design was cycled back to the original thrust-to- means that do not require detailed aircraft definition,
weight ratio and wing loading of the baseline, which and the establishment of dependencies among subsys-
represented unchanged point performance require- tem level and aircraft level parameters. A salient fea-
ments. Additionally, the stabilizers were re-sized to ture of this integrated approach is that information is
maintain the baseline horizontal tail volume ratio iteratively exchanged between aircraft sizing and sub-
(Vh ¼ 1.04) and vertical tail volume ratio (Vv ¼ 0.09), system sizing processes until simultaneous convergence
which represented unchanged stability and control- occurs. This allows the so-called ‘‘snowballing’’ effects
lability requirements. With the dependencies between arising out of subsystem architecture changes to be
system (aircraft) level and subsystem level parameters effectively captured. While the results presented in the

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

16 Proc IMechE Part G: J Aerospace Engineering 0(0)

paper were obtained using a specific set of tools, 4. Cronin MJJ. The all-electric aircraft. IEE Rev 1990; 36:
another salient feature of the proposed approach is 309–311.
that it may be implemented using any set of tools 5. Sinnett M. Boeing 787 no-bleed systems: Saving fuel
that collectively provide a stated set of capabilities. and enhancing operational efficiencies, 2007. http://
www.boeing.com/commercial/aeromagazine/
The approach was demonstrated through a compara-
6. Van den Bossche D. The A380 flight control electro-
tive assessment of an ESA relative to a CSA. A sequen-
hydrostatic actuators, achievements and lessons
tial three-step approach was used to successively isolate learnt. In: 25th International congress of the aeronaut-
and identify the effects of changing the form and mag- ical sciences, ICAS 2006, Hamburg, Germany, 2006,
nitude of secondary power extraction, the effect of the pp.1–7.
change in subsystem weights, and the effect of re-sizing 7. Recksiek M. Advanced high lift system architecture
the vehicle in response to a defined rule-set. In terms of with distributed electrical flap actuation. In: Aviation
reduction in mission fuel consumption, the magnitude system technology workshop, Hamburg, Germany,
of the first effect was seen to be the most prominent. A 2009. http://www.transport-research.info/Upload/Doc
further fuel burn benefit was obtained due to the pre- uments/201210/20121030_122633_15453_Advanced_Hi
dicted reduction in subsystem weights. Finally, the full gh_Lift_System_Architecture_with_Distributed_Electri
cal_Flap_Actuation.pdf.
effect of the transition was identified when the vehicle
8. Bennett JW, Mecrow BC, Jack AG, et al. A prototype
and associated subsystems were cycled according to a
electrical actuator for aircraft flaps and slats. In: 2005
rule-set that ensured that the same operational and IEEE international conference on electric machines and
point performance capabilities were maintained. drives, San Antonio, TX, USA, 2005, pp.41–47.
The demonstrated approach is directly applicable 9. Atallah K, Caparrelli F, Bingham CM, et al. Permanent
to jet-engined and turboprop tube-and-wing aircraft magnet brushless drives for aircraft flight control
of varying sizes. It is also extendable to unconven- surface actuation. In: IEE Colloquium on electrical
tional aircraft configurations such as blended / machines and systems for the more electric
hybrid wing body (BWB/HWB); however, given the aircraft (Ref. No. 1999/180), London, UK, 1999,
redundant control surfaces, the peak actuation loads pp.8/1–8/5.
are not as easily identifiable from FAR §25. A differ- 10. Fronista GL and Bradbury G. An electromechanical
actuator for a transport aircraft spoiler surface. In:
ent approach that considers control allocation would
Proceedings of the 32nd intersociety on energy conversion
have to be employed in such a case. Avenues for fur-
engineering conference, IECEC-97, vol. 1, 1997,
ther work include exercising the proposed approach pp.694–698.
to investigate a larger number of subsystem architec- 11. Nicolas Y. eTaxi - Taxiing aircraft with engines
tures (to further explore the architectural design stopped; 2013. Flight Airworthiness Support
space) for aircraft of different sizes. Technology (FAST) 51, Airbus Technical Magazine,
http://www.airbus.com/support/publications/
Acknowledgments 12. Raymer DP. Aircraft design: A conceptual approach. 4th
The authors would like to acknowledge Mr Christopher ed. Reston, VA: American Institute of Aeronautics and
Perullo (Research Engineer II at the ASDL) for supporting Astronautics Inc, 2002.
this work through his expertise with the NPSS tool, Mr 13. Roskam J. Airplane design, Part V - Component weight
David Trawick (PhD student at the ASDL) for his assis- estimation. Design Analysis & Research, 1999.
tance in reviewing the power generation and distribution 14. Torenbeek E. Synthesis of subsonic airplane design: An
network and estimating the fixed cabin loads, and Dr introduction to the preliminary design of subsonic general
Elena Garcia (Research Engineer II at the ASDL) for pro- aviation and transport aircraft, with emphasis on layout,
viding valuable technical feedback during the preparation of aerodynamic design, propulsion and performance.
the manuscript. Netherlands: Delft University Press, 1976.
15. Tagge G, Irish L and Bailey AR. Systems study for an
Conflict of interest integrated digital/electric aircraft (IDEA). NASA
Contractor Report 3840, 1985.
None declared.
16. Cronin MJ, Hays AP, Green FB, et al. Integrated
digital/electric aircraft concepts study. Burbank, CA:
Funding Lockheed-California Company, 1985.
This research received no specific grant from any funding 17. Lytle JK. The numerical propulsion system simulation:
agency in the public, commercial, or not-for-profit sectors. An overview. Glenn Research Center, Cleveland, OH,
2000. NASA/TM-2000-209915, http://ntrs.nasa.gov/
References archive/nasa/casi.ntrs.nasa.gov/20000063377.pdf.
1. Faleiro L. Beyond the more electric aircraft. AIAA 18. Blanding D. Subsystem design and integration for the
Aerospace America, September 2005, pp.35–40. more electric aircraft. In: 25th International congress of
2. Jones RI. The more electric aircraft - Assessing the bene- the aeronautical sciences (ICAS 2006), Hamburg,
fits. Proc IMechE, Part G: J Aerospace Engineering 2002; Germany, 2006, pp.3732–3739.
216: 259–270. 19. Rudolph PKC. High lift systems on subsonic
3. Wheeler PW, Clare JC, Trentin A, et al. An overview of commercial airliners. National Aeronautics and
the more electrical aircraft. Proc IMechE, Part G: J Space Administration, 1996. CONTRACT
Aerospace Engineering 2012; 227: 578–585. A46374D(LAS).

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

Chakraborty et al. 17

20. Roskam J. Airplane design, Part VI - Preliminary calcu- 37. Thomas SK, Cassoni RP and MacArthur CD. Aircraft
lation of aerodynamic, thrust and power characteristics. anti-icing and de-icing techniques and modeling.
Design Analysis & Research, 1999. J Aircraft 1996; 33: 841–854.
21. Scholz D. Development of a CAE-tool for the design of 38. Krammer P and Scholz D. Estimation of electrical
flight control and hydraulic systems. In: Institution of power required for deicing systems. Hamburg
Mechanical Engineers (ed.) Avionics systems, design and University of Applied Sciences (HAW), Germany, 2009.
software, 1996, pp.1–22. 39. Al-Khalil K, Horvath C, Miller D, et al. Validation of ther-
22. Federal Aviation Regulations (FAR), Part mal ice protection computer codes: Part 3 - The validation
25 - Airworthiness standards: Transport category air- of ANTICE. In: 35th Aerospace sciences meeting and exhi-
planes. Federal Aviation Administration (FAA), bit, AIAA-97-0051, Reno, Nevada, USA, 1997, pp.1–12.
http://www.ecfr.gov/ 40. Pourbagian M and Habashi WG. Surrogate-based opti-
23. Moraris VM, Lawson NJ and Garry KP. Aerodynamic mization of electrothermal wing anti-icing systems.
and performance characteristics of a passive leading J Aircraft 2013; 50: 1555–1563.
edge Krueger flap at low Reynolds numbers. Aeronaut 41. Bu X, Lin G, Yu J, et al. Numerical analysis of a swept
J 2012; 116: 759–769. wing hot air ice protection system. Proc IMechE, Part
24. Kelly JA and McCullough GB. Aerodynamic loads on a G: J Aerospace Engineering 2013; 227: 1608–1622.
leading-edge flap and a leading-edge slat on the NACA 42. Papadakis M and Wong S. Parametric investigation of
64A010 airfoil section (NACA TN 3220). National a bleed air ice protection system. In: 44th AIAA aero-
Advisory Committee for Aeronautics (NACA), 1954. space sciences meeting and exhibit, AIAA-2006-1013,
25. Chakraborty I, Mavris DN, Emeneth M, et al. A meth- Reno, Nevada, USA, 2006, pp.1–28.
odology for vehicle and mission level comparison of 43. Hua J and Liu HH. Fluid flow and thermodynamic
more electric aircraft subsystem solutions - Application analysis of a wing anti-icing system. Can Aeronaut
to the flight control actuation system. Proc IMechE, Space J 2005; 51: 35–40.
Part G: J Aerospace Engineering 2014. 44. Meier O and Scholz D. A handbook method for
26. Li W and Fielding JP. Preliminary study of EMA land- the estimation of power requirements for electrical de-
ing gear actuation. In: 28th International congress of the icing systems. In: Deutscher Luft-und Raumfahrtkongress,
aeronautical sciences (ICAS 2012), Brisbane, Australia, Hamburg, Germany, 2010. http://www.mp.haw-ham
2012. http://www.icas.org/ICAS_ARCHIVE/ICAS201 burg.de/pers/Scholz/MOZART/MOZART\_PUB\_DL
2/PAPERS/109.PDF. RK\_10-08-31.pdf.
27. Jenkins SFN. Landing gear design and development. 45. Wright W. An evaluation of jet impingement heat trans-
Proc IMechE, Part G: J Aerospace Engineering 1989; fer correlations for Piccolo tube application. In: 42nd
203: 67–73. Aerospace sciences meeting and exhibit, AIAA-2004-
28. Collins A. EABSYS: Electrically actuated braking system. 0062, 2004, pp.1–12.
In: IEE Colloquium on electrical machines and systems for 46. Liscouet-Hanke S, Pufe S and Mare JC. A simulation
the more electric aircraft, London, UK, 1999, p. 4. framework for aircraft power management. Proc IMechE,
29. Bennett J, Mecrow B, Atkinson D, et al. Fault-tolerant Part G: J Aerospace Engineering 2008; 222: 749–756.
electric drive for an aircraft nose wheel steering actu- 47. Giraud X, Piquet H, Budinger M, et al. Knowledge-
ator. IET Electric Syst Transport 2011; 1: 117–125. based system for aircraft electrical power system recon-
30. Botten SL, Whitley CR and King AD. Flight control figuration. In: Electrical systems for aircraft, railway
actuation technology for next-generation all-electric air- and ship propulsion (ESARS), Reno, Nevada, USA,
craft. Technol Rev J Millenium Issue 2000; 55–68. 2012, pp.1–6.
31. Hunt EH, Reid DH, Space DR, et al. Commercial air- 48. Biedermann O and Geerling G. Power control units
liner environmental control system - Engineering aspects with secondary control hydraulic motor - A new con-
of air quality. In: Aerospace medical association annual cept for application in aircraft high lift systems. In:
meeting, Anaheim, California, USA, 1995, pp.1–8. Proceedings of the conference on recent advances in aero-
32. Nelson T. 787 Systems and performance, 2009. http:// space hydraulics, Toulouse, France, 1998, pp.1–5.
myhres.com/Boeing-787-Systems-and- 49. Cronin MJ. All-electric vs. conventional aircraft: The pro-
Performance.pdf. duction/operational aspects. J Aircraft 1983; 20: 481–486.
33. Pérez-Grande I and Leo TJ. Optimization of a commer- 50. Allen T, Miller T and Preston E. Operational advan-
cial aircraft environmental control system. Appl Therm tages of carbon brakes; 2009 Quarter 3. http://
Eng 2002; 22: 1885–1904. www.boeing.com/commercial/aeromagazine/
34. Liscouet-Hanke S, Mare JC and Pufe S. Simulation 51. Honeywell. Boeing 737 NG Wheel and
framework for aircraft power system architecting. CERAMETALIX Brake - Proven braking performance
J Aircraft 2009; 46: 1375–1380. and reliability with low operational costs. http://aero
35. Al-Khalil K. Thermo-mechanical expulsion deicing sys- space.honeywell.com/ /media/UWSAero/common/do
tem - TMEDS. In: 45th AIAA aerospace sciences meet- cuments/myaerospacecatalog-documents/ATR_Brochu
ing and exhibit, AIAA-2007-692, Reno, Nevada, 2007, res-documents/737NG_Data_Sheet_US.pdf.
pp.1–13. 52. Simsic CJ. Electric actuation system duty cycles. In:
36. Goraj Z. An overview of the deicing and anti-icing Proceedings of the IEEE 1991 National conference on
technologies with prospects for the future. In: aerospace and electronics, Dayton, Ohio, USA, 1991,
Proceedings of 24th international congress of the aero- pp.540–545.
nautical sciences (ICAS), Yokohama, Japan, 2004. 53. de Tenorio C. Method for collaborative conceptual
http://www.icas.org/ICAS_ARCHIVE/ICAS2004/PAP design of aircraft power architectures. Atlanta, GA:
ERS/547.PDF. Georgia Institute of Technology, 2010.

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

18 Proc IMechE Part G: J Aerospace Engineering 0(0)

Appendix Table 8. Summary of flight control surface actuation

requirements for baseline aircraft.
Actuation subsystems
.Environmental control system and ice protection system Actuating
Control surface load Rate Controls Actuators
. Atmospheric density , freestream pressure p1, / Aircraft / Control
dynamic viscosity , thermal conductivity k0, and
specific heat of air at constant pressure Cp were Ailerons 4,200 Nm 60 /s 2 2
evaluated based on flight condition. Ratio of spe- Elevators 7,600 Nm 60 /s 2 2
cific heats,
¼ 1.4. Rudder 8,200 Nm 60 /s 1 3
. Characteristic lengths were taken as x ¼ ðfuselage Flight spoilers 4,200 Nm 60 /s 8 1
lengthÞ=2, x ¼ ðslat chordÞ=2, x ¼ nacelle length Ground spoilers 3,800 Nm 40 /s 4 1
. Similarity parameters: Re ¼ V1 x=, Pr ¼ Trailing-edge 51,000 N 102 mm/s 4 2
Cp =k0 , Nu ¼ 0:0296 Re4=5 Pr1=3 flaps
. Temperature recovery factor: Rc ¼ 1  0:99 Leading-edge 6,300 N 60 mm/s 8 1
ð1  Pr0:5 Þ slats
. External convection coefficient: hext ¼ Nu k0 =x Krueger flaps 5,600 Nm 16 /s 4 1
. Internal convection coefficient: hint ¼ 5 W/m2K
(free convection assumption)
. Fuselage thermal resistance: Rskin ¼ 0.2 m2/WK
(insulation, metal skin, and paint – in series) Table 9. Main and nose landing gear retraction mechanism
. The heat transfer surface area for a cabin zone: parameters.
Ai ¼  ðfuselage diameterÞ  ðzone lengthÞ
Parameter Nose gear Main gear
. Total cabin-to-ambient thermal resistance:
Rtot ¼ 1=ðhint Ai Þ þ Rskin þ 1=ðhext Ai Þ BC (a) 0.09 Llg 0.09 Llg
. Total heat flux: q_tot ¼ q_ conv þ q_sens þ q_ evap þ q_ kin CA (b) 0.28 Llg 0.28 Llg
. Convective
 heat
 flux: q_ conv ¼ hext fTskin  T1 CG (‘cg ) 0.5 Llg 0.5 Llg
1 þ Rc
1
2 M 2
g  80 80
. Sensible heat flux: q_ sens ¼ m_ local ðTskin  T1 Þ Wlg 0.5% MTOW 1.25% MTOW
ðnci þ ð1  nÞ cw Þ þ n Lf g, n ¼ cw ðTref  T1 Þ=Lf m 0.80 0.80
. Evaporative heat flux: q_ evap ¼ minðm_ local Le ,
up , dn 20 , 90 25 , 90
0:7 hext Le Rh pe11 Cep skin Þ
t 15 s 15 s
. Kinetic heat flux: q_kin ¼ m_ local v21 =2
. Specific heat of water cw, specific heat of ice ci,
latent heat of fusion of ice Lf, latent heat of eva-
poration of water Le, reference melting point of ice Table 10. Brake system parameters assumed for CSA and
Tref ¼ 273 K, saturation vapor pressure at free- ESA designs.
stream and skin temperatures e1 , eskin
Carbon–Carbon
Parameter Steel (CSA) (ESA)

Specific heat (c) 544.3 J/kg/K 1297.9 J/kg/K

Permissible temp. 500 C 500 C
rise (Tmax)
Friction coefficient () 0.3 0.3
Dissipated kinetic max. of D1
energy (KE) & D2.

Table 11. Nose landing gear (NLG) and nose-wheel steering (NWS) parameters.

NLG parameters NWS parameters

Max. static load es 15% Deflection max , min  75

Rake angle r 7 Rate !max , !min  20 /s
Tire radius rtire 38 cm
Tire trail xtire 18% of tire radius

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

Chakraborty et al. 19

Table 12. ECS pack parameters.

Component characteristics HX characteristics

Compressor efficiency c 0.75 Pack air P/P h 0.18

Turbine efficiency t 0.80 Ram air P/P c 0.06
Ram diffuser efficiency rd 0.95 HX max effectiveness emax 0.80

Table 13. Wing ice protection system (WIPS) and cowl ice protection system (CIPS) parameters.

Geometric parameters Operational parameters

Common parameters
Protected area / wing Aprot 3.45 m2 Skin temp. (running-wet) Tskin 6 C
Protected area / nacelle ’’ 4.01 m2 Skin temp. (evaporative) Tskin 37 C

Pneumatic IPS Bleed air entry temp. Tsupply 200 C

(post-regulation)
Electrothermal IPS
Parting strip area ratio ps 20 % Cyclic heating activity ratio cyc 5%
Heat ‘‘on’’ time theat 9s
Ice thickness to melt ‘ice 0.5 mm
Overall heater efficiency heat 70%

Downloaded from pig.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016