AIAA JOURNAL

Vol. 52, No. 5, May 2014

Aeropropulsion for Commercial Aviation in the Twenty-First

Century and Research Directions Needed

Alan H. Epstein∗
Pratt & Whitney, East Hartford, Connecticut 06108
DOI: 10.2514/1.J052713
Two driving imperatives of 21st century commercial aviation are improving fuel consumption and reducing
environmental impact. The research important to aeropropulsion’s advancing these goals is shaped both by physics of
the design space and by design choice. As fuel becomes increasingly more expensive, engine architectures and design
details evolve to reflect the new balance between engine fuel consumption, weight, and manufacturing and
maintenance costs. The evolution of engine architectures changes the relative value of specific technologies. The
Downloaded by UNIVERSITA DEGLI STUDI DI ROMA on October 29, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J052713

engines of the future will be advanced gas turbines due to their superior fuel burn at the aircraft level. They will be
fueled by sustainable liquid hydrocarbons. Both the thermal and propulsive efficiency of the gas turbine can be
significantly improved. The need to improve propulsive efficiency has driven engine bypass ratio up, to 12 recently,
and higher in the future. This is a different, less familiar design space than the 5 to 8 bypass ratio, which characterized
the last 40 years of engine experience. Realignment of research priorities is required to address 21st century
challenges, such as the knowledge needed to realize efficient engines at very small core sizes. The new challenges open
up new opportunities for both designers and researchers.

I. Introduction Needs and opportunities peculiar to military applications or general

I T HAS been more than 70 years since the flight of the first jet
airplane and over 50 years since the first successful commercial
jet airliner, the Boeing 707, entered service. Reflecting R&D
investments of tens of billions of dollars over this period, the jet
engine has improved enormously: efficiency up by three times,
power-to-weight ratio up by of two to four times, and reliability and
life improved 100–200 times and 5–10 times, respectively.
The turbofan jet engine is now the aeropropulsion system of choice.
It is appropriate now to ask how much further jet engines can be
improved. Will continued investments here be fruitful, and if so, what
should they be? In a broader sense, the gas turbine is now the aircraft
engine of choice because of its high efficiency, low weight, low
emissions, and extraordinary reliability. How much longer will this
continue?
This paper considers these questions with the aim of identifying
and prioritizing research paths relevant for advancing aeropropul-
sion. There are very diverse applications for airplanes, including
commercial, military, and general aviation. Commercial aviation is
focused on the transportation of people and goods and represents the
majority of the economic value that aircraft bring to the world. It is
also responsible for the majority of the environmental impact of
aviation and most of the business revenue associated with aviation
engines. This discussion is thus focused on commercial aviation.
aviation are not considered here.
II. Defining Aeropropulsion
Aircraft propulsion can be considered as consisting of two
necessary elements. The first is a motor to convert stored energy to
mechanical power, typically in the form of a rotating shaft. The
second is the conversion of mechanical power into propulsive power.
Excluding rockets, to date we have identified only two methods
of propelling an airplane: flapping wings or spinning a propeller. The
flapping of wings has not been notably successful for airplanes and
so may be safely neglected here. Indeed, theoretical analysis
suggests that flapping is less efficient than a propeller in converting
mechanical power into propulsive power [1]. A propeller may be
operated in free air, installed in a duct to produce a jet and called a fan,
or canted to the flight direction and called a rotor (as in a helicopter).
Herein, we will adopt the term propulsor as referring to a device
which converts shaft power to propulsive power, inclusive of
propellers, fans, and rotors.
Propulsors are turned by motors: internal combustion in the old
days, gas turbines for the past half-century. Recently, there has been
consideration of using electric motors, so care must be taken
to distinguish between power and energy. Power and energy
requirements for a wide variety of land, sea, and air vehicles are

Alan H. Epstein is Vice President of Technology and Environment at the Pratt & Whitney Division of United
Technology Corporation. He leads Pratt & Whitney’s efforts to identify and evaluate new methods to improve engine
performance, fuel efficiency, and environmental impact. He also provides strategic leadership in the investment,
development, and incorporation of technologies that reduce the environmental impact of Pratt & Whitney’s
worldwide products and services. Before joining Pratt & Whitney, Dr. Epstein was the R. C. Maclaurin Professor of
Aeronautics and Astronautics at the Massachusetts Institute of Technology (MIT) and the Director of the MIT Gas
Turbine Laboratory. He currently holds an appoint there as Professor Emeritus. Dr. Epstein is a member of the U.S.
National Academy of Engineering and is a Fellow of AIAA and of the American Society of Mechanical Engineers. He
received his B.S., M.S., and Ph.D. degrees from MIT in aeronautics and astronautics.

Presented as Paper 2013-0001 at the AIAA Aerospace Sciences Meeting, Grapevine, TX, 7–10 January 2013; received 4 April 2013; revision received 17 October
2013; accepted for publication 31 October 2013; published online 28 March 2014. Copyright © 2013 by United Technologies Corporation. Published by the
American Institute of Aeronautics and Astronautics, Inc., with permission. Copies of this paper may be made for personal or internal use, on condition that the copier
pay the $10.00 per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923; include the code 1533-385X/14 and $10.00 in
correspondence with the CCC.
*Vice President, Technology and Environment, 400 Main Street, M/S 162-24. Fellow AIAA.
901
 902
Fig. 1 Energy and power of air, land, and sea vehicles.
Downloaded by UNIVERSITA DEGLI STUDI DI ROMA on October 29, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J052713
shown in Fig. 1. Clearly, large aircraft flying long distances at high
speed require prodigious amounts of both power and energy.
III. Energy Sources and Energy Storage
Will we use the same fuel in the future as we use now? Engineering
criteria for jet aircraft fuel selection changed little in the 20th century.
In the last decade, increased concern for the environment, climate
change in particular, has added a new imperative for aviation:
reduction in greenhouse gases, especially CO2. Thus, it is prudent to
consider whether we will continue to use the same aircraft fuels in the
21st century as we did in the 20th. Current jet fuel is chemically
similar to kerosene. The technical attribute of fuel most important to
airplane design and performance is energy density, both gravimetric
and volumetric. Cost and emissions are very important as well, with
additional concerns of thermal stability, lubricity, etc. Over the past
70 years, research on improved fuels has yielded relatively minor
gains, mainly in slightly increased density (JP-10) and thermal
stability (JP-7, JP-8
100).
The energy density and energy cost for a variety of “fuels” are
shown in Table 1 [2]. In terms of room-temperature liquids, Jet A has
the highest energy density and lowest cost. Although the gravimetric
energy density of methane is close to that of Jet A, and hydrogen is 2.7
times greater than Jet A; these are gases at room temperature and thus
must be stored as cryogenic liquids or at high pressure. The weight of
high-pressure containment makes the latter option impractical given
tank materials available today. Cryogenic storage as liquid is possible
but introduces many questions including routine handling and safety,
especially in accidents. Liquid hydrogen has less than 10% the
volumetric density of Jet A. For equivalent onboard energy, liquid
hydrogen fuel requires storage volume 10 times greater than today’s
liquid fuel with a concomitant increase in aircraft weight, drag, and
energy consumption. This suggests that liquid hydrogen is not an
attractive fuel for high-speed aircraft, a lesson first learned in the
1950s [3]. Hydrogen might have a role for low-speed surveillance
applications when persistence is a dominant design criterion [4].
Combustion motors derive their energy from chemical energy
stored in fuel. Electric motors need electric power. Conceivably, this
can be generated by combustion motors driving generators, by fuel
Table 1 Gravimetric (GED) and volumetric energy density
(VED) and cost of liquid fuels
Fuel type GED, MJ∕kg VED, MJ∕l Cost, $∕MJ
Li battery (rechargeable) 0.3 0.3 0.03
Li Battery (primary) 0.6 0.6 170
Honey 14 20 0.29
Goose fat 38 35 0.26
Kerosene (Jet A) 44 36 0.018
Natural gas 45 19a 0.005
Hydrogen 117 8.3a 0.44
a
Volume of liquid only, not accounting for cryotank.
EPSTEIN
cells from chemical energy in fuel, from chemical energy in batteries,
or from solar cells on the vehicle. The latter is impractical on this
planet for transport aircraft given the level of solar irradiance falling
on the Earth. This irradiance is insufficient to support aircraft wing
loading above 20 lb∕ft2 (98 kg∕m2 ), far below that needed for all but
the slowest-speed flight. Thus, we can safely rule out solar cells
powering commercial airplanes. This means that aircraft of the
future, as those of the past, must be fueled.
Batteries are a different approach to energy storage and present
their own challenges. The theoretical energy density of lithium
chemistry is about 10% that of kerosene. When batteries are
engineered with current technology for such practical considerations
as safety, performance over a wide temperature range, and long life,
their energy density is 10% of the theoretical maximum and thus only
about 1% that of kerosene. Thus, even a several hundred percent
improvement in battery technology would still leave batteries many
times inferior to hydrocarbon fuels in terms of energy density.
Furthermore, the weight decrease during flight of fueled aircraft is an
important factor in establishing aircraft range (the Breguet range
equation [5]). Fuel weight decreases during a mission but battery
weight does not, implying an additional penalty for a battery-
powered vehicle. Given the previous considerations, pending the
discovery of as yet unknown battery chemistry, it is unlikely that
batteries will replace fuel on commercial aircraft.
To significantly improve its climate change impact, aviation must
reduce both the amount of fuel burned on each flight and the net CO2
produced by that fuel. This means a switch from fossil fuel. Currently,
hydrocarbon fossil fuel serves as both the energy source and the
energy storage medium on the airplane. This must change. Rather
than depend on fossil energy, aviation must move to a sustainable
energy source such as solar, wind, or nuclear. Whatever the energy
source, the previous discussion implies that energy is best delivered
to and stored on the aircraft as a liquid hydrocarbon. Current focus is
on capturing solar energy in the form of renewable biofuels. Here, the
CO2 exhausted by the engine is that absorbed from the atmosphere by
plants or algae. With current technology, the growing, processing,
and transportation of fuel produces an amount of CO2 somewhat less
than that in the engine exhaust, and so the net reduction from a biofuel
is greater than 50% [6]. Biofuel supply chain technology should be
able to improve this considerably.
Many ground and flight tests have shown that drop-in biofuels are
technically feasible, and a blend of up to 50% of a biofuel is now
approved for use on commercial aircraft. The fuels approved to date
are in relatively short supply and expensive. One reason is that they
use expensive feedstock, basically vegetable oil. With current crop
yields and processes, the net efficiency of the conversion of solar
energy to jet fuel in this manner is only about 0.05%, implying that
there is considerable room for improvement. Improvement requires
research in such areas as increasing crop yields, new or modified
organisms engineered for biofuel production, and new processes
suited to low cost feedstock. Promising avenues include cellulosic
biomass, algae, and halophytes. Also, as society greens, the CO2
overhead associated with the growth, processing, and transportation
of biofuels should improve.
IV. Motors to Power Propulsors
Energy will continue to be supplied to and stored on aircraft as
liquid fuel, but will the gas turbine continue as the device of choice to
convert that energy to shaft power? Other candidates might be fuel
cells powering electric motors, different thermodynamic cycles
(Otto, Rankin, Sterling, etc.), or some hybrid combination. This
question can best be addressed by considering why gas turbines are
the current motor of choice, physical constraints and limitations, and
metrics by which aircraft motors are now and will be assessed. To
potentially replace an existing approach, a new approach must be
significantly better than the incumbent or at least appear to be. The
metrics by which these are evaluated include efficiency, weight,
safety and reliability, emissions, and cost.
 EPSTEIN
A. Thermal Efficiency
The limit to the ideal thermal efficiency of Brayton cycles such as a
gas turbine is readily estimated at about 80% for flight in the lower
stratosphere. How close to that theoretical maximum these devices
can practically reach is not as simple a question. In the gas turbine
industry, there are several definitions of efficiency that are defined for
different uses. Koff [7] defined the thermodynamic efficiency of the
core as the fluid power available at the core exit divided by the heat
added from the fuel’s chemical energy and plotted that against the
propulsive efficiency times the transmission efficiency (transmission
includes the losses in the turbine driving the fan, the fan itself, the fan
duct, etc.). This is shown in Fig. 2 at cruise and illustrates the progress
to date. The product of the two efficiencies is shown as arcs, which
represent the total efficiency of the conversion of chemical energy in
the fuel to propulsive power. Since Whittle’s first engine, this
thermodynamic efficiency has improved from about 10% to over
50%. When weight and drag are not an issue, as in ground-based
Downloaded by UNIVERSITA DEGLI STUDI DI ROMA on October 29, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J052713
power plants, then gas turbine combined cycle plants (a gas turbine
whose exhaust heat runs a steam cycle) can now deliver efficiencies
above 60%. Propulsive efficiency has improved as well, from 50 to
70%. Overall, gas turbine aeroengine total efficiency has climbed
from 10% to almost 40%.
Koff’s definition of thermal efficiency is useful for comparing
among jet engines. Another useful definition of gas turbine thermal
efficiency for comparing with other engines or motors is one used for
turboprops that accounts for all of the core fluid power as shaft power
deliverable to a propulsor, designated here as “motor efficiency”. The
evolution of commercial aircraft gas turbine motor efficiency is
shown in Fig. 3. This efficiency has improved by about 16 points over
four decades and now approaches 55%. (The considerable scatter
implies that thermal efficiency has not always been the primary
design driver.) By contrast, diesel engines now range from 30 to 50%
motor efficiency, with the higher efficiencies at the largest sizes,
10–60 MW [8]. A practical advantage of diesels over gas turbines in
Fig. 2 Core thermal and propulsive efficiencies for commercial aircraft
engines.
Fig. 3 Evolution of commercial turbofan motor efficiency.
903
some applications is that diesels retain relatively more of their peak
efficiency at part power. Because most transport aircraft engines are
designed for peak efficiency at cruise, where most of the fuel is
burned, this attribute has much less importance for airplanes than for
ground vehicles or power generation.
Current fuel cells combine H2 and O2 to generate electricity. How
the H2 is generated varies widely. If the fuel cell is to operate from a
complex hydrocarbon fuel, then the definition of efficiency must
include all of the reforming processes that convert the fuel into H2 .
Current ground power-generation systems [9] operate at about 40%
overall efficiency. In a practical aviation application, the efficiency
implications of the electric motors, drive train, and their cooling
would need to be considered as well, consistent with the definition of
motor efficiency.
In summary, modern, large gas turbine engines are the most
efficient devices in service to convert hydrocarbon chemical energy
to mechanical power. They are by no means mature, and so
considerable improvement in efficiency can result from focused
research, as is discussed later.
B. Weight
Airplanes are all about weight, and so airplane engines must be as
well. The Wright brothers built their own engine out of aluminum for
just this reason, even though aluminum was a very expensive material
in those days. Thirty-seven years later in 1940, American technical
luminaries were very skeptical of the concept of gas turbines for this
same reason [10]:
“The gas turbine could hardly be considered a feasible
application to airplanes mainly because of complying with
the stringent weight requirements imposed by aero-
nautics : : : The present internal-combustion engine used in
airplanes weighs about 1.1 pounds per horsepower, and to
approach such a figure with a gas turbine seems beyond the
realm of possibility with existing materials.”
This report was issued a year after the first jet plane had flown in
Germany, unbeknownst to the authors. The designers of the German
engine used air cooling to circumvent “ : : : the realm of possibility
with existing materials”. This illustrates both the role that materials
play in determining engine weight and the skill of engineers
and designers in circumventing what scientists may regard as
fundamental barriers, such as material properties.
Since the early days of turbofan development, commercial
turbofan power-to-weight ratios have improved by a factor of 4 or
more, to 9 hp∕lb (15 kW∕kg). In contrast, a 10–60 MW diesel
engine is more than 400 times heavier. Part of this weight difference is
a result of aeroengine applications favoring light weight over low cost
and thus embracing relatively expensive materials such as titanium.‡
However, the most important factor influencing the relatively low
weight of a gas turbine is that the average air velocity through a gas
turbine is very much higher than that through other combustion or
electrochemical (fuel cells) motors. At the same thermal efficiency,
motors consume the same fuel and thus need the same air for
combustion. To first order, the motor with the higher average through
flow velocity will have the smaller cross section and weight. For a
consistent comparison, the weight of an electrochemical motor must
include the complete fuel cell system, electric drive train, cooling
system, and structure needed to produce shaft power at all altitudes.
On a weight basis alone, fuel cells appear to be highly unattractive for
commercial aircraft propulsion.
C. Emissions and Noise
Since the 1960s, both the chemical emissions and the noise of jet
engines have been regulated to improve well being around airports,
with regulations becoming increasingly stringent over time. Noise
‡
To reduce weight, the Wright brothers used aluminum for their first engine.
Aluminum was then 30 times more expensive than steel. Titanium is now
about 30 times more expensive than steel.
 904 EPSTEIN

has been the bane of aviation from its inception over 100 years ago
[11] and continues to be so to this day. Takeoff and landing noise in
the immediate vicinity of an airport is regulated; cruise noise is not.
Lack of viable noise-reduction technology has been a recognized
barrier to the introduction of commercially viable supersonic
transportation since the 1960s.
Currently, oxides of nitrogen (NOx), particulates, and unburned
hydrocarbons are regulated during landing and takeoff. Gas turbines
inherently produce much less NOx than internal combustion (IC)
engines, so that IC engines need considerable exhaust treatment such
as catalytic converters. Fuel cells that operate from hydrogen or
methanol produce no regulated emissions. Fuel cells that internally
produce hydrogen from hydrocarbons, such as solid oxide cells or Fig. 5 Fuel price and wide-body airplane cash operating cost in then-
year USD.
fuel reformers, operate at higher temperature and may produce NOx
and unburned hydrocarbons, but this area has yet to see much study.

D. Reliability and Maintainability

Downloaded by UNIVERSITA DEGLI STUDI DI ROMA on October 29, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J052713

Reliability and maintainability are important measures of airplane

engine value. The first influences safety, while both influence
operating cost. One measure of reliability is in-flight shutdown
(IFSD) rate. This metric has improved dramatically, by a factor of
200, over the past 50 years; see Fig. 4. Extended operations requires
an IFSD rate better than 0.020 shutdowns per 1000 h of operation.
Today’s state of the art (SOA) is better than 0.002. Time between
overhauls and time on wing are useful measures of maintainability.
These, too, have improved from 400–800 h in the days of the large
piston engines to 6000–14,000 h today. Now, engines may stay on the
wing seven to 10 years before they need be removed for overhaul.

E. Engine Economics: Cost, Price, and Value

Engine-related costs are one of the most important factors affecting
the economics of aircraft ownership and airline operations, and so
these costs are an important consideration in engine selection.
Engines account for about 15–20% of the list price of a new aircraft,
over 50% of the maintenance cost, and of course, they determine the
amount of fuel burned. Therefore, operators’ cost is always a major
design criterion for engine designers. Researchers often do not
consider product cost because of the difficulty of connecting it to
engineering fundamentals and the paucity of available data.
For many decades, fuel was significantly less than $1 per gallon.
One widely used airline cost measure is cash airplane-related
operating cost (CAROC), which includes fuel, airframe and engine
maintenance, crew costs, fees, and ground handling but excludes
capital-related costs. Figure 5 shows the spot jet fuel price over the
past 20 years and illustrates wide-body aircraft operating cost as a
function of that fuel price. At $0.50 per gallon, the fraction of
CAROC attributable to engines is 22%. This rises to 60% at $4 per
gallon. In the past five years, fuel has been as high as $5 per gallon. An
extrapolation of CAROC to prices well above the historical record is
shown in Fig. 6, suggesting that high fuel prices may overwhelm
other considerations. Prediction of future aircraft fuel prices is well
beyond the capability of this author. However, if fuel prices continue
Fig. 6 Wide-body cash operating cost as a function of fuel price in 2012
USD (“other” costs include flight crew, insurance, and landing,
navigation, and ground fees).
above $3 per gallon, then the historical balance among operating
costs remains disrupted, and fuel consumption will continue as the
overriding economic concern.
The cost of manufacturing a jet engine and its list price scales with
engine sea-level static (SLS) thrust. Figure 7 shows an estimate of the
list price per unit of thrust of commercial jet engines over an order of
magnitude in engine size. Prices range from about $200 to $400 per
pound of thrust. The smaller engines are more expensive because
items such as an electronic fuel control are needed independent of
engine thrust. At the very high thrust size, mechanical scaling is
unfavorable such that engine weight per unit thrust rises. To keep the
weight of large engines under control, more expensive construction is
used, such as hollow metallic or composite fan blades. Also, the
largest engines power very long-range aircraft, which are most
sensitive to fuel price, so that a reduction in fuel burn may offset an
increase in engine cost to the owner.

Fig. 4 Evolution of aero engine reliability.

 EPSTEIN
Fig. 7 Bare engine list price per pound of thrust.
Downloaded by UNIVERSITA DEGLI STUDI DI ROMA on October 29, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J052713
The economics of aircraft manufacturing constrain the fraction of
the airplane cost that aircraft manufacturers have been willing to
allocate to the engines. Cross plotting engine list price with airplane
list price reveals that the engines on current commercial aircraft are
about 15 to 22% of the price of the airplane. This ratio has been
constant since at least the 1980s.
The price of new aircraft has been constrained by competition, the
availability of used aircraft, and airline economics. Since the mid-
1960s, corrected for fuel price variation, new technology has dropped
the operating cost of narrow-body aircraft by 30 ∼ 40%, but net
aircraft selling prices have remained constant. Technology has added
value but this value has not been recovered by aircraft manufactures
and their engine suppliers in the form of higher prices. Therefore,
innovations and technology that raise engine production cost are
avoided by industry. This suggests that new technology should avoid
adding net cost to manufacturing an engine.
V. Aircraft Engines of the Future
Appropriate metrics for aircraft engines are efficiency, weight,
emissions, noise, and reliability. In all of these, the large aircraft gas
turbine is unmatched, with no successor on the horizon. Thus,
hydrocarbon-fueled, Brayton cycle-driven propulsors appear to be
the most promising approach for commercial aeropropulsion over the
next few decades. What this means for specific aeropropulsion
research directions is dependent on application, engine thrust class,
and design choices. Indeed, the interplay between the clever designer
and the insightful researcher is perhaps the least appreciated dynamic
in propulsion. Design approaches can determine the relative value of
a research topic. Designers can obviate, or at least delay, the need for
fundamental understanding. The World War II German designers
who used turbine air cooling in the Jumo 004 because they did not
have access to high-temperature materials illustrate this point.
Another example is that a fundamental understanding of nacelle drag
is much more important for a high-bypass-ratio turbofan than for a
turboprop of the same thrust because the turboprop’s much smaller
nacelle is a relatively minor factor in propulsive performance. Thus,
the relative importance of a technology is often very dependent on
design approaches and engine architecture. The converse is true as
well; a good designer designs from strength and eschews approaches
that are poorly understood.
One example of how design approach can influence research
directions concerns takeoff and landing approach noise. The exhaust
jet has been the major takeoff noise source, and so it has been the
focus of considerable research effort and resulting literature since the
1960s. Although this research has resulted in greater understanding
of the physical processes involved, it has not resulted in significant jet
noise-reduction technology. Nevertheless, jet noise is no longer the
dominant noise source. Figure 8 illustrates the relative magnitude and
direction of important turbofan engine noise sources as they have
evolved over 40 years. This evolution resulted from technologies that
have enabled low fan pressure ratios and the resulting high bypass
ratios (BPRs). For the most modern designs in the 10 to 12 BPR
range, the jet exhaust velocity is reduced so that it is largely irrelevant
905
Fig. 8 Turbofan noise source evolution.
to takeoff noise, which is now dominated by fan noise. Thus, research
on jet noise is no longer warranted for this purpose. On approach and
landing, the engine noise is now less than that of the airframe in some
cases, suggesting that noise researchers may be wise to focus mainly
on fan and airframe noise.
In light of this background, the following sections consider the
current state of the art and speculate on future design directions and
the research necessary to realize them. Although predicting the future
is an inexact art, thermodynamics is quite clear. We know that
improved thermal efficiency will demand higher cycle pressures and
temperatures, improved component efficiency, and reduced cooling
and secondary air. We know that increasing propulsor efficiency
requires low-pressure-ratio propulsors with low drag nacelles and
perhaps variable geometry blading or exhaust nozzles. All of this
must be accomplished at weights and overall costs that do not
outweigh the advantages of improved efficiency. We know where we
must go with some clarity. How to get there requires research.
A. Propulsors
Commercial aircraft built over the last 50 years have been gas
turbine-powered, and either turbofan- or propeller-propelled. At the
most basic level, the differences are the total fan pressure ratio (FPR)
produced across the rotor (FPR) and whether the rotor operates in a
duct or in free air. The pressure ratio determines the propulsor exhaust
velocity and therefore the propulsive efficiency. It also sets the
propulsor diameter. For example, at the 25,000-30,000 lb takeoff
thrust level, a currently flying turbofan engine with a FPR of 1.7 has a
rotor diameter of about 1.6 m. Reducing the FPR to 1.2 at constant
thrust grows the rotor diameter to 2.3 m. A two-rotor, contra-rotating
propeller is 4.3 m in diameter, while a single-rotation propeller needs
a 5.2 m diameter to produce the same thrust. Clearly, engineering
considerations for these configurations may be different in detail.
We define propulsion efficiency, as is commonly done for
propellers, as the thrust power delivered to vehicle (thrust FNfan times
flight velocity V 0 ) divided by the mechanical power input to the shaft,
SHPfan . Figure 9 shows the variation in fan stream propulsive
efficiency, nPfan , with fan pressure ratio, FPR, at a flight Mach
number of 0.80 [12]. Three curves are shown: the ideal relation
between FPR and propulsive efficiency (“ideal” solid line), a curve fit
to practical designs for which the overall propulsor geometry was
optimized for each pressure ratio (“actual” dashed line), and a curve
Fig. 9 Propulsive efficiency scales with fan pressure ratio.
 906
with fixed component losses in the fan stream. For these purposes,
FPR is defined as the total pressure ratio across the rotor and stator of
a fan and across both rotors of a contra-rotating prop. Ducted
configuration total pressure losses (subsequently referred to as
“loss”) include those from the rotor, stator, inlet, duct, and nozzle, as
discussed next. Effects of external nacelle drag are not included in
this plot. Propeller losses are principally in the rotors and, in the case
of single-rotation props, residual swirl. Propeller rotor adiabatic
efficiencies are well below those of ducted fans, but the overall
propulsive efficiency is higher, mainly due to lower exhaust velocity
but also because there is no inlet, stator, duct, or nozzle to add their
losses. Note that the diameter of the propulsor must grow as FPR
drops to maintain constant thrust.
To illustrate the importance of internal component losses on ducted
propulsor design, the line labeled “fixed component losses” in Fig. 9
is an example of impractical designs. On this curve, as FPR is reduced
from the reference design point; the propulsive efficiency first
Downloaded by UNIVERSITA DEGLI STUDI DI ROMA on October 29, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J052713
increases due to reduced fan nozzle exhaust velocity. At a sufficiently
low FPR (about 1.4 in this example), this benefit is overwhelmed by
the fixed internal component losses (inlet, rotor, stator, duct, leakage,
and nozzle), and so propulsive efficiency begins to drop.
For practical designs (Fig. 9, dashed line), the dominant loss
mechanisms change with FPR. Figure 10 shows the percentage
change in net thrust, Fnet , attributable to different loss sources in the
fan flow path. As the design FPR is reduced, the rotor tip speed can be
reduced. This reduces rotor total pressure loss, which scales with tip
speed to a power greater than 2. Although the rotor loss reduces with
decreasing pressure ratio, the other losses (inlet, duct wall, nozzle,
etc.) do not if their geometry remains the same. Thus, nonrotor losses
become proportionately more important as FPR drops. The designs
underlying the actual curve of Fig. 9 assume that, as rotor pressure
ratio drops, 1) fan rotor efficiency increases, 2) inlet and exhaust duct
length-to-diameters reduce, and 3) improved nacelle sealing reduces
leakage. This is illustrated in Fig. 10, which shows that, at high fan
pressure ratios characteristic of turbofan engines in service today
(FPR ∼ 1.7), rotor total pressure loss dominates. As the FPR is
dropped, however, other flow-path losses become more important.
The importance of internal flow path loss relative to external nacelle
drag is also sensitive to the FPR, with nacelle drag dominating at very
low FPRs (Fig. 11).
With the previous background, we can see that there are several
engineering challenges associated with lowering fan pressure ratio to
improve propulsive efficiency. At FPRs below approximately 1.45,
fan operability is reduced because the fan exhaust nozzle unchokes
over part of the flight regime, which challenges fan operability,
flutter, and stall [13]. Solutions can include variable geometry
(variable nozzle area or rotor-blade pitch change) and improved
damping for aeromechanics and stall and surge. A geared turbofan
engine scheduled to enter service in 2014 will be the first commercial
engine designed with a variable-area fan nozzle [14]. Variable pitch
approaches have been demonstrated as well [15].
Another class of challenges is the reduced propulsor flow-path
length needed to keep wall viscous loss, nacelle drag, and engine
weight tolerable. Shortening both the inlet and fan exhaust duct
reduces this loss, as may wall treatments or geometries that reduce
drag. But shortening inlets can increase fan inlet distortion.
Fig. 10 Loss source variation with design FPR.
EPSTEIN
Fig. 11 Nacelle loss and drag trends with FPR.
Therefore, improved inlet geometries and distortion tolerant fans are
needed. Shortening the inlet and exhaust also reduces the area
available for the acoustic liner used for noise attenuation. However,
fan noise correlates with fan-tip speed to the fitth power, all else
remaining the same. Thus, as the FPR is reduced and the fan-tip speed
drops, less noise is generated. Less noise requires less acoustic
attenuation. In implementations to date, lower rotor tip speed also
reduces weight of the rotor, rotor case, and rotor mounting system,
even taking into account the growth in rotor diameter that results from
lowering FPR at constant thrust.
The third challenge for lower FPR is thrust reversing. Although not
required by regulation, reverse thrust is a desirable feature. As the
exhaust duct is shortened concomitant with low-pressure-ratio fan
design, current cascade reverser designs become the length limiting
element. New approaches are needed here, including perhaps
fuselage-mounted reversers.
In summary, improved propulsive efficiency will require low
fan pressure ratio designs. Enabling technologies needed include
1) fan aeromechanics enhancement, 2) fan surge and stall stability
improvement, 3) fan rotor and stator loss reduction, 4) duct wall loss
reduction, 5) inlet distortion mitigation and tolerance, 6) fan noise
reduction, 7) thrust reversing, 8) nacelle drag and weight reduction,
and 9) weight reduction. The last item, weight reduction, can stem
from lower specific weight materials such as composites and advance
alloys, as well as design innovations that reduce nacelle, case, rotor,
and rotor mount weights. Except for the omission of jet noise, this list
is little different from what it would be had it been written 30
years ago.
B. Propellers
Propulsors without ducts are propellers. The importance of
propeller efficiency was understood early on, and props with
efficiencies over 81% were offered commercially by 1911 [16]. Since
then, we have greatly increased the flight speed and loading
capability of propellers, introduced variable pitch, and increased
service life but have only modestly improved efficiency. Propellers
have been used in two configurations over the past century.
Single-rotation propellers have one blade row, while contra-rotating
propellers (CRPs) consist of two blade rows turning in opposite
directions. The contra-rotating configuration has the advantages of
requiring a smaller diameter for the same thrust and of removing the
swirl left in the flow by a single propeller, thus improving the
propulsive efficiency when mounted behind a wing (the wing
removes most of the swirl in the most common tractor configurations
used today). The advantages of the contra-rotating configuration
come at the cost of increased complexity, weight, noise, and cost.
Although single rotation is much more common now, contra-rotating
approaches have been used on many airplanes since 1930 [17].
Contra-rotating propellers for high flight Mach numbers were first
explored in the 1980s and were then called propfans or unducted fans.
The term currently in favor is open rotor.
Propellers eliminate many of the nacelle loss and weight concerns
associated with ducted systems but introduce their own challenges.
There is still a nacelle, but it is much smaller in diameter, and so
 EPSTEIN
weight and drag are less of an issue. There are no duct or nozzle losses
because there are no ducts or nozzles, but propellers have lower
adiabatic efficiencies than ducted fans due to aero and mechanical
constraints. A ducted fan commonly operates with a choked fan exit
nozzle. This adds the loss of a nozzle but enables the optimization of
the airfoils for aerodynamic performance because the rotor inlet axial
Mach number is largely independent of flight speed. Over its flight
regime, a propeller sees a much larger variation in flow angle because
both the axial and tangential velocity change significantly with
flight condition, and so high-speed propellers require a pitch change
mechanism. Propeller blades must be thicker than fan blades,
especially at the root, which impacts aerodynamics. One reason is
that the propeller blades must turn the flow axially when an engine is
at an angle of attack as at rotation, and thus the propeller rotor and its
support system must carry that load. (In a turbofan, the inlet turns the
flow before it reaches the fan.) Another mechanical concern is blade
retention and failure because there is no containment as there is in a
Downloaded by UNIVERSITA DEGLI STUDI DI ROMA on October 29, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J052713
turbofan. These factors combine to drop the adiabatic efficiency of a
propeller compared to that of a fan rotor.
Figure 12 compares the variation of propulsive efficiency with
flight speed for ducted and unducted systems, using the efficiency
definition of Fig. 9. The three bottom curves are ducted fans,
representing ducted FPRs characteristic of current, near term, and
longer-term designs. Their propulsive efficiency increases mono-
tonically with flight Mach number. The performance of a
representative contra-rotating propeller is also shown. The CRP
efficiency improves with speed up to its design point, M  0.75, and
then drops quickly with increasing flight speed [18]. A CRP can be
designed for a higher flight speed than shown, but its efficiency will
be less at all lower speeds. The performance of an in-service, large,
SOA single-rotation prop is included for reference. Figure 12 clearly
shows that there is a trade between efficiency and design flight speed
for propellers.
It is important to note that, in the example illustrated, the efficiency
advantage of the CRP decreases as flight speed is increased, from
27% at takeoff (M  0.25) to 12% at cruise (M  0.74). During a
500 n mile flight, about half the fuel is burned during takeoff and
climb. This implies that, all other things being equal, a prop solution
will be more attractive at short ranges than long ranges where high
flight Mach numbers are desired and most fuel is consumed in cruise.
There are several challenges for propellers. The first is noise,
which is a concern both for regulated takeoff and landing operations,
and en route, where it affects passenger comfort and structural
fatigue, in addition to impact on communities under the flight path. It
is not that propellers necessarily make more noise than fan rotors; it is
that they have no duct to block and attenuate the sound they produce.
The high-speed contra-rotating props first demonstrated in the 1980s
generated very high noise levels at both takeoff and cruise [19].
Recent research suggests that it may be possible to reduce CRP
takeoff/landing noise levels to 5–10 EPNdB cumulative below
current Chapter 4 [20] regulations. However, these levels are about 10
EPNdB higher than that of modern ducted engines, levels that should
go down as FPRs are reduced. Also, cruise noise remains an
unresolved issue. Mounting the props at the rear of the aircraft,
behind the cabin, keeps the prop wakes off the wing and reduces
Fig. 12 Variation of fixed propulsive efficiency with Mach number.
907
cabin noise but introduces additional noise sources. One such source
is the engine pylon wake as it passes through the pusher props. Active
noise suppression has been used to reduce cabin noise but has
not been shown to be effective for community noise. All things
considered, one would expect that the propellers will continue to
incur noise penalties compared to ducted engines of contempora-
neous design.
The second suite of challenges is the weight, performance, and cost
penalties that may accrue when meeting upgraded propeller safety
standards. As the number of aircraft increases, the safety-relevant
incident rate must decrease to maintain and improve the now very
impressive safety record of well run airlines. To this end, authorities
are reviewing propeller certification standards [21]. Issues under
study include robustness to blade release and to foreign object
ingestion, including birds, ice, and tire debris. The latter two are of
particular concern for pusher aircraft configurations in which contra-
rotating props are mounted at the aircraft’s tail. The solution to the
blade release challenge most likely lies in the realization of prime-
reliable blading that meets all other constraints. Historically,
propeller pitch change mechanisms have been a major source of in-
flight problems and a maintenance cost issue. A CRP approach with
two rotors doubles the number of pitch change mechanisms, and so
progress in this area would be useful.
The relative merits of ducted verses unducted configurations are
highly dependent on both application and technology. One would
expect propellers to continue to be an attractive, fuel-efficient
solution for short-range aircraft at smaller sizes where speed is less of
an issue and the absolute noise levels are less. Aerodynamic,
acoustic, and mechanical technologies can increase the range and size
at which a propeller solution is attractive.
C. Motors
Because there are no candidates on the horizon that have the
promise to displace the gas turbine as the aircraft motor (e.g.,
producer of shaft power) of choice for commercial aviation, the
continuing evolution of the gas turbine is of particular concern. In
addition to safety, the most important motor design considerations are
thermodynamic efficiency, weight, and cost.
Brayton cycle thermodynamics are very well understood [5].
Improving thermodynamic efficiency centers mainly on increasing
the cycle overall pressure ratio (OPR), turbine rotor inlet temperature,
and component efficiencies as well as on reducing compressor
bleed air used mainly for turbine cooling. The latter can be eliminated
as an independent parameter if compressor efficiency is defined as
flow work used for propulsion divided by the compressor shaft work.
Then, the bleed air appears as a debit to compressor adiabatic
efficiency.
Authors over the years have discussed challenges to improved
thermodynamic efficiency, which relate to increasing OPR, and
component efficiency (and reducing bleed air) [13]. First, the
increase in the compressor exit air temperature T 3 as OPR increases
leads to mechanical design challenges at the rear of the compressor
and in the high-pressure turbine because turbine cooling air
temperature also rises with OPR. Second, increases in compressor
and turbine adiabatic efficiencies are needed to justify increased
OPR. Third, combustor emissions such as NOx increase with OPR,
all else remaining the same. Over the past 50 years, T 3 has increased
by about 5°C per year as illustrated in Fig. 13. Such progress is
becoming increasingly challenging.
Research needed to address these challenges now focuses new
attention to traditional solutions, including new and improved high-
temperature structural materials including metals and ceramics;
improved cooling and thermal barrier coatings; design or architecture
approaches such as disk design details and cooled cooling air; and
reduced emissions combustor schemes.
Researchers are also revisiting intercooling and recuperation,
both of which can be used to improve the thermodynamics of an
engine but which bring challenges associated with necessarily large
heat exchangers: weight, packaging, reliability, and cost [13].
Heat exchangers suitable for aircraft engines have been intractable
 908
Fig. 13 Evolution of commercial engine compressor exit temperature
Downloaded by UNIVERSITA DEGLI STUDI DI ROMA on October 29, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J052713
T3 .
problems over the past decades. The previous challenges to
increasing thermal efficiency have not changed over the past 30 to 40
years. Only quantitative values and solutions are different.
What is new is the challenge of size now felt at the narrow-body
end of the commercial aircraft design space. As overall airplane and
engine efficiencies improve, the power needed from a motor for a
given payload and range aircraft drops. Also, as overall pressure ratio
increases, air density at the exit of the compressor increases. These
two factors combine to significantly reduce the core size, expressed
as a compressor exit flow area function. This is shown Fig. 14, whose
curves are fits to data. Engines entering service in 2014–2016 will
have about 10 times less compressor exit flow area than their
predecessors from the early 1970s.
Because turbomachinery efficiency is known to scale with size
[22], this introduces a new challenge for 20,000–30,000 lb thrust
class engines for future narrow-body aircraft: how to realize high-
pressure compressor and turbine efficiencies at very small sizes. This
problem is exacerbated if the current engine architecture of
concentric spools and shafts is maintained. Smaller, more efficient
motors are associated with larger, slower-turning, more efficient
propulsors, which means that increased torque must be transmitted
through the compressor to the fan, necessitating relatively large bores
in the rear compressor disks. This constraint implies a narrow annulus
at the back of the compressor with very short blades compared to
those in today’s engines. Engine architectures other than those in
common use today may be attractive [23].
Today, very small engines do not operate at the overall pressure
ratios needed for state-of-the-art thermal efficiency because current
small compressors do not have high pressure ratios. The correlation
between size and OPR for some in-service engines is illustrated in
Fig. 15, which spans two orders of magnitude in thrust. A simple
explanation for the difference in OPR between small and large
engines is that the largest engines are used in longer-range wide-
body aircraft for which operating cost (especially fuel consumption)
is the primary driver. The smallest engines are used in general and
commuter aviation for which purchase price has been relatively more
Fig. 14 Technology increased efficiencies and reduced core size.
EPSTEIN
Fig. 15 Core size of some current engines.
important, and so the construction and material choices are much
more constrained by manufacturing cost.
Small turbomachinery with performance at the state of the art
of large machines does not now exist. Can it? Certainly, higher
OPR machines can be built at small sizes if cost constraints are
relaxed. But, to improve thermal efficiency, the turbomachinery
adiabatic efficiency must be increased concomitantly. This is an area
in which research may yield considerable benefit. A design system
prediction of the change in turbomachinery efficiency with size
shows a 3% change in efficiency for a factor of 7 change in
turbomachine size (Fig. 16). This change reflects factors including
design, manufacturing, and technology-dependant constraints such
as clearances, manufacturing tolerances, etc. The primary aerody-
namics constraint of size is viscous scaling, Reynolds number, which
is only 10% of the total small size penalty. The design constraints on
turbomachinery for advanced narrow-body engines will be different
than for business jet engines of equivalent core size because, in the
case of the narrow-body engines, the small compressor exit area
comes from high OPR (50–70), and so airfoil Reynolds number is not
a problem for these machines. Rather, most of the challenges lie in the
realm of mechanical design and manufacture. This implies that
research on topics focused on achieving high-efficiency, high OPR,
small turbomachinery may be of great value.
Technology goals for high-pressure-ratio compressors are
examples of how design choices, engine architecture in particular,
can set the agenda for research direction. Commercial engines now in
service consist of either two or three concentric spools running at a
different rotational speeds (Fig. 17). For two-spool engines, the inner
spool is referred to as the high-pressure (HP) spool, or the core, or the
gas producer, while the outer spool is known as the low-pressure (LP)
spool. In three-spool engines, the inner is the HP spool, the next is the
intermediate-pressure (IP) spool, and the outer is the LP spool. Most
engines are direct-drive, i.e., the low-pressure spool components
(fan, LPC, and LPT) all turn at the same rotational speed. In a direct-
drive two-spool engine, generation of the fluid power to run the
propulsor is split between the HP and LP spools. Because the LP
spool is constrained to run at low speed by the directly coupled,
relatively large-diameter fan, it is most economical to place as much
of the engine pressure rise on the HP spool as compressor technology
permits. No such economy exists for a direct-drive three-spool engine
Fig. 16 Component efficiency and core size trend.
 EPSTEIN
Downloaded by UNIVERSITA DEGLI STUDI DI ROMA on October 29, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J052713
Fig. 17 Turbofan architectures.
where both the IP and HP operate at high speed, and so here pressure
ratio is moved off the HP spool.
The geared two-spool turbofan engines entering service in the
2014–2016 period place a reduction gear on the low spool between
the fan and LPC so that the LPC and LPT can turn three to four times
faster than the fan. By operating at higher speeds, the LPC and LPT
require fewer stages, the deletion of which more than compensates for
the weight and cost of the gear. The optimum work split between the
spools of such an engine is different from that for a direct-drive two-
spool arrangement because as much work can be done per stage with
a high-speed LPC as with a HPC. Aerodynamically, the LPC in a
geared engine more closely resembles an IPC in a three-spool engine
than it does the LPC in a direct-drive two-spool engine. It is desirable
to move work from the HP spool to the LP spool to reduce overall
stage count, to improve operability, and to increase efficiency
(because the LPT is uncooled, it has higher efficiency than the HPT).
These considerations imply that the optimum pressure ratio split for
geared engines lies between that of the direct-drive two- and three-
spool engines (Fig. 18). Thus, research enabling the increase in HP
pressure ratio is of much greater value for direct-drive two-spool
engines than for three-spool or geared engines.
D. Some System-Level Considerations
Aircraft engines were considered as consisting of a motor and a
propulsor in the previous discussion. In practice, these are very
highly integrated and optimized systems, and so many important
issues are properly considered at the overall system level. One such is
throttle range, or operating point variation. Although defined
differently by different organizations, throttle range generally refers
to the variation in throttle angle (old definition), corrected thrust,
Fig. 18 Engine architecture influence on spool pressure ratio.
909
turbine inlet temperature, or rotor corrected speeds over an aircraft
mission (often excluding idle). It is a measure of the range of
operating conditions through which an engine must operate and is an
important consideration in the aerodynamic design of components,
part cooling, setting part life aeromechanics, and rotor dynamics.
Currently, commercial gas turbines designs are optimized around one
operating point, typically cruise efficiency, because that is where
most of the fuel is burned. The optimum cruise efficiency may not be
reached due to compromises needed for operation at “off-design”
(i.e., other throttle positions). For example, maximum turbine cooling
air is often needed at takeoff, implying that valves in the cooling
system to reduce cooling air at cruise could increase cruise efficiency.
Although considered for decades, modulated cooling air has yet to
see wide use due to cost, weight, and reliability concerns. Variable-
area turbines are another technology that offer off-design benefits but
whose benefits have yet to outweigh debits.
Operating-point variation is a function of engine cycle. As the
design fan pressure ratio (FPR) drops, the needed throttle range
of a commercial engine also drops. This is illustrated in Fig. 19,
which shows the change in turbine inlet temperature (TIT) for engine
cycles of equivalent technology optimized for different FPRs and
normalized to the same TIT at aircraft rotation (sea level, M  0.25).
At FPR  1.7, typical of many aircraft flying today, sea-level static
takeoff power and rotation power settings are the same and are the
highest temperatures the engine must endure. Climb and cruise
powers are 130 and 300°C cooler. In terms of TIT, the throttle range
for this engine is 300°C. For a 1.3 FPR that an advanced geared
turbofan might adopt a decade from now, takeoff rotation is the
highest cycle temperature, while sea level static is 60°C cooler. Climb
and cruise powers are 100 and 240°C cooler. Overall, the throttle
range requirement of the low FPR cycle is about 20% less than that
needed by the high FPR cycle (240°C versus 300°C). This implies
that technologies such as variable turbine area A4 and modulated
cooling air decrease in importance as FPR drops, all else remaining
the same.
Another example of a system-level consideration is waste heat
rejection. An aircraft generates significant waste heat from sources
such as electric generators, avionics, and environmental control. An
engine generates waste heat from bearings, seals, disk windage, etc.
Rejecting this heat has and will become increasingly challenging
because 1) there is more waste heat per unit of cruise thrust, and 2) it is
more expensive in terms of performance and weight to reject each
unit of heat.
Heat is rejected from the engine and the airplane it powers through
heat transfer to either fuel or air. Heat rejection to fuel is preferred
because such heat exchangers are relatively compact, do not add drag,
and the otherwise waste heat reduces the amount of fuel burned
(albeit only slightly). But the heat loads do not change very much over
an aircraft mission compared to the fuel flow. Cruise and idle descent
fuel flows are only 30 and 5% of max takeoff, and so even when there
is adequate fuel flow at takeoff to absorb all heat, over much of the
flight regime there is not. Both the engine and the aircraft must share
the available fuel heat capacity. One approach used is to circulate
Fig. 19 Throttle range required for engines optimized at different FPR.
 910 EPSTEIN
sufficient fuel through the heat exchangers to absorb the heat and
return that fuel not consumed by the engine to the aircraft tanks. This
technique is less effective on composite aircraft because the thermal
conductivity of organic composite is lower than the aluminum it
replaces, and so less waste heat can be dissipated from the tanks to
ambient air. Advanced fuels such the military JP8
100 and some
new biofuels [24] can help because they are stable to higher
temperatures and so can absorb more heat, but currently engines must
be designed for the worst fuel they may encounter, not the best.
When there is more heat to reject than the fuel system can
accommodate, then waste heat must be rejected to air. For a turbofan
engine, oil–air heat exchangers are typically placed in the fan duct.
This increases fan exhaust temperature but generates total pressure
loss in the fan flow. Overall, the exhaust velocity and thus the thrust
are reduced. Propellers have very low pressure ratios, and so
turboprops may add auxiliary blowers or ejectors to push air through
their heat exchangers. They therefore incur the largest heat rejection
Downloaded by UNIVERSITA DEGLI STUDI DI ROMA on October 29, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J052713
penalties. Heat rejection on modern aircraft is no longer a secondary
design concern.
Trends in aircraft and engine design will challenge both the
generation and rejection of waste heat. Heat sources scale differently
with engine size. All else the same, a smaller engine generates more
waste heat per unit of thrust than a larger engine because it turns
faster. Bearing and viscous losses increase linearly with rotational
speed; turbulent windage scales with speed to the 1.8 power. Also,
increasing engine efficiency drops engine core size, and so a more
efficient engine will generate relatively more waste heat. A more
efficient engine consumes less fuel, and so more heat must be rejected
to air. A more efficient airplane requires a smaller engine, which
further exacerbates the challenge. Figure 20 illustrates the variation
of waste heat per unit of thrust variation with core size (compressor
exit corrected flow as used in Figs. 14 and 15). The curve shown is a
fit to engines spanning a five-to-one thrust range. Also shown is the
heat capacity of the fuel flow at cruise. Large wide-body engines have
sufficient fuel heat capacity to absorb the waste heat; smaller narrow-
body engines do not. Overall, engine and aircraft heat rejection will
become increasingly challenging as we build more efficient airplanes
and engines.
VI. Concept Whose Time Has Come?
The Oxford English Dictionary defines “innovates” as “makes
changes in something established, especially by introducing
new methods, ideas, or products”. Are there new challenges and
opportunities for commercial propulsion or only old problems still
unsolved? Each year for the last six decades, thousands of engineers
have spent billions of dollars a year to make better jet engines. Even
so, there has certainly been considerable research work over the
decades that, for one reason or another, has not made its way into jet
engines. But engineering capabilities generally improve with time,
and what was an unsuccessful concept a few years or decades ago
may be enabled by new technologies or needed to meet new
requirements.
Fig. 20 Waste heat rejection capability.
Harry S. Truman is quoted as saying, “There is nothing new in the
world except the history you do not know.” The need for technologies
to enable high-performance small cores as discussed previously
implies that an old concept may come back. In 1957, Edkins and
Thorson proposed using several small engines in place of one large
engine [25]. They noted that history up to that point was going in the
opposite direction (i.e., engines were getting larger). The authors
enumerated their view of the advantages of small engines, including
higher thrust-to-weight, lower overall cost for smaller airplanes, and
synergies with boundary-layer control and supersonic flight. Among
the challenges they noted then were weight of accessories and overall
performance. Thus far, not much has come from their suggestion.
Engines have continued to grow. In 1957, 10,000 lb of thrust was
considered large. Now a large engine may produce over 100,000 lb.
Also, the number of engines on a commercial jetliner has generally
shrunk, from four on the B-707 to the much more common twin-
engine aircraft of today. The drivers for this trend have been both
performance and cost. More recently, researchers have pointed out
the synergy between distributed propulsion and multiple-engine
architectures [26].
It may be time to re-examine the concept of many small engines,
not so much as a technical innovation but as a business one enabled by
some of the technology trends discussed earlier. For many years, the
commercial aviation community has been wrestling with the question
of engines as a commodity. If there is a choice of two or three different
engines on a particular airplane, is the selection strictly a commercial
decision or is it possible to technically differentiate one engine
from another? Now, engines are designed for specific thrust levels
for specific airplanes. This is not the only alternative. Consider an
approach in which, at its limit, thrust is truly a commodity in that there
is only one size commercial engine manufactured, in the range
of 5,000 to 20,000 lb of thrust. Two engines power the smallest
commercial airplanes, while many are used to power the largest.
Aircraft manufacturers could design to a common interface
specification that could be met by different engine producers, so that
engines were interchangeable on a given airplane. Is this a viable
concept, an attractive concept?
For commercial aviation, attractiveness is about value for the
purchaser: an airline or lessor. Value can be generated in many ways.
The most important may be via reduced purchase price and reduced
cost of ownership: reduced fuel consumption, improved reliability,
and reduced maintenance cost. Purchase price reductions would
come from reduced cost of manufacture. Now, a few tens to a few
hundreds of any one engine model are manufactured per year, with
many models produced to satisfy the market demands of various
aircraft types. If one small engine size satisfied all aircraft, then
30,000–100,000 engines of that type would be manufactured each
year. This production volume justifies different manufacturing
paradigms that could reduce unit cost. Also, current engine
development costs are quite large and production quantities small,
factors that favor incumbents and inhibit the entry of new players.
The change from fragmented design niches to a unified market might
attract new manufacturers.
Fuel is the major cost of commercial airliners. Current large
engines have better fuel consumption than small engines. Much of
this difference is design intent. Large engines are designed for long-
range aircraft for which fuel consumption is the overriding
consideration. Small engines designed for short range commercial
aircraft must deliver the higher number of flight cycles needed for
acceptable overhaul intervals, which constrains overall pressure
ratio. Small engines designed for business aircraft are constrained by
purchase price, which is a more important consideration for these
relative low-utilization aircraft than fuel cost. Simply put, we do not
now design small engines to the same efficiency as large engines.
Could we?
Small engines could be designed for much better fuel consumption
with current technology if manufacturing cost were not a constraint.
With current approaches, cooling is more difficult at small turbine
size, and so the high temperatures needed in the core would need to be
met with better (and more expensive) materials and cooling
technologies. Smaller parts can yield superior material properties and
 EPSTEIN
are also better suited to brittle materials such as high-temperature
ceramics. The lower weight advantage from the cubed-squared law
that Edkins and Thorson discussed [25] could be put to advantage to
increase fan diameter and drop fan pressure ratio to improve
propulsive efficiency. A conceptual study showed this to be a
promising approach [27]. A full preliminary design including
installation would be needed to assess the fuel burn potential relative
to a large engine. There is considerable scope for research to improve
the performance of these small engines. Indeed, including installation
effects, it may be possible for a small-engine-powered aircraft to burn
less fuel than a more conventional design. In addition to the topics
discussed earlier, manufacturing technologies capable of meeting
design requirements both at small sizes and in large volumes could be
fruitful.
Maintenance cost is another major expense for the aircraft
owner. Overhaul is most of that cost and most of the overhaul cost is
parts. The financial margin on parts is considerable. Currently, the
Downloaded by UNIVERSITA DEGLI STUDI DI ROMA on October 29, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J052713
selling price of engine parts is set by overall enterprise financial
considerations, which reflect that the high development cost is
amortized over relatively few units and that many engines are sold at
or below unit manufacturing cost. To first order, selling more parts at
a lower margin could raise the same revenue. Also, as is clear from
Figs. 5 and 6, should fuel prices stay high or rise further, maintenance
cost will be much less important than it has been historically. New
financial models for the engine business are beyond the scope of this
discussion.
Another aspect of operator value is reliability. At the same level of
reliability, more engines would result in more in-flight shutdowns, all
else being equal. But all else need not be equal. With sufficient
numbers of engines, a mission can be completed or even dispatched
with an engine inoperative. This question needs careful analysis.
Multiple small engines may yield aircraft-level benefits, especially
for larger aircraft. An increase in engine numbers lessens the impact
that an engine failure places on aircraft design considerations such as
tail sizing. Multiple engines may also be beneficial for airframes
configured for boundary-layer ingestion.
The major point of this discussion is that the need for improved fuel
economy will continue to drive down the size of the engine cores for
narrow-body commercial aircraft. The enabling technology for this
imperative may transform an old idea in to a new reality, answering
the question, “Why small engines?”
VII. Conclusions
The future of commercial aircraft will be driven by fuel
consumption and environmental concerns. There is nothing on the
technical horizon that threatens to displace the gas turbine as the
engine of choice. These engines will continue to be fueled with liquid
hydrocarbons due to their superior energy density and safety, but the
source of the fuel will most likely switch from fossil deposits to
sustainable biological sources.
As fuel becomes increasingly more expensive, engine archi-
tectures and design details will evolve to reflect the new balance
between engine specific fuel consumption, weight, and manufac-
turing and maintenance costs. As both the thermal and propulsive
efficiencies of the gas turbine are improved, bypass ratios will
grow, and cores will shrink. Although this could be considered the
continuation of a long-term trend, the reality is that engines are
moving into a different, less familiar design space than the 5 to 8 BPR
that characterized the last 40 years of engine experience, with
different needs and constraints. Thus, realignment of research
priorities is needed to address 21st century challenges, dropping
some topics and introducing new ones. Changing research goals and
requirements mean that the future for aeropropulsion enabling
research is rich with challenges and technical opportunities.
Acknowledgments
The author thanks W. K. Lord, K. L. Hasel, J. A. Glahn, and E. M.
Greitzer for many stimulating discussions on the concepts advanced
911
in this paper as well as some of the figures. E. Mitchell helped with
many aspects of data gathering and preparation.
References
[1] Hall, K. C., and Hall, S. R., “Minimum Induced Power Requirements
for Flapping Flight,” Journal of Fluid Mechanics, Vol. 323, No. 1, 1996,
pp. 285–315.
doi:10.1017/S0022112096000924
[2] Tennekes, H., The Simple Science of Flight, Revised and Expanded
Edition: From Insects to Jumbo Jets, MIT Press, Cambridge, MA, 2009,
p. 37.
[3] Garrison, P., “The Head Skunk,” Air & Space, Vol. 25, No. 7,
March 2010.
[4] Norris, G., “Phantom Eye Makes First Flight,” Aviation Week, Vol. 175,
No. 6, June 2012.
[5] Kerrebrock, J. L., Aircraft Engines and Gas Turbines, 2nd ed., MIT
Press, Cambridge, MA, 1992, pp. 464–466.
[6] Carter, N., Stratton, R., Bredehoeft, M., and Hileman, J., “Energy and
Environmental Viability of Select Alternative Jet Fuel Pathways,” 47th
AIAA/ASME/SAE Joint Propulsion Conference, AIAA Paper 2011-
5968, Aug. 2011.
[7] Koff, B. L., “Spanning the Globe with Jet Propulsion,” 21st Annual
Meeting and Exhibit, AIAA Paper 1991-2987, May 1991.
[8] Wärtsilä Rt-Flex96c and Wärtsilä Rta96c Technology Review, Wärtsilä
Switzerland Ltd., Winterthur, Switzerland, 2012
[9] Thorstensen, B., “A Parametric Study of Fuel Cell System Efficiency
Under Full and Part Load Operation,” Journal of Power Sources,
Vol. 92, Nos 1–2, 2001, pp. 9–16.
doi:10.1016/S0378-7753(00)00497-3
[10] Committee on Gas Turbines, National Academy of Sciences, “An
Investigation of the Possibilities of the Gas Turbine for Marine
Propulsion,” U.S. Navy Dept., Technical Bulletin No. 2, Jan. 1941.
[11] Editorial “On the Fitting of Silencers,” The Aero, Vol. V, No. 100,
July 1911, p. 94.
[12] Lord, W., and Sepulveda, D., “Ultra-High BPR Ducted Propulsor,”
Proceedings of the ICAO/CAEP Fuel Burn Reduction Technology
Workshop, London, March 2009.
[13] Cumpsty, N. A., “Preparing for the Future: Reducing Gas Turbine
Environmental Impact,” Journal of Turbomachinery, Vol. 132, No. 4,
Oct. 2010, Paper 041017.
[14] MacIssac, B., and Langton, R., “Gas Turbine Propulsion Systems,” John
Wiley and Sons Ltd., West Sussex, U.K., 2011, p. 263.
[15] Norris, G., “P&W Uncovers ADP Gear Design,” Flight International,
Vol. 141, No. 4320, 1992, p. 11
[16] Advertisement, The Aero, Vol. V, No. 98, May 1911, p. 3.
[17] “Two Propellers in Tandem Turning in Opposite Directions Give
Warplanes New Speed,” Popular Science Monthly, Vol. 139, No. 6,
Dec. 1941, p. 111.
[18] Lesieutre, D., and Sullivan, J., “The Analysis of Counter-Rotating
Propeller Systems,” Society of Automotive Engineers TP-850869,
Warrendale, PA, 1985,
[19] Hager, R., and Vrabel, D., “Advanced Turboprop Project,” NASA Rept.
SP-495, 1988.
[20] Trimble, S., “NASA Positive on Open Rotor But Cautions on Noise
Problem,” Flight International, Vol. 181, No. 5324, Jan. 2012, p. 11.
[21] “Open Rotor Engine & Installation,” European Aviation Safety Agency,
Rept. ToR-MDM.092, Cologne, Germany, March 2011
[22] Cumpsty, N. A., Compressor Aerodynamics, 2nd ed., Krieger, Malabar,
FL, 2004, pp. 11–45.
[23] Van Nimwegen, R., “Design Features of the Garrett ATF3 Turbofan
Engine,” Society of Automotive Engineers TP-710776, Warrendale,
PA, 1971.
[24] Kinder, J. D., and Rahmes, T., “Evaluation of Bio-Derived Synthetic
Paraffinic Kerosenes (Bio-SPK),” The Boeing Company, Sustainable
Biofuels Research & Technology Program, Chicago, June 2009.
[25] Edkins, D. P., and Thorson, M. H., “Why Small Engines,” Society of
Automotive Engineers TP-570224, Warrendale, PA, Jan. 1957.
[26] Sehra, A., and Whitlow, W. Jr., “Propulsion and Power for 21st Century
Aviation,” Progress in Aerospace Sciences, Vol. 40, No. 4, 2004,
pp. 199–235.
doi:10.1016/j.paerosci.2004.06.003
[27] Chan, N., “Scaling Considerations for Small Aircraft Engines,” S.M.
Thesis, Dept. of Aeronautics and Astronautics, Massachusetts Inst. of
Technology, Cambridge, MA, 2008.
P. Givi
Associate Editor