Reducing Energy Demand: What Are the Practical Limits?

Abstract

Concern over the global energy system, whether driven by climate change, national security, or fears of shortage, is being discussed widely and in every arena but with a bias toward energy supply options. While demand reduction is often mentioned in passing, it is rarely a priority for implementation, whether through policy or through the search for innovation. This paper aims to draw attention to the opportunity for major reduction in energy demand, by presenting an analysis of how much of current global energy demand could be avoided. Previous work led to a "map" of global energy use that traces the flow of energy from primary sources (fuels or renewable sources), through fuel refinery, electricity generation, and end-use conversion devices, to passive systems and the delivery of final energy services (transport, illumination, and sustenance). The key passive systems are presented here and analyzed through simple engineering models with scalar equations using data based on current global practice. Physically credible options for change to key design parameters are identified and used to predict the energy savings possible for each system. The result demonstrates that 73% of global energy use could be saved by practically achievable design changes to passive systems. This reduction could be increased by further efficiency improvements in conversion devices. A list of the solutions required to achieve these savings is provided.

Full text

rXXXX American Chemical Society Adx.doi.org/10.1021/es102641n |Environ. Sci. Technol. XXXX, XXX, 000–000
ARTICLE
pubs.acs.org/est
Reducing Energy Demand: What Are the Practical Limits?
Jonathan M. Cullen, Julian M. Allwood,* and Edward H. Borgstein
Department of Engineering, University of Cambridge, Trumpington Street, Cambridge CB2 1PZ, United Kingdom
b
SSupporting Information
ABSTRACT: Concern over the global energy system, whether driven by climate change, national security, or fears of shortage, is
being discussed widely and in every arena but with a bias toward energy supply options. While demand reduction is often mentioned
in passing, it is rarely a priority for implementation, whether through policy or through the search for innovation. This paper aims to
draw attention to the opportunity for major reduction in energy demand, by presenting an analysis of how much of current global
energy demand could be avoided. Previous work led to a “map”of global energy use that traces the flow of energy from primary
sources (fuels or renewable sources), through fuel refinery, electricity generation, and end-use conversion devices, to passive systems
and the delivery of final energy services (transport, illumination, and sustenance). The key passive systems are presented here and
analyzed through simple engineering models with scalar equations using data based on current global practice. Physically credible
options for change to key design parameters are identified and used to predict the energy savings possible for each system. The result
demonstrates that 73% of global energy use could be saved by practically achievable design changes to passive systems. This
reduction could be increased by further efficiency improvements in conversion devices. A list of the solutions required to achieve
these savings is provided.
’INTRODUCTION: THE IMPORTANCE OF ENERGY
DEMAND REDUCTION
Three issues drive current interest in present and future global
use of energy: the combustion of fossil fuels as a driver of global
warming, national security related to energy trade, and global
scarcity due to anticipated supply shortages. Possible responses
to these concerns are summarized in the Kaya identity
1,2
CO2emissions ¼Population GDP
Population Energy
GDP
CO2emissions
Energy ð1Þ
Responses to concerns about energy patterns arising from the
first two terms of eq 1, involving population control or wealth
reduction, would be unprecedented and are far from current
thinking. Most discussion is therefore limited to the third and
forth terms, the technical options of energy efficiency (demand
reduction) and carbon efficiency (supply substitution).
Supply substitution is politically the more attractive of these
two options. This is reflected in the International Energy Agency
(IEA) (ref 3, p 173) figures on worldwide research and devel-
opment expenditure on energy, where less than 10% of spending
has been on energy efficiency in comparison with 40% for nuclear
fission and fusion alone. The well-known catalogue of viable
decarbonization options in the energy supply includes: switching
to less carbon intensive fuels (e.g., from coal to gas), increased
use of renewable energy sources (wind,hydro,solar,andbiomass),
increased nuclear power, and development of electricity genera-
tion with carbon capture and storage (CCS). However, as MacKay
4
demonstrates, for many countries (such as the U.K.) there is insuf-
ficient land to allow a meaningful switch to renewable supplies,
CCS is still unproven at scale, and the costs of investment in
renewables, nuclear, or CCS solutions are high. As a result, in
anticipating energy scenarios for 2050, the IEA in their most
aggressive BLUE scenario (ref 3, p. 65), which targets a 50%
global reduction in CO
2
emissions from current levels, estimate
that 21% of emissions savings will come from renewable energy,
19% from CCS, 18% from fuel switching, and 6% from nuclear,
while energy efficiency measures account for 36% over and above
their anticipated annual 0.9% baseline efficiency improvement.
Are these predictions of future savings from energy efficiency
sensible, conservative, or physically possible? The aim of this
research is to provide a rational basis for assessing the potential
for future developments in energy efficiency. In particular, the
paper aims to answer the question: Given known physically
proven options, how much global energy could be saved while
still delivering current levels of final energy services? This research is
needed because no previous study has provided a physical basis
for examining the efficiency of the global energy system.
Four strands of previous work have been identified, each
contributing to the understanding of global energy efficiency
potential, but none providing an integrated view:
•Comparative methods compare energy use across countries,
sectors, or products. Several reports provide means to com-
pare energy use between different activities, attempting to
prioritize larger energy users for more urgent action. For
example, Summers (ref 5, p 150), the IPCC (ref 6, p 259),
and the World Resources Institute (7, pp 4-5) provide
Received: August 3, 2010
Accepted: December 20, 2010
Revised: December 14, 2010

Bdx.doi.org/10.1021/es102641n |Environ. Sci. Technol. XXXX, XXX, 000–000
Environmental Science & Technology ARTICLE
Sankey diagrams
8
showing the flow of energy from fuel to
economic sectors, the United States Department of Energy
(USDOE)
9
,andEuropeanUnion(EU)
10
publish best-prac-
tice energy technology case-studies, and the United Kingdom
(U.K.) Carbon Trust
11
attributes national CO
2
emissions to
end-use consumer activities. These methods prioritize by
scale but do not attempt to quantify the potential for energy
saving -and cannot be used for this purpose, as they have
insufficient detail on the technologies and do not fully repre-
sent the flow of energy from fuel to final energy service.
Comparative methods, which attribute energy use to economic
sectors rather than to technologies, fail to direct attention to the
technical areas where efficiency gains can be made.
•Top-down models aim to track historical trends in efficiency
indicators and extrapolate forward. For example, based on a
20 year trend, the IEA predicts a baseline 1.7% annual
improvement in global energy intensity from structural and
efficiency effects between 2005 and 2050,
3
while Pacala and
Socolow
12
predict a 1.96% annual improvement in carbon
intensity for the next 50 years. However, such extrapolations
based on past data ignore physical limits and assume that the
underlying components of energy demand are stable, where,
for example, Raupach et al.
13
show how the declining trend
in global energy intensity from 1980-2000 was reversed
during the period 2001-2005. In practice, it seems that
these top-down models are useful only for short-term
predictions.
•Bottom-up models survey best-practice efficiency measures
and estimate their combined potential for reducing energy
demand. Recent examples of bottom approaches include
those from the EU,
14
International Energy Agency (IEA),
15
and U.K. Treasury.
16
Bottom-up approaches have the attrac-
tion of being technologically rooted but may be misleading
because they only report known or emerging technologies,
so they fail to capture the potential for innovation and
because they often ignore the connections between devices
in the energy network. Efficiency gains at different points in
the energy network cannot simply be added because a saving
in one technology may often reduce the potential for savings
in a connected device. For example, an improvement to the
aerodynamics of a vehicle will reduce the engine power
requirement, allowing the engine size to be reduced. How-
ever, a reduction in engine size also reduces the possible
energy savings from efficiency improvements to the engine.
Vattenfall’s“abatement curve”,
17
McKinsey’s similar work,
18
and Pacala and Socolow’s“wedges”
12
are all potentially
misleading for this reason.
•Physical models avoid the emphasis on known innovations
implicit in bottom-up models by calculating theoretical
efficiency limits based on engineering principles. Examples
include the “energy bandwidth studies”of the USDOE
Industrial Technologies Programme
9
and 15 societal exergy
studies reviewed by Ertesvag,
19
including an estimate by
Nakicenovic et al.
20
that the global efficiency of energy con-
version in 1990 was around 10% of the theoretical limit.
Often theoretical analyses are extremely technical, so they
have been overlooked in broader studies aiming to identify
energy efficiency options. It can be difficult to anticipate how
much of the theoretical efficiency is practically achievable.
Compared to the aims of this paper, we have not been able to
identify any previous work that takes a technical view of the global
energy system, using top-down data to ensure complete cover-
age, but with appropriate and practically based bottom-up phys-
ical models to ensure the practically achievable potential effi-
ciency is identified. This paper aims to fill this gap. In addition, no
existing work distinguishes the possible energy savings from
improvements to conversion devices (e.g., engines, furnaces, and
light-bulbs) from the savings in the surrounding passive systems
(e.g., the building shell, the vehicle body structure, and the space
into which the light shines).
’GLOBAL ENERGY FLOW
In ref 21, we created a Sankey diagram for current global
energy use, which traces the flow of energy from primary sources
(fuels or renewable sources), through energy conversion devices
and passive systems, to final energy services (for example, trans-
port, illumination, and sustenance). Central to our understand-
ing of the flow of energy through the global system is the dis-
tinction between conversion devices and passive systems.
Conversion devices transform or upgrade energy into more
useful forms and include examples such as fuel refineries, electricity
generators, combustion engines, electric motors, and lightbulbs.
An efficient device converts energy from one form to another
quickly and with minimal loss. Conversion devices are normally
linked together in chains, with the final conversion being to
`useful energy’, mainly in the form of heat, motion, and light.
Useful energy is delivered to passive systems, the last technical
component in each energy chain, where it is eventually “lost”or
dissipated as low-grade heat to the environment in exchange for
the provision of final energy services. Unlike conversion devices,
passive systems do not actively or intentionally convert energy to
another form but instead “hold”or “trap”the useful energy for a
time, to provide a level of final service -hence the term passive.
Thus the building envelope traps heat (or coolth) to deliver a
service of thermal comfort, the car body holds kinetic energy to
provide a transport service, and the room traps light (albeit
briefly) to provide illumination.
Useful energy is eventually dissipated from the system, mostly
as low-grade heat to the environment. A passive system that has
lost energy must be topped up with more energy, in order to
maintain a constant level of service. An efficient passive system is
one that limits the loss of useful energy and therefore traps the
energy for a longer period. For example, a well-insulated and
sealed house traps heat for longer, an aerodynamic car body
maintains kinetic energy, and a white painted room increases
surface reflectance to prolong light.
Thus, a passive system can be thought of as a reservoir or tank
of stored energy -kinetic energy for the car, thermal energy for
the house, and light for the room. The level in the tank equates to
the final service level, and any output of energy from the tank
must be balanced by an equal input of energy to maintain
a constant service level. Fluctuations in the service level waste
energy -for instance, when a car is starting and stopping fre-
quently or when a furnace is not operated continuously -as
moving from a high to low service level requires the tank to be
drained of energy. Conversely, if all outputs from the tank are
sealed, then a constant service level can be maintained without
any additional energy input. In the idealized case of a car without
friction or drag, motion would continue indefinitely, and if a
building is perfectly sealed and insulated, a constant temperature
can be maintained without the addition of heat.
The production of materials and goods in industry deserves
special attention. For industrial materials, most of the useful

Cdx.doi.org/10.1021/es102641n |Environ. Sci. Technol. XXXX, XXX, 000–000
Environmental Science & Technology ARTICLE
energy is delivered to passive systems such as furnaces and motor
driven systems, where it is trapped as heat and motion and after
some time lost as low-grade heat. However, a portion of the
energy remains embedded in the material, resulting in a change in
chemical or physical properties. This embedded energy is still
trapped and provides a constant level of final service but, in this
case, for a very long period of time. For example, energy is required
to change iron ore into steel, with a portion of this energy be-
coming embedded in the steel, while delivering mechanical strength
over many years in a structure.
Eventually, after a considerable time period, the energy embedded
in the material will be lost, not as low-grade heat but as the material
returns back to its natural environmental state, for example, the
oxidation of steel to iron oxide. Analysis in the Supporting In-
formation (summarized in Table S.13) shows that the total
embedded energy (shown in the column Theoretical Minimum)
in the global production of steel, cement, paper, and aluminum
(11 EJ) is small compared to its total processing energy (66 EJ).
In addition, the embedded energy in these materials is more than
balanced by the release of embedded energy when making industrial
chemicals (-15 EJ), as the energy in fossil fuels is liberated during
exothermic chemical reactions. Thus, for this analysis, embedded
energy in materials is ignored in order to focus attention on the
passive systems of furnaces, driven, and steam systems, where
losses could be more easily avoided.
The importance of the distinction between devices and systems is
to draw attention to the different technical options each offers when
seeking future energy efficiency: Conversion device efficiencies
relate to more effective energy transformations and ensuring devices
are sized for the output required of them; and passive system
efficiencies are related to the avoidance of unintended losses and
reducing mechanical and thermal inertia when the service level is
required to fluctuate. The failure to distinguish between these
two stages of global energy flow -energy conversion in devices
and service delivery from passive systems -leads to the misleading
concept of additive efficiencies implicit in abatement curves.
Our work in ref 21 led to a Sankey Diagram in which the
exergy value of primary energy sources (the maximum work that
can be obtained from the energy source) was allocated through
the elements of Figure 1 to deliver final energy services. In order
to understand the efficiency of this flow, in addition to alloca-
tions, we must also examine losses, and in ref 22, we calculated
the theoretical efficiencies of the energy system and attributed
losses to 10 mechanisms. The paper calculated a global conver-
sion efficiency of 11% for all devices, including fuel transforma-
tion, electricity generation, and end-use conversion, by comparing
the current energy use in devices to the theoretical minimum
energy requirement. However, this approach proved unsuitable
for passive systems because, as discussed above, a final service
level can be maintained without the input of any energy if all
unintended losses are eliminated -the theoretical minimum
energy requirement for service provision is zero, and defining a
theoretical efficiency limit is meaningless.
The work of ref 22 therefore calculated a theoretical conver-
sion efficiency value for global energy flow only as far as the useful
energy output by devices. Hence, this paper aims to extend the
analysis by providing physical-based efficiency estimates for passive
systems. By calculating the savings in energy use that can be
achieved by existing physical solutions, it will be possible to make
more informed comparisons between the investments required
to address key energy concerns across both supply- and demand-
side options.
The analysis in this paper then is intended to assess the efficiency
of passive systems, permitting a global energy flow diagram with
losses to be created, with a structure like that of Figure 1.
’PRACTICAL LIMITS TO EFFICIENCY IN PASSIVE SYS-
TEMS
In order to estimate the practical limits to energy efficiency of
passive systems, scalar equations are identified for each passive
system to predict the useful energy required to deliver the final
energy service. These equations are dependent on design para-
meters (mass, size, and so on), and by considering the limiting
values of these design parameters, we can predict the practical
minimum energy limits required to provide each service. The
complete work of developing these models is described in the
Supporting Information and summarized in the remainder of this
section.
In aiming to estimate the practical limits to energy require-
ments in passive systems, three groups of passive systems are
examined (as identified in ref 22): Vehicles (car, truck, plane,
ship, and train), Industry (furnace, driven system, and steam
system), and Buildings (heated/cooled space, hot water system,
six types of domestic appliance, and illuminated space). For each,
a model is developed based on simple scalar relationships familiar
to all undergraduate engineers: Fourier’s law of thermal conduc-
tion, Burnet-Gerrans equation of heat capacity, Newton’s law of
cooling applied to heat exchange, extended Bernouilli equation
for mechanical pumping, Darcy Weisbach equation for pressure
loss in piping, Coulomb’s friction law, Rayleigh’s equation for
drag, and Newton’s second law of motion. Where simple equa-
tions are unable to provide sufficient insight, for instance in deter-
mining energy requirements for planes and trains, reference is
made to major studies examining future efficiency options.
A key concern in developing our estimates has been to ensure
representative system descriptions. For example, in estimating
the potential for energy saving for all cars, the analysis depends
on an estimate of current fleet composition, and this has been
based on analysis of European Commission data on vehicle use in
15 EU countries in 2000. Similarly, estimates depend on assump-
tions about the behavior of the users of passive systems. For
example, without global statistics on cooking behavior, it is
assumed that a typical domestic oven is used to cook food at
200 °C for 1 h per day. When statistics are available, they are
used, for instance, average vehicle driving cycles are used to
attribute fractions of engine power to overcoming inertia, mechan-
ical,and aerodynamic drag. The context in which passive systems
operate is also important, particularly for buildings, where heating
Figure 1. Schematic Sankey diagram for provision of energy and material
services showing losses, reproduced with permission from ref 22.

Ddx.doi.org/10.1021/es102641n |Environ. Sci. Technol. XXXX, XXX, 000–000
Environmental Science & Technology ARTICLE
and cooling requirements are dependent on regional tempera-
ture and sunshine patterns. In this case, the required heating and
cooling loads were determined using representative daily mean
temperatures and solar irradiation cycles for two cities in each of
the four global climatic zones selected.
Armed with appropriate equations and representative data on
current use, the practically achievable energy savings are esti-
mated by examining how model parameters could feasibly be
changed. Where parameters relate to the design of the passive
system -coefficient of friction, mass, and thermal conductivity,
for example -the values are selected from examples of existing
best practice design, while aiming to avoid extremes that could
not occur in practice. So, although the world’s most efficient car
(the PAC-II) has a total unloaded mass of just 29 kg, the practical
limit is defined here as 200 kg based on scaling the design for the
Rocky Mountain Institute “2000 Revolution Hypercar”,
23
an
SUV, to the size of a typical small car. Low friction tires (tubeless
Michelin 45-75R16) and minimal aerodynamic drag for the car
body (based on the PAC-II) are assumed. For trains, the design is
based on existing high-speed passenger trains, which have very
low drag. Similarly, the estimate of practical limits in house design
is based on established Passivhaus buildings standard, with annual
heating loads of less than 15 kWh/m
2
. The standard specifies low U
values for the building shell (typically >300 mm of cavity wall
insulation and triple glaze windows) and tightly controlled ven-
tilation rates with heat recovery.
In addition to these design choices, some limited behavior
change is assumed to allow changes to the “set-points”of current
use without otherwise assuming any reduction in service provi-
sion. Thus, on the basis of the analysis of human adaptability for
comfort in buildings, a wider range of acceptable temperatures
within buildings is used to reduce the need for heating and cool-
ing; a lower maximum speed for trains and ships is assumed
(though not cars or planes); and domestic hot water systems are
allowed to operate at lower temperatures (50 °C rather than
65 °C), as current settings are based on the need to avoid growth
of legionella bacteria in hot water tanks, which are eliminated in
the best practice design.
Our analysis is summarized in Table 1, which shows the
current energy supply allocated to each passive system taken
directly from global analysis in ref 21 equations for each simple
model with key design parameters and their current and practi-
cally limiting values, an estimate of the fraction of current energy
use that could be saved solely by changes to the passive system.
The results suggest that we have engineering options to reduce
current energy demand by 73%. A full description of each passive
system model and all assumptions used to calculate the energy
savings are in the Supporting Information.
Figure 2 shows these results, with the width of each bar
representing the scale of primary energy supply allocated to each
passive system based on ref 21. Color is used to distinguish bet-
ween the practical minimum energy required (blue) and what
could practically be avoided (gray). The passive systems within
each sector are ranked according to achievable energy savings
(gray), and this ranking could be use to direct priorities for future
research development and policy.
’DISCUSSION
The physical basis of our analysis in Table 1 can be translated
into specific actions that will lead to a step change reduction in
energy demand. Table 2 summarizes the practical steps required
to achieve this in the passive systems discussed in this paper.
The greatest energy savings can be achieved in the passive
systems of buildings, both absolute (179 EJ) and as a percentage
of total demand (83%). This is dominated by the savings in
heating and cooling spaces (85 EJ), which could be avoided by
designing buildings to the Passivhaus standard, and appliances
(59 EJ). Several factors currently inhibit the deployment of more
efficient technical solutions in buildings spaces: the variety of
building designs, fragmented ownership, the disconnect between
landlords and tenants, a reluctance to change the appearance of
historical buildings, an overemphasis on initial capital cost instead of
costing over the entire building lifetime, and the slow turnover of
building stock. The practical minimum energy required for appli-
ances is comparatively large (33%, 29 EJ) because the energy
required to heat and cool food or display information is an essential
part of the final energy service, and it is difficult to recovery ener-
gy from washing applications at low temperature.
Transport passive systems are already more efficient than
buildings (avoidable loss averages 68%) because the need to carry
fuel creates a natural driver to reduce frictional drag and vehicle
weight. This is particularly pronounced for planes, where the weight
of any additional fuel must be supported by the lift force generated
bytheenginethrustandwings.Forthisreason,theplaneisthemost
optimized of the passive systems examined, with only 46% practical
energy savings available. The largest savings for transport occur in
passenger cars (91%, 37 EJ), where reducing the mass of the car and
the rolling resistance are important. Cars are also less optimized
because of the diversity of models and the need to compromise
efficiency to satisfy the fashion element of car design. Light-
weighting strategies for ships and trains are limited because the
transported goods make up a large proportion of the total vehicle
mass (typically 60% in trucks versus 5-10% in cars).
The passive systems in industry are on average the most
efficient (avoidable losses averages 62%), although the absolute
savings are higher than for transport (95 EJ versus 72 EJ). This
reflects the existence of economies of scale in industrial systems,
standardization of equipment, and the natural commercial drive
for efficiency improvement in energy intensive industries. The
fraction of available energy savings across all three industrial
systems -furnaces, driven systems, and steam systems -is
similar (59-66%). Advances in industrial systems, despite the
long operating life of equipment, give hope that similar improve-
ments to building systems are also possible.
Table 1 demonstrates a saving of 73% of global energy use by
practically achievable design changes to passive systems. How-
ever, the idea of global energy flow from source to service dis-
cussed above and summarized in Figure 2 shows that total energy
savings will be greater than this if equivalent steps are also taken
in energy conversion devices (including fuel transformation,
electricity generation, and end-use conversion). We have already
in ref 22 provided calculations of theoretical efficiency potentials
for conversion devices, so to illustrate the extent of practically
achievable global energy savings, we estimate the practical effi-
ciency limits for conversion devices by halving the theoretical
gains calculated in ref 22.
This crude estimate is derived from two sources. First, Blok
(ref 24, p 89) suggests that “it is possible to bridge about half of
the gap between the present best technologies and the thermo-
dynamic minimum; that is, it is possible to cut the avoidable
energy use in half”using known energy efficiency technologies in
industry. Industry is already the most efficient sector, so this
reduction may be a conservative estimate of possible savings in
the building and transport sectors.

Edx.doi.org/10.1021/es102641n |Environ. Sci. Technol. XXXX, XXX, 000–000
Environmental Science & Technology ARTICLE
Table 1. Summary Analysis of Passive System Savings
a
system Eequation design parameters CPF
Building 215 83
heated/cooled
space
86 QHþQZþQI¼QSþQVþQTU
roof
(W/m
2
K) 1.2 0.1 98
QS¼P
i
ðUiAiÞðTinside -ToutsideÞU
walls
, heat transfer coeff. 1.3 0.1
U
floor
1.2 0.1
QV¼qFCpðTinside -ToutsideÞð1-RÞU
windows
4.0 0.8
q(l/s/per), ventilation rate 30 7
QT¼mbuildingCpC
s
, solar heat gain coeff. 0.5 0.5
QZ¼AwindowCSHθR, heat recovery fraction 0 0.8
hot water systems 23 QH¼CpðTout -TinÞþQL-QXU
tank
(W/m
2
K) 1.1 ∼80
QL¼P
i
UiAiðTstore -TairÞA
tank
(m
2
) 2.2 0
U
X
(W/m
2
K) 1200 1500
QX¼UXAXΔTLM A
X
(m
2
), heat exchange ∼3.1
illuminated space 18 LFdevice ¼EA
UFLLF E(Wm
2
), illuminance 780 130 95
UF, utilization factor 0.3 0.9
appliances 88 see SI for 6 appliance types ∼∼∼67
Industry 154 62
furnace 67 QF¼QSþQVþQTw
shell
(mm), insulation depth 100 250 62
QS¼P
i
ðUiAiÞðTinside -ToutsideÞk
shell
(W/mK), thermal cond. 1.0 0.5
QV¼qFCpðTinside -ToutsideÞð1-Rflue ÞR
flue
, heat recovery fraction 0 0.8
QT¼mproductCpTð1-Rproduct ÞR
product
, heat recovery fraction 0 0.4
driven system 56 FWshaft ¼FWloss þΔPD, pipe diameter D1.3D59
FWloss ¼8fLq2
πgD5see SI for nonpumping applications
steam system 31 same as furnace model
but with distribution
replacing ventilation loss
Q
S
and Q
T
improve as
furnace, Q
D
eliminated
∼∼66
Transport 106 70
car 40 W¼RFνdtm(ton), loaded mass 1.3 0.3 91
ν(m/s), average velocity 19 19
F¼μmg þ1
2Fv2CDAfþmμ, tire friction 0.015 0.001
C
D
, drag coefficient 0.41 0.10
A(m
2
), frontal area 2.0 1.5
truck 38 W¼RFνdtm(ton), loaded mass 14 13 54
F¼μmg þ1
2Fv2CDAfþmν(m/s), average speed 18 18
μ, tire friction 0.012 0.005
C
D
, drag coeff. 0.95 0.31
A(m
2
), frontal area 7 7
plane 10 SFB ¼W1-W2
WPR¼c2
X
1-Re-Z
ZðRe-Z-c1Þη
p
(%), propulsive efficiency 81 95 46
c
1
, structural constant 0.32 0.38
c
2
, structural constant 2.0 1.9
Z¼R=X¼InðW1=W2ÞA, wing aspect ratio 10 5
k, vortex drag factor 1.2 1.1

Fdx.doi.org/10.1021/es102641n |Environ. Sci. Technol. XXXX, XXX, 000–000
Environmental Science & Technology ARTICLE
Second, a large proportion of the thermodynamic losses from
conversion devices result from heat exchange across a finite tem-
perature difference. The theoretical efficiency limit for these
“heat engines”is defined using the ideal Carnot cycle: η=1-
T
2
/T
1
, where T
2
and T
1
are the temperatures in Kelvin of the
cold and hot reservoirs. However, Curzon and Ahlborn
25
explain
that to achieve this ideal efficiency requires the infinitely slow
transfer of heat between the working substance and reservoirs,
resulting in an actual power output of zero. Instead, if the rate of
heat exchange is taken into account, a maximum power output
can be calculated using: η=1-(T
2
/T
1
)
1/2
. In typical heat
exchange applications this reduces the avoidable losses by
approximately half and serves as a realistic practical minimum
efficiency.
Applying this logic, a halving of the avoidable losses and using the
calculated theoretical efficiency of 11% from ref 22 implies a
practical efficiency limit for conversion devices of 56% (89%/2 þ
11%). Multiplying the practical efficiency limit for devices (56%) by
that for passive systems (27%) provides an overall estimate of global
efficiency of 15%; that is 85% of energy demand could be practically
avoided using current knowledge and available technologies. A more
detailed prediction of the practical energy savings in conversion
devices requires another extended analysis of the type we have
described in this paper, and we plan this as future work.
Table 1. Continued
system Eequation design parameters CPF
X¼Hηeηpffiffiffiffiffiffiffiffiffi
πA
4kCDO
qC
DO
, profile drag coeff. 0.021 0.003
L/D, lift/drag ration 18 37
R(km), range 5000 9000
ship 10 summary from survey of empirical testing of scale models ∼∼∼63
train 8 W¼RFνdtm(ton), loaded mass 300 400 57
ν(m/s), average velocity 42 42
F¼AþBv þCv2A, (N), drag coeff. 3300 2300
Aμmg, B¼Bðm,lÞB(Ns/m), drag coeff.2858
c¼1
2FAfðCDþλl=dÞC(Ns
2
/m
2
), drag coeff.118
Total 475 73
a
Notes.E: global energy supply allocated to this system (EJ). C: current setting for design parameters. P: practical limit to design
parameter settings. F: fraction of current energy use that could be saved (%). Parameter settings have been rounded in some cases.
Building parameters: For materials in the building fabric and the heating fluid (air or water): mass, m; density, F; specific heat capacity,
C
p
; temperature, T; and overall heat transfer coefficient, U. Heat transferred into building by heater, Q
H
; solar, Q
Z
; internal heating, for
instance by occupants bodies, Q
I
; heat transferred out of building through shell, Q
S
; by ventilation, Q
V
; or by transfer into thermal mass,
Q
T
. Other building parameters: A
i
, surface area of type, i; and H
q
, geographically specific solar irradiation. Water heating requirements:
Q
H
, depends on mass flow rate, set point T
out
; losses Q
L
, from storage tank and pipes (reduced to zero for a tankless point-of-use system),
and Q
X
, if a waste water heat recovery system is used; lighting device supplying luminous flux, LF
device
, required to provide average
illuminance, E, over area, A, with utilization factor, UF; and light loss factor, LLF (assumed constant). Six appliances are discussed in the
Supporting Information, SI so only the total savings are reported. Industry parameters: Heat provided to the furnace, Q
F
, is balanced by
heat loss from the shell, Q
S
; leakage through door, vent, and flue openings, Q
V
; and thermal losses from heating the shell (intermittent
use) and the product, Q
T
. For the practical limit, the furnace is continuously operated with no loss from air leakage, reduced shell thermal
conductivity, k; thickness, w; and heat recover, R, from the flue gases and product. All driven systems are related to pumping;either for
transport or pressure increase, DP, with losses related to the dimensional Darcy friction factor, f; the effective pipe length, L; pipe
diameter, D; and q, the volumetric flow per unit area. The pipe diameter, D, is increased to find the practical limit. See the Supporting
Information (SI) for solid materials, compressed air, and refrigeration. For steam generation, the furnace model is modified with
distribution losses, Q
D
, replacing ventilation losses, Q
V
. For the practical limit, increased insulation reduces boiler losses and heat is
recovered from the condensed steam, while distribution losses are eliminated with point-of-use systems. Transport parameters: Cars
and trucks use the work provided by the engine to overcome mechanical drag (rolling), aerodynamic drag, and inertia (the resistance of
mass). The practical limit is determined by reducing the loaded mass, m; tire friction, μ; drag coefficient for air, C
D
; and the frontal area,
A. Inertia is assumed to be recovered by coasting (without braking) or regenerative brake systems. For planes, the specific fuel burn (SFB)
of an airplane in kg/t km is calculated from the fuel consumed in flight (initial weight, W
1
, less final weight, W
2
) over the payload, W
P
,
multiplied by the range, R, via the key performance factor X; two structural constants, c
1
and c
2
; the fuel heating value, H; the propulsive,
η
p
, and thermal, η
e
,efficiency of the engine (the latter of which is in our language a device, so invariant within the passive system); the
wing aspect ratio, A; vortex drag factor, k; factor for non-crusing fuel, R; and profile drag coefficient; C
DO
. The practical limit for the plane
is derived for a laminar flying wing aircraft with unducted fan propellers, and both current and limit designs are compared for their
optimum ranges. The model of drag forces for ships is nonlinear and difficult to solve, so instead a survey of experimentally tested hull
designs is used to estimate practical savings. For trains, the drag forces comprise three terms of which Ais related to mechanical drag,
Brelates to both drag and air intake acceleration, and Cis the aerodynamic drag. Full details of each passive system model, including
equations, design parameters, current and limiting values, energy savings, and assumptions, are given in the Supporting Information.

Gdx.doi.org/10.1021/es102641n |Environ. Sci. Technol. XXXX, XXX, 000–000
Environmental Science & Technology ARTICLE
The estimates made in this paper are subject to uncertainty,
but by providing a practical basis for recommended change, we
hope that the numbers are at least a pragmatic guideline to
what could be achieved. Our calculation of potential energy
savings is technically achievable, and yet significant economic
and socio-cultural barriers are still present that limit the
realizable improvement. Human choices are rarely driven by
a sole desire to use energy efficiently, but instead include
complex drivers such as social status (e.g., desire for larger cars
and houses), security (e.g., stockpiling of armaments), and
fashion (e.g., throwing away perfectly functional products
because they appear outdated). We have not considered the
social and economic implications of these changes, but the fact
that they are based on existing designs gives a basis for future
work in evaluating costs and predicting the effects of these
changes on demand patterns.
The analysis in this paper ought to be a reminder to policy
makers and thought leaders on energy futures worldwide: res-
ponses to the threat of climate change and to concern over energy
security have to date mainly focused on developing new energy
sources or decarbonising existing energy supply; those that ad-
dress demand-side energy typically fail to isolate the potential
reductions from passive systems. This analysis predicts that we
could currently live with 73% less energy supply by applying
known engineering best practice to passive systems that transform
useful energy to services. An initial assessment of the practical effi-
ciency limits in conversion devices provides an overall global and
practical estimate for the reduction potential in demand-side op-
tions, of 85%. National and international energy policy and the
scenario analysis that support its development appears to pay
insufficient attention to demand reduction, presumably in the
fear that it implies service reduction. We hope that by describing
a physical basis for reducing global energy demand to one-quar-
ter of current levels by modifying passive systems (one-sixth if
conversion devices are also included), without reduction in service
Table 2. Practical Actions Required To Save 73% of the
Energy Used by Passive Systems
system practical actions to reduce energy demand
Building
heated/
cooled
space
>300 mm wall and roof cavity insulation with minimum thermal
bridging and triple glazed windows, controlled ventilation with
heat exchange and no leaks, maximum use of solar gains in cold
climates, controlled solar gains and in hotter areas, and high
thermal mass and passive ventilation to average out external
temperature variations and avoid cooling
hot water
systems
eliminate hot water tanks, reduce set point to 50 °C, and apply
drain-water heat recovery
illuminated
space
avoid excess illuminance, use task lighting to focus light as
needed, and improved luminaire design
appliances stove-top cooking in pans: use lid, halve thermal mass of pan, and
add 30 mm thick fiberglass insulation to pot
oven: seal oven throughout cooking time, 100 mm thick fiber-
glass insulation, and avoid steel casing to reduce thermal mass
fridge-freezer: defrost frozen food in fridge, mount compressor
at top of fridge, 200 mm thick insulation, and improved com-
partmentalization to reduce ventilation
washing machines, dishwashers, and dryers: reduce temperature
set-point, recover heat from wastewater, use less water, hori-
zontal axis in washing machines, and heat recovery from dryers
Industry
furnace increaseinsulationthickness by 150mm, reduce thermalconductivity
to lowest value for known refractory bricks, improved heat recovery
from exhaust flue, and recover heat from discharged heated products
driven
system
increase pipe diameter by 25%, reduce corners in pumped systems,
and similar increases for conveyor and other handling systems
steam
system
avoid distribution by point of use generation, increase insula-
tion, and reduced thermal conductivity
Transport
car reduce loaded car mass to 300 kg, reduce tire rolling resistance to
0.001 (Michelin 45-75R16), and reduce drag coefficient C
D
to 0.1
and frontal area from 2 m
2
to 1.5 m
2
truck reduce tire rolling resistance to 0.005 and drag coefficient to 0.31
plane laminar fixed wing aircraft with unducted fan engine design
ship improve propeller and hull design, and reduce fleet speed by 10%
train adopt Swedish X2 high-speed passenger train design, but limit to
current average speed of 100km/h for freight and 150 km/h for
passengers
Figure 2. Practically available energy savings from global passive
systems.

Hdx.doi.org/10.1021/es102641n |Environ. Sci. Technol. XXXX, XXX, 000–000
Environmental Science & Technology ARTICLE
provision, we can contribute to rebalancing this skewed empha-
sis. True sustainability depends on living within an ecologically
acceptable budget, and the key to achieving that is to require
fewer input resources.
’ASSOCIATED CONTENT
b
SSupporting Information. A detailed explanation of the
calculations and assumptions used to predict the practical energy
savings available in passive energy systems. Suppporting Information
is available free of charge via the Internet at http://pubs.acs.org/
’AUTHOR INFORMATION
Corresponding Author
*Phone: þ44 1223 338181; fax: þ44 1223 332662; E-mail:
jma42@cam.ac.uk.
’ACKNOWLEDGMENT
The first two authors are supported by a Leadership Fellow-
ship provided by the U.K. Engineering and Physical Sciences
Research Council (EPSRC) Reference EP/G007217/1.
’REFERENCES
(1) Kaya, Y. Impact of Carbon Dioxide Emission Control on GNP
Growth: Interpretation of Proposed Scenarios; Response Strategies Work-
ing Group, Energy and Industry Subgroup of the Intergovernmental
Panel on Climate Change: Paris, 1990.
(2) Yamaji, K., Matsuhashi, R., Nagata, Y. Kaya, Y. An Integrated
System for CO
2
/Energy/GNP Analysis: Case Studies on Economic
Measures for CO
2
Reduction in Japan. In Workshop on CO
2
Reduction
and Removal: Measures for the Next Century, International Institute for
Applied Systems Analysis, Laxenburg, Austria, March 19-21, 1991.
(3) Energy Technology Perspectives 2008; International Energy Agency:
Paris, 2008.
(4) MacKay, D. J. Sustainable Energy: Without the Hot Air; UIT
Cambridge: Cambridge, U.K., 2009.
(5) Summers, C. M. The conversion of energy. Sci. Am. 1971,225,
148–160.
(6) Sims, R.; Schock, R.; Adegbululgbe, A.; Fenhann, J.; Konstantina-
viciute, I.; Moomaw, W.; Nimir, H.; Schlamadinger, B.; Torres-Martinez,
J.; Turner, C.; Uchiyama, Y.; Vuori, S.; Wamukonya, N.; Zhang, X.
Energy Supply. In Climate Change 2007: Mitigation. Contribution
of Working Group III to the Fourth Assessment Report of the Inter-
governmental Panel on Climate Change;Metz,B.,Davidson,O.R.,
Bosch, P. R., Dave, R., Meyer, L. A., Eds.; Cambridge University
Press: Cambridge, U.K., 2007.
(7) Baumert, K. A., Herzog, T. Pershing, J. Navigating the Numbers:
Greenhouse Gas Data and International Climate Policy; World Resources
Institute: Washington, DC, 2005.
(8) Schmidt, M. The Sankey diagram in energy and material flow
management. J. Ind. Ecol. 2008,12 (1), 82–94.
(9) ITP Best Practices. Industrial Technologies Program, U.S. De-
partment of Energy. http://www1.eere.energy.gov/industry/bestprac-
tices/.
(10) Action Plan for Energy Efficiency. European Union. http://
europa.eu/legislation_summaries/energy/energy_efficiency/.
(11) The Carbon Emission Generated in All That We Consume;
Carbon Trust: Whitney, U.K., 2006.
(12) Pacala, S.; Socolow, R. Stabilization wedges: Solving the climate
problem for the next 50 years with current technologies. Science 2004,
305, 968–972.
(13) Raupach, M. R.; Marland, G.; Ciais, P.; Le Quere, C.; Canadell,
J. G.; Klepper, G.; Field, C. B. Global and regional drivers of accelerating
CO
2
emissions. Proc. Natl. Acad. Sci. U.S.A. 2007,104, 10288–10293.
(14) Action Plan for Energy Efficiency: Realising the Potential; Com-
mission of the European Communities: Brussels, 2006.
(15) Energy Technologies at the Cutting Edge; International Energy
Agency: Paris, 2005.
(16) Energy Efficiency Innovation Review; HM Treasury: London, 2005.
(17) Nelson B. Vattenfall’s Global Climate Impact Abatement Map,
2007. Vattenfall. http://www.iea.org/work/2007/priority/Nelson.pdf.
(18) Bressand, F.; Farrell, D.; Hass, P.; Morin, F.; Nyquist,
S.; Remes, J.; Roemer, S.; Rogers, M.; Rosenfeld, J.; Woetzel, J. Curbing
Global Energy Demand Growth: The Energy Productivity Opportunity;
McKinsey Global Institute, 2007. http://www.mckinsey.com/mgi/pub-
lications/.
(19) Ertesvag, I. Society exergy analysis: A comparison of different
societies. Energy 2001,26, 253–270.
(20) Nakicenovic, N.; Gilli, P. V.; Kurz, R. Regional and global
exergy and energy efficiencies. Energy 1996,21, 223–237.
(21) Cullen, J. M.; Allwood, J. M. The efficient use of energy:
Tracing the global flow of energy from fuel to service. Energy Policy
2010,38 (1), 75–81.
(22) Cullen, J. M.; Allwood, J. M. Theoretical efficiency limits for
energy conversion devices. Energy 2010,35 (5), 2059–2069.
(23) Lovins, A. B.; Cramer, D. R. HypercarsÒ, hydrogen, and the
automotive transition. Int. J. Veh. Des. 2004,35,50–85.
(24) Blok, K. (2005). Improving energy efficiency by five percent
and more per year? J. Ind. Ecol. 2005,8(4), 87-99.
(25) Curzon, F. L.; Ahlborn, B. Efficiency of a Carnot engine at
maximum power output. Am. J. Phys. 1975,43,22–24.