Documente Academic
Documente Profesional
Documente Cultură
P Mock and SA Schmid, German Aerospace Center (DLR), Institute of Vehicle Concepts, Stuttgart, Germany
& 2009 Elsevier B.V. All rights reserved.
Introduction
In view of a possible climate change and increasing
oil-prices, there is a need to change our rate of using
carbon-intensive fossil energy resources in the future.
Fuel cells and batteries have been proposed as technical
solutions for energy conversion and storage, allowing for
paving the way for a prospective world relying mainly on
renewable low-carbon intensive energy sources.
For manufacturers considering investing into research
and later on into production facilities for these technologies as well as for public planners and concerned politicians, several questions with regard to a possible launch
of the technologies arise:
Taking into account all relevant influencing factors,
do fuel cells and batteries have a realistic chance to
gain a noteworthy share of the market in the future?
Are there any supplementary economical, political, or
infrastructural measures available and possibly required to enhance introduction of fuel cells and batteries into the market?
What will be the overall effect of introduction of fuel
cells and batteries in large numbers with regard to
future fossil energy consumption as well as greenhouse gas emissions?
Patent
analysis
Publication
analysis
R&D
activity
analysis
Integrated modeling
Competitiveness
Future market scenarios
Figure 1 Scheme of a workflow for a techno-economic assessment including tools for analysis and general result descriptions for
individual assessment steps.
566
567
568
Technological
progress
Technical
performance
Research
activities, time
Introduction
Growth
Maturity
Research
activities, time
Introduction
Growth
Maturity
Figure 2 Idealized technology life cycle with schematic identification of major important phases during the lifetime of a technology.
When plotting annual changes of technical performance indicating a technological progress, the result is a normal distribution. When
plotting cumulative changes, an s-curve shape is derived.
569
Limit 1
Technology 1
Research
activities, time
570
offices are included in every database), but lately generally show an average growth of approximately 4.75%
per year according to the Worlds Intellectual Property
Organization. As the Organization states, the number of
patent applications per year has doubled from 1985 to
2004 to approximately 1.6 million patent applications per
year. The growth rate closely follows the increase in
world gross domestic product (GDP) and is most evident
for developing countries like China and India. Only a
fraction of patent applications really lead to a final issuing. In 2004, there were 0.6 million patents issued
worldwide, resulting in more than 5.4 million patents in
force.
For the field of scientific publications, a similar growth
is observable. The total number of articles or books
published annually is increasing steadily, again total
numbers depending on the database analyzed. For example, ScienceDirects, the online database of Elsevier
B.V., has indexed approximately 400 000 files for the year
2006 compared to approximately 150 000 in 1985.
This has to be kept in mind when analyzing patent or
publication statistics over a long period of time. Just like
taking inflation into account when looking at business
statistics, it might be necessary to adjust for increasing
total numbers in annual patent applications or scientific
publications, and therefore looking at relative numbers in
percent of all patent applications/publications of a specific year instead of absolute values.
Fuel Cells and Batteries
The total number of annual patent applications worldwide is increasing continuously. Absolute figures depend
on the database used for analysis (as usually not all patent
0.25
0.20
0.15
0.10
0.05
0.00
1950
1960
1970
1980
1990
2000
2010
Figure 4 Historical analysis of worldwide annual patent application numbers for various fuel cell types. To account for generally
increasing annual patent application numbers, values are given as a proportion of relevant applications in comparison to the sum of
patents for all topics.
During the 1960s, the proportion of patent applications related to four major important types of fuel cells
compared to all patent applications increased from close
to zero to more than 0.2% before decreasing again down
to o0.05% in 1975. The major driver of this development was the AFC technology.
It can be assumed that the strong increase of fuel cell
research activity was mainly due to aerospace research at
that time. Fuel cells were developed to be applied during
space flights producing energy and water for the spacecraft. They were selected over other competing power
systems as they better met defined requirements with
regard to power, efficiency, weight, lifetime, reliability,
and safety. Furthermore, they offered the advantage of
being able to produce drinking water for crew consumption and cabin air humidification.
For example, NASA Apollo missions between 1968 and
1972 made use of an AFC onboard. Other types of fuel
cells played a less decisive role at that time. The PEMFC
technology was first used during NASA Gemini missions
from 1962 to 1965, but later on the AFC type was preferred
as it was regarded more suitable and reliable.
After the decrease in the 1970s, it was not before the
mid-1990s until a new boom of fuel cell technology research activity started. In about 1995, the proportion of
patent applications for the four most relevant fuel cell
types compared to all patent applications started increasing from approximately 0.05% to 0.5% about 10
years later.
The main driver for the currently ongoing boom is
PEMFCs. The number of patent applications for this
type of fuel cells started increasing earlier than it did for
other types and has by far the highest proportion of all. It
can be expected that the use of PEMFCs in automotive
0.20
0.15
571
Leadacid-based batteries
High-temperature batteries
Lithium-ion-based batteries
Nickelmetal-based batteries
0.10
0.05
0.00
1970
1975
1980
1985
1990
1995
2000
2005
Figure 5 Historical analysis of worldwide annual patent application numbers for various battery types. To account for generally
increasing annual patent application numbers, values are given as a proportion of relevant applications in comparison to the sum of
patents for all topics.
572
of annual patent applications for this specific type compared to all patent applications is approximately 0.17%.
The strong growth in patent application numbers began
during the early 1990s, shortly before the increased research activity for fuel cells, and continues until now.
Analysis of scientific publications in the area of fuel
cell technology leads to similar results as patent analysis.
The number of publications on fuel cells increased about
10-fold from 1995 to 2005 and no end of the growth trend
is yet to be observed.
A majority of approximately 65% of todays publications on fuel cells are on the PEMFC technology type,
followed by a significant number of publications on direct
methanol fuel cells (DMFCs). On the contrary, DMFC is
only in position four when differentiating patent application statistics between types of fuel cell technologies.
This might be an indication of the number of research
activities on DMFC that do not lead to the same number
of patent applications yet.
Depending on the database used for analyzing publication data, it is furthermore possible to differentiate
the regional distribution of scientific publications. For
fuel cells in general, it is evident that regional distribution of the origin of scientific articles on fuel cell
technology has changed. Up to 1984, a large proportion
of publications originated from the United States,
whereas in the years after Asian countries first Japan
and later on China and South Korea gained an increasing share in the number of publications on fuel cells
(Figure 6). Other countries with a notable proportion in
consideration of their size are Canada, Germany, Italy,
France, Great Britain, and the Netherlands. Especially
for China, a significant increase in the number of
100
Other
South Korea
India
China
Japan
France
Sweden
Netherlands
Italy
Germany
UK
Canada
USA
80
60
40
20
0
1975
1979
1980
1984
1985
1989
1990
1994
1995
1999
2000
2004
2005
2006
Figure 6 Geographical distribution of the origin of the authors of scientific publications on fuel cells in general, differentiated by a set of
time periods. Numbers are given as a proportion of publications for a specific country in comparison to the sum of publications for all
countries.
573
1200
Idealized s-curve
Historical data
900
600
300
0
1950
1955
1960
1965
1970
1975
1980
Figure 7 Historical cumulative patent application numbers for alkaline fuel cell (AFC) type. An idealized mathematical s-curve was
fitted to the data points given.
20 000
Idealized s-curve
Historical data
16 000
12 000
8000
4000
0
1990
1995
2000
2005
2010
2015
2020
Figure 8 Patent application numbers for proton-exchange membrane fuel cell (PEMFC) type. An idealized mathematical s-curve was
fitted to the data points known so far to estimate likely future progress.
This analysis suggests that todays PEMFC technology is still in its growth phase of technology life cycle
but might reach maturity phase by approximately
2015.
It is important to realize that even after this plateau of
maturity is reached, it does not mean that there are no
further technological advancements possible for fuel cells
anymore. On the contrary, it is possible and likely that a
new type of technology with different technological
limits will start a new s-curve.
An example for this could be the high-temperature
PEMFC technology, which offers some advantages in
application but today is still in its introduction phase
with only limited research activities and severe hurdles
for mass production and application.
574
700
R 2 = 0.99
600
2004
500
2006
2005
2003
400
300
2002
200
2001
2000
100
1999
0
0
100
200
300
400
500
600
700
800
Figure 9 Cumulative worldwide patent applications of Ballard Power Systems versus overall cumulative research and development
(R&D) expenditures of the company.
575
100
Lithium based
Nickel based
Lead based
Sodium based
80
60
40
20
0
1970
1975
1980
1985
1990
1995
2000
2005
2010
Figure 10 Historical development of annual public research and development (R&D) funding on batteries in Germany. Values are
given as a proportion of funding for a specific battery type compared to the sum of funding on batteries in general.
public German R&D funding for batteries. Patent application numbers for this type of technology increased
dramatically; therefore, the input/output ratio regarding
R&D is significantly better for lithium-based batteries
than it was for the sodiumsulfur battery type.
576
100 000
Mk 1100
Mk 902
Mk 7
Mk 513
Mk 5 (2)
Mk 5 (1)
Mk 4
Mk 3
10 000
1000
100
10
Mk 1*
*Single cell
1
0
100
200
300
400
500
600
700
800
Figure 11 Historic development of net power of Ballard Power Systems proton-exchange membrane fuel cell (PEMFC) stacks versus
cumulative research and development (R&D) expenditures of the company.
Power Density
The parameter power density indicates the power output
of a stack with regard to its volume or weight. As both
volume and weight are limited, especially in passenger
cars, power density is a key parameter for the development of fuel cell stacks.
The first PEMFC stack application for NASA Gemini
mission in the 1960s achieved a power density of
20 W L1. During the 1980s and 1990s, Ballard Power
Systems for instance was impressively successful in improving this value so that by the year 2000 their fuel cell
stacks had a volumetric power density of approximately
1300 W L1. Today, the Mk 1100 module, which in contrast to the stack does include some auxiliary equipment,
reaches about 1340 W L1. Other manufacturers, for instance Honda, claim to have already reached higher
values up to about 1900 W L1 for the latest stacks. Ballard Power Systems own target value for 2010 is at
2500 W L1, and the target set by the US Department of
Energy (DOE) is 2000 W L1 (Figure 12).
Gravimetric power density, indicating the ratio of
weight of the fuel cell stack in comparison with its power
output, improved similarly over time. Development
started with approximately 15 W kg1 for the NASA
Gemini PEMFC stack and now is at approximately
1000 W kg1 for the Mk 1100 module or 1500 W kg1 for
the current Honda fuel cell stack. The DOE target value
for the gravimetric power density of the stack is
2000 W kg1 for 2010.
Although formal differences between manufacturers
in the definition of power density complicate a comparison of the current status, after analyzing the historical
development of volumetric and gravimetric power
577
3000
2500
2000
Ballard target
DOE target
Honda
1500
Mk 900
Mk 7
1000
DOE target
Mk 1100
Mk 902
500
NASA Gemini
0
1965
1975
Mk 4
1985
Mk 513
Mk 5
1995
2005
2015
Figure 12 Historic development of volumetric power density of proton-exchange membrane fuel cell (PEMFC) stacks including
various targets set for the future.
578
120
Gasoline passenger cars
100
80
60
R 2 = 0.91
40
R 2 = 0.93
20
0
1945
1955
1965
1975
1985
1995
2005
2015
Figure 13 Power-to-weight ratio over time for conventional gasoline- and diesel-driven passenger vehicles as well as fuel cell-driven
cars.
700
600
500
400
300
200
100
0
1990
1995
2000
2005
2010
2015
Figure 14 Historic development of volumetric and specific energy for specific single high-energy lithium-ion cells. A mathematical
s-curve shape was fitted to the data points given.
579
580
Cost Assessment
Once fundamental technical hurdles have been overcome, the most important parameter for market success
of a new technology is cost. In order to be competitive,
costs have to be at least almost as low as for established
technologies already available in the market.
Focus of the next paragraphs will be on the production costs of the technology itself, from the perspective of a manufacturer. The concept of total cost of
ownership (TCO), including other major cost factors
from a customer point of view, will be discussed later on.
Learning Curves
Production costs for a new technology usually are very
high at the beginning of a technology life cycle. Only few
prototypes exist, being assembled manually using raw
materials and preliminary products, which can be
ordered only at high prices owing to low volumes.
As production increases, production costs tend to
decrease. One reason is advantages caused by the higher
production volume itself (economies of scale), for example when ordering raw materials or preliminary
products. Another reason is improvements of production
techniques (learning by doing), for instance by using
special assembly lines for the new product. Furthermore,
scientific advancements with regard to product attributes
(learning by searching), for example requiring lower
amounts of expensive materials, help to decrease production costs. These effects are usually summarized
under the term learning or experience curve.
Historically, learning curves were used to describe the
reduction in man-hours or costs per unit of a product
manufactured by an individual company. One of the first
to analyze this effect in detail was Henry Ford for its
Model T being produced between 1910 and 1921. Today,
the concept of learning curves is not only used for individual companies, but also as a trend analysis tool for
the entire product groups produced on a global scale.
Numerous studies have been carried out in order to
determine learning curve rates for all kinds of products.
It was found that the average learning curve rate is at
approximately 85%. This means that whenever the cumulative production volume doubles, the overall production costs are reduced by 10085% 15%.
However, major differences between commodities and
sectors exist. For more mature products, for example
from the area of mechanical engineering, rates from 85 to
95% were found, whereas less mature products, for instance from the sector of electrical and electronic engineering, often showed learning curve rates at around
6080%. It was furthermore suggested that cost reduction depends on the size of a product, as costs of small
modular products usually decrease faster than production
costs of large nonmodular elements.
Generally, deriving an overall learning curve rate is
more difficult for complex products consisting of a
number of different part groups. For example, for passenger cars, no reduction of the overall production costs
and sales prices could be found for the past few years.
Even as many of the parts used in a car did experience
cost reduction owing to improved effectiveness of production methods, these cost reductions were compensated by adaptations in other areas, for example materials
ensuring a higher safety for the passengers or exhaust
after treatment systems for meeting stricter air quality
regulations.
When analyzing historical cost development for
technical products, it can be a problem that often only
sales prices are known rather than production costs for
the manufacturer. Yet, it was shown that although there
might be differences between cost and price development
during certain time periods, in the long run development
of sales prices is a good indicator for changes of costs for
the manufacturer.
Top-Down Cost Assessment Approach
If production costs of fuel cell stacks and batteries for
todays low production volumes are known, this information can be used together with learning curve factors
from the literature to derive future costs at higher production volumes.
For a more precise result, this can be carried out
separately for individual parts of a fuel cell system or a
battery. For example, costs for the membrane in a fuel
cell stack today could be combined with a typical
learning curve rate known from chemical industry
products in order to estimate a future decrease of production costs. On the contrary, values for an air compressor of a fuel cell system could be derived using a
more conservative learning curve rate from the segment
of mechanical engineering.
Nevertheless, this approach is classified as a top-down
type of method as it does not analyze any specific step of
production in great detail. To account for uncertainties,
calculations can be carried out using several different
learning curve factors to reveal possible variations in the
results. For an assessment of future production costs of
PEMFC stacks, this type of approach has been used for
example by H. Tsuchiya and coworkers.
Bottom-Up Cost Assessment Approach
Another methodology to assess future production costs
for fuel cells and batteries relies on analyzing every necessary step of production in detail and making
581
582
1000
Average
90% Minimum
90% Maximum
Tsuchiya, (worst)
100
Tsuchiya, 2004
DTI, 2003
Carlson, 2005
Tsuchiya, (best)
10
DOE target 2015
1
0
1 000 000
2 000 000
3 000 000
4 000 000
5 000 000
Figure 15 Cost estimates for fuel cell stack production at high production volumes. Market price for crude platinum was assumed to
be fixed at 20 000 $ kg1. Target values set for the future by the US Department of Energy have been included. FC, fuel cell; DOE, US
Department of Energy.
1500
Average 88% learning curve
Bosch, 2008
1250
1000
750
Argonne, 2000
CARB, 2007
500
CARB, 2007
DLR, 2007
250
Argonne, 2000
DLR, 2007
0
0
250 000
500 000
750 000
1 000 000
Figure 16 Production cost estimate for lithium-ion batteries at medium and high production volumes of small battery electric vehicles
(BEVs).
Integrated Modeling
Evaluation of an innovative technology alone might result in misleading conclusions, as it is essential to know
about any competing technologies, whether they are established conventional ones or innovative new ones, and
about their possible advancements in the future. Furthermore, usually there are a variety of external parameters that may have a significant influence on future
competitiveness of the technology and therefore should
also be incorporated in any analysis.
Customer Behavior
Assuming a rather rational and mostly economically
based behavior of customers when choosing a certain
type of propulsion concept for their vehicle allows for
representing the process of decision by using a mathematical analysis tool. Basic idea of such a model would
583
Conclusion
Even the most sophisticated method does not allow for a
definite prognosis of the future. Yet it is possible to develop a set of supposable scenarios for future developments and to discuss any outcome and possibly necessary
actions using a set of methods for techno-economic
assessment.
Fuel Cells
The assessment of technological parameters indicated
that for fuel cells impressive advancements have been
achieved within the past few years. Many key parameters
already satisfy customer requirements or are close to
doing so. Others are still challenging for engineers, even
as analysis of patents, batteries, and R&D activities has
shown that further developments are to be expected.
Fuel cells as an application in passenger cars do have
the potential for reaching a significant market share in
the future, especially for purposes where pure battery
electric-driven cars do not offer a sufficient driving range.
However, production costs of fuel cell systems will have
to decrease considerably in order to be competitive
compared to conventional propulsion systems. As cost
584
Nomenclature
Abbreviations and Acronyms
AFC
BEV
CARB
DLR
DMFC
DoD
DOE
FC
FCV
GDP
ICE
IPC
PEMFC
PHEV
TCO
USABC
WTT
Further Reading
Archibugi D (1992) Patenting as an indicator of
technological innovation: A review. Science and Policy 19(6):
357--368.
Carlson EJ, Kopf P, Sinha J, Sriramulu S, and Yang Y (2005) Cost
Analysis of PEM Fuel Cell Systems for Transportation. Cambridge:
National Renewable Energy Laboratory.
Chakrabarti AK (1989) Technology indicators: Conceptual issues and
measurement problems. Journal of Engineering and Technology
Management 6(2): 99--116.
Foster RN (1986) Innovation: The Attackers Advantage. New York:
Summit Books.
Foussier P (2006) From Product Description to Cost: A
Practical Approach. Volume 1: The Parametric Approach. London:
Springer.
Ghemawat P (1985) Building strategy on the experience curve. Harvard
Business Review March-April: 143149.
Griliches Z (1990) Patent statistics as economic indicators: A survey.
Journal of Economic Literature 12: 1661--1707.
585