Hydrogen production scenarios in Italy
Giorgio Ballardin'
Scuola Mattei — Eni Corporate University
Via S. Salvo 1 San Donato Milanese
20097 Milan
Italy
tel. +39 02 52047298
fax +39 02 52057908
giorgio.ballardin@enicorporateuniversity.eniit
Abstract
Hydrogen, an energy vector, displays remarkable versatility with regards to the ways it can be
produced. State-of-the-art technologies allow almost every energy source to be converted into
hydrogen. What is more challenging, however, is the feasibility of building a new infrastructure
to overlap with and, possibly, substitute existing one. This investigation aims to assess what it
would entail to add 5% of hydrogen fuel to road transport energy consumption through 2050.
The comparison spans five technologies: steam methane reforming, coal gasification, and water
electrolysis where power is generated from wind, solar, and nuclear sources. The simulation
provides two sets of estimates: calculations on physical infrastructure requirements and its
related variable and fixed costs. With regards to facility requirements, the considered
technologies show different degrees of feasibility. Coal and nuclear power are not as land-
intensive as solar and wind power, but bear problems with pollution and waste disposal,
respectively. Economically, coal is least expensive, followed by wind. Natural gas loses
competitiveness because of high hydrocarbon prices. The sheer economic rank of preferable
energy sources for generating hydrogen should be put into question when internalizing
environmental impact of the considered options.
1 The usual disclaimer applies.
Abbreviations: RES, Renewable Energy Sources; kwh, kilowatt-hour; SMR, steam methane reforming; Mtep, million
tons of equivalent petroleum, GHGs, green-house gases.
1, Introduction
Hydrogen has received a great deal of attention from both scientists and policy makers over the last
decade. The energy revolution carries the promise of tackling some old energy industry issues:
reliable supplies, independence from foreign oil, global and local emission reductions. Hydrogen’s
versatility displays manifold solutions: a sizeable amount of feedstock can indeed be converted into
it, just as it is the case with electricity. Home production and reliable trade partners can help secure
energy supply for the domestic economy.
Of course, this view abstracts a number of challenges that still exist. New technologies ought to
be better than old ones in several aspects. First of all, in terms of energy efficiency, higher well-to-
wheel performances must be achieved. Similarly, global and local emissions along the whole
energy chain must be lowered. From a financial perspective, hydrogen has to make business sense
for both suppliers and end users. Finally, new technologies must be easy to use and safe. The bulk
of these questions have been faced, and partly addressed, by the scientific community. Feasibility,
though, still has a long way to go.
The viability of hydrogen is, indeed, the question. This work is a top-down scenario analysis
aiming at investigating what would be the additional infrastructure requirement if 5% of the energy
in road transport were to be substituted by hydrogen. Data refer to Italy spanning up to 2050, and
the focus is on fuel production. In particular, the study consists of two sections. One deals with the
physical and technological issues: how many square kilometers of solar panels would be needed to
produce the foreseen amount of hydrogen, for example. The other deals with the economics of
infrastructure development. Each feedstock gives rise to fixed and variable costs. The ultimate goal
is to collect evidence about how the adoption of hydrogen on a broad scale could possibly be
buffered by the whole energy system. Can hydrogen be absorbed smoothly or does it require major
infrastructure work?
Such evaluations are useful for policy-makers to estimate the investment’s magnitude of
hydrogen-related technologies. It is vital to not just develop knowledge on single technologies but
also on how the whole energy system can be shaped.
2. Literature review
The growing body of research that surrounds hydrogen economy-related issues can be divided into
two strands. Scientists have focused on applying technological data on actual countries, states, or
continents. This kind of exercise becomes instrumental in seeing how feasible it would be to
employ hydrogen on an extensive scale. One can also find state-of-the-art technology descriptions
and analyses on how scientific progress could make seminal solutions marketable. This group of
contributions is location-free, that is facilities are not often meant to be built in a specific
geographical area.
When considering location-based contributions, analogies with the current work are strong. For
example, Wietschel et al. (2006) study the construction of an hypothetical hydrogen infrastructure
in Europe, up to 2030. Though they analyze the whole hydrogen chain, great emphasis is placed on
production. Interestingly, they choose the same technologies considered in this paper, except that
they also consider also nuclear power. In an attempt to reconcile business sense and the
environment, they find that natural gas reforming would be preferable to coal gasification and RES
at an early stage of development. On a narrower scope, Ramesohl and Merten (2006) discuss only
Germany. They stress that no matter which simulation model is used, great care should be taken to
account for how hydrogen penetration would affect the remaining share of the energy system. In
particular, they point out that every kwh used to produce hydrogen would be inevitably subtracted
from the electric stationary application. This ultimately raises a policy issue on how to best use the
limited RES potential. Ramesohl and Merten (2006) conclude that RES for fueling hydrogen would
be preferable at a later stage, after 2030, and before that car efficiency should be improved. Still
keeping the focus on Germany, Fischedick et al. (2004) underline once again the importance of
pondering how to allocate RES potential among its alternative uses. Moreover, they step back and
stress the paramount importance of reducing the demand of energy to make a RES-based system
viable. Kruger et al. (2003) make the case of hydrogen fuel introduction in New Zealand from 2010
to 2050. With data on demographics, technology, and economics, they analyze what it would mean
4
to produce hydrogen in centralized plants which would then fuel on-board devices. With respect to
feedstock, they take nuclear power into great consideration. Similar issues regarding were raised by
Ramesohl and Merten (2006). Finally comes California, a state that has taken building a cleaner
automotive fleet into great consideration. In Lipman et al. (2004), a team of well-renowned
scientists provides suggestions on how California should continue supporting hydrogen-related
technologies. Since sustainability is key in their view, RES and biomasses must have a pivotal role
in the production of hydrogen.
Since the issue of how to best allocate RES potential between electricity and hydrogen is
mentioned with regularity, Winter (2005) becomes an essential reading. When considering
production feedstock, both electricity and hydrogen share the same versatility. With regards to
storage issues, instead, the two energy vectors face similar limits. In both cases, technological
solutions to store energy are still quite demanding in both financial and energetic terms. Transport
and distribution are not issues for electricity, while they still are for hydrogen. Along these lines, the
overall efficiency from production until the final use of the two energy vectors happens to be a key
factor in determining which system would be preferable. Needless to say, gaseous emission
considerations closely track energy efficiency ones.
on
3. The model
This section is organized as follows. First comes a description of the considered technologies. Then
the reader gets introduced to how energy demand in transport is generated. Next follows a depiction
on how hydrogen share penetration is modeled. Finally, and conveying the previous results, the
model’s outcomes are wrapped up.
3.1 Technological framework
Hydrogen can be derived from a variety of feedstock. These include fossil resources, such as natural
gas and coal, as well as RES, such as biomass, sunlight, and water. Local availability of feedstock,
technology maturity, market applications and demand, policy issues, and costs will all affect the
choice and timing of these various options.
Along the lines of Wietschel et al. (2006), five technologies are considered within Italy: steam
methane reforming, coal gasification, and water electrolysis where power is generated by wind,
solar, and nuclear power.
Steam methane reforming
Steam reforming involves the endothermic conversion of methane and water vapor into hydrogen
and carbon monoxide. The heat is often supplied from the combustion of some of the methane feed-
gas. The process typically occurs at temperatures of 700 to 850 °C and pressures of 3 to 25 bar. The
gaseous product contains approximately 12% CO, which can be further converted to CO, and Hz
through the water-gas shift reaction.
CH, + H20 + heat > CO + 3H> @)
CO + HO — CO; + Hz + heat 2
Partial oxidation of natural gas is the process whereby hydrogen is produced through the partial
combustion of methane with oxygen gas to yield carbon monoxide and hydrogen. In this process,
heat is produced in an exothermic reaction, and hence a more compact design is possible as there is
no need for any external heating of the reactor.
6
Coal Gasification
Hydrogen can be produced from coal through a variety of gasification processes (e.g. fixed bed,
fluidized bed, or entrained flow). In reality, high-temperature entrained flow processes are
preferred, so as to maximize carbon conversion to gas. This avoids the formation of significant
amounts of char, tars, and phenols. A typical reaction for the process is given in equation number
(3), in which carbon is converted to carbon monoxide and hydrogen.
Cy) + H2O + heat > CO + Hy 3)
Since this reaction is endothermic, additional heat is required, as with methane reforming. The
CO is further converted to CO2 and H: through the water-gas shift reaction, as described in equation
(2).
Hydrogen production from coal is commercially mature, but it is more complex than the
production of hydrogen from natural gas. Since coal is abundant in many parts of the world and it
will likely be used as an energy source, regardless, it is worthwhile exploring the development of
clean technologies for its use.
Water electrolysis
Water electrolysis is the process where water is split into hydrogen and oxygen through the
application of electrical energy.
H20O + electricity > Hy + 1/20) 4
The total energy that is needed for water electrolysis is a slight increase in temperature, while the
required electrical energy decreases.
A high-temperature electrolysis process might, therefore, be preferable when high-temperature
heat is available as waste heat from other processes. This is especially important globally, as most
of the electricity produced is based on fossil energy sources with relatively low efficiencies.
Inevitably, the model makes some assumptions on the more complex issues that pertain to future
technological development. In particular, how technical progress will change the chain efficiencies
7
of the five considered technologies is not modeled. Moreover, the future evolution of energy prices
will impact their competitiveness, and, therefore, the amount of investment each technology will
receive. Provided the difficulty in predicting energy prices, the model is oblivious to these market
dynamics.
3.2 Primary energy in road transport forecast
The analysis takes into account the aggregated road transport sub-sectors. It includes the following
groups, according to the PRIMES model in European Commission (2003), which is the data
source:
* — Private (cars, motorcycles)
* Public (buses, taxis)
* Commercial (trucks)
The data have been manipulated, combining them with the percentage of each source on the total
energy demand in the transport industry. These last shares have been extrapolated from Bianchi and
Bianchi and Di Giulio 2005 CEPRIG model, another data source. Bianchi and Di Giulio (2005)
consider the following energy sources for transportation: oil, natural gas, electricity, and renewable.
The trend of total energy demand forecasted by Bianchi and Di Giulio (2005) is as follows:
Year | Tot Energy Oil Natural gas Renewable __ Electric
2005 | 40.46Mtep 98.05% 1.19% 0.57% 0.20%
2010 | 41.70Mtep 97.62% 1.40% 0.62% 0.39%
2015 | 42.10Mtep 96.16% 1.87% 0.67% 0.69%
2020 | 43.20Mtep 95.48% 2.80% 0.75% 0.96%
2025 | 43.64Mtep 94.54% 3.14% 0.82% 1.50%
Figure (1). Bianchi and Di Giulio (2005) energy demand forecast in the transport sector.
Since the energy consumption estimates of the two mentioned models do not stretch until 2050,
three scenarios with different functional specifications were regressed’,
With regard to this section, the assumptions are:
* Hydrogen share equal to 5% in 2050;
2 "Italy: baseline scenario”.
3 See annex 2.
* Hydrogen substitutes only oil-derived products;
* Mtepy=2.5 Mtep,i : equivalence for efficiency between hydrogen and oil;
Adopting Bianchi and Di Giulio (2005) findings, macroeconomic variables such as GDP and
population dynamics are those considered in that model. Similarly, variables related to the motor
fleet such as average mileage, energy efficiency, and number of vehicles are not modeled. How
congestion could affect energy use is also not modeled.
Data sources adopted, both technical and economical, can be found in annex 1.
3.3 Hydrogen share forecast
Consistent with the forecast on energy employed in the transport sector, three scenarios are drawn.
Each scenario consists of a different pattern according to which hydrogen share is supposed to
evolve. The functional form of the pattern determines the speed of hydrogen’s adoption.
Low scenario
This optimistic scenario shows a relevant decrease in energy demand starting from 2030. The
decrease could be due not only to substantial technical improvements, with a subsequent reduction
in energy intensity, but also to real changes in peoples’ behavior with respect to energy
consumption. This scenario embraces a quadratic functional form, so that hydrogen’s share is very
small in the first decades and then increases swiftly with the introduction of new hydrogen
technologies.
%MtepH=a: +b: t+c (5)
Medium scenario
In this case, the assumption is a linear growth of hydrogen’s share, considering also a linear growth
of all other sources.
%MtepH=d: t+e (6)
9
High scenario
Due to the fact that this scenario foresees a continuous growth in the total energy demand for
transport, a logarithmic function serves the purpose best. In this way, hydrogen’s growth is stronger
in the first decades and then its growth rate drops. Such a profile could be explained, for example,
by a strong public incentive-based energy policy, stimulated by a greater care taken by citizens on
transport-related environmental issues.
%MtepH=f -Log(t)+ )
An important point to model is the gain in energy efficiency when hydrogen substitutes oil-
derived products. Since hydrogen is more efficiently produced than traditional oil-derived products,
its energy demand in transport gets proportionally reduced. The actual energy needed after the when
hydrogen gets employed is:
Mtep yyy Mteprn
Mt =
Pra 17 5. %Mtepy.)
(8)
There is therefore a savings in total energy consumed due to the introduction of a new and more
efficient process.
3.4 Model’s outcomes
As previously mentioned, the model provides two outcomes: one relates to physical, infrastructural
requirements and the other extrapolates the cost estimates of building and running such plants. All
the estimates refer to the three scenarios: low, medium, and high. Each scenario embeds different
pattems of both energy demand in transport and hydrogen penetration shares. The low scenario is
the most conservative of the simulations. The high scenario provides the most disruptive case for
the energy system.
Costs
With regard to costs, each energy source has both fixed and variable costs. Fixed costs refer to the
initial investment required to set up a plant. Their magnitude depends mainly on the size of the plant
10
which, in tum, depends on the amount of energy required. Combining standard plant installed
power with energy demand forecasts determines the number of plants to build. This, in turn gives
fixed investment costs.
The amount of energy to deliver each year as hydrogen determines the size of the variable costs.
Each of the five energy sources has specific outlays, such as maintenance, fuel, and
decommissioning, which add up to the global variable cost.
Total costs are computed as follows. Variables costs clearly belong to the year in which they
arise. Fixed costs are charged only to the year in which the plants are built. This may be seen as a
naive choice, but clearness benefits from consistency across energy sources in such a decision. This
is the reason why some of the following graphs, figures (2)-(4), experience bumps.
3,000 5
2,500 +
2,000 +
—— Wind
g —x— Solar
2 1,500 + Nuclear
= ——— Natural gas
—e— Coal
1,000 +
500 +
0
2010 2015 2020 2025 2030 2035 2040 2045 2050
Figure (2). Total costs in the low scenario.
11
4,500 5
—+— Wind
a) —x— Solar
<
2 Nuclear
= - —--~-—Natural gas
—e— Coal
0 r r r r r r r r 1
2010 2015 2020 2025 2030 2035 2040 2045 2050
Figure (3). Total costs in the medium scenario.
12,000 -------------------------------------------------------
10,000 +
8,000 +
—+— Wind
4 —x— Solar
2 6,000 4 Nuclear
= — —--—Natural gas
—e— Coal
4,000 +
2,000 +4
0
2010 2015 2020 2025 2030 2035 2040 2045 2050
Figure (4). Total costs in the high scenario.
12
Solar power under-performs all of the other energy sources almost every year. Sunlight-related
technologies still have some way to go before becoming competitive. Wind, another RES, shows
itself to be even more cost-effective than natural gas. This finding may be partly explained by the
high prices of natural gas experienced in the 2000s. The other hydrocarbon fuel, coal, benefits from
low feedstock costs and contained installation costs. This results in the lowest costs the same
amount of hydrogen over almost all the considered years. Finally, nuclear power, as anticipated,
fluctuates heavily depending on the occurrence of set-up costs. If it was only about variable costs,
nuclear power would challenge coal’s position as the most cost-effective measure for producing
hydrogen.
This simulation does not include any modeling effort that takes emissions into account. In
particular, it seems clear that coal’s ranking, for example, would come into question whenever one
would monetize the GHGs emissions compared to, say, nuclear or solar power.
Infrastructure
The estimates of infrastructure requirements follow along the same lines.
Wind Solar Nuclear Natural gas Coal
number of | square numberof |numberof |billionm? ) numberof | million
windmills | kilometers | 1,000 MW _ |plants (in 2050) | plants hard coal
(cumulative) | (cumulative) | plants (cumulative) (cumulative) | ton (in
(cumulative) 2050)
Low 5,832 67 5 5 2.0 8 6.1
Medium 6,646 TP 5 5 3.3 9 6.9
High 6,829 79 5 5 3.4 9 pt
Figure (5). Infrastructural requirement.
Even considering the scenarios at the two ends of the spectrum, low and high, estimates do not
differ much. If the goal were only to reach 5% of energy demand in transport, however, some of the
energy sources become demanding in terms of infrastructural requirement. For example, five
hypothetical nuclear plants would pose serious troubles in a country, Italy, that dismissed nuclear
power as an energy source with a referendum in 1987. Solar, as well as wind power, would call for
a sizable land surface. Natural gas, in turn, would be less troublesome. The Italian natural gas
13
consumption in 2004 was 78.74' billions m?, which would make 3 billions m? of natural gas easily
buffered by the energy system. Coal would demand more plants than nuclear and natural gas, since
Italian hard coal consumption in 2005 was 48,4 million tons?,
* Source: Eni’s World Oil and Gas Review 2006.
5 Source: Intemational Energy A gency 2007 database, expressed as sum of import and domestic supplies.
14
4, Conclusions
From an economic standpoint, solar and nuclear power suffer from sizable total and fixed costs,
respectively. Natural gas, though promising, gives up a lot to oil market fluctuations, to which its
price is benchmarked. Wind is rather competitive in all of the considered scenarios. Coal
gasification appears to be the most cost-effective solution from a financial perspective. Since this
work is limited to outlining the economic framework, the environmental impact of the different
energy sources ought to be included to draw a final conclusion on which feedstock would be
preferable as policy choice.
With regards to the infrastructure needed, a moderately positive message arises. Except for solar
and nuclear power, the other three alternatives seem to impact the energy system minimally. This
means that there are ways to slowly adopt hydrogen, even without revolutionizing the current
energy infrastructure. Nevertheless, the viability of hydrogen introduction is limited to energy
production.
As pointed out earlier, how RES potential would be allocated between electricity and hydrogen
is key. In particular, what remains to be determined is how investing in RES instead of hydrogen
would impact the energy system differently in terms of global and local emissions. Pairing up these
results is necessary to understand which option is best.
Not only does the competition with electricity needs to be considered, but also transport and
distribution infrastructure. Since this work is concerned with fuel production, it provides only a
partial view that ought to be pondered in a broader context. Research on the downstream side of the
hydrogen chain could well complement the production cost and infrastructure estimates.
Completing the current work would include an analysis on the energy efficiency differential
between internal combustion engines and fuel cells. The tank-to-wheel is already known story: fuel
15
cells can be up to three times more efficient than intemal combustion engines in employing energy’.
Because the difference so striking, a comprehensive analysis could not be oblivious of such a fact.
The current model could be developed further along two lines. One is computing global and
local emissions under different scenarios. This way, the more environmentally-friendly technologies
would get proportional credits they deserve. With respect to global emissions, market values of CO2
per ton provide a good reference of the financial benefit of emitting less. As the Kyoto protocol will
become effective starting in 2008, the European Emission Trading Scheme gives the evaluation a
sound indicator. Local emissions impact is a bit more complicated to assess, but the Externality
Theory can help in this sense.
Another improvement could be modeling technology. In the past, simulation models suffered
from having systematically downplayed the technology potential that could unfold in the future.
Technology’s role, however, is pivotal and influences forecasts significantly. Instead of taking a
conservative stand on scientific development, it would be interesting to draw more dynamic and
realistic evolution pattems.
° Source: http://www.fuelcells.org/basics/benefits_transp.html.
16
5. References
Barreto, L. and Yamashita, K. 2003, “Integrated Energy Systems for the 21 Century: Coal
Gasification for Co-producing Hydrogen, Electricity and Liquid Fuels”, IIASA Interim Report
IR-03-039.
Bianchi, A. and Di Giulio, E. 2005, “Domanda di energia ed emissioni di CO2 in Italia al 2020”
Energia, 4, 50-61.
Committee on Alternatives and Strategies for Future Hydrogen Production and Use, National
Research Council, National Academy of Engineering 2004, The Hydrogen Economy:
Opportunities, Costs, Barriers, and R&D Needs, National Academies Press, 202-204.
Energy Intelligence A gency 2006, Assumptions to the Annual Energy Outlook, Washington DC.
European Commission 2003, European Energy and Transport trends to 2030, Luxembourg.
Fischedick, M., Nitsch, J., and Ramesohl, S. 2004, “The role of hydrogen for the long term
development of sustainable energy systems--a case study for Germany”, Solar Energy, 78, Issue
5, 678-686.
Kruger, P., Blakeley, J., and Leaver, J. 2003, “Potential in New Zealand for use of hydrogen as a
transportation fuel”, International Journal of Hydrogen Energy, 28, Issue 8, 795-802.
Lipman, T., Kammen, D., Ogden, J., and Sperling, D. 2004, "dn Integrated Hydrogen Vision for
California", Institute of Transportation Studies, paper UCD-ITS-RR-04-43, Davis, California.
Mintz, M., Molburg, J., Folga, S., and Gillette, J. 2003, “Hydrogen distribution infrastructure”,
Proceedings of Hydrogen in materials & vacuum systems: First International Workshop on
Hydrogen in Materials and Vacuum Systems - AIP Conference, Newport News, Virginia, 11%
13" November 2002, 671, 119-132.
Ramesohl, S. and Merten, F. 2006, “Energy system aspects of hydrogen as an alternative fuel in
transport”, Energy Policy, 34, Issue 11, 1251-1259.
17
Stoddard, L., Abiecunas, J., and O'Connell, R. 2006, “Economic, Energy, and Environmental
Benefits of Concentrating Solar Power in California’, National Renewable Energy Laboratory,
Overland Park, Kansas.
The European Wind Energy Association 2004, Wind power economics, Brussels.
Wietschel, M., Hasenauera, U., and de Groot, A. 2006, “Development of European hydrogen
infrastructure scenarios--CO. reduction potential and infrastructure investment”, Energy Policy,
34, Issue 11, 1284-1298.
Winter, CJ. 2005, “Electricity, hydrogen--competitors, partners?”, International Journal of
Hydrogen Energy, 30, Issues 13-14, 1371-1374.
World Nuclear Association 2005, The New Economics of Nuclear Power, London.
18
6. Appendix
Annex 1. Data sources.
Data sources
Road transport — Bianchi and Di Giulio 2005
scenario forecast European Commission (2003)
Nuclear power — Energy Intelligence Agency (2006)
World Nuclear Association (2005)
Solar power Stoddard, L. et al. (2006)
Wind power The European Wind Energy Association
(2004)
Natural gas Barreto, L. and Y amashita, K. (2003)
reforming Mintz, M. et al. (2003)
Coal gasification Mintz, M. et al. (2003)
Annex 2. Energy demand in road transport forecast.
Low scenario
Mtep =-0.171 t? +1.745 t + 38.816; R?=0.98
44,
Medium scenario
Mtep = -0.064 t + 1.172 t + 39.41; R7=0.99
455
44 J
234
o
=
424
41
40
S > © o ) S o 9
Ss Ss w § 3 “s S
SS ST LK KL LF LF LK KS
Year
High scenario
Mtep = 0.549 t + 40.46; R?=0.96
20
