Hydrogen mobility from wind energy – A life cycle assessment focusing on the fuel supply

Applied Energy 181 (2016) 54–64

Contents lists available at ScienceDirect

Applied Energy journal homepage: www.elsevier.com/locate/apenergy

Hydrogen mobility from wind energy – A life cycle assessment focusing on the fuel supply Jörg Burkhardt a,⇑, Andreas Patyk a, Philippe Tanguy b, Carsten Retzke c a

Karlsruhe Institute of Technology (KIT), Institute for Technology Assessment and Systems Analysis (ITAS), Karlstraße 11, 76133 Karlsruhe, Germany Total S.A., La Défense 6, 2 Place Jean Millier, 92078 Paris La Défense Cedex, France c Total Deutschland GmbH, Jean-Monnet-Straße 2, 10557 Berlin, Germany b

h i g h l i g h t s

g r a p h i c a l a b s t r a c t

a r t i c l e

i n f o

Article history: Received 9 March 2016 Received in revised form 21 July 2016 Accepted 25 July 2016

Keywords: Hydrogen refueling station Water electrolysis Wind power Life cycle assessment

System Boundary

on production and provision of hydrogen.
Primary data collected from a 700 bar refueling station incl. alkaline electrolyser.
Construction of facilities dominates the primary energy demand and emissions.
Refueling station contributes to same extent to GHG emissions as electricity supply.
Remarkably high expenditures for provision of supplies.

5.4 kgH2/h 15 bar

Electric Grid

Electrolyser

System Boundary


Environmental performance, focusing

900 bar

Compressor

Storage

GH2700 bar

Wind Turbine

Hydrogen Refueling Station

Fuel Cell Vehicle

Vestas 2 MWel 2,500 full-load h/a

67 kWhel/kgH2 3,000 full-load h/a ca. 15 vehicles/day

Mercedes B220 F-Cell 0.975 kgH2/100km

a b s t r a c t In the current debates on reducing greenhouse gas emissions in the mobility sector, hydrogen produced via water electrolysis from renewable electricity is commonly regarded to be a sustainable energy carrier with large potential for decarbonisation of the mobility sector. Directly produced at the refueling stations site, hydrogen greenhouse gas emissions are presently defined to be zero in e.g. the Directives of the European Union since emissions arising from the facilities construction are defined to be negligible. In order to check the validity of this assumption with respect to the latest technical developments in hydrogen supply, the present article aims to report the environmental performance of hydrogen being produced and compressed for mobility purposes. To this end, a state-of-the-art hydrogen refueling station (HRS) with an on-site alkaline electrolyser is assessed, which was built and operated in Berlin. Assuming electricity supply from wind energy generation, a life cycle assessment for the complete value chain was carried out where primary data for the build-up of electrolyser and HRS were obtained during decommissioning of the station. The results show that the construction of HRS and on-site electrolyser requires higher material and energy expenditures compared to previous investigations on similar but technically less advanced systems. These expenditures generate a significant footprint in the specific e.g. greenhouse gas emissions if the electrolyser is operated at a reduced load factor as it may be foreseen for grid stabilisation purposes. To ensure a strong reduction of emissions compared to conventional fuels, this load factor should be

Abbreviations: CED, cumulative energy demand; FCEV, fuel cell electric vehicle; GHG, greenhouse gas; HHV, higher heating value; HRS, hydrogen refueling station; ICE, internal combustion engine; LCA, life cycle assessment; LCIA, Life Cycle Impact Assessment; NR, non-regenerative. ⇑ Corresponding author at: c/o Anna-Louisa-Karsch-Straße 2, 10178 Berlin, Germany. E-mail addresses: [email protected] (J. Burkhardt), [email protected] (A. Patyk), [email protected] (P. Tanguy), [email protected] (C. Retzke). http://dx.doi.org/10.1016/j.apenergy.2016.07.104 0306-2619/Ó 2016 Elsevier Ltd. All rights reserved.

J. Burkhardt et al. / Applied Energy 181 (2016) 54–64

55

sufficiently high and should be defined to not fall under a certain threshold in upcoming directives. Besides, excessive use of supplies should be avoided and the refueling station should be operated with renewable electricity to the largest extent. Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction To fulfil the European Union goals for reduction of greenhouse gas (GHG) emissions by 80% in 2050, Germany foresees to have at least 80% of its electricity production being supplied by renewable resources [1,2]. In 2014, electricity from renewables already contributed to 25.8% of Germanys gross electricity production [3], leading to temporarily production of electricity that exceeded the capacity of the electric grid in some regions. In this case, electricity production from wind turbines and PV modules is curtailed. In 2014, this amount of excess electricity was 1581 GWhel [4] which corresponds to the yearly electricity production of ca. 320 state of the art wind turbines (nominal power of 2 MWel, 2500 full-load hours). 77% of this excess electricity was actually produced by wind turbines [4]. There is a wide consensus throughout the German energy economy that the amount of excess electricity is going to increase with the further transition of the energy system. Strongly depending on the assumptions made, a negative residual load in the electric grid is predicted to occur between 2100 and 4000 h/a by 2030–2050 in several studies [5–7]. Having electrolysers installed to balance the loads in the electric grid, this negative residual load is discussed to be used for water electrolysis. To decarbonize the mobility sector, the use of the produced hydrogen in the mobility sector is discussed. Even though the penetration of fuel cell cars in Germany is limited nowadays, there are 50 hydrogen refueling stations (HRS) to be installed by the mid of 2016 [8]. By 2023, the installation of 400 HRS is foreseen by the H2 Mobility initiative [9]. 1.1. Former investigations on wind electrolysis and hydrogen supply The use of hydrogen in energy applications does not cause direct emissions, of e.g. greenhouse gases, since hydrogen reacts to water when it is oxidised. However, emissions occur during the production, compression and storage of hydrogen, e.g. due to the use of fossil fuels for operation and construction of the whole facilities. A considerable amount of studies have been carried out to determine these emissions occurring from hydrogen mobility with hydrogen being provided from water electrolysis. Regarding the direct production of hydrogen at the HRS site, interesting results have been published in life cycle assessments carried out by Spath and Mann [10], Maak [11] and Patterson et al. [12]. Spath et al. consider a system comprising an electrolyser (30 Nm3/h) being fed by three 50 kWel wind turbines and a HRS delivering hydrogen at 200 bar. They showed that the majority of emissions originate from the systems construction, with the wind turbines having the largest contribution (78%) to the systems specific GHG emissions of 0.97 kgCO2-equ/kgH2. Patterson et al. consider a similar system (30 kWel wind turbines, 350 bar HRS). They use however a different kind of methodology making a direct comparison of results quite complex. Maak et al. assume their electrolyser (60 Nm3/h) to be fed with a European electricity mix and to be operated with a capacity factor of nearly 100%. They conclude that emissions are mainly driven by the electricity supply and could be shifted to the construction phase if electricity was supplied by renewable sources.

Additional studies on wind electrolysis have been published in [13–19]. The majority of these studies refer to results or material inventories from [10,11] or use restricted (or not traceable) databases. Regarding technical parameters of hydrogen refueling stations in the reviewed literature, hydrogen is mostly considered to be compressed and dispensed at 200–400 bars. In contrast, hydrogen is nowadays mostly foreseen to be compressed to up to 900 bars to achieve a pressure of 700 bars inside the vehicles tanks [20,21]. Even though hydrogen mobility is widely investigated for different HRS sizes and pressure levels in major mobility studies, e.g. in [22,23], construction of plants and facilities is mostly neglected in these studies. A specific hydrogen refueling station is also not considered in [24] or [25]; conventional stations are assumed instead. Additionally, electric output powers of often regarded wind turbines (30–50 kWel) may not reflect the state of currently installed wind turbines, having e.g. average electric powers of about 2500 kWel in Germany (2012) [26]. 1.2. Goal of the study The above shows clearly that the availability of primary data on wind electrolysis systems is very limited; the majority of investigations rely on data from few investigations. A similar outcome is found in [27], carrying out a literature review on 21 electrolysis studies and stating that Spath and Mann [10] is ‘‘the major data source for every paper discussing LCA of wind based electrolysis”. Additionally, environmental burden for providing hydrogen by 700 bars HRS are incompletely addressed in literature since expenditures arising from the construction of these HRS are not explicitly regarded due to a lack of primary data. This is covered by [25]. The aim of this paper is to assess the environmental impact related to the production of hydrogen and its dispensing to state of the art fuel cell electric vehicles. Being embedded in an energy system with a high share of intermittent electricity production, hydrogen is assumed to be produced from water electrolysis fed by excess electricity from wind turbines. To catch up with the current state of technology, a modern wind turbine and a 700 bar HRS including an on-site electrolyser being installed and operated in Berlin, Germany, are assessed. In order to provide a valid database, data were collected from the HRS documentation and measurements during the HRS decommissioning. Since primary data on the build-up of the electrolyser and HRS were obtained in the framework of this study, the actual work contributes to fill an important data gap. Besides addressing the scientific community, the article addresses policy makers to classify environmental implications being predominantly related with the facilities construction, which is mostly neglected in major mobility studies. Environmental burdens are assessed with a special view on the influence of embedding the regarded technology in an energy system with high shares of renewable electricity production, which is currently not covered in literature. 2. Methods 2.1. Environmental life cycle assessment (LCA) For investigating the environmental impact, a life cycle assessment of the system has been carried out. Life cycle assessments (LCAs) are compilations and interpretations of

56

J. Burkhardt et al. / Applied Energy 181 (2016) 54–64

– inputs and outputs between the techno sphere and the environment and – potential environmental impacts of a product or product system along the total product life cycle. If, as in this study, technologies are to be assessed, they are represented by their products. The procedure for conducting life cycle assessments is well described by ISO standards 14040/44 [28,29]. According to the dominating practice, the LCA is performed as an attributional, process-based LCA. One typical parameter in an LCA is the cumulative energy demand (CED). It sums up the primary energy demand and can be distinguished in the type of energy resources, i.e. fossil, nuclear or renewable. According to VDI 4600 [30], the CED can additionally be differentiated by the LCAs life cycle phases:

CED ¼ CEDCon þ CEDOp þ CEDDeco CEDCon Cumulative energy demand for construction. Besides efforts for provision of materials and production of components, installation and transportation are considered. CEDOp The cumulative energy demand for operation sums up all primary energies needed to operate a system during its entire life cycle (electricity, fuels, supplies, etc.), including maintenance of components, spare parts, and transportation. In this study, the kinetic energy of the wind is not included in CEDOP. CEDDeco Cumulative energy demand for decommissioning, including deconstruction, recycling and disposal. A special issue is the assessment of solar, wind and water power. Usually, the real conversion of radiation (photovoltaic) or kinetic energy (wind, water) to power is not calculated. Rather, the converted energy is set equal to the produced electric power (e.g. 3.6 MJ wind primary energy per kWh wind electricity). 2.2. Evaluated system Fig. 1 shows a sketch of the evaluated system and its boundaries. The system consists of the following main components: wind

Energy and Material Resources

2.3. Scope and procedure Life cycle assessments for the main components (wind turbine and electrolyser/HRS) are carried out in a modular way. LCA results from the wind turbine are related to its total electricity production. Herewith, specific impacts per kWhel wind electricity (e.g. MJ/ kWhel, gCO2eq/kWhel, etc.) are obtained and used as input for the electrolyser and HRS’s electricity demand. This way, only the

Electrical Grid Wind Turbine

Water Chiller

turbine, water electrolyser, hydrogen compressor, hydrogen storage and dispenser. In the course of this article, the compressor, storage vessels and dispenser are referred to as HRS. A horizontal axis 2.0 MWel wind turbine is considered, which is assumed to be connected to the electric grid and the electrolyser/ HRS. Whereas the electrolyser and HRS are assumed to be solely supplied by wind electricity, electricity not used for electrolyser/ HRS operation is assumed to be used by unspecified electric consumers also connected to the electric grid. The electric grid itself is not considered and excluded from the investigation. The assessed electrolyser and the HRS were installed at the same site and have been operated from May 2010 until December 2012 in Berlin, Germany. The alkaline electrolyser (production capacity of 5.4 kgH2/h) was designed to withstand short load changes that occur when wind or solar power is used [31]. It was built as a compact unit with the electrolyser stack, the transformer/rectifier and the process section (gas separators, pumps, gas conditioning, etc.) being mounted in a steel container. A second container included a piston compressor that compressed hydrogen from 15 bars (electrolyser outlet) to 450 bars or 900 bars before it was stored in carbon fiber vessels located underground. These vessels were stored in steel ‘‘coffins” being recessed in a pool containing 25 m3 of 70% water/30% glycol mixture. Altogether, an amount of about 360 kgH2 could be stored. Via two dispensers, hydrogen was refueled to reach pressures of 350 bars or 700 bars in the vehicles tanks. For the investigations in this study, only 700 bar refueling is considered. Due to the heating of hydrogen during the refill process, an additional cooling to 40 °C is required to limit the temperature inside the vehicles tanks. The fuel cell vehicle is excluded from the results presented in chapter 3. To draw a comparison to mobility from conventional fuels in chapter 4, the vehicles hydrogen consumption and construction efforts are derived from literature.

System Boundary

20 kV

Cable (1km) C 20 kVAC

Waterconditioning 200 VDC

Process Part

Electrolyser Stack

Compressor

Cooling Unit

Wind Turbine

Transf./Rectifier

Electrolyser Container

Compressor Container H 2 Storage

Storage Tanks

Dispenser

Dispenser System Boundary

Fuel Cell Vehicle Emissions Fig. 1. Flow chart of the evaluated system and its boundaries.

J. Burkhardt et al. / Applied Energy 181 (2016) 54–64

energy and environmental burdens effectively arising from the electrolyser and HRS’s electricity demand are considered. Electricity produced but not consumed by the electrolyser and the HRS is fed into the electric grid and not considered in the LCA. The HRS must not be seen as being supplied by only one wind turbine. Since electrolyser and HRS only consume as much electricity as needed for their operation, both components can also be seen as being generally supplied by state of the art wind turbines, located in windy regions. However, grid losses are assumed to be negligible since the wind turbine is assumed to be located nearby the electrolyser and HRS (see Section 2.4.1, Wind Turbine). The amount of the wind turbines full load hours is assumed to be 2500 h/a, which corresponds to an average onshore site in Brandenburg, Germany. The electrolyser is assumed to be operated for grid stabilisation purposes. According to predictions for future electrolyser load factors in systems with high shares of intermittent electricity production in [5–7], an average value of 3000 full-load hours (corresponding to a load factor of 34.2%) is used. With a production capacity of 5.4 kgH2/h, an average mass of 44.4 kgH2 is produced and dispensed per day. Due to the higher electric power of the wind turbine and the unsteady availability of wind, the wind turbines lower amount of full-load hours (2500 h/a) is not in contradiction with the higher amount of full-load hours of the electrolyser (3000 h/a). The fullload hours only reflect the time span that the turbine is running at nominal power. For a better understanding, Fig. 2 shows the classified electric output power for a Vestas V90 being operated at Kahler Asten, Germany. Hourly averaged wind data from 2012 are taken from measurements [32], the power characteristic of the turbine (depending on wind velocity) is taken from the manufacturer [33]. Assuming an availability 100%, the turbine produces electricity for ca. 8000 h/a, but only for ca. 900 h/a at nominal load. For ca. 6000 h/a, the turbines output power exceeds the combined input power of electrolyser and HRS (364.6 kWel, compared to Table 1).

2.3.1. Functional unit The functional unit of investigation is 1 kgH2 at gaseous state, 700 bars and approximately 40 °C (39 °C to 33 °C) for refueling fuel cell cars according to SAE J2601 [21], which meets the requirements of next generation fuel cell cars in series production from e.g. Toyota and Hyundai [34,35]. Additionally, GHG emissions for 1 km driven by a middle class passenger car are considered to provide a comparison to conventional fuels.

2.3.2. Temporal and geographic reference The period of 2010–2015 seems plausible regarding the real operation time and the specification of the regarded system. Irrespective of this fact, energy and material flow data are taken from

Powe er [kWel]

2000

1000 500 Electrolyser and HRS 364

8760

0 0

2000

4000

a database (Ecoinvent, [36]) with different reference periods, as is common practice. Geographic reference: The site of the HRS is Berlin; the wind turbine can be located nearby in Brandenburg, i.e. close to Berlin. However, this reference only applies strictly to the operation phase. The production of the equipment, the used materials, etc. extend the reference location to Europe or supplying countries outside Europe. Further, the system can be installed in a similar way in all regions of the world having similar climatic conditions. Therefore, it is not mandatory to specify the geographic reference. Beyond that, many energy and material flow data taken from a database have different or no specific references. Irrespective of this fact, they are used as surrogate data, as is common practice. The results are largely valid on a global scale. 2.3.3. Environmental indicators and impacts To evaluate the global warming, equivalents of CO2 (CO2equ) are assessed, which sum up the global warming potential of all greenhouse gases (CO2, CH4, etc.) and relate their differing impacts to equivalents of CO2. Additionally, terrestrial acidification, freshwater eutrophication, human toxicity and terrestrial ecotoxicity are considered. Life Cycle Impact Assessment (LCIA) is carried out in accordance with the ReCiPe 2008 method [37] and using the cumulative energy demand method [30]. In order to provide a more illustrative presentation of results, the kinetic energy of the wind (3.6 MJ/kWhel) is excluded from the CED. 2.4. Data generation The setup of an LCA requires a life cycle inventory (LCI), which lists up the amount of resources, materials, wastes and emissions related to the system under study. The LCI is based on lists of components of an evaluated system and, respectively, the amount of materials needed for their construction, operation and deconstruction. Fundamental distinctions are made between process or component categories and the corresponding data categories for the system being studied:
Specific processes, plants or components for which data are available from data sheets, name plates, own measurements or assessments based on measurements, interviews with technical staff, etc.
Specific processes, plants or components where data were adopted from other LCAs, technical studies or assessments based on them. These were used for the wind turbine (type, power and material composition), full-load hours, lifetimes, etc.
Generic processes: production processes for the wind turbine and HRS (surrogate data), upstream processes such as provision of materials, auxiliary and operating materials and fuels that are used. These data are taken from established, validated database systems [36] or are generated on that basis. The following sections give explanations on assumptions and data used to assess the three main life cycle phases construction, operation and decommissioning.

Wind Turbine

1500

57

6000

8000

Time [h/a] Fig. 2. Example of a wind turbines classified electric output power.

2.4.1. Construction phase The construction phase considers material and energy efforts for the construction of components. In addition to concrete, metals contribute most to the mass balances of the system. For provision of metals, German shares for the use of secondary metals in 2012 are applied, which are 45% for iron, steel and copper, and 60% for aluminium [38]. The total masses of the required individual materials are linked to respective Ecoinvent datasets for material provision.

58

J. Burkhardt et al. / Applied Energy 181 (2016) 54–64

Since many components (electric motors, pumps, valves, heat exchangers, etc.) of the considered system require sophisticated construction, energy expenditures in manufacturing processes of metals are seen to not be negligible and are therefore considered via the Ecoinvent database. According to [39,40], construction of components is considered in two steps. The first step is rolling the metal to a semi-finished product. The second step is finishing the product individually. For the latter, average metal construction data sets are applied, which have been established by several manufacturers in Europe between 2002 and 2005 [39]. The manufacturing process has a considerable influence on results since for e.g. low alloy steel, manufacturing of the component (6.5 + 31.5 MJ/kg) demands a larger primary energy input as compared to the provision of the steel itself (27.9 MJ/kg). For high alloy steel, expenditures increase by 70% (from 76.6 to 129.9 MJ/kg). An exception from this method is made for the wind turbines steel tower which is assumed to be rolled twice and welded afterwards [41,42]. Also the amount of steel used to build the coffins containing the hydrogen storage tubes is assumed to be only rolled to sheets. Carbon fiber materials (C3) are assumed to be produced in a pyrolysis process from acrylonitrile (C3H3N) from Sohio process (at plant). Regarding the ratio of molar masses (53 g/mol for C3H3N and 36 g/mol for C3), 1.47 kgC3H3N have to be spent for 1 kgC3. The process energy (5 MJ/kgC3) for carbonization is delivered by natural gas (at industrial furnace >100 kWth). Electronics, such as control units, and displays are considered by using the Ecoinvent dataset ‘‘electronics for control units”. 2.4.1.1. Wind turbine. For the 2 MWel wind turbine assessed in this study, a detailed material inventory is taken from the manufacturer [43]. The hub height was increased from the initial value of 80–105 m in order to meet average values of newly built wind turbines in Germany in 2012 [26]. Based on specific tower masses derived from [33,44], the steel mass of the tower is increased by 2 t/m. The wind turbines electricity outlet voltage being 20 kV, its connection to the electrolyser and HRS is considered by a 1 km mid-voltage cable. 2.4.1.2. Electrolyser and HRS. Even though the electrolysers (threephase) transformer/rectifier (total mass: 2000 kg) only transform electricity from 400 V to ca. 200 V, an additional transformer (20 kV to 400 V) is not accounted for in the LCA. At the original HRS plants site, this (20 kV–400 V) transformer was used by several devices and can therefore not solely be accounted to the electrolyser and/or the HRS. Data from the electrolyser and the HRS were obtained from a commercial refueling station including an on-site electrolyser being installed in Berlin, Germany. To evaluate efforts arising during the construction of the plant, its decommissioning was attended and a material inventory – based on measurements, documentation and interviews with technical staff – was established. Individual components, especially of the electrolyser container, are determined to the largest possible extent. Material masses of uniform components (steel trays, water and hydrogen piping, tanks, etc.) are calculated using their dimensions (from measurements and documentation) and the specific material density. Masses of complex components (e.g. transformer/rectifier, electrolyser stack, water chiller, heat exchangers, etc.) are taken from operation manuals, technical drawings and nameplates. Weights of foundations are calculated via their dimensions and the density of concrete. Specific material shares in the total mass of the components are calculated via material shares derived from literature data. This way, the cumulated masses for the main components (electrolyser container, compressor container, storage, dispenser and

piping) have been calculated. These cumulated masses are validated with the component masses that have been measured by a crane (accuracy of ±250 kg) during decommissioning. Taking the main components overall masses and the subcomponents mass shares into account, the plants material inventory is obtained and given in Table 3 and Tables A.1–A.4. While the calculated weight fits with the measured one for the hydrogen storage, deviations occur for the electrolyser and the compressor containers. These deviations originate from the construction of these containers and their sub-components, which are far more complex than the hydrogen storage. Therefore, the measured weight was taken as the reference value. Its difference to the calculated weight is caused by an unspecified mass which is assumed to consist of 25% highly and 75% lowly alloy steel. This assumption has been derived on basis of component’s material shares being determined within the containers, mainly steel-based. 2.4.2. Operation phase Except for the spare parts, the operational life of all main components is assumed to be 20 years. 2.4.2.1. Wind turbine. For the wind turbine, energy expenditures during operation are mainly caused by maintenance and spare parts: the gear box is assumed to be changed after 10 years and 150 L of hydraulic oil are changed every year according to [41]. 2.4.2.2. Electrolyser and HRS. The electrolyser and HRS relevant operational parameters have been derived from the manufacturer’s documentations and are defined in Table 1. The evaluated efficiencies/specific electricity consumptions are in good agreement with those defined in [45]/[46]. A list of supplies needed for operation was also available from the manufacturer. Besides a small amount of lye for the electrolyser, mainly hydraulic oil for the compressor as well as glycol for cooling circuits and the hydrogen storage were needed. The complete need for supplies is allocated to the operation phase. The compressor was designed to run simultaneously to the electrolyser. The hydraulic aggregate driving the pistons of the compressor uses a mineral oil that is usually changed after approx. 2500 h of operation [47]. In the present study, an exchange after 3000 h is assumed. After its use, the hydraulic oil is assumed to be burned in an industrial furnace (1 MWel); the substituted priTable 1 Operational parameters of electrolyser and HRS [31,48]. Electrolyser Nominal capacity Max. electric power Specific electricity consumption Efficiency (HHV) Operating temperature Outlet pressure Working range

60.0 288.0 53.4 73.8% 80.0 15.0 5–100%

Nm3/h kWel kWhel/kgH2 % °C bar %

Electrolyser peripherals Nominal Power Specific Electricity Consumption

26.3 4.9

kWel kWhel/kgH2

Compression (900 bar) incl. H2 cooling Nominal capacity Specific electricity consumption

50.3 9.4

kWel kWhel/kgH2

360.0 38 2

kgH2 – –

364.6 67.6 58.3%

kWel kWhel/kgH2 %

H2 storage Total storage capacity Amount of storage tubes, 450 bar Amount of storage tubes, 900 bar Total (H2 at p = 900 bar; T = 40 °C) Nominal electric power Specific electricity consumption Efficiency

59

J. Burkhardt et al. / Applied Energy 181 (2016) 54–64 Table 2 Fuel consumption and GHG emissions from vehicle construction.

Fuel consumption Vehicle construction

Gasoline

Diesel

Hydrogen

6.2 L/100 km 30.53 gCO2-equ/km

4.6 L/100 km 30.78 gCO2-equ/km

0.975 kgH2/100 km 57.0 gCO2-equ/km

mary energy (light fuel oil) is credited to the process. The water/ glycol mixture of the hydrogen storage is assumed to be changed after 10 years. With respect to the degradation of the electrolyser stack, its complete replacement is assumed after 10 years, which corresponds to a stack durability of 8–15 years defined in [49]. Based on experience with strong attrition of valves during operation of the plant as well as other power-to-gas facilities [50], the valves and piping of the hydrogen conditioning unit are assumed to be replaced after 10 years. Hydrogen losses (e.g. from purging of the electrolyser) and electrolyser degradation strongly depend on the electrolysers operation and are hard to determine. For the investigations at hand, both parameters mainly affect the electricity demand of electrolyser/HRS and are assumed to be already covered by the electricity demand defined in Table 1. Regarding the fact that the electricity demand is nearly 40% higher as defined in e.g. [25], this assumption appears to be conservative. A dependency between the load factor and the electrolysers/ HRSs operational life is not considered. However, operation of several electrolysis and HRS projects by Total Deutschland showed that a constant operation is preferred to an interrupted operation with long standstills in terms of reliability and maintenance frequencies. 2.4.2.3. Vehicle. The vehicle itself is not considered in the results given in chapter 3. For a comparison to conventional fuels, specific GHG emissions (per km) of a middle-class vehicle are assessed in chapter 4. The fuel consumption from a Mercedes B-Class [51] is considered, being applicable for different power trains (see Table 2). Additionally, the vehicle construction and operation is given using data from [24], neglecting efforts for the construction and maintenance of roads. The comparably higher GHG emissions of the fuel cell electric vehicles construction are related to ‘‘additional burdens from fuel cell production” [24]. 2.4.3. Decommissioning phase The end of life phase is represented by the deconstruction, which comprises craning and transportation of the machinery and decommissioning of concrete. Recycling of metals is not explicitly modelled but implicitly considered by using supply mixes primary/secondary for the metals used in the construction phase. Disposal of supplies is considered in the operation phase. 3. Results This chapter shows the determined material inventory and results from the systems LCIA. Results from the wind turbine are not depicted since its material inventory is adopted from [43]. The assessed wind turbines specific GHG emissions of 11.6 gCO2equ/ kwhel are in good agreement with similar investigations [52]. 3.1. Mass and material inventory, electrolyser and HRS Fig. 3 gives an overview about electrolysers and HRSs mass shares. The foundations of the components are mainly made of concrete and contribute 64% to the total HRS mass of 163.3 t. Apart from the foundation concrete, the material shares are dominated

Total Mass: 163.3t Dispenser and Piping 0%

11% Electrolyser 8% Foundation Electrolyser

Foundation H 2 Storage 50%

9% Compressor

7% Foundation Compressor

15% H 2 Storage Fig. 3. Mass shares of electrolyser and HRS.

Table 3 Material inventory, electrolyser and HRS.

Low alloy steel High alloy steel Cast iron Copper Aluminium Nickel Polymer Carbon fibers Resin Zeolith Electronics Concrete

[t]

[%]

37.0 6.2 1.2 1.3 0.3 0.4 0.4 11.3 0.1 0.1 0.1 105.1

22.6 3.8 0.7 0.8 0.2 0.3 0.2 6.9 0.1 0.0 0.1 64.3

163.3

100.0

by iron metals (27.1%) and carbon fiber materials (6.9%) used for hydrogen storage, see Table 3. Due to the underground storage of hydrogen, the foundation for the hydrogen storage is by far the heaviest component (50% of the whole mass). Since the tanks were stored in a glycol/water mixture, not only the ground needed to be covered with concrete (as for the electrolyser and compressor) but also the walls. Electrolyser and compressor containers both contribute to ca. 10% to the whole mass whereas the dispenser and piping at the site of the HRS have a negligible contribution. Related to the weight of the whole plant, the electrolyser stack itself only accounts for ca. 1%. The measured container weight of the electrolyser is in good agreement with the mass specification (16 t) of a state-of-the-art alkaline electrolyser from another manufacturer, having the same production capacity (60 Nm3H2/h) and output pressure (15 bar) [53]. Compared to the material inventory defined in [11], also comprising a HRS incl. a 60 Nm3/h electrolyser, the masses of electrolyser and compressor are more than twice as high. For both components, the share of high alloy steel defined in [11] is 50– 60% higher (compressor, electrolyser) than determined in the study at hand. In contrast, the demand for nickel (electrolyser electrodes) is nearly identical in both studies.

60

J. Burkhardt et al. / Applied Energy 181 (2016) 54–64

3.2. Life cycle impact assessment

100%

Transport: installation and maintenance Manufacturing: process heat Manufacturing: nonferrous metals Manufacturing: steel and ferrous metals Supplies and disposal

90%

Table 4 shows results for the environmental indicators regarded in the systems life cycle assessment. All indicators are distinguished in the life cycle phases construction, operation and decommissioning. Regarding the CED, a further differentiation into fossil, nuclear and renewable primary energy is made whereas the GHG emissions are split into contributions of the systems main components: wind turbine, electrolyser and HRS. The results of the CED and GHG emissions are further discussed in Sections 3.2.1 and 3.2.2. Fig. 4 relates to the same indicators as given in Table 4, showing the contributions of material and energy expenditures to the particular indicators. A distinction is made for the provision of materials, manufacturing and transport processes. The CED and the GHG emissions are mainly defined by provision of materials and the disposal of supplies, altogether contributing to 70–71%. Manufacturing processes, predominantly requiring steel and ferrous metals, contribute to 26–27% whereas transport plays a minor role (3%). Regarding the provision of materials, steel and ferrous metals contribute most (29–32%). The remarkable share of supplies and their disposal (17–18%) is mainly related to the glycol demand for the underground hydrogen storage. In contrast to CED and GHG emissions, the remaining indicators are largely defined by the provision of nonferrous metals, e.g. contributing to 38% to the terrestrial acidification and mainly arising from the nickel electrodes of the electrolyser. The provision of nonferrous metals also contributes to 20–25% to the freshwater eutrophication and human toxicity. For these indicators, the provision of electrics contributes to 11–12%, mainly caused by the construction of the wind turbine. The terrestrial ecotoxicity is largely affected by the provision of steel and ferrous metals (36%).

80% 70% 60% 50% 40% 30%

Concrete

20% 10%

Electrics

0%

Carbon fibers, polymers, miscellaneous Nonferrous metals Steel and ferrous metals

Fig. 4. Shares of material and energy expenditures, whole system.

CED = 11.1 TJ Decommissioning 2%

32% Wind Turbine

Operation

Construction

24%

74%

13% Electrolyser 10% Compressor 18% H2 Storage 1%

Dispenser and piping

Fig. 5. Cumulative energy demand, whole system.

3.2.1. Cumulative energy demand The total primary energy input during construction, operation and decommissioning of the whole system is 11.1 TJ. Relating the CED to the total hydrogen production in 20 years of operation, the specific CED is obtained which is 34.3 MJ/kgH2. Its fossil energy content is 28.1 MJ/kgH2 (82% of the CED), corresponding to the energy content of ca. 0.9 l gasoline. Fig. 5 shows the systems CED and the respective contributions of its three lifecycle phases. Considering these phases, the systems construction has by far the largest contribution (74%), followed by the operation phase (24%) and a nearly negligible decommissioning phase (2%). The construction phase is further differentiated according to energy expenditures for producing the main components. Regarding these components, the wind turbine contributes to the largest extent (32%). The hydrogen storage accounts for the second highest share in the construction phase, which is mainly caused by the need for carbon fiber materials for the storage tanks, and not so by the significant

mass of the concrete footing (compared to Fig. 3). Electrolyser and compressor contribute to roughly the same extent, whereas dispenser and piping at the site play a minor role. Expenditures of the operation phase are caused by provision and disposal of supplies, transports and spare parts. In here, the provision and disposal of glycol, mainly needed for the underground hydrogen storage, takes the largest part (13%).

3.2.2. Green-house gas emissions The specific GHG emissions for producing and refueling hydrogen to FCEV (1.92 kgCO2-equ/kgH2) correspond to the direct emissions of burning 0.8 l gasoline. As can be seen in Table 4, the HRS approximately contributes to the same extent to the systems overall GHG emissions as the wind turbine does. Regarding the shares of the operation phase, electrolyser and HRS show a greater

Table 4 Specific results, whole system. Construction phase (%)

Operation phase (%)

Decommissioning phase (%)

Specific CED Fossil + prim. forest Nuclear Renewable

34.3 28.1 4.3 1.8

MJ/kgH2 MJ/kgH2 MJ/kgH2 MJ/kgH2

74 73 76 79

25 25 23 20

2

Climate change Wind turbine Electrolyser HRS

1,919.3 787.0 368.4 763.9

gCO2-equ /kgH2 gCO2-equ /kgH2 gCO2-equ /kgH2 gCO2-equ /kgH2

74 87 70 64

24 11 28 35

2

Terrestrial acidification Freshwater eutrophication Human toxicity Terrestrial ecotoxicity

11.3 1.2 1,772.1 0.3

gSO2-equ/kgH2 gP-equ/kgH2 gDCB-equ/kgH2 gDCB-equ/kgH2

70 84 85 77

29 16 15 21

1 0 0 2

2 1 0 2 2 1

61

J. Burkhardt et al. / Applied Energy 181 (2016) 54–64

Specific GHG emissions: 1.92 kg CO2-equ /kg H2 41%

19%

Electrolyser

12%

27%

Compressor

1%

Storage

Dispenser

Fig. 6. GHG emissions, whole system.

3.2.3. Energy amortization time and physical harvest factor The energy amortization time tA is used to assess the energy sustainability of renewable energy systems compared to conventional ones. According to [30], tA is determined by relating the non-regenerative (NR) CEDs for construction and decommissioning of a renewable system (CEDCon;NR and CEDDeco;NR ) to the substituted amount of primary energy from a conventional reference system (CEDRef ;NR ) minus CEDOp;NR of the renewable system. The definition of tA is shown in Eq. (1):

tA ¼

CEDCon;NR þ CEDDeco;NR CEDRef ;NR  CEDOp;NR

ð1Þ

The amount of energy being substituted by hydrogen is the pro_ H2 ) and hydrogen higher heating duct of dispensed hydrogen (m value (HHV H2 ). CEDRef ;NR is obtained by dividing this value by the efficiency of a comparable reference system (gRef Þ, see Eq. (2):

CEDRef ;NR ¼

_ H2  HHV H2 m

gRef

ð2Þ

The conventional reference energy carrier used is gasoline. Referring to data from GEMIS [54], the efficiency of its provision to stations gRef is 84.8%. According to this, the energy amortization time (or energy payback time) of the considered system is 3.1 years. In other words: after 3.1 years the system will have saved the same amount of non-renewable primary energy as needed for its construction, operation and decommissioning. By relating the higher heating value of hydrogen (142 MJ/kgH2) to the non-regenerative specific CEDNR (32.5 MJ/kgH2), the physical harvest factor is obtained [30], which is approximately 4.4. This means that the energy benefit is more than four times higher than the non-regenerative energy input. 3.2.4. Increased electrolyser workload As explained in Section 2.3, the electrolyser load factor is chosen according to its operation with excess electricity. Assuming a higher availability of electricity or a direct connection of the electrolyser to a wind park, the load factor can be increased. Taking the classified power curve in a windy German region in Fig. 2, electrol-

2500

GHG Emissionss [gCO2-equ/kgH2]

contribution as compared to the wind turbine. For the electrolyser, this is mainly caused by the replacement of the stack whereas the supply/disposal of glycol mainly defines the expenses during the operation of the HRS. Regarding the contribution of the single components to the overall GHG emissions, a further differentiation is given in Fig. 6. For each component, the emissions arising during the construction, operation and decommissioning phase are taken into account. With a share of 41%, the wind turbine contributes to the largest extent to the overall GHG emissions. The remaining 59% are emitted by the electrolyser and the HRS whereby the same order is found as for the components masses (compared to Fig. 3) and the CED (compared to Fig. 5). With 27%, the storage contributes to the largest extent, which can be explained by the large storage capacity of ca. 360 kgH2 and the corresponding amount of carbon fiber material.

HRS Electrolyser

2000

Wind Turbine

1500

1000

500

0 2000

3000

4000

5000

6000

7000

8000

Electrolyser full-load hours [h/a] Fig. 7. Sensitivity of specific GHG emissions from the electrolyser workload.

yser and HRS could theoretically be operated with up to 6000 fullload hours per year by only being connected to one 2 MWel wind turbine. Fig. 7 shows the sensitivity of the specific GHG emissions from the load factor of the electrolyser. The system’s specific emissions decrease by 30% if the workload is increased from 3000 to 6000 full-load hours per year or by 50% if the workload is increased from 2000 to 8000 h/a. The reason for the reduction is that expenditures arising from the construction of electrolyser and HRS are allocated to a larger hydrogen production (and use). This can clearly be seen from Fig. 7 where the systems specific GHG emissions are split into the contributions of wind turbine, electrolyser and HRS. Since emissions from the wind turbine are independent from the electrolyser’s workload, the wind turbines contribution (787 gCO2-equ/kgH2) to the total specific emissions does not change. In contrast, emissions arising from electrolyser and HRS decrease by 74% (1694 to 434 gCO2-equ/kgH2) if the electrolyser is operated with 8000 instead of 2000 full-load hours. Relating the electrolyser and HRSs emissions to the fuel consumption of the fuel cell electric vehicle (0.975 kgH2/100 km, see Section 2.4.2), a span from 4.2 to 16.5 gCO2-equ/km is obtained.

4. Comparison to mobility from conventional fuels To obtain a comparison to mobility from conventional fuels, results are compared on basis of specific GHG emissions per km (see Fig. 8). Data for conventional fuels are taken from [55], including upstream processes (exploration of mineral oil, refinery processes, etc.). Fuel consumptions for the studied power trains – internal combustion engines (ICE) for gasoline and diesel, fuel cell electric vehicle (FCEV) for hydrogen – are based on manufacturer information, emissions arising from the vehicles construction are taken from [24] (see Section 2.4.2). Emissions from the provision of gasoline and diesel are summarised in ‘‘Fuel supply” whereas emissions for the provision of hydrogen are broken down in contributions of Electricity supply,

62

J. Burkhardt et al. / Applied Energy 181 (2016) 54–64

GHG Emissions ons [gCO2-equ /km]

220 200

Vehicle Fuel supply HRS Electrolyser Electricity supply

203.2

180

168.8

160 140

117.3

120 100 75.7

80

125.7

89.0

60 40 20 0

ICE - Gasoline

ICE - Diesel

FCEV - Base Case

FCEV Offshore Wind

FCEV - PV

FCEV - Base Case, HRS Grid

Fig. 8. Specific GHG emissions of different power trains.

Electrolyser and HRS. Whereas FCEV-Base Case represents results from chapter 3, the other FCEV cases show results from the same system and an identical electrolyser workload but with alternative electricity supplies from an offshore wind park (Offshore Wind, 32 gCO2-equ/kWhel, [40]) and from photovoltaic panels (PV, 75 gCO2-equ/ kWhel, [52]). In FCEV – Base Case, HRS Grid, the HRS is assumed to be fed by electricity from the electric grid (German electricity mix 2012, 562 gCO2-equ/kWhel, [56]) like applied to most HRS being operated in Germany nowadays. For the FCEV, emissions occurring from the vehicle are nearly twice as high as compared to those from the ICEs (compare to Table 2). With 57 gCO2-equ/km, the vehicle contributes to the largest extent to the total emissions of all FCEV scenarios, causing e.g. three times higher emissions than the provision of hydrogen (wind turbine, electrolyser and HRS) does in the FCEV-Base Case Scenario. However, all FCEV cases reduce emissions compared to the conventional fuels. Due to the lowest GHG footprint of electricity from onshore wind turbines, the FCEV-Base Case shows the strongest potential for reduction. Compared to the fuel supply of gasoline and diesel ICEs, emissions are reduced by 89% and 86%. Additionally regarding the vehicle, ‘‘only” a total GHG emission reduction of 63% (gasoline) and 55% (diesel) is achievable. In FCEV-Base Case, electrolyser and HRS contribute 15% to the overall emissions. As can be seen from Fig. 8, the FCEV’s emissions are also quite sensitive on the origin of electricity. Taking the FCEV-PV scenario, 30% of GHG emissions are reduced compared to diesel. Electrolyser and HRS contribute 9% to FCEV-PVs overall emissions. According to the relatively high GHG emissions of the German electricity mix, the HRS Grid-Scenario shows the highest emissions of all FCEVscenarios.

5. Summary, discussion and conclusion Results from the life cycle assessment show that hydrogen being supplied by the considered system bears a strong potential to lower GHG emissions as compared to conventional fuels. Solely regarding the vehicles fuel consumptions, FCEV reduce emissions by 86–89% as compared to ICEs driven by gasoline or diesel. The major (74%) primary energy input for providing hydrogen arises from the construction of wind turbine, electrolyser and the refueling station. Due to the relatively low load factor of the electrolyser (3000 full-load hours/year), the construction phase leaves a remarkable footprint in the systems specific CED (34.3 MJ/kgH2) and GHG emissions (1919 gCO2-equ/kgH2). Specific emissions for hydrogen provision can be reduced considerably by increasing the electrolyser workload, e.g. GHG emissions decrease by 30% if the electrolysers amount of full-load hours is increased from 3000 h/a to 6000 h/a. However, an opera-

tion with a higher load factor may be in conflict with the purpose of only using excess electricity. Nevertheless, expenditures for hydrogen provision are relativized if the construction of the vehicle is additionally included. In the basis scenario, specific emissions for producing and refueling hydrogen correspond to 1/3 of the emissions occurring from the vehicles construction and maintenance. The construction of electrolyser and HRS also plays a rather minor role for the total GHG emissions if the electricity supply is more CO2 intense as compared to the onshore wind turbine being assessed in the investigations base case. Regarding the whole value chain (including the vehicle), electrolyser and HRS contribute to 9–15% to the total GHG emissions of the considered scenarios. However, emissions occurring from hydrogen provision are found to be twice as high as compared to those emissions being assessed in a study on a similar but older system (Spath and Mann [10]: 970 gCO2-equ/kgH2). Deviations are attributed to mainly arise from the systems different and technically more challenging parameters (H2 at 900 vs. H2 at 200 bar), requiring higher component masses and higher expenditures for construction. This is indicated by the wind turbines being the main driver (78%) for GHG emissions in [10] whereas the majority of emissions (59%) arise from electrolyser and HRS in the study at hand. The only material balance found to be available from literature (Maak [11]), defines component masses for electrolyser and compressor that are less than half as high as evaluated from the HRS assessed in the study at hand. Altogether, the construction phase causes the majority of the systems primary energy demand and GHG emissions. Even though the contributions of electrolyser and HRS may be outnumbered by the impact of the vehicle at first glance, they can still leave a remarkable footprint in the total emissions, especially if the workload of the electrolyser is rather low. Regarding plant applications it can be concluded that – the demand for supplies can have a significant footprint. Refraining from e.g. underground hydrogen storage in water/ glycol mixtures may help to reduce emissions. – the workload of the system should be as high as possible to reduce specific emissions arising from the construction of facilities. Operating the considered system solely by excess electricity may not only endanger the economics but also the environmental performance of the system. – the hydrogen storage contributes most to the HRSs emissions and should therefore be adjusted to the hydrogen demand and the availability of excess electricity for electrolyser operation. Limitations and outlook: – In the study at hand, the HRS is assumed to be supplied by electricity from renewables. In contrast, HRS are nowadays usually operated by electricity from the electric grid since compression and hydrogen cooling demand a reliable electricity supply. To reduce emissions, investigating the possibility to operate HRS also with electricity from renewables seems meaningful. – Since only few material inventories from HRS and electrolysers are found to be available from literature, further research on different plant sizes and supply chains seems to be required to catch up with the latest state of technology. In the same way the construction of the cars should be considered as it is done up to now in a limited number of studies only, e.g. [24,57].

Acknowledgement The authors are grateful for the precious support by the Total Deutschland retail direction team Hydrogen/E-Mobility.

J. Burkhardt et al. / Applied Energy 181 (2016) 54–64

Appendix A See Tables A.1–A.4.

Table A.1 Material inventory, electrolyser (excl. foundation). [kg] Low alloyed steel High alloyed steel Cast iron Copper Aluminium Nickel Polymer Resin Zeolith Electronics

%

11,410.4 3,553.0 463.1 1,004.7 175.3 432.9 307.1 106.8 71.0 85.0

64.80 20.18 2.63 5.71 1.00 2.46 1.74 0.61 0.40 0.48

17,609.4

100.00

Table A.2 Material inventory, compressor (excl. foundation). [kg] Low alloyed steel High alloyed steel Cast iron Copper Aluminium Polymer Electronics

[%]

11,927.1 2,085.4 739.8 279.5 77.3 64.7 26.0

78.47 13.72 4.87 1.84 0.51 0.43 0.17

15,200.0

100.00

Table A.3 Material inventory, hydrogen storage (excl. foundation). [kg] Low alloyed steel Monomere and Polymer

[%]

12,278.3 11,284.5

52.11 47.89

23,562.8

100.00

Table A.4 Material inventory, dispenser and piping. [kg] Low alloyed High alloyed Electronics

158.70 521.25 6.00 685.9

[%] 23.14 75.99 0.87 100.00

References [1] European Commission, editor. ‘‘A Roadmap for moving to a competitive low carbon economy in 2050”. European Commission, Communication from the Commission to the European Parliament, COM (2011) 112 final; March 2011. [2] The Federal Government, editor. ‘‘Energiekonzept für eine umweltschonende, zuverlässige und bezahlbare Energieversorgung”. The Federal Government, Germany; September 2010. [3] German Association of Energy and Water Industries, editor. ‘‘Renewables’ Share Expected to Rise to 33 Percent in 2015”. German Association of Energy and Water Industries, Press release, March 2015, , availability checked in December 2015. [4] Bundesnetzagentur, editor. ‘‘Monitoring Report 2015”. Bundesnetzagentur für Elektrizität, Gas, Telekommunikation, Post und Eisenbahnen, Bundeskartellamt; November 2015. [5] Stolzenburg K et al. ‘‘Integration von Wind-Wasserstoff-Systemen in das Energiesystem”. Nationale Organisation Wasserstoff- und Brennstoffzellentechnologie (NOW GmbH); March 2014.

63

[6] Henning H-M, Palzer A. ‘‘Energiesystem Deutschland 2050”. Fraunhofer Institute for Solar Energy Systems ISE; November 2013. [7] Höflich B et al. Integration der erneuerbaren Energien in den deutscheuropäischen Strommarkt. Deutsche Energie-Agentur GmbH (dena), Final Report; August 2012. [8] CEP, editor. ‘‘Hydrogen, Joint the Movement”. Clean Energy Partnership (CEP). Brochure; 2015. [9] Feige A et al. ‘‘H2 Mobility initiative – Leading industrial companies agree on an action plan for the construction of a hydrogen refueling network in Germany”. Press statement on H2 Mobility initiative from September 2013, www.media.daimler.com, availability checked in December 2015. [10] Spath P, Mann M. ‘‘Life cycle assessment of renewable hydrogen production via wind/electrolysis”. National Renewable Energy Laboratory (NREL), NREL/ MP-560-35404; February 2004. [11] Maak M. Generation, of the energy carrier Hydrogen In context with electricity buffering generation through fuel cells. Icelandic New Energy 2008;8-5 RS1a. [12] Patterson T et al. Life cycle assessment of the electrolytic production and utilization of low carbon hydrogen vehicle fuel. Int J Hydrogen Energy 2014;39:7190–201. http://dx.doi.org/10.1016/j.ijhydene.2014.02.044. Elsevier. [13] Cetinkaya E et al. Life Cycle Assessment of various hydrogen production methods. Int J Hydrogen Energy 2012;37(3):2071–80. http://dx.doi.org/ 10.1016/j.ijhydene.2011.10.064. Elsevier. [14] Bartolozzi I et al. Comparison between hydrogen and electric vehicles by life cycle assessment: a case study in Tuscany, Italy. Appl Energy 2012. http://dx. doi.org/10.1016/j.apenergy.2012.03.021 [Elsevier]. [15] Granovskii M, Dincer I, Rosen MA. Life cycle assessment of hydrogen fuel cell and gasoline vehicles. Int J Hydrogen Energy 2006;31(2006):337–52. http:// dx.doi.org/10.1016/j.ijhydene.2005.10.004 [Elsevier]. [16] Hacatoglu K, Rosen MA, Dincer I. Comparative life cycle assessment of hydrogen and other selected fuels. Int J Hydrogen Energy 2012;37 (2012):9933–40. http://dx.doi.org/10.1016/j.ijhydene.2012.04.020 [Elsevier]. [17] Koroneos C, Dompros A, Roumbas G, Moussiopoulos N. Life cycle assessment of hydrogen fuel production processes. Int J Hydrogen Energy 2004;29 (2004):1443–50. http://dx.doi.org/10.1016/j.ijhydene.2004.01.016 [Elsevier]. [18] Dufour J et al. Life Cycle Assessment of alternatives for hydrogen production from renewable and fossil sources. Int J Hydrogen Energy 2012;37 (2):1173–83. http://dx.doi.org/10.1016/j.ijhydene.2011.09.1352011 [January]. [19] Koj JC, Schreiber A, Zapp P, Marcuello P. ‘‘Life Cycle Assessment of improved high pressure alkaline electrolysis”. In: The 7th international conference on applied energy – ICAE2015, http://dx.doi.org/10.1016/j.egypro.2015.07.576, 2012. [20] Eichlseder H, Klell M. ‘‘Wasserstoff in der Fahrzeugtechnik”. Springer Vieweg, 3rd ed., ISBN 978-3-8348-2196-6; 2012. [21] Anonymous. ‘‘70 MPa hydrogen refueling station standardization -functional description of station modules”. H2 Mobility, V1.1 07-06-2010; 2010. [22] Edwards et al. ‘‘Well-to-wheels analysis of future automotive fuels and powertrains in the european context, WTT, Appendix 2”. European Commission, Joint Research Centre, Institute for Energy and Transport; 2011. [23] Wietschel M, Bünger U, Weindorf W. ‘‘Vergleich von Strom und Wasserstoff als CO2-freie Endenergieträger”. Ludwig-Bölkow-Systemtechnik GmbH, Final Report; May 2010. [24] Bauer C, Hofer J, Althaus H-J, Duce A, Simons A. The environmental performance of current and future passenger vehicles: life cycle assessment based on a novel scenario analysis framework. Appl Energy 2015;157 (2015):871–83. http://dx.doi.org/10.1016/j.apenergy.2015.01.019, February. [25] Wulf C, Kaltschmitt M. Life cycle assessment of hydrogen supply chain with special attention on hydrogen refuelling stations. Int J Hydrogen Energy 2012;37:16711–21. http://dx.doi.org/10.1016/j.ijhydene.2012.03.028 [Elsevier]. [26] Wallasch A-K, Eckert M, Rehfeldt K. ‘‘Status des Windenergieausbaus in Deutschland”. Deutsche WindGuard; December 2012. [27] Bhandari R, Trudewind CA, Zapp P. Life cycle assessment of hydrogen production via electrolysis a review. J Clean Prod 2014;85(2014):151–63. http://dx.doi.org/10.1016/j.jclepro.2013.07.048. [28] International Organization for Standardization, editor. ‘‘ISO 14040. Environmental management – life cycle assessment – principles and framework”; 2006. [29] International Organization for Standardization, editor. ‘‘ISO 14044. Environmental management – life cycle assessment – requirements and guidelines”; 2006. [30] Association of German Engineers, editor. ‘‘VDI 4661 – Energetic characteristics, definitions, terms, methodology”. Association of German Engineers, Beuth Verlag GmbH; September 2003. [31] NEL Hydrogen, editor. ‘‘If the Future could choose”. Brochure from NEL, , availability checked in December 2015. [32] DWD, editor. ‘‘WebWerdis – weather request and distribution system”, Deutscher Wetterdienst (DWD), 2013: , availability checked in March 2016. [33] Vestas, editor. V90-1.8/2.0 MW. Vestas, technical specification. , availability checked December 2015. [34] Toyota, editor. ‘‘Toyota Mirai”. Brochure from Toyota, , availability checked in February 2016. [35] Website Hyundai ix35 Fuel Cell. , availability checked in February 2016.

64

J. Burkhardt et al. / Applied Energy 181 (2016) 54–64

[36] Swiss Centre for Life Cycle Inventories, editor. ecoinvent Database V. 2.2 (CD ROM); 2010. [37] Goedkoop MJ, Heijungs R, Huijbregts M, De Schryver A, Struijs J, van Zelm R. ‘‘ReCiPe 2008 – a life cycle impact assessment method which comprises harmonised category indicators at the midpoint and the endpoint level”. Ministerie van Volkshuisvesting, Report I: Characterization, . January 2009, availability checked in March 2016. [38] Babies H-G et al. ‘‘Deutschland – Rohstoffsituation 2011”. Deutsche Rohstoffagentur (DERA), ISSN: 2193-5319; December 2012. [39] Steiner R, Frischknecht R. ‘‘Metals processing and compressed air supply – data v2.0 (2007)”. Swiss Centre for Life Cycle Inventories, ecoinvent Report No. 23; December 2007. [40] Wagner H-J, Baack Ch, Eickelkamp T, Epe A, Kloske K, Lohmann J, et al. Life cycle assessment of the offshore wind farm alpha ventus. Energy 2011;36 (5):2459–64. http://dx.doi.org/10.1016/j.energy.2011.01.036, May. [41] Burger B, Bauer C. ‘‘ecoinvent Teil XIII – Windkraft Data v2.0 (2007)”. Swiss Centre for Life Cycle Inventories, ecoinvent Report No. 6-XIII; December 2007. [42] Hau E. ‘‘Windkraftanlagen – Grundlagen, Technik, Einsatz, Wirtschaftlichkeit”. Springer, ISBN 978-3-540-72150-5; 2008. [43] Garret P, Ronde K. ‘‘Life Cycle Assessment of Electricity Production from a V90–2.0 MW Gridstreamer Wind Plant”. Vestas, , Version 1.0. December 2011, availability checked in March 2016. [44] Guezuraga B, Zauner R, Pölz W. Life cycle assessment of two different 2 MW class wind turbines. Renewable Energy 2012;37(1):37–44. http://dx.doi.org/ 10.1016/j.renene.2011.05.008, January. [45] Körner A et al. Technology roadmap hydrogen and fuel cells. International Energy Agency (IEA); 2015. [46] Hydrogenics, editor. ‘‘Hystat Hydrogen Fueling Stations”. Brochure from Hydrogenics, , availability checked in February 2016. [47] Personal message from Daniel O’Meara, Total Deutschland GmbH; November 2013.

[48] Clean Energy Partnership, editor. ‘‘Clean Energy Partnership – Bericht 2002– 2007”. Clean Energy Partnership Berlin, ; 2007. [49] Wenske M. ‘‘Wasserstoff – Herstellung per Elektrolyse”. Fachhochschule Stralsund, 15. In: Symposium Nutzung Regenerativer Energiequellen und Wasserstofftechnik, ISBN 3-9809953-8-0; November 2008. [50] Gahleitner G. Hydrogen from renewable electricity: an international review of power-to-gas pilot plants for stationary applications. Int J Hydrogen Energy 2013;38(5):2039–61. http://dx.doi.org/10.1016/j.ijhydene.2012.12.010. [51] Daimler, editor. ‘‘Fuel consumption of vehicle type ‘B-Class’ with 100 kW engine”. , availability checked in February 2016. [52] Burkhardt J, Feck N, Große Böckmann T, Koch MK, Kruse P, Wagner H-J. CO2 emissions in nuclear power – studies of nuclear electricity generation compared to other power generation technologies. Atw – Int J Nucl Power 2007;52(7):488–96. ISSN 1431-5254. [53] Hydrogenics, editor. ‘‘HYSTAT Hydrogen Generators”. Hydrogenics, brochure; 2011 , availability checked in December 2015. [54] IINAS, editor. GEMIS Version 4.81 (Globales Emissions-Modell Integrierter Systeme). International Institute for Sustainability Analysis and Strategy (IINAS), February 2013. [55] Knörr W, Kutzner F, Lambrecht U, Schacht A. ‘‘Fortschreibung und Erweiterung ‘Daten- und Rechenmodell: Energieverbrauch und Schadstoffemissionen des motorisierten Verkehrs in Deutschland 1960–2030’ (TREMOD, Version 5)”. IFEU – Institut für Energie- und Umweltforschung Heidelberg, FKZ 3707 45101; March 2010. [56] Icha P. ‘‘Entwicklung der spezifischen Kohlendioxid-Emissionen des deutschen Strommix in den Jahren 1990 bis 2013”. Umweltbundesamt. ISSN 1862-4359; July 2014. [57] Troy H, Singh B, Majeau-Bettez G, Strømman A. Comparative environmental life cycle assessment of conventional and electric vehicles. J Ind Ecol 2012;17 (1). http://dx.doi.org/10.1111/j.1530-9290.2012.00532.x.