Impact of energy supply infrastructure in life cycle analysis of hydrogen and electric systems applied to the Portuguese transportation sector

Impact of energy supply infrastructure in life cycle analysis of hydrogen and electric systems applied to the Portuguese transportation sector

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 1 0 9 7 3 e1 0 9 8 5 Available online at www.sciencedirect.co...

868KB Sizes 2 Downloads 84 Views

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 1 0 9 7 3 e1 0 9 8 5

Available online at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/he

Impact of energy supply infrastructure in life cycle analysis of hydrogen and electric systems applied to the Portuguese transportation sector Alexandre Lucas a,*, Rui Costa Neto b, Carla Alexandra Silva b a

Massachusetts Institute of Technology Portugal Program, Avenida Prof. Cavaco Silva, Campus IST e TagusPark, Room e 16.6, 2780-990 Porto Salvo, Portugal b Department of Mechanical Engineering, IST-Technical University of Lisbon, Av. Rovisco Pais 1, Pav. Mec. I, 2 andar, 1049-001 Lisboa, Portugal

article info

abstract

Article history:

Hydrogen and electric vehicle technologies are being considered as possible solutions to

Received 8 November 2011

mitigate environmental burdens and fossil fuel dependency. Life cycle analysis (LCA) of

Received in revised form

energy use and emissions has been used with alternative vehicle technologies to assess the

25 April 2012

Well-to-Wheel (WTW) fuel cycle or the Cradle-to-Grave (CTG) cycle of a vehicle’s materials.

Accepted 26 April 2012

Fuel infrastructures, however, have thus far been neglected. This study presents an

Available online 7 June 2012

approach to evaluate energy use and CO2 emissions associated with the construction, maintenance and decommissioning of energy supply infrastructures using the Portuguese

Keywords:

transportation system as a case study. Five light-duty vehicle technologies are considered:

Hydrogen

conventional gasoline and diesel (ICE), pure electric (EV), fuel cell hybrid (FCHEV) and fuel

Electric vehicle

cell plug-in hybrid (FC-PHEV). With regard to hydrogen supply, two pathways are analysed:

Life cycle analysis

centralised steam methane reforming (SMR) and on-site electrolysis conversion. Fast,

Infrastructure

normal and home options are considered for electric chargers. We conclude that energy

Uncertainty

supply infrastructures for FC vehicles are the most intensive with 0.03e0.53 MJeq/MJ emitting 0.7e27.3 g CO2eq/MJ of final fuel. While fossil fuel infrastructures may be considered negligible (presenting values below 2.5%), alternative technologies are not negligible when their overall LCA contribution is considered. EV and FCHEV using electrolysis report the highest infrastructure impact from emissions with approximately 8.4% and 8.3%, respectively. Overall contributions including uncertainty do not go beyond 12%. Copyright ª 2012, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.

1.

Introduction

In the last two decades, world energy use has risen 45%, driven mostly by the development in countries such as China, India and the middle east [1,2]. From 1990 to 2008, the same

growth in Europe and in the USA has been 8% and 20%, respectively [1,3]. In Portugal, this increase has been approximately 57%. In 2009, primary energy imports represented 81.2%, making Portugal one of the most exterior energydependent countries in Europe. In the same year, the

* Corresponding author. Tel.: þ35 1961741327. E-mail addresses: [email protected], [email protected] (A. Lucas), [email protected] (R.C. Neto), [email protected] utl.pt (C.A. Silva). 0360-3199/$ e see front matter Copyright ª 2012, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2012.04.127

10974

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 1 0 9 7 3 e1 0 9 8 5

transportation sector in Portugal accounted for 38.2% of final energy use. From this percentage, 86% was due to road transportation [4]. Without substantial changes in policy and practice, this unsustainable tendency will evidently continue. In 2009, both the European Union (EU) and G8 leaders agreed on the long-term goal that CO2 emissions should be cut by 80% by 2050. However, achievement of this goal means that 95% decarbonisation must be achieved in the road transportation sector [5]. The European 20-20-20 agreement set ambitious goals for 2020, cutting greenhouse gases by 20% (compared with 1990 levels), reducing energy use by 20% and meeting 20% of European energy needs from renewable energy sources (RES) [3]. Portugal has set its goals even higher, with 20% efficiency, 60% electricity production from RES and 31% of final energy from RES, with 10% energy use reduction within the transportation sector [6]. Hydrogen and electricity have been regarded as possible solutions to mitigate fossil fuel scarcity and environmental impact. However, hydrogen and electricity alone cannot directly and unconditionally solve the problem. Hydrogen in particular, when compared to electricity, has good filling time and provides a good range of autonomy, despite the considerable amount of energy associated with its compression. Previous life cycle analysis (LCA) has focused on the fuel cycle divided into its upstream impact (Well-to-Tank, WTT) and downstream impact (Tank-to-Wheel, TTW) stages and has included only few materials typically related to the vehicle itself (manufacturing, maintenance, end-of-life). The final report of the Hysociety Project [7] shows an extensive amount of research work related to LCA studies of hydrogen applied to the transportation sector. No work refers to the energy supply infrastructure, including final refuelling stations. Expected energy use and emissions from passenger car transportation with different engine technologies in 2020 have been analysed by [8]. The study covered pure combustion engines (gasoline, diesel), hybrids combining combustion engines with electrical engines and fuel cells and battery-powered electric cars. The study concluded that hybrids with a diesel combustion engine combined with an electrical engine had the lowest energy use, followed by a hybrid combining a CNG combustion engine with an electrical engine and a similar hybrid using a gasoline combustion engine. Reference studies from [9] present a comprehensive report of WTW, clearly defining the term WTT as the direct energy use for producing and distributing the fuel required for propulsion of the vehicle, while TTW corresponds to the term for direct energy use or propulsion energy. Estimates are given for several hydrogen pathways, electricity, and conventional fuels, as well as local or distributed philosophies. Regarding plug-in hybrids, possible scenarios of infrastructure implementation and associated costs are presented by [10]. A similar study for hydrogen fuelling infrastructure assessment was performed by [11], elaborating future costs of hydrogen fuel to develop largescale facilities. They demonstrate that a hydrogen infrastructure that could support volume deployment of fuel cellelectric vehicles can be commercially viable. Another study by [12] compares and assesses the WTW impact of plug-in technology with hybrid and conventional technologies considering different AER (All Electric Range), with pure CD (charge depletion), blended and CS (charge sustaining) mode.

They conclude that the PHEVs offer reductions in the use of petroleum energy as compared with regular hybrid electric vehicles (HEVs). More petroleum energy savings and Greenhouse Gas (GHG) emissions reductions were realised as the AER increased except when the marginal grid mix was dominated by oil-fired and coal power generation. The driving cycle considered in the different studies has also been an issue for review, as the driving cycle is crucial for a proper comparison between vehicles and estimations. A consensus remains to be reached among academia, industry and policy makers because driving behaviours differ from region to region. Because public policies or decisions are also often at a regional level, such behaviours have been interesting to assess. Differences should, nevertheless, be well understood and described. Using a Portuguese driving cycle [13], estimate WTW and Cradle-to-Grave (CTG) for vehicle materials impacting a large spectrum of vehicle technologies in the same study, enhancing the importance of plug-in vehicle penetration to mitigate both energy and emissions problems. In another WTW study [14], an energy analysis of electric vehicles using batteries or fuel cells using the ECEEUDC driving cycle was developed. The investigators conclude that the best results are achieved by pure EVs for only very limited driving range requirements, while the fuel cell solutions achieve best performance for more extended driving ranges where the battery weight becomes too high. The energy use, emissions and cost of plug-in hybrid vehicles were evaluated comparing different driving cycles (CAFE, FTP75, NEDC and JC08), different driving distances and different user behaviours regarding battery recharging, and the main characteristics between them were described [15]. In terms of infrastructures, several passenger transport systems, such as cars, buses, trains and air travel, were analysed [16]. Energy use and emissions arising from construction, operation and maintenance of the transport system infrastructures, in addition to the WTW and CTG stages, were included in the authors’ analysis. For road transportation, these infrastructures dealt mostly with roadway construction, operation and maintenance, which can also be found in [17] LCA and parking lots. For on-road transportation, the authors found that the life cycle component contributions are roughly the same as the GHG contributions and produce 1.4e1.6 times larger life cycle factors than the operational components. Infrastructures alone ranged from approximately 7 to 18% in MJ/PKT and CO2eq/PKT in the total LCA for a conventional gasoline-powered sedan. Regarding the actual fuel infrastructure, few studies address direct and indirect energy and emissions contributions in vehicle LCA [18e20]. A study has identified the environmental and economic aspects of hydrogen production through water electrolysis, using wind power and the Korean electricity mix, applying wind power and the Korean electricity mix to FCHEV and has compared them to conventional fuels (gasoline and diesel) [18]. This study makes a scale-up inventory from a small H2 station used in California and also accounts for the wind structure. The authors conclude that although the wind station may be economically viable, it is extremely dependent on wind availability. The first study to address energy supply infrastructures applied to vehicles, comparing conventional fossil fuels and

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 1 0 9 7 3 e1 0 9 8 5

electricity supply chain facilities, was performed by [19]. The study focuses on EV technology using estimations for other LCA stages to assess the weight of infrastructure. Using the 2006 WTT average value from the EU electricity mix, the impact of the EV energy supply infrastructure did not go beyond 8%, while conventional fuel infrastructure remained below 2% impact. The study uses estimated service ratios based on the foreseen use of public charging facilities and identifies these variables as sensitive to the results. Nevertheless, this study develops inventories for petrol stations and charging points and presents a methodology to estimate energy supply infrastructure impacts, contemplating other scenarios of WTW and CTG stages. A brief literature summary was performed and is presented in Table 1, identifying values for the most highly influential parameters and differences between the studies. Due to the use of different vehicle references within a given technology, one may argue that it would be reasonable to obtain different results of TTW, WTT or CTG. However, other factors influence the results as well. One is the lifetime parameter of kilometres driven, a parameter which can vary from 150,000 km [13] to 300,000 km [8]. This parameter directly affects the CTG stage because the absolute values of the LCA are divided by the lifetime driving distance of the vehicle to obtain a value per kilometre. Another important parameter is the Driving Cycle considered in the study, fundamental to determine the TTW impact. Europe has been adopting the New European Driving Cycle (NEDC), while many car manufacturers still have their own driving cycles. In the USA, the most common Driving Cycle used is the Federal Test Procedure (FTP). The main differences between cycles are acceleration, velocity and mobility dynamics. Other local or specific cycles found in the literature can also be used to assess real scenarios [13]. Within the scope and aim of LCA, although all studies tend to address energy and emission impacts, not all studies include the CTG stage [9]. Given the identified lack of assessments of the whole energy supply infrastructure impacts in the literature, the aim of this study is to perform a life cycle analysis to estimate energy use and greenhouse gas emissions associated with the construction, maintenance and decommissioning of fuel support facilities for ICE vehicles, EV, FCHEV, FC-PHEV and their impact in overall vehicle LCA. This LCA is particularly important when considering alternative technologies with different pathways to provide policy and decision makers the most complete set of information about each choice. Despite being applied to the specific case of Portugal, the methodology could be extended to other countries, similar to the case of fuel WTW studies.

2.

Methodology

Life cycle analysis, particularly in the case of infrastructure, raises two major methodological concerns. The first concern is about system boundaries, where one should, on the one hand, be cautious not to double-count efforts or activities already assessed in other stages, and, on the other hand, not disregard a relevant activity. Another concern deals with how to allocate burdens to different products with different

10975

energy contents that may share the same facilities. This study addresses energy infrastructure according to each technology and pathway. For that reason, roads, traffic signs, parking lots, repair shops, and car factories were not considered in this study as they are not a distinctive characteristic between technologies. The principles of ISO 14040 [22] were followed. The methods used were Global Warming Potential for 100 years (GWP100) and Cumulative Energy Demand (CED). GWP is based on the heat-absorbing ability of each gas relative to that of CO2, as well as the decay rate of each gas, i.e., the amount removed from the atmosphere over a given number of years [23]. Because infrastructures are being considered in this study, and given the contribution of different gases and their ability to linger in the atmosphere for different time ranges, a 100 year time horizon was considered. CED methodology was used to investigate the energy use throughout the life cycle of a good, in this case infrastructure. The CED methodology is especially suited to determine and compare the energy intensity of processes, including the direct as well as the indirect uses of energy. Indirect energy inputs consider all inputs that are used for purposes other than manufacturing products, such as infrastructure and equipment. The Simapro tool [24] allows the comparison between contributions of different sources. However, because the goal of the study is to obtain comparable values, the output unit of MJeq was used, assuming Simapro primary energy conversion factors. Supply chain scope and boundaries were outlined for the technologies considered until the vehicle was reached. Three major groups could be identified: primary fuel handling infrastructure, end fuel transport system and distribution facilities.

2.1.

Scope

Fig. 1 systematises the boundaries for the base case supply chain. For the extraction of raw material, an oil well and an offshore platform were used. For the exploration well, the oil hydrocarbon probability of concentration in depth formation varies according to a normal distribution, so the mean value was considered to be approximately 2,600 m [25], to represent the depth and the corresponding drilling efforts to explore a North Sea well. Despite the existence of two refineries in Portugal, only Sines refinery was used, as this refinery represents almost the entire share of refining products [26]. Refinery efficiency was considered to be 85% [9]. As Simapro does not account for construction of office buildings in the refinery, 50% of the area was taken as a measure for inventory of buildings. Roads inside the refinery were also accounted for. Storage facilities were omitted. For fuel transportation, the main existing infrastructure is a multiuse pipeline, so allocations were made to account only for the gasoline and diesel share. There were 2.52 thousand refuelling stations in 2010, and these stations were assumed to satisfy the demand of all the fleet perfectly. The fleet was composed of 5.63 million vehicles, of which approximately 3.00 million were gasolinepowered and 2.63 million were diesel-powered [27]. Maintenance was considered only for the refinery. The electric infrastructure and H2 production supply chain using the electrolysis pathway are represented in Fig. 2.

10976

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 1 0 9 7 3 e1 0 9 8 5

Table 1 e Brief LCA Literature Overview. Study Scope and Aim Software/ database Technologies Pathways

Power/Weight t ratio (W/kg)

Driving Cycle Uncertainty TTW

- MJeq/km - gCO2eq/km

CTG

[9] CO2eq (100y); Energy Use; WTW e 2010þ E3 database by LBST ICE; FCHEV Petrol and diesel EU Mix SMR:yes Electrolysis:yes Others PISI- 65 DISI- 54 DICI-54 FC:54 NEDC; Yes PISI:1.90 DISI:1.88 DICI:1.61 FCEV:0.94 PISI:140.3 DISI: 138.8 DICI: 119.7 FCEV: 0 e

- MJeq/km

- gCO2eq/km

e

WTT

Crude Oil to Petrol: 0.32 (PISI); 0.32 (DISI) Crude Oil to Diesel: 0.31 SMR with 4000 km pipeline: 0.677 EU-mix electricity, on-site electrolysis: 3.62 (MJx/MJf) Petrol: 26.98 (PISI); 26.68 (DISI) Diesel: 25.61 SMR with 4000 km pipeline: 92.87; EU-mix electricity, on-site electrolysis: 209.1(gCO2/MJf) 200,000

- MJeq/km

- gCO2eq/km

Lifetime (km)

[8]

[13]

CO2eq (100y); Energy Use; WTW þ CTG for 2020

CO2; Energy Use; WTW þ CTG

ADVISOR; TEDB; EPA; NHTSA; FHWA 2005; ETH Zurich Matlab Simulink ICE; FCHEV; EV Petrol and diesel US Mix SMR decentralized for H2 conversion; US average mix for electricity

CTG e GREET Copert Advisor ICE; FCHEV; FCPHEV; EV EU fossil fuel Mix EU Electric Mix SMR centralized (a) Electrolysis decentralized (b); Others Petrol:75 diesel:55 EV:54 FCHEV:54 FCPHEV:57 Lisbon to Cascais; No

Base SI:75 Adva. SI ICE:75 Adva. CI ICE:75 EV:75 FC:75 FTP; Yes Base SI:1.75 Adva. SI ICE:1.54 Adva. CI ICE:1.35 EV:0.51 FCHEV :0.81 Base SI:126.0 Adva. SI ICE:111.0 Adva. CI ICE:103.0 FCHEV and EV:0.0 Base SI:0.22 Adva. SI ICE:0.22 Adva. CI ICE:0.23 EV:0.28 FCHEV:0.27 Base SI: 15.8 Adva. SI ICE (Petrol):15.4 Adva. CI ICE (diesel):16.1 EV:19.1 FCHEV :18 Petrol: 0.37 (Base SI); 0.32 (SI ICE); diesel: 0.19 Electricity Mix:1.10 On site SMR:0.62

Petrol:1.96 diesel:1.67 FCHEV:1.08 FCPHEV:0.55 EV: 0.57 Petrol:143.0 diesel:124.4 FCHEV, FCPHEV and EV:0.0 Petrol:0.48 diesel:0.50 FCHEV:0.73 FCPHEV:0.77 EV: 0.77 Petrol:30.7 diesel:32 FCHEV:48.4 FCPHEV:49.5 EV: 47.8 Petrol and diesel:0.27 FCHEV(a):0.62 FCHEV(b):3.89 FCPHEV(a):0.31 FCPHEV(b):1.97 EV: 1.06

Petrol: 31.5 (Base SI); 27.7 (SI ICE); diesel:16.5 Electricity Mix:101.1 On site SMR:107.0

Petrol:24.5 diesel:23.7 FCHEV(a):95.4 FCHEV(b):223.3 FCPHEV(a):56.7 FCPHEV(b):96.4 EV:72.9

300,000

150,000

[21] CO2eq (100y); Energy Use; Criteria pollutants; WTW þ CTG þ end of life Ecoscore Ecoinvent ICE; FCHEV; EV SMR:yes Electric Mix BE and EU Mix

Base SI:75 Adva. SI ICE:75 Adva. CI ICE 75 EV:75 FCHEV:59 NEDC; Yes

Petrol: 1.98 diesel: 1.76 FCHEV: 1.25 EV: 0.54 Petrol: 144.0 diesel: 128.0 FCHEV and EV:0.0 CTG-End of life treatment Petrol:0.28 diesel:0.28 FCHEV: No ref. EV:0.56 CTG-End of life treatment Petrol:11.2 diesel:12.0 FCHEV: No ref. EV:16.1 Petrol: 0.76 diesel: 0.53 FCHEV: No ref. EV: 1.18

Petrol: 37.0 diesel: 26.0 FCHEV:110.1 EV: 56.5

230,500

Note: (a) centralized Steam Methane Reforming (SMR) pathway; (b) On-site Electrolysis pathway. Port injection spark ignition (PISI), direct injection spark ignition (DISI), and direct injection compression ignition (DICI). Net energy expended (MJx) (excluding the energy transferred to the final fuel. Energy of final fuel (MJf).

Each of the power plants composing the electric mix of 2010 in Portugal [28] was assessed. Infrastructure for extraction of raw materials such as coal, oil and gas was ignored. Due to the high use of natural gas as a feedstock for power

plants, only the pipeline from Algeria crossing Spain and Portugal was considered. For allocation purposes, only the portion sold to Portugal and used by the power plants was considered [29]. For the power plants, with the identification

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 1 0 9 7 3 e1 0 9 8 5

10977

Fig. 1 e System boundaries for conventional fossil fuel supply chain.

of the feedstock types, the installed capacity of the power plants and their sold energy, their capacity factor and mix were determined. For the electric grid stage, 8% losses were assumed. All 50.6 TWh of 2010 energy [28], without imports and pumping, transited through the electric grid, so all high, medium and low voltage lines were considered. For FC-PHEV and EVs, a charging infrastructure had to be considered, so they were divided into three categories according to charging options in Portugal: normal charging, fast charging and home charging, according to Table 4 of supporting information (SI). While addressing EV technology [19] only considered public charging service rates for diurnal and nocturnal areas. Furthermore it assumes that the total charging needs would be fulfilled the same way by normal and fast chargers. However, private infrastructure must also to be considered, and a better understanding of the use of charging options must also be investigated further. According to comprehensive studies for France performed by [30,31], public and private service rates are suggested to be 1.1 (chargers or plugs/car) at a stable penetration scenario. This ratio can be even higher if installation of chargers at working places is also considered. Regarding maintenance, electricity charging points have low routine maintenance, requiring only periodic inspections, cleaning, repairs and communication systems and lighting testing, so only power plant maintenance activities were considered [32]. Also represented in Fig. 2 is the centralised SMR pathway, which includes the natural gas (NG) pipeline

and a hydrogen plant based on [33]. A compressor exiting the plant was also considered to normalise the pressure of hydrogen to be transported. For the H2 pipeline estimations, the Simapro database was considered, assuming similarity with the transport of CH4 in terms of technical specifications and construction characteristics even though compression is different. For this reason, the compressors used in the study are, in fact, for H2 compression and not CH4. Compression stations and required buildings were also included in the pipeline analysis. In the Inventory stage, Simapro 7.1 was considered as the reference source and used in triangulation with the GEMIS [34] database and existing literature. Scaling techniques were used to adjust raw data from the databases to the Portuguese installation characteristics (for example, installed power, capacity factor or mix). An inventory for H2 refuelling stations had to be reformulated based on [35,36], to contemplate an average-size station comparable to a conventional Portuguese fuel station and also to make the distinction between an incoming pipeline H2 station and an onsite electrolysis production station. Charging points were also missing from the literature or databases with few exceptions [32,37] where, although small inventories were conducted, these inventories were somehow scarce in information and quantities regarding the three charging methods: normal, fast and home charging developed in Portugal. Based on earlier work [19] and with further contact

Fig. 2 e System boundaries for electricity and Hydrogen (Electrolysis and SMR pathways) supply chains.

10978

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 1 0 9 7 3 e1 0 9 8 5

with the leading charging stations manufacturer in Portugal (EFACEC), a more detailed inventory was developed. In addition, petrol stations also had to be inventoried because values from Simapro, although existing, did not correspond to the size of an average fuel station in Portugal. Furthermore, because fuel stations offer services other than just fuel distribution, only the facilities attributable to that activity had to be considered; thus, an adjusted inventory was required. The same was conducted for H2 stations with the necessary scale and technical adjustments based on [33], whether SMR or electrolysis pathways were being considered. The H2 plant from the SMR inventory was followed according to [33]. Regarding the fleet, the 2010 Portuguese numbers were considered as a reference for light-duty diesel and gasoline vehicles only. Energy allocations were made according to the 2005 refining mix [26] for conventional fuels and the 2010 electric mix [28] for the EV, FCHEV(b) and FC-PHEV(b). Maintenance was considered only for power plants, the refinery and the H2 station. Gross values were calculated, taking into account [38], and then adjusted according to the installed capacity and capacitor factor for each plant. Energy use in power plant maintenance was considered to be mainly from heavy duty diesel trucks or machinery operation. Estimations were made using two functional units. First, the energy and carbon intensity of construction, maintenance and decommissioning of all facilities was calculated and divided by the total output energy during its lifetime, so units of MJeq/MJ were obtained. Second, those values were multiplied by the TTW stage of each technology, adjusted according to each facility’s efficiency along the supply chain. This procedure allows other investigators to use these energybased estimations with different TTW values or technologies. A choice of similar characteristics between vehicle technologies should be taken into account when evaluating options. The power/weight ratio is a determinant characteristic to be described when comparing technologies. Vehicles used should belong to the same performance class, otherwise an error of judgment could occur due to performance issues. For the purpose of this study, values from [21] were used for the TTW stage and WTT estimations were calculated based on [9] adapted to the 2010 Portuguese emissions and efficiency. For

non-existent values of TTW for FC-PHEV technologies, estimations were made based on the [13] study, using as a reference gasoline vehicles from both studies and applying the same difference ratio between technologies. All used values can be observed in Table 6 of (SI). For FC-PHEV technology, a 0.21 MJ/km electricity use and 0.34 MJ/km of hydrogen were considered. For the CTG values [21], adjustments were also made keeping in mind the difference of kilometres driven between the studies. A lifetime range of 150,000 km, corresponding to an average of 12,800 km per year in Portugal, was used. Values from EURO 5 legislation were considered, and the VW Golf was used as a reference vehicle by [21]. Another factor that influenced the choice of the study was the use of NEDC. On the one hand, use of NEDC allows a generalisation for future studies, and on the other hand, NEDC is closer to a Portuguese driving cycle than FTP. Energy use and emissions LCA estimations are given by Eqs. (1) and (2) in terms of their final units: The infrastructure value is directly proportional to the TTW stage given by MJ/km:   MJWTT MJInfra: MJ MJ MJCTG þ  TTW þ TTW þ ; MJ MJ km km kmtotal

LCAMJeq ¼

 gCO2eq:WTT

LCACO2eq ¼

þ

MJ gCO2eq:CTG

þ

gCO2eq:Infra MJ

 

(1)

MJTTW gCO2eq TTW þ km km (2)

kmtotal

Relevant losses for all facilities, presented in Table 5 of the SI were added to the corresponding load (TTW) to be demanded from each facility along the chain.

2.2.

Inventory and data assembly

2.2.1.

Construction and decommissioning

Details of inventory and complementary considerations regarding the overall lack of data are described in Table 2 of the SI section. Lifetimes considered can be found in Table 1 and Table 2 of the SI based on Simapro and the literature [39e41]. Values simulated by Simapro for the transport and distribution grid were validated by the literature [42,43]. The power plant characterisation mix can be found in Table 1 of

Table 2 e Inventory for construction and installation of Charging Points. Material Copper (kg) Aluminium (kg) Iron (kg) Concrete (kg) Stainless Steel (kg) PVC - Polyvinyl chloride (kg) ABS, Glass Fibre, PVC (kg) Characteristics Lifetime (years) Service rate (socket) Impact Values per charger

Home Charger

kgCO2eq 3.2 Eþ01

Normal Charger*

Quick Charger

e e e e 3.5 7.5 e

2 2 e 960 2.5 e 81

206 34 228 2400 90 12 380

15 0.941

6 0.164

12 0.003

MJeq 6.17 Eþ02

kgCO2eq 2.9 Eþ02

MJeq 6.20 Eþ03

kgCO2eq 2.9 Eþ03

Source: Own inventory based on EFACEC product line. *Average values per satellite, considering 1 central column per 10 satellites.

MJeq 5.96 Eþ04

10979

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 1 0 9 7 3 e1 0 9 8 5

the SI, excluding values of import/export or pumping. Upstream from the power plants, only the natural gas pipeline was considered to be representative of the main source of natural gas importation in Portugal. All of the 1638 km [29] length of the pipeline was taken into consideration. However, because only 8.2% was found to be used for electricity conversion at Portuguese plants, only this portion was added to natural gas power plants. Regarding hydro plants, taking into account that Simapro data are based on five Swiss and Austrian plants and that no installed power data are provided, according to the UDI World Electric Power Plants Database, a mean value of 544.77 MW was used as a reference for scaling up the Portuguese installed capacity. For the charging point inventory, Table 2 presents a brief list of the main materials and quantities considering installation requirements. Normal charger average values consider a central column and 10 satellites with 2 sockets each. Using a total service ratio of 1.1 [30,31], partial ratios by charging option were estimated. Although [44] present some estimation to what these may be, the study never refers an aggregate service ratio, thus estimations were followed according to [30,31] expectations. They define that from the 1.1 service ratio, 0.105 are secondary public chargers, 1 primary chargers and 0.003 fast chargers. According to those studies, 90% of the charging is expected to be 3 kVA primary charging. Because not all users will be able to charge in private garages, we consider that 6% of level 1 chargers will be on the streets, which added to the 0.105 given by the study means a total of 0.164 of normal chargers and 0.941 home chargers. We also assume that on average 20.5% of all charging (15% normal chargers and 5.5% fast chargers) will be performed using public chargers. This means that from the 90% of primary charging, 79.5% will be on home chargers, and 10.5% on normal chargers. The other 4.5% of normal charging and 5.5% using fast chargers will be in other public spaces. Table 2 of SI presents a summary of these characteristics. According to the same studies 3 kVA chargers are physically identical to level 2 normal chargers, only with minor differences in connections and cabling. Every two normal charger connections correspond to one satellite. Lifetimes in years were estimated according to the manufacturer’s expectations, taking into account technological transition, usage degradation and the payback period. For FC-PHEV technology, the EV MJ-based value of emissions and energy use of all charging points was considered before calculating the final km-based functional unit. Table 3 presents the inventory for a conventional refuelling station based on [19], making the necessary adjustments to contemplate only the services attributable to fuel distribution. The H2 station inventory shown in Table 4 was based on the Grjo´tha´ls station, situated on Reiquejavique Island. This station was built by HydroStatoil, and its life cycle was studied by [36]. Values were inventoried by quantity to be simulated in Simapro and separated by modules, to distinguish the two pathways analysed with or without electrolyser. To obtain a relationship between distributed energy and the LCA estimations for the H2 station, a service ratio (stations/car) had to be estimated according to each pathway. Regarding SMR, and from the demand side, considering that the FCHEV has been increasing its autonomy range and even

Table 3 e Inventory for construction and installation of a conventional refuelling station. Total Gravel ETH U Area covered with gravel equivalent Total Concrete, sole plate and foundation Foundation for 4 dispensers m3 Foundations for building support m3 Foundations for metallic structure m3 Total Building, hall/CH/I U Payment building with storage room m2 Total Steel I Support Gantries Covering metallic structure 500m2 Interior metallic supports Total Bitumen, at refinery/RER U Interior accesses Mechanic Componentsa Pipelines, tubes and accessories Storage Tanks and reservoirs Other components Storage Diesel Storage Tanks Gasoline Storage Tanks Total

MJeq 1.05 Eþ05 8.90 Eþ05 1.32 Eþ04

31.6 ED04 31.6 Eþ04 1.31 ED03 4.80 Eþ02 3.50 Eþ02 4.80 Eþ02 2.00 ED02 2.00 Eþ02 8.00 ED04 3.00 Eþ04 2.50 Eþ04 2.50 Eþ04 2.19 ED05 2.19 Eþ05 kgCO2eq 3.0 Eþ03 4.5 Eþ04 6.0 Eþ02

4.81 Eþ06 6.58 Eþ06 2.70 Eþ07

4.4 Eþ04 1.7 Eþ04 5.5 Eþ05

a Data from contractor company (Petroassist). Values adjusted for 4 dispensers. Electrical components were not considered.

though it is below conventional fuels range, it can be considered somewhat similar in refuelling time and frequency of refills. As for the supply side, because production is centralised and the distribution is made by pipeline, no limitations exist, so the same service rate as a conventional station was considered. As for the on-site electrolysis scenario, there is a supply limitation, and that limitation is the hydrogen production capacity of the electrolyser. Because the installed capacity of the station is 47,250 kg H2/year [36] and its capacity factor is 92.88%, due to maintenance activities and general idle hours, taking into account the TTW [21] energy use and the vehicle lifetime driving range of 12,800 km per year, a value of approximately 329 cars per station can be reached.

2.2.2.

Maintenance

The maintenance stage was considered for the refinery, for all power plants and for the H2 refuelling station. Concerning power plants, maintenance data from [38] were used. Because operation values are part of WTW assessments, only data for expendables and maintenance were considered. Data reported were given in g CO2/kWh units according to the specific lifetime, capacitor factor and installed power. So that it could be adapted to Portuguese reality, the gross value of CO2 was calculated applying the values presented in the 2010 electric mix characterisation shown in Table 1 of the SI section, and new values of g CO2/kWh were found and are presented in Table 5 of the SI. Because maintenance activities involve mostly transportation and electricity use, and given that no specific data were found, the CO2 released was considered to be due to burned diesel fuel in the operation of trucks and machinery. Regarding refinery data, no literature data were found regarding quantities of materials used in passive or active maintenance. However, the main activities include tube restoration, painting and galvanisation, scaffolding,

10980

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 1 0 9 7 3 e1 0 9 8 5

Table 4 e Inventory for construction and installation of a H2 refuelling station. Materials/Energy

Quantity (kg)

Buildings or foundations Steel for concrete, installation Smooth coated Glass, installation Plaster fibre plate, installation Silica sand, installation Concrete, installation m3 Resistant concrete, installation m3 Gravel, not specified, installation Lubricant oil, installation Electricity (kWh) Diesel (MJ) Transport lorry 32t (tkm) Storage module Stainless Steel 18/8, installation Electricity Diesel Transport lorry 32t (tkm) Other components Steel for concrete, installation Nitrogen, liquid, installation Stainless Steel 18/8, installation Polypropylene, granulated, installation Transport, lorry 32t (tkm) Compressor Steel for concrete, installation Total Without Electrolyser With Electrolyser

Compressor (cont.)

kg

Stainless steel 18/8, installation Casting iron, installation Etilenoglicol, installation Lubricant oil, installation Aluminium, production mix, installation Insulation tube, installation Cooper, regional storage Electricity (kWh) Heat, NG, ind. furnace > 100 kW (MJ) Transport lorry 32t (tkm) Additional Electrolysis module Electrolyser Chromium steel 18/8 at plant Nickel 99,5% at plant Synthetic rubber, at plant Reinforcing steel at plant Cooper at regional storage Tube insulation, at plant Aluminium, production mix, at plant Acrylonitrile-butadiene-styrene copolymer, ABS, at plant Polyethylene, LDPE, granulate, at plant Glass fibre, at plant Cast iron, at plant Nylon 66, glass-filled, at plant Transport lorry 32t (tkm) kgCO2eq 1.8 Eþ04 1.9 Eþ04

278.68 100.5 3.09 1781.77 0.31 4.02 55735.22 0.62 15.49 1325.36 6846.22 2602.44 29.71 26.51 260.24 50.91 4.43 12.51 0.31 3028.13 76.8

58.81 18.56 0.22 0.56 1.86 0.47 1.39 30.94 111.47 15.89

262.88 30.94 1.55 82.07 23.7 10.53 6.8 2.48 6.19 6.19 2.11 0.77 43.49 MJeq 3.03 Eþ05 3.38 Eþ05

Source: Based on [36].

insulation works, welding of materials and equipment lubrication. The main materials used and the expendables are rock and glass wool, aluminium, copper, stainless steel, paint, steel SAE 1020, zinc bath in the galvanisation process, diesel and electricity. With the cost of construction of a similar refinery, a direct relationship was established between the cost and CO2eq emissions and energy use assessed by Simapro, and the same relationship was attributed to the annual cost of

maintenance. Knowing that 40% of the cost is spent in material acquisition and the rest in hand labour [41], values for maintenance were obtained. As for the H2 refuelling station, compressor equipment has to be subjected to a significant overhaul every few years (typically assumed to be at 5-year intervals) [33]. Because the lifetime of the station is 15 years [36], calculations were made for two total replacements.

Table 5 e Complete LCA values of fossil fuels and EV technologies.

Gasoline Diesel FCHEV (a) FCHEV (b) FCPHEV (a) FCPHEV (b) EV

MJeq/km gCO2eq/km MJeq/km gCO2eq/km MJeq/km gCO2eq/km MJeq/km gCO2eq/km MJeq/km gCO2eq/km MJeq/km gCO2eq/km MJeq/km gCO2eq/km

WTT

TTW

Vehicle Material

Infrastructure min-max (best est.)

0.34 28.1 0.33 28.0 0.90 123.5 4.09 141.2 0.57 48.4 1.45 53.3 0.83 36.60

1.98 144.0 1.76 128.0 1.25 0.0 1.25 0.0 0.56 0.0 0.56 0.0 0.54 0.00

0.43 17.7 0.43 18.5 0.65 28.0 0.65 28.0 0.69 28.6 0.69 28.6 0.86 24.80

0.03e0.06 (0.05) 0.5e1.2 (1.0) 0.02e0.06 (0.05) 0.6e1.3 (1.2) 0.06e0.17 (0.07) 1.5e3.5 (2.0) 0.20e0.41 (0.23) 13.0e22.0 (15.4) 0.05e0.09 (0.06) 2.4e3.4 (2.7) 0.10e0.17(0.10) 5.6e8.7 (6.5) 0.09e0.14 (0.10) 4.9e7.0 (5.6)

a Difference in percentage when compared to gasoline base case.

Infrastructure Impact min-max (best est.) 0.9%e2.3% 0.3%e0.6% 0.9%e2.2% 0.3%e0.7% 2.0%e6.2% 1.1%e2.5% 3.2%e6.5% 7.0%e11.9% 2.9%e5.0% 3.1%e4.5% 3.4%e6.0% 6.4%e10.0% 3.8%e6.2% 7.3%e10.3%

(1.8%) (0.5%) (1.8%) (0.7%) (2.4%) (1.4%) (3.7%) (8.3%) (3.0%) (3.5%) (3.7%) (7.3%) (4.3%) (8.4%)

% of base casea 100% 100% 89% 112% 127% 194% 452% 1489% 111% 262% 203% 627% 197% 544%

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 1 0 9 7 3 e1 0 9 8 5

2.3.

Uncertainty

A total of 166 variables were submitted to uncertainty analysis, according to the fit test distribution, observable in Table 3 of the SI. Uncertainty Simapro data are based on the Ecoinvent’s pedigree matrix, which refers to basic uncertainty and under-specification uncertainty characterised by 6 indicators, each with 5 levels. These values were subjected to 1000 iterations, and variations were matched with a curve fit tool into Crystal Ball. Regarding inputs for own inventories such as chargers and values used for H2 stations, because a case study is being considered, only basic uncertainty, which relates to measurements of quantities, was taken into consideration and assumed to vary according to normal distributions (Table 3 of the SI). Attempting to generalise the study will certainly increase uncertainty, especially for the inventories developed. Issues such as the sample size, reliability, geographic correlation, underspecification and impact uncertainty will become relevant issues to deal with [45]. A 10,000-trial simulation was then run in Crystal Ball with all 166 variables. Maximum and minimum values were obtained for each technology assessed. WTW and CTG uncertainty stages were not considered. Calculations were based on their best estimated values.

3.

10981

decommissioning and maintenance. Energy use and emissions uncertainty are represented in Fig. 3 for each of the technologies studied. In total, five parameters are represented: median, 1st percentile, minimum, maximum and 3rd percentile. The median value was preferred to the mean, so that it could represent the central trend. The uncertainty value range of infrastructures is higher in FCHEV(b) when compared to other technologies in several aspects. Higher amplitude of deviation from the median value, both in minimum and maximum extremes, is reported. Furthermore, the interquartile range is higher in both FCHEV than other technologies, reducing the probability of the best estimate value falling far from the median. Overall, lower amplitude of uncertainty in both emissions and energy use of ICE vehicles can be explained by a higher accuracy in measurements due to standardised practices and processes of construction, appropriate for a mature technology. Regarding alternative vehicle support infrastructures, construction techniques, scale, and geographic positioning, raw materials processing and technology used can differ widely, meaning higher uncertainty. Regarding median positioning between extremes, EV presents a midpoint location, expected when most probability functions, or those that represent the most weight, are normal distributions. Other technologies present median values closer to the minimum extreme, which is influenced by the use of lognormal distributions. To understand the nature of this positioning, although the uncertainty behaviour in terms of ranges and probabilities was identified,

Results and discussion

All infrastructure values simulated in Simapro were corroborated by GEMIS or the literature. Regarding distribution facilities, inventories were conducted for all technologies according to the lifetime driving range characteristics and use of energy. When comparing distribution infrastructures of conventional with hydrogen, electric or plug-in vehicles, caution should be exercised regarding refuelling or charging times. In typical petrol or H2 stations, the time required for filling a tank may vary between 4 and 7 min, whereas an electric charge varies, according to Table 4 of the SI, from 20 to 480 min. A question may be raised regarding levelling the time of refuelling and only then comparing technologies and facilities. However, because refuelling times are different between technologies, one may assume that the service rate will change accordingly. That is, if one facility is occupied for more than a certain period, another installation should be provided to the market. However, if charging/refuelling times are sufficiently low to accommodate more vehicles per time or kilometres driven, fewer facilities are required, compensating in terms of materials and corresponding construction, maintenance and decommissioning efforts. When comparing the amount of energy supplied by distribution facilities, a refuelling station provides more MJ than a charging facility. However, electric vehicles are more efficient than an ICE, resulting in a benefit to the system. This adjustment is assumed to be linear in this study, even though we recognise that further work should be performed. Total results for all facilities are presented by technology in Table 5 of the SI, in final energy use units, and for each infrastructure group analysed, including construction,

Fig. 3 e Energy Use (a) and emissions (b) uncertainty analysis.

10982

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 1 0 9 7 3 e1 0 9 8 5

it is still crucial to know what variables contribute the most to such uncertainty. Contribution of the main variables to emission and energy use variance is shown in Fig. 4. Only variables with contribution to variance higher than 4% were distinguished. Variables lower than 4% were added and placed under the category “others”. In Fig. 4, variables that contribute the most for EV, emissions and energy use uncertainty are the normal charger lifetime with 57.9% influence and overall power plant construction, especially hydro plants. FC-PHEV(a) emissions uncertainty also depends on 42.7% of the normal charger lifetime and the H2 pipeline’s energy use with 52.4% influence, as is FCHEV(a) with 75.7% and 37.3% for emissions. Both FC technologies using the electrolysis pathway depend on approximately 50% of power plants’ construction activities, mostly hydro plants (30%) and H2 stations. Regarding both ICE vehicle technologies, two variables report higher impact both

in emissions and energy use. The first variable is the lifetime of refuelling stations with an influence of approximately 60% in the case of emissions and 83% in energy use for both technologies. The second variable is civil works activities during refinery construction, with an influence, for both emissions and energy use, of approximately 25% and 7%, respectively. Fig. 5 presents the absolute values of infrastructure per each technology, showing values per unit of final energy. Absolute values report higher energy use and CO2 emissions from construction, maintenance and decommissioning for alternative vehicle technologies. Higher uncertainty in alternative fuels can be explained by a longer supply chain and a variety of efficiencies, as well as the reason for higher energy and carbon intensity for these technologies, as hybrid solutions need complementary infrastructures for both fuels used. The energy per kilometre travelled, relative to the infrastructure, is necessarily higher, as this energy will depend on its TTW stage, as illustrated in Figs. 1 and 2 of the SI. Even though some technology infrastructure best estimate values are inferior to others, they can have a higher impact on their LCA when presented in their final functional units. Table 5 shows the results for all stages and the corresponding impact of infrastructure using minimum, maximum and best estimated values. While the impact of conventional

Fig. 4 e Contribution to energy use (a) and emissions (b) variance.

Fig. 5 e Energy use (a) and emissions (b) of energy supply infrastructures per technology.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 1 0 9 7 3 e1 0 9 8 5

fuel can be neglected under the assessed conditions because its contribution is below 2.5%, this is not the case in alternative vehicle technologies. The difference between the contribution of fossil fuel infrastructures and alternative technologies is illustrated mostly by end distribution facilities, H2 pipelines and power plants (see Table 5 of the SI). Charging point facilities are about ten times higher in energy use and five times higher in emissions than a conventional refuelling station. In the H2 supply chain, H2 refuelling stations, when using the SMR pathway, are very similar to conventional refuelling stations. However, when considering the on-site electrolysis system, the H2 station emits about 6 times more CO2eq per unit of distributed MJ than conventional stations. To compare the contribution of fuel infrastructures in WTW and CTG LCA, all stages are presented in Fig. 6. Uncertainty corresponds only to infrastructure. The impacts of infrastructures relate only to each technology LCA. Values are presented in this way to highlight the weight of this stage on the total LCA of each technology. Percentage weights among technologies should not be compared. The best estimated impact of infrastructure on total EV LCA was 4.3% for energy use and 8.4% for emissions. The impact of diesel infrastructure in its total LCA is 1.8% of

energy use and 0.7% of emissions. Gasoline energy supply facilities represent 1.8% of energy use and 0.5% of emissions in total LCA. H2 energy supply infrastructure represents 2.4% of energy use and 1.4% of emissions using the SMR pathway and 3.7% and 8.3% using electrolysis. FC-PHEV infrastructures account for 3.0% of energy use and 3.5% of emissions using the SMR pathway and 3.7% and 7.3% using electrolysis. Pipelines and end distribution facilities are extremely significant for their overall contribution; thus, definition of their service ratios is extremely important. In this study, we have broadened the technologies assessed in terms of their energy supply infrastructure and used service ratio estimates of 0.941 for home chargers, 0.164 for normal chargers and 0.003 for fast chargers. A higher level of attention was given to the input materials for all chargers. Being two times higher than fast chargers, normal chargers report the highest values of energy use and emissions per kilometre among all chargers, being very sensitive to their lifetime usage. If the effective use increases (i.e., if the energy flow increases in the normal charger use), its impact would decrease. In addition, if the overall service ratio of 1.1 estimated by [30] should decrease, then the absolute value per kilometre of normal chargers would be reduced the most. If a decrease should occur in the CTG and WTT stages, both TTW and infrastructures will have a higher impact on the total LCA. If the TTW stage decreases, the contribution of infrastructures per kilometre will also decrease due to the direct proportional relationship between the absolute value of infrastructures and the TTW energy use value. The more efficient vehicles become, the less impact TTW and infrastructure per kilometre will have (see Eqs. (1) and (2)).

4.

Fig. 6 e LCA energy use (a) and emissions (b) per stage and technology.

10983

Conclusions

The study goal was to provide energy use and CO2eq estimations related to energy supply infrastructure for several vehicles and to estimate the impact on their overall LCA. A methodology [19] was used with Portugal as a case study and applied to conventional diesel and petrol vehicles as well as alternative technologies such as EV, FCHEV and FC-PHEV. With the case study being analysed, alternative vehicle technologies energy supply infrastructures are more carbon- and energy-intensive per unit of supplied fuel than conventional energy supply infrastructures. For the current scenario of 2234 vehicles per conventional and H2 (SMR) refuelling station and the foreseen service rate scenarios of 333.3 vehicles per fast charger, 6.1 vehicles per normal charger (socket), 1.1 per home charger and 329 vehicles per H2 (electrolysis) station, the LCA of conventional fuel infrastructure potentially represents an energy use of 0.01e0.04 MJeq/MJfuel for both gasoline and diesel fuels. Emissions represent 0.3e0.9 g CO2eq/MJfuel for gasoline and 0.3e1.0 g CO2eq/MJfuel for diesel. For electricity, these values are 0.15e0.30 MJeq/MJfuel with emissions of 8.5e14.7 g CO2eq/MJfuel. For hybrid vehicles, FCHEV(a) shows 0.03e0.12 MJeq/MJfuel and 0.7e2.9 g CO2eq/MJfuel, the FCHEV(b) shows 0.08e0.23 MJeq/MJfuel and 5.5e12.0 g CO2eq/MJfuel. For plug-in technologies, FC-PHEV(a) shows 0.18e0.36 MJeq/MJfuel and 8.7e16.6 g CO2eq/MJfuel, and FC-PHEV(b) shows 0.23e0.53 MJeq/ MJfuel and 13.6e26.7 g CO2eq/MJfuel.

10984

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 1 0 9 7 3 e1 0 9 8 5

EV reports the highest infrastructure LCA contribution per kilometre with 8.4%, followed by the FCHEV (b) with 8.3% in emissions. Conventional fuel facilities report the lowest values, contributing impacts below 2.5%. EV is mostly affected by the charging infrastructure with approximately 60% weight followed by power plants and maintenance with about a 33.2% contribution in terms of energy use and 42.6% in emissions, while FCHEV (b) is mostly affected by power plant construction and maintenance with approximately 68%, with 30% from H2 stations. End distribution facilities tend to be the higher contributors among energy supply infrastructures. Overall contributions with uncertainty do not go beyond 12% with the assessed conditions. A more friendly choice of materials or higher efficiency in H2 stations, H2 pipelines and electric chargers is advisable. Total and partial service ratios as well as usage from each type of chargers are crucial issues and should be further investigated. Increasing the energy flow homogeneously in normal chargers by promoting the vehicle to grid (V2G) concept would decrease the impact of chargers. The set environmental goals within the electricity generation mix and an increase in the lifetime driven distance will lower the WTT and CTG stages. These facts make the contribution of energy supply infrastructure pertinent to analysis in future studies, particularly in the alternative vehicle technologies studied. A generalisation of the study is, however, advisable. Due to a large range of uncertainty, closer attention should be paid to how the impacts vary according to different scenarios of behaviours, geographies or material inputs.

Support information Supplementary data attached in Appendix A.[46e48]

Acknowledgments The authors would like to thank the MIT-PP and the support of all companies that made this study possible, in particular EFACEC for the financial support. Special thanks to Elsevier Language Editing Services.

Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ijhydene.2012.04.127.

Nomenclature

AER CAFE CD CED CS GHG

All Electric Range; Corporate Average Fuel Economy; Charge Depleting; Cumulative Energy Demand; Charge sustaining; Green House Gas;

CTG Cradle-to-Grave of vehicle materials; DICI Direct Injection Compression Ignition; DISI Direct Injection Spark Ignition; ECE-EUDC European Driving Cycle Extra-Urban Driving Cycle; EV Full Electric Vehicle; FCHEV Fuel Cell Hybrid Electric Vehicle; FC-PHEV Fuel Cell Plug-in Hybrid Electric Vehicle; FTP Federal Test Procedure; GWP Global Warming Potential; ICE Internal Combustion Engine; LCA Life Cycle Analysis; NEDC New European Driving Cycle; PISI Port Injection Spark Ignition; PKT Passenger Kilometre travelled; RES Renewable Energy Sources; SI Supporting Information; SM Steam Methane Reforming; TTW Tank to Wheel; WTT Well to Tank; WTW Well to Wheel.

references

[1] Key world energy 2010 statistics. Paris: OECD/IEA Energy Statistics Division; 2010. 82 pp. [2] European Environment Agency [Internet]. Primary energy consumption by fuel in the EU-27, 1990-2008, [cited 2011 July] Available from: http://www.eea.europa.eu/. [3] Communication from the commission to the European council and the European parliament: an energy policy for Europe. Brussels: COM; 2007 Jan. 28 pp. [4] DGGE [Internet] Balanc¸o energe´tico - Direcc¸a˜o Geral de Energia e Geologia. [updated 2010 Oct.; cited 2010 Oct.] Available from: http://www.dgge.pt/. [5] McKinsey and Company. A portfolio of power-trains for Europe: a fact-based analysis. The Role of Battery Electric Vehicles, Plug-in Hybrids and Fuel Cell Electric Vehicles; 2010. 68. [6] Ministe´rio da Economia e da Inovac¸a˜o. A Estrate´gia Nacional para a Energia com o horizonte de 2020. Lisboa (Portugal); 2010 Apr.. 25 pp. Portuguese. [7] Geerken T, Timmermans V, Lassaux S. Hydrogen and it’s applications: review of life cycle assessment studies and well-to-wheel studies, 10th European Roundtable on sustainable consumption and production, Antwerp (Belgium), Hysociety project, 5-6-7; 2005 Oct.. [8] Weiss M, Heywood J, Drake E, Schafer A, AuYeung F. On the road in 2020-A life-cycle analysis of new automobile technologies. Cambridge (MA): MIT Energy Laboratory; 2000 Oct. 160 pp. Report No.: MIT EL 00e003. [9] Edwards R, Larive JF, Beziat JC. Well-to-wheels analysis of future automotive fuels and powertrains in the European context. Version 3c. Concawe, JRC, EUCAR; 2011 Jul. 74 pp. [10] Morrow K, Karner D, Francfort J. Plug-in hybrid electric vehicle charging infrastructure review. Idaho Falls (ID): Battelle Energy Alliance; 2008. 40 pp. Report No.:58517. Contract No.: DEAC0705ID14517. Supported by the Department of Energy. [11] Gross B, Sutherland I, Mooiweer H. Hydrogen fuelling infrastructure assessment. Mound Road (MI): General Motors Research Development Center; 2007 Dec. 43 pp. Report No.:R&D11,065.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 1 0 9 7 3 e1 0 9 8 5

[12] Elgowainy A, Burnham A, Wang M, Molburg J, Rousseau A. Well-to-Wheels energy use and greenhouse gas emissions analysis of plug-in hybrid electric vehicles. Chicago (IL): DOE; 2009 Feb. 70 pp. Argonne Laboratories, Energy Systems Division Report No.: ANL/ESD/09e2. Contract No.: DE-AC0206CH11357. Supported by U.S. Department of Energy. [13] Baptista P, Toma´s M, Silva C. Plug-in hybrid fuel cell vehicles market penetration scenarios. Int J Hydrogen Energy 2010;35: 10024e30. [14] Campanari S, Manzolini G, Garcia de la Iglesia F. Energy analysis of electric vehicles using batteries or fuel cells through well-to-wheel driving cycle simulations. J Power Sources 2009;186:464e77. [15] Silva C, Ross M, Farias T. Evaluation of energy consumption, emissions and cost of plug-in hybrid vehicles. Energy Convers Manage 2009;50:1635e43. [16] Chester MV, Horvath A. Environmental assessment of passenger transportation should include infrastructure and supply chains. Environ Res Lett 2009;4:024008. [17] Kendall A, Keoleian A, Helfand E. Integrated life-cycle assessment and life-cycle cost analysis model for concrete bridge deck applications. J Infrastruct Syst 2008;14:214e22. [18] Lee JY, An S, Cha K, Hur T. Life cycle environmental and economic analyses of a hydrogen station with wind energy. Int J Hydrogen Energy 2010;35:2213e25. [19] Lucas A, Silva C, Neto R. Life cycle analysis of energy supply infrastructure for conventional and electric vehicles. Energ Policy 2012;41:537e47. [20] Granovskii M, Dincer I, Rosen M. Life cycle assessment of hydrogen fuel cell and gasoline vehicles. Int J Hydrogen Energy 2006;31:337e52. [21] Siddikou B, Vincent W, Nele S, Heijke R, Maarten M, Mierlo JV. LCA and policy measures report. CLEVER Clean Vehicles Research ETEC and MOBI. Brussels (Belgium): Vrije University Brussels; 2011. 83 pp. [22] International Organization for Standardization. ISO 14040:2006 and ISO 14044:2006: environmental management e life cycle assessment e principles and framework. [23] Forster P, Ramaswamy V, Artaxo P, Berntsen T, Betts R, Fahey D, et al. Changes in atmospheric constituents and in radiative forcing. Climate Change 2007: the Physical Science Basis. IPCC Fourth Assessment Report. 106 pp. Working Group I: [Chapter 2]. UK, New York, USA: Joint publication of Cambridge University Press; 2007 Feb. [24] PRe´ Consultants [Internet]. Simapro tool [updated 2012 Apr.; cited 2012 Apr.] Available from: http://www.pre.nl/simapro [25] Silva A. Energy risks and safety. Risk Management in oil and gas activities [slides], vol. 133. Lisbon (Portugal): Partex Oil and Gas Company (President): ISEG Risk Management meetings; 2010 June 6. slides: color 2 x 2 in. [26] Data Book de seguranc¸a sau´de e ambiente - Refinaria de Sines 2009. Lisboa (Portugal): Galp Energia; 2009. 76 pp. Portuguese. [27] ACAP [Internet]. Associac¸a˜o Automo´vel de Portugal. Parque Automo´vel em [updated 2012 Fev.; cited 2012 Apr.] Available from: http://www.acap.pt; 2010. [28] Dados Te´cnicos 2010. Lisboa (Portugal): Redes Energe´ticas Nacionais; 2011. 1 pp. Portuguese. [29] DGGE [Internet]. Direcc¸a˜o Geral de Geologia e Energia. InfraEstruturas de Petro´leo e Ga´s Natural [updated 2010 Oct.; cited 2010 Oct.] Available from: http://www.dgge.pt/. [30] Groupe de travail sur le: Livre vert pour le de´veloppement des infrastructures de charge e´lectrique Paris. MEEDDM; 2010 Apr.. 198 pp. French.

10985

[31] Depoorter S, Assimon P. Les ve´hicules e´lectriques en perspective: analyse couˆts-avantages et demande potentielle Paris. Commissariat ge´ne´ral au de´veloppement durable (France); 2011 May. 64 p. Etudes et documents No.: 41. French. [32] Electric Transportation Engineering Corporation. Electric vehicle charging infrastructure deployment guidelines for Greater San Diego area. U.S. Department of Energy; 2010 May. 62 pp. Document 024293.1, Contract No. DE-EE0002194. [33] Spath PL, Mann MK. Life cycle assessment of renewable hydrogen production via natural gas steam reforming. Golden (CO): NREL; 2001. 33 pp. Report No.: NREL/TP-57027637. Contract No.: DEA3699GO10337. Supported by the Department of Energy. ¨ ko-Institut and Gesamthochschule Kassel [34] OEKO [Internet]. O [updated 2012 Apr.; cited 2012 Apr.] Available from: http:// www.oeko.de/service/gemis/en/. [35] Spath PL, Mann MK. Life cycle assessment of renewable hydrogen production via wind/electrolysis. Golden (CO): NREL; 2004. 13 pp. Report No.: NREL/MP-560-35404. Contract No.: DE-AC36-99-GO10337. Supported by the Department of Energy. [36] Maack M. Generation of the energy carrier hydrogen in context with electricity buffering generation through fuel cells. NEEDS: New Energy Externalities Developments for Sustainability; 2008. 47 pp. Report No.: 8-5 RS1a. Project No.: 502687 for Icelandic New Energy. Co-sponsored by the European Commission within the Sixth Framework. [37] Nansai K, Tohno S, Kono M, Kasahara M, Moriguchi Y. Lifecycle analysis of charging infrastructure for electric vehicles. Appl Energ 2001;70:251e65. [38] Hondo H. Life cycle GHG emission analysis of power generation systems: Japanese case. Energy 2005;30:2042e56. [39] Ministe´rio das obras pu´blicas, transportes e comunicac¸o˜es (PT), Dia´rio da Repu´blica; 2009. Set 4 Dec. L. No. 172-5975, artigo 5 , 1.a se´rie: 5972e5978. Portuguese. [40] Varun Bhat IK, Prakash R. LCA of renewable energy for electricity generation systems e a review. Renew Sust Energ Rev 2009;13:1067e73. [41] Vestas Wind Systems. Life cycle assessment of offshore and onshore sited wind power plants based on Vestas V90-3.0 MW turbines. Denmark; 2006 Jun. 60 pp. [42] Working Group B2.15. Life Cycle Assessment (LCA) for overhead lines. Paris: Cigre; 2005 Jan. 168 pp. Report No.:265. [43] Harrison GP, Maclean ENJ, Karamanlis S, Ochoa LF. Life cycle assessment of the transmission network in Great Britain. Energ Policy 2010;38:3622e31. [44] Frost and Sullivan. Strategic technology and market analysis of electric vehicle charging station infrastructure; 2011 Mar. 152 pp. Report No.: M616e18. [45] Frischknecht R, Jungbluth N, Althaus HJ, Doka G, Dones R, Heck T, et al. Overview and methodology. Swiss Centre for Life Cycle Inventories; 2007c. Ecoinvent report No. 1, 2007 Dec. 60 pp. vol. 2.0. Du¨bendorf, Switzerland. [46] Total.com [Internet]. Total Group media information: oil processing, projects and achievements. Normandy refinery [updated 2011 Jan; cited 2011 Fev 11]. Available from: http:// www.total.com/. [47] Silva C. Raffinage, ajustement a` un environnement en mutation. Lyon (France): Panorama annual; 2010 Jan. 7 pp. Report No.: Innovation Energy Environnement 2010. [48] Edpsu.pt [Internet]. Energias de Portugal: Origens da electricidade, emisso˜es 2009 e 2010. [updated 2011 Mar; cited 2011 Oct. 11]. Available from: http://www.edpsu.pt.