Energy balance and global warming potential of ...

Fuel Processing Technology 132 (2015) 74–82

Contents lists available at ScienceDirect

Fuel Processing Technology journal homepage: www.elsevier.com/locate/fuproc

Energy balance and global warming potential of biogas-based fuels from a life cycle perspective Elham Ahmadi Moghaddam a, Serina Ahlgren a, Christian Hulteberg b, Åke Nordberg a a b

Department of Energy and Technology, Swedish University of Agricultural Sciences (SLU), P.O. Box 7032, 750 07, Uppsala, Sweden Department of Chemical Engineering, Lund University, P.O. Box 124, 221 00, Lund, Sweden

a r t i c l e

i n f o

Article history: Received 19 May 2014 Received in revised form 19 September 2014 Accepted 9 December 2014 Available online xxxx Keywords: Biogas GTL fuels LCA Energy performance Global warming potential

a b s t r a c t Biogas is a multifunctional energy carrier currently used for co-generation or compressed biomethane as vehicle fuel. Gas-to-liquid (GTL) technology enables conversion of biogas into other energy carriers with higher energy density, facilitating fuel distribution. The energy efficiency and global warming potential (GWP) for conversion of biogas to compressed biogas (CBG), liquefied biogas (LBG), Fischer–Tropsch diesel (FTD), methanol and dimethyl ether (DME) were studied in a life cycle perspective covering the technical system from raw biogas to use in city buses. CBG, methanol and DME showed the best specific fuel productivity. However, when fuel distribution distances were longer, DME, LBG and methanol showed the best energy balance. Methanol, FTD and DME emitted half the GWP of LBG and CBG. Choice of electricity mix had a large impact on GWP performance. Overall, taking into account the different impact categories, combustion properties and fuel yield from raw biogas, DME showed the best performance of the fuel conversion scenarios assessed. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Fossil fuels currently comprise 80% of global primary energy consumption, 58% of which is consumed by the transport sector alone [1]. Biofuels are renewable alternatives and, owing to their origins in natural bioresources, they are geographically more evenly distributed than fossil fuels. Biogas (approximately 60% methane (CH4), 40% carbon dioxide (CO2) and some trace gases) produced during anaerobic digestion of organic matter (organic waste, sewage, manure etc.) is a renewable energy carrier, which can be used for e.g. combined heat and power (CHP) production. However, if biogas is cleaned and upgraded in order to increase the energy content, the resulting biomethane can be used as a renewable substitute for natural gas. By the end of 2012, 221 biogas upgrading plants were in operation worldwide, of which 55 units were located in Sweden. Water scrubbing, pressure swing adsorption and chemical scrubbing are the choice of technology in 90% of these plants [2]. The biomethane is uploaded to the natural gas grid or directly compressed to CBG (compressed biogas) for vehicle use. Due to the limits of grid infrastructure in certain regions and problems relating to storage and distribution systems for CBG, interest in technologies which convert biomethane to even higher energy density and more feasible transportability has increased. The option of converting biomethane to liquid biofuels would facilitate the supply of biofuels to geographically broader and larger markets. Furthermore, the potential for blending with liquid fossil fuels would be very useful. Today there are different routes for exploiting biogas energy as liquid biofuel. Liquefied biogas (LBG) is a form of upgraded biogas that has

http://dx.doi.org/10.1016/j.fuproc.2014.12.014 0378-3820/© 2014 Elsevier B.V. All rights reserved.

been cooled and liquefied at temperatures around − 161 °C under atmospheric pressure by cryogenic technology. LBG is three times more space-efficient than CBG (stored at 200 bar), while the fuel is in the gas phase when it reaches the engine [3]. A novel route of biogas conversion to vehicle fuel is gas-to-liquid (GTL) technology, a means to exploit gaseous energy sources as fuel, higher hydrocarbons (e.g. ethylene, a-olefins, paraffin, wax) and chemical products [4,5]. Existing GTL technology includes conversion of methane (from natural gas or upgraded biogas) to syngas (a mixture of carbon monoxide (CO) and hydrogen gas (H2)) and subsequent synthesis to e.g. Fischer–Tropsch Diesel (FTD), methanol and dimethyl ether (DME) through catalytic synthesis. Interest in producing GTL fuels from biomass and biogas as available renewable feedstock is increasing [6,7]. Moreover, the current situation of high oil prices and anticipation of increased market share for diesel fuels presents an entry point for GTL alternatives to the biofuel market. FTD is interchangeable with conventional diesel fuels and fully compatible with existing diesel engines and infrastructure, which is conducive to implementation in the short term [8]. FTD has a high cetane number, does not contain sulphur or nitrogen and has the potential for blending with diesel at any ratio with little to no modification of diesel engines [4]. In addition, synthetic fuels have emissions benefits in the reduction of hydrocarbons (HC), CO, nitrous oxides (NOx) and particulate matter (PM). Synthetic fuels can satisfy many of the ideal fuel requirements of modern diesel engines [9]. Methanol is another GTL product, which has been produced for many decades for manufacturing of high-value chemicals and fuel

E. Ahmadi Moghaddam et al. / Fuel Processing Technology 132 (2015) 74–82

additives. Methanol is the most basic alcohol and is a desirable choice as a transportation fuel due to its efficient combustion and ease of distribution. Methanol can be used directly as fuel or blended with petrol, converted to DME as a diesel replacement, or used in the biodiesel production process [10–12]. Methanol is a high-octane fuel that enables very efficient and powerful engine performance. However, methanol is toxic, has an affinity to water and has half the energy content of petrol on a volumetric basis [13]. Since methanol fuel is corrosive to certain materials commonly used in engines and fuel lines, it is blended with other fuel. Small modifications must be made to engines to include methanol-compatible components and to permit running on highlevel blends such as M-85 (a mixture of 85% methanol and 15% petrol). However, low-level blends of methanol do not cause adverse effects on car engines and can be used in cars today where available without any problems. Methanol is indeed also considered an alternative to marine gas oil or liquefied natural gas for ship propulsion and is claimed to have advantages in this application [14]. However, the substitution of marine gas oil with methanol is not within the scope of this paper. DME is the simplest ether primarily produced directly from syngas or indirectly by dehydration of methanol. Due to the chemical structure of DME, the possibility of forming carbonaceous PM and NOx emissions during combustion is limited [15]. DME combustion does not produce soot and is considered a clean fuel. Unlike methanol, DME is a gas at ambient temperature and pressure, so it is stored under pressure as a liquid similar to liquefied petroleum gas (LPG) and can use the same existing infrastructure as LPG [16]. The most challenging aspects of a DME engine relate to its physical properties and not its combustion characteristics. A DME fuel storage tank must be twice the size of a conventional diesel fuel tank due to the lower energy density of DME compared with diesel fuel, in order to achieve an equivalent driving range to CIDI (Compression-Ignition Direct-Injection) diesel. Modifications to pumps and fuel injectors are needed due to the 20-fold lower viscosity of DME compared with diesel [17]. Today, innovations in GTL technology, e.g. micro-channel technology, have led to improvements in efficiency of productivity and infrastructure. Introduction of the micro-channel technology will enable transformation of energy and chemical processing industries by greatly reducing the size of chemical reactor hardware. The main characteristic of micro-channel technology is parallel arrays of micro-channels, with typical diameter dimensions in the 0.1–5.0 mm range. Processes are accelerated by reducing heat and mass transfer distance, whereby system volumes can be reduced 10-fold or more compared with the conventional hardware [18–20]. Thus, the development of small-scale GTL technology offers future possibilities for converting biogas from anaerobic digestion to liquid fuels, facilitating distribution and flexible use. However, when nominating novel systems there is a need to analyse the energetic and environmental performance in a systems perspective and to compare it with that of conventional techniques. Life cycle assessment (LCA) is an internationally accepted method for measuring environmental performance and a useful tool for analysing products or services. LCA enhances the understanding of how alternative systems compare with each other, but also how different sub-processes in a system affect the overall results [21]. LCA methodology aims at change, or improvement: sometimes in more direct ways (decision-making) and sometimes in more indirect ways (influencing market behaviour, identifying improvement possibilities) [22]. Reduction of greenhouse gas (GHG) emissions is one of the main reasons behind introducing biofuels as alternatives to fossil fuels. In order to ensure that these GHG emissions are not excessive, emission limits and methods for calculating the emissions have been introduced into biofuel standards and legislation. In 2009, the European Union introduced sustainability criteria for biofuels in two directives; the Renewable Energy Directive (Directive 2009/28/EC) [23] and the Fuel Quality Directive (Directive 2009/30/EC) [24]. In these sustainability criteria a methodology to account for GHG savings compared to fossil

75

fuels is described. The sustainability criteria have greatly influenced the biofuel producers and the biofuel market in the EU. The objective of this study was to assess alternative biogas processing routes in terms of their energy efficiency and global warming potential (GWP) in a life cycle perspective. The study included conversion to liquid and gaseous fuels, such as LBG, FTD, DME and methanol, as well as conventional conversion to CBG. The assessment covered the technical system from raw biogas to use of the biofuel in public city buses. 2. Methodology The energy and environmental performance of the biofuel production chain, including raw biogas upgrading, fuel production, storage, distribution, fuelling and final conversion in bus engines was included in the study. The study was based on the LCA methodology described in ISO standards 14040 and 14044 [25,26], however some important deviations from the standards were made; the assessment was limited to only two impact categories and only energy allocation was included. This is similar to the methodology described in the sustainability criteria for biofuels in the EU [23]. Further the study had an attributional modelling approach, i.e. accounts for the immediate physical flows in a life cycle. This can be compared to consequential LCA-modelling, which examines the environmental consequences of change in a life cycle, often with a market-oriented approach [27]. The energy performance was based on the energy output (LHV; lower heating value) of the biofuel produced, compared to the required primary energy (PE) input. Factors used for conversion of data on electricity, heat and diesel to PE are presented in Table 1. The PE factor is defined as the ratio between PE and delivered useful energy. Included in PE are extraction of fuel, transportation and conversion, transmission and distribution losses [3]. The environmental impact included was GWP considering the direct emissions of the greenhouse gases CO2, CH4 and N2O during the life cycle of biofuel production. Direct emissions were defined as emissions occurring inside the system boundary, connected to the fuel production chain, an example being emissions from production of input electricity. Emissions occurring outside the system boundary, such as emissions occurring from market induced changes, were not included in this study. The emissions were calculated as CO2-equivalents (CO2-eq.) using characterisation factors for a 100-year perspective based on IPCC, 2007 [28]. According to this, 1 kg of CO2, CH4 and N2O is weighted as 1, 25 and 298 kg CO2-eq., respectively. Biogenic carbon was not included in the GHG accounting. In the GTL scenarios (FTD, methanol and DME), fuel synthesis was modelled in flow sheet software (AspenTech's Aspen Plus 7.3.2). The simulations performed are applicable to the micro-channel concept described above. The operating parameters used for the unit operations are summarised in Appendix A. 2.1. Functional unit and system boundaries A common basis for calculation had to be defined in order to compare different scenarios. Each scenario was analysed based on energy balances and GHG emissions. Since the aim of this study was to assess alternative processing routes for raw biogas, an input-based functional unit (FU) was deemed appropriate. Thus, the FU was defined as 1 Nm3

Table 1 Primary energy (PE) factors for different energy carriers (MJ PE/MJ energy carrier). Energy carrier

Specifications

Primary energy factor

Electricity Fuel Heat

NORDEL Diesel, low-sulphur District heating

2.17a 1.27a 0.79b

a b

[24]. [25].

76


raw biogas produced by anaerobic digestion. The anaerobic digestion process (handling and use of substrate, plant operation and use of digestate) was assumed to be identical for all the scenarios assessed and therefore not included in the analysis. However, the fuel production and use of the fuel were included; thus the study had a gate-to-grave modelling approach. The system boundary encompassed the following processes: • Biogas cleaning and upgrading • Compression and production of fuel (CBG, LBG, FTD, methanol and DME) • Storage, transportation, distribution and fuelling of fuel • Combustion of fuel in the internal combustion engine of public buses in urban conditions

In the GTL scenarios in this study there were three output products: biofuel, heat and steam. We opted to base our allocation on the LHV of the products. This is a commonly used methodology when performing LCA of biofuel systems [29] and is also used in the EU regulations on biofuel GHG performance [23]. The amounts of output products on which the allocation factors were based were calculated in flow sheet software modelling of fuel synthesis. The choice of method for dealing with multi-functionality has a strong influence on the results, as shown for example by Luo et al. [30]. Therefore we explored the impact of other allocation methods in a sensitivity analysis. 3. System description 3.1. General overview

The alternative biofuel production plants studied were assumed to apply micro-channel technology and to be located in connection to the biogas production unit, reflecting regions with limitation in gas grid infrastructure, e.g. Sweden. Demand for electricity for fuel stations such as lightning, control systems etc. was included, and was considered the same for all fuels. Production of capital goods such as machinery and buildings was not included in the calculations, as it was estimated that this would have only slight effects on the overall results [28]. Moreover, processes related to the production, maintenance and disposal of the buses and processes connected to road building and infrastructure were not included.

2.2. Allocation procedure Multi-functionality problems arise when two (or more) products share or partly share a production system. To handle such multifunctionality, the ISO standards recommend avoiding allocation in the first instance by increasing the level of detail in the study, or by performing a system expansion where the co-products are assumed to replace other products on the market, thus subtracting the environmental impact of substitution from the main production system. However, multi-functionality can also be handled by allocating the environmental impacts over the products based e.g. on their physical or economic properties [25].

The assessment was based on production of 1140 Nm3/h raw biogas (1 atm, 0 °C) with a composition of 60% CH4, 39.8% CO2 and 2000 ppm hydrogen sulphide (H2S). The annual gross energy production was approximately 60 GWh based on 9.97 kWh/Nm3 CH4. Five different scenarios for conversion of biogas to vehicle fuel quality were compared (Fig. 1). One scenario was based on the current conventional conversion of biogas to CBG and four scenarios on the conversion to liquid fuels, i.e. LBG as an emerging technology and FTD, methanol and DME as potential future technologies. In the CBG scenario, the raw biogas was assumed to be cleaned, upgraded and compressed to CBG. In the LBG scenario, the raw biogas was cleaned and upgraded including a CO2 polishing step to a quality allowing cryogenic liquefaction of biomethane. The three remaining scenarios (FTD, methanol and DME) were based on cleaning and upgrading the biogas and conversion to syngas, which was assumed to be synthesised to the respective fuels through catalytic reactions. In all scenarios, the fuel produced was stored and then transported 100 km to a filling station and used to fuel city buses. 3.2. CBG scenario In the CBG scenario, raw biogas was assumed to be upgraded by water scrubbing technology for removal of H2S and CO2. The electricity requirement was 0.45 kWh/Nm3 upgraded gas (UG) [31]. Direct emissions from biogas upgrading relate to methane losses of 1.2 g

Fig. 1. Scenarios for biofuel production from raw biogas.


CH4/kWh UG during the process [32]. The upgraded biomethane (97% CH4) was assumed to be compressed to 200 bar, requiring 0.18 kWhel/Nm3 UG [32], and stored in pressurised vessels. The CBG was transported using a hook lift container with a total capacity of 5160 Nm3 and average diesel consumption (round trip) of 0.4 L/km [3]. At the CBG fuelling station, a fuel pump dispenses the compressed fuel to vehicles [33], with electricity consumption of 0.07 kWh/Nm3 UG and methane emissions of 0.0038 g CH4/kWh UG [31,32]. 3.3. LBG scenario In the LBG scenario, the raw biogas was assumed to be first upgraded by the same technology as in the CBG scenario. Thereafter, the remaining CO2 was removed by a chemical scrubber. The biomethane was chilled to − 161 °C and liquefied by the closed Brayton cycle method, where N2 is the refrigerant fluid. Liquefaction requires 0.63 kWhel/Nm3 UG [31]. LBG was stored in a vacuum insulated vessel and delivered by semi-trailer with a tank capacity of 33,000 Nm3. The average fuel consumption (round trip) of a LBG semi-trailer is 0.46 L/km [3]. At the fuelling station, the LBG was transferred from the vessel by transfer pumps through vacuum insulated pipes to the dispenser, which dispenses fuel at a pressure of 5–8 bar [3]. The electricity consumption at the fuelling station is 0.025 kWh/m3n UG and the methane emissions are 0.00076 g CH4/kWh [31,32]. 3.4. GTL fuel scenarios Biogas upgrading and syngas production are two common stages in GTL fuel scenarios. In the present study, biogas upgrading was assumed to be performed by a chemical scrubber based on an activated methyldiethanolamine (MDEA) system, with an electricity requirement of 0.12–0.14 kWh/Nm3 UG and a heat requirement of 0.55 kWh/Nm3 UG [2]. The chemical scrubber was chosen since the surplus heat generated in the GTL scenarios can be utilised. Methane emissions related to the chemical scrubber are 0.2 g CH4/kWh UG [32]. Upstream of the scrubber unit, an activated carbon filter is used for removing the bulk of the H2S, while traces of H2S are removed in the chemical scrubber. The resulting biomethane is reformed to syngas by the steam reforming method, modelled here as Gibbs free energy reactors. The steam reformer is externally heated by burning part of the feed. The burner was modelled as a combustion reactor in Aspen Plus setup to provide the energy required in the steam reformer. Temperature in the steam reformer was set to 800 °C and with a pressure drop of 1 bar. After reforming, the gas contains H2, CO, CO2, nitrogen gas (N2) and water vapour (H2O). To set the ratio between CO and H2 for synthesising the fuel, CO is shifted with H2O to H2 in the water–gas shift reaction (Eq. (1)). CO þ H2 O↔CO2 þ H2

ð1Þ

The water–gas shift reactors were modelled in Aspen Plus as equilibrium reactors operating adiabatically with outlet temperatures of 450 °C and 220 °C, respectively [35]. The water–gas shift reaction is an

Syngas

77

equilibrium reaction that is pushed towards H2 at low temperature. In order to achieve maximum H2 content in the gas, two water–gas shift reactors are required, a high temperature and a low temperature reactor. The systems are heat integrated using pinch technology. There is therefore energy conservation within the system boundary such that the exported values of steam and heat are the lowest ones that can be achieved with sound engineering. The syngas produced is guided to different paths for fuel (FTD, methanol and DME) production. Data on direct emissions for GTL scenarios were taken from the literature [35–37]. In all cases, the steam generated is medium pressure steam at 26 bar (a) and 275 °C. The hot water was assumed to be at 130 °C and hence compatible with district heating. In all cases, the internal heat requirement was satisfied first and the values reported are net export. 3.4.1. FTD scenario The resulting syngas was assumed to be converted via FT synthesis into liquid hydrocarbons suitable for the manufacture of fuels and chemicals. In order to achieve the highest yield of FT fuel from the syngas, it is important that the maximum amount is converted in the downstream FT reactors. This requires the composition of the syngas to match the overall usage ratio of the reactions. The FT reactor was modelled as an isothermal plug reactor with varying length and, as the reactor is operating at 250 °C, both iron (Fe) and cobalt (Co) are reasonable catalyst materials. The reaction kinetics were modelled as power law reactions following the Anderson–Schulz–Flory distribution (Eq. (2)), with α-value chosen (α = 0.85) to yield as much diesel fuel (C10–C15) as possible.

2

Wn ¼ nð1–αÞ αðn−1Þ

ð2Þ

The refining step can include processes as severe as hydrocracking and/or fluid catalytic cracking. In our simulations, FT synthesis and hydrocracking were considered. The main energy source in this step is electricity, 2.18 MJ/Nm3 raw biogas. During the refinery stage, lowpressure steam and heat are produced. The hydrocarbons that are heavier than diesel fuels are good feedstock for a cracker to produce more vehicle fuel. The length of the reactor can be varied to achieve a once-through conversion of 80% of the ingoing CO. The hydrocarbons produced in the reactor contain alkanes and alkenes from methane to C20 compounds. The breakpoint between useful fuel and tail-gas is around C6, with most of the C6 compounds ending up in the tail-gas and most of the C7 in the diesel fuel. Part of the tail-gas is burnt to produce heat and power for the plant. The remaining tail-gas is recirculated to the reformer (Fig. 2). From the refinery, the fuel is transported via road tankers to intermediate storage and distribution centres (tank farms) for distribution to refuelling stations. FT fuels may be distributed pure, or blended with conventional diesel to improve fuel characteristics. In all cases, the storage, distribution and station (pump) infrastructure is identical to the existing diesel network and does not require modification [8].

Tail-gas recycled to reformer

Separator

Syngas compressor Pre-heater

Fischer-Tropsch reactor

Condensor

Fig. 2. Fischer–Tropsch diesel (FTD) production process.

Diesel fracon

78


Tail-gas

Recirculation compressor Syngas

Methanol reactor Knock-outdrum

Syngas compressor Pre-heater

Condensor Methanol

Fig. 3. Methanol production process.

3.4.2. Methanol scenario The methanol reactor was simulated using a steam-raising type of reactor, where the reactor is cooled by boiling feed water. The water circulation is provided by a thermosyphon effect and therefore no pumps are required on the water side. Methanol synthesis recycle ratio is 5:1 [34,38]. The product stream is condensed and methanol and water are removed. The gas is recycled to the inlet of the reactor except for a small part that is removed as tail gas (Fig. 3). Methanol and water are distilled to yield 99.9% pure methanol [39]. Unlike the reforming process, the synthesis of methanol is highly exothermic, occurring over a catalyst bed at moderate temperatures. Most plant designs make use of this extra energy to generate the electricity needed in the process, but in this study the energy was assumed to be used for steam generation.

3.4.3. DME scenario Methanol is the feedstock for DME, which utilises the methanol synthesis described above as input. The methanol is dehydrated in the presence of a catalyst, resulting in the production of DME. The product is cooled and methanol, DME and water are separated in a two-step distillation process. Methanol is recycled back to the reactor inlet. The purge stream contains methanol, DME, water and trace amounts of H2, CO etc. The purge stream is burnt to produce heat and power (Fig. 4). The simulations were based on pressurised storage of DME, requiring no energy for storage. This form of storage was chosen because the DME is separated in the liquid phase under pressure, so keeping it there is the most efficient storage method in the system. The properties of DME are similar to those of LPG fuels for a number of different physical parameters. Therefore storage and numerous distribution infrastructures for LPG could be used for DME, and transitioning to DME could be less costly than building completely new infrastructure. The fuelling facility contains a storage tank of raw DME and a consumption tank of DME and additive mixture, plus two small storage tanks and dosage pumps for lubrication and odour additives [40].

3.5. Fuel combustion in engines (city buses) Fuel combustion and emissions were studied in the engine of public buses. Based on different fuel scenarios, engines differ. The CBG engines used in buses are monovalent and dedicated for CBG usage only. CBG fuel tank installations for buses include four 320-L aluminium tanks, which are mounted in a package on the roof. The operating pressure is 200 bar and the typical inner city operating range is around 500 km. LBG vehicles are equipped with a cryogenic tank and here a similar engine to LNG was considered [41]. FTD can be used as a compression ignition (CI) fuel and would require no engine modifications. The operation of a DME engine requires a new storage system and a new fuel delivery system, while the engine itself does not need modification. For FTD and DME, similar engines to diesel fuel are required. Methanol can be used in engines similar to petrol engines [42].

4. Results and discussion In the present study, the GTL fuel scenarios were modelled based on maximum fuel production, i.e. it was assumed that all non-diesel hydrocarbons were reconverted to syngas and reprocessed. However, the syngas production step results in a large purge gas stream which can be utilised only for generation of steam or electricity. The amounts of fuel, heat and steam generated in the different scenarios and the allocation factors used, based on the LHV, are presented in Table 2. CBG and LBG gave the highest fuel yield, followed by DME and methanol, while FTD gave the lowest fuel yield. There is rather substantial production of steam and heat in the various GTL scenarios, with the FTD case having the highest production. The heat generated internally in the process is used for regeneration of the chemical scrubber (0.62 MW). This is normally done by consumption of raw biogas. The steam generated can be exported as steam or used for CHP production with a small steam turbine. For the purposes of this study, the steam stream was assumed to be exported as steam.

DME Feed methanol Pre-heater

DME reactor

Methanol pump

Distillation

Cooler Purge

Recirculation pump

Fig. 4. Dimethyl ether (DME) production process.

Water

E. Ahmadi Moghaddam et al. / Fuel Processing Technology 132 (2015) 74–82 Table 2 Amount of fuel, heat and steam (LHV) produced (MJ/Nm3 raw biogas) in the different scenarios and the resulting allocation factors. Fuel

CBG LBG FTD Methanol DME

Heat

Steam

Total energy

Fuel

Heat

Allocation factor (%)

21.19 21.19 9.24 16.23 19.31

100 100 41.28 83.22 86.73

2.86 1.49 1.65

10.29 1.78 1.30

and LBG have the same fuel output, the specific fuel productivity varies due to the high energy input during liquefaction in the LBG scenario and the lower input for compression in the CBG scenario.

Steam

MJ produced/Nm3 raw biogas 21.19 21.19 22.39 19.50 22.26

79

4.2. Emissions analysis 12.77 7.65 7.42

45.96 9.12 5.85

The CBG and LBG scenarios only generated fuel, to which 100% of the PE input and the GWP were allocated. In the DME and methanol scenarios, these were mainly allocated to the fuel (87% and 83%, respectively). The PE input and GWP in the FTD scenario were mainly allocated to fuel and steam (41% and 46%, respectively). In all the GTL scenarios, the allocation factors were applied to the upgrading and syngas/fuel synthesis step. Since the transport and fuelling station steps are only related to the fuel, the PE use and GWP were only allocated to the fuel part. 4.1. Energy analysis The PE input per FU for different stages in each fuel scenario is presented in Table 3. Considering the allocated values for fuel production, the total PE input per FU for the different fuel chains showed the following order of consumption (highest to lowest): LBG, CBG, FTD, DME and methanol. LBG requires a high PE input, mainly due to the liquefaction and the upgrading processes of raw biogas, which together represent 96% (53% and 43%, respectively) of the total PE input. The PE inputs for upgrading and compression in the CBG scenario together represent 74% (53% and 21%, respectively) of the total PE input. For the GTL fuels, the syngas/fuel synthesis process step corresponds to 86%, 67% and 62% of the total PE input for FTD, DME and methanol, respectively. Upgrading of LBG involves a separate CO2 polishing step for the removal of excess CO2 in addition to the primary upgrading with a water scrubber. The PE input for the chemical scrubber in the GTL scenarios is lower, since there is no need for compression and the heat for regeneration of the liquid is generated internally. Transport and fuelling of CBG require a high PE input in comparison with the other scenarios, mainly due to low density of gas in CBG transportation (~ 200 bar) and distribution to mobile storage units. During CBG transport, a high proportion of steel is transported in comparison with the amount of gas, while the liquid fuel scenarios are of high efficiency in terms of transportation. The PE input at fuelling stations is mostly related to electric pumps. The specific fuel productivity, described as the ratio of output fuel to total PE input, was relatively high in the CBG, methanol and DME scenarios, while it was lower for LBG and FTD (Table 3). Although CBG

The total GWP included direct emissions and indirect emissions in the different stages of the fuel scenarios (Table 4). Total GWP was highest for the LBG and CBG scenarios, while methanol showed the lowest GWP. In the LBG and CBG scenarios, the high GWP was mainly related to the upgrading (64% and 80% of total GHG emissions, respectively). For the upgrading in the CBG and LBG scenarios, direct methane emissions represented approximately 60% of the GWP, with the remainder from electricity use. The water scrubbing technology used as the upgrading technology in the CBG and LBG scenarios emits higher levels of methane than the amine scrubber used in the GTL fuel chain, due to the higher pressure used in the water scrubber [2]. The higher pressure increases the solubility of methane in the water, which is released in the CO2 recovery step. The amine scrubber, on the other hand, is operated at near-ambient pressure, resulting in low methane solubility and hence losses. The syngas/fuel synthesis processes contributed most of the GWP within the GTL fuel chain: 82% of total GWP for FTD, 62% for DME and 57% for methanol. These emissions are mainly related to electricity consumption. There were no significant direct emissions related to the GTL scenarios except for N2O emissions from the FTD syngas/fuel production step, which accounted for 7% of the emissions in the syngas/fuel production step (5.7% of total GHG emissions in the scenario) [36]. The calculations and assessments revealed considerable differences in PE input, specific fuel productivity and GHG emissions between the different scenarios. This was generally due to the fuel processing stage, while the transport and fuelling stages made a minor contribution.

4.3. Fuel combustion in engine In order to have a tangible benchmark, fuel efficiency in terms of distance driven by different fuels is presented. Table 5 shows the GWP caused by engine combustion and distance driven by public buses for the different fuels studied [43,44]. GTL fuels had different ranges in kilometres driven per FU. DME travelled the longest and methanol the shortest distance per unit volume for the FU used. A methanol bus consumes 2.3–2.6 times as much fuel as a DME and FTD diesel bus [41]. CBG and LBG gave approximately the same distance travelled per FU. Engine combustion among the GTL fuels resulted in less GHG emissions, especially for FTD, compared with LBG and CBG. It should be borne in mind that for the purposes of this study, fuel consumption and GHG emissions were calculated using 1 Nm3 raw biogas as the FU. Therefore the low emissions in the FTD scenario were also due to poor yield of fuel.

Table 3 Primary energy input (MJ/Nm3 raw biogas) for the scenarios studied.a CBG

LBG

FTD

Methanol

DME

MJ primary energy/Nm3 raw biogas Upgrading Compression Liquefaction Syngas/fuel synthesisc Transport Fuelling station Total PE Specific fuel productivityd a b c d

2.14 0.86 0 0 0.72 0.33 4.05 5.23

2.41b 0 2.99 0 0.08 0.12 5.59 3.79

0.43 (1.04) 0 0 3.42 (8.30) 0.05 0.09 3.99 (9.47) 2.31 (0.98)

0.86 (1.04) 0 0 2.01 (2.42) 0.19 0.16 3.22 (3.81) 5.03 (4.26)

0.90 (1.04) 0 0 2.58 (2.97) 0.20 0.19 3.87 (4.41) 4.99 (4.38)

Values not allocated to fuel for the GTL scenarios are given in brackets. Includes upgrading and a purification step. No external heat was included in the GTL fuel production process. Specific fuel productivity is described as output fuel/total PE in the fuel chain.

Table 4 Global warming potential (g CO2-eq./Nm3 raw biogas) for the fuel scenarios studied.a CBG

LBG

FTD

Methanol

DME

70 (84) 0 0 106 (127) 2 8 187 (222)

73 (84) 0 0 136 (156) 2 10 221 (253)

(g CO2-eq./Nm3 raw biogas) Upgrading Compression Liquefaction Syngas/fuel synthesis Transport Fuelling station Total GWP a b

289 45 0 0 8 18 360

298b 0 157 0 1 6 462

35 (84) 0 0 185 (449) 1 5 226 (540)

Values not allocated to fuel for the GTL scenarios are given in brackets. Includes upgrading and a purification step.

80


Table 5 Emissions (g CO2-eq.) and distance travelled (km) per functional unit (1 Nm3 raw biogas).

g CO2-eq./Nm3 raw biogasa km/Nm3 raw biogasb g CO2-eq./km a b

CBG

LBG

FTD

Methanol

DME

57.20 1.88 30.38

67.79 1.90 35.75

17.56 1.03 16.98

32.44 0.7 46.44

32.44 2.16 15.02

[42]. [43].

Table 7 Change (%) in PE input per functional unit when selected input parameters were changed.


Allocation factors based on equivalent electricity

1000 km transport

Hard coal electricity mix

Swedish electricity mix

+39 +13 +10

+160 +13 +11 +53 +47

+37 +43 +44 +42 +43

+4 +5 +5 +5 +5

4.4. Sensitivity analysis Sensitivity analyses should be performed in order to understand the impact of uncertainties on results and potential decisions [45]. In this study a sensitivity analysis was performed to evaluate the influence on total PE input and GHG emissions of allocation method, transport distance and choice of electricity mix used in the LCA calculations. Choice of allocation method had a large influence on the LCA results. As mentioned in the Introduction, allocation can be based on products according to e.g. their physical or economic properties. In the base case in this study, allocation was based on LHV of outgoing products. However, allocation based on LHV does not take the exergy of the various energy sources into account. Therefore, a sensitivity analysis was performed based on equivalent electricity produced instead of LHV. Electricity equivalents based on power generation efficiency have previously been used to represent the overall exergy [46,47]. Allocation factors for GTL scenarios based on electricity equivalents are presented in Table 6. The results indicated that using electricity equivalents as the allocation base gave higher values of input PE (Table 7) and GWP (Table 8) for the GTL scenarios. However, the GTL scenarios still resulted in total lower allocated emissions compared with the CBG and LBG scenarios. Another method to deal with multi-output systems is to use system expansion. In this case, co-products are assumed to replace marginal alternatives on the market. While this is mainly a method used in consequential LCAs, some authors argue system expansion can also be done in attribution modelling if using average data [48]. Performing a system expansion usually requires further modelling and data collection and is not within the scope of this study, however it is reasonable to assume that a system expansion would give different results especially for FTD that has a large co-production of steam. The sensitivity of transport distance was analysed in order to determine the influence on transport of gaseous and liquid fuel for larger distances. For comparison, 1000 km transport was assumed instead of 100 km. Comparison of the results showed a large increase in PE input and GHG emissions per FU for CBG compared with the liquid fuels studied. FTD and LBG showed a minor increase compared with the other fuels, mainly due to the higher energy density of the fuel (Tables 7 and 8). Increasing the transport distance to 1000 km reduced the specific fuel productivity of CBG to only 2.01 output fuels per total PE input, which was the lowest of all the different fuel chains, followed by FTD (2.08). DME, LBG and methanol resulted in the highest specific fuel productivity, 3.41, 3.35 and 3.29, respectively. However, the total allocated GWP was still higher for CBG and LBG compared with the GTL fuels.

Table 6 Fuel, heat and steam yield (MJ electricity equivalents/Nm3 raw biogas) in the different scenarios including the allocation factors. Fuel


Heat

Steam

Total energy

Fuel

Heat

Allocation factor (%)

12.2 12.2 5.17 11.03 10.79

100 100 57.77 95.25 95.91

0.29 0.15 0.17

3.49 0.40 0.29

3.24 1.30 1.51

4.5. Overall discussion Based on specific fuel productivity and total PE input, the best choices for utilisation of raw biogas for GTL fuels were methanol and DME. These two fuels had also lower GWP in comparison with the other fuels studied. Assessment of fuel combustion in engines showed that although methanol has an overall high energy and performance low GWP, owing to its lower heating value, larger volumes of fuel must be consumed for similar distances, resulting in larger fuel tanks and heavier tanks. FTD showed a relatively low fuel yield and a high steam yield. Thus, the localisation of an FTD-plant close to an industry or CHP-plant would facilitate an efficient utilisation of the produced steam. Furthermore, the results showed that converting biogas to CBG leads to a high specific fuel productivity if the CBG is utilised on a local market. However, in order to reach a regional or national market with relatively high specific fuel productivity, methanol, DME and LBG would be a better choice. It is important to consider that the present study is reflecting a situation with limitations in the gas grid infrastructure, meaning that the anaerobic digesting plant cannot utilize the gas grid for injection of upgraded biogas in order to reach a regional or national market. In a situation with a well-developed existing gas grid, the option of gas injection from several anaerobic digestion processes would facilitate CBG to reach a larger market, but also to reach a larger scale for the GTLprocess with other options for utilisation of steam and heat.

Table 8 Change (%) in GWP per functional unit when selected input parameters were changed.

Steam

MJ electricity produced/Nm3 raw biogas 12.2 12.2 8.95 11.58 11.25

In the base case, an electricity mix for the Nordic countries (NORDEL) was used for the calculations, based on 35% water, 11% biomass, 32% nuclear, 20% fossil and 2% wind, solar and geothermal sources. The electricity mix assumed in LCA calculations can be discussed. In the sensitivity analysis, we compared the results for the Nordic mix with results obtained using a higher share of fossil fuel (Hard coal NORDEL) and the Swedish electricity mix. The PE factors and GWP used in the sensitivity analysis for the hard coal NORDEL and Swedish electricity mix are presented and compared with the base NORDEL values in Table 9. The hard coal NORDEL was chosen to represent a fossil-based electricity mix, with 98% hard coal [49]. The sensitivity analysis showed that hard coal NORDEL increased the input PE and GHG emissions by an average of 42% and 629%, respectively (Tables 7 and 8). Thus, the relative increase was in the same range for all scenarios. The Swedish electricity mix is mainly based on nuclear energy (N 50%) and therefore increased PE input by an average of 4.8%, while GWP decreased to an average of 72.2%.

38.99 3.42 2.58


Allocation factors based on equivalent electricity

1000 km transport

Hard coal electricity mix


+39 +14 +10

+20 +2 +4 +8 +10

+621 +624 +632 +628 +629

−71 −73 −73 −72 −72

E. Ahmadi Moghaddam et al. / Fuel Processing Technology 132 (2015) 74–82 Table 9 PE factors and CO2-equivalents for different electricity mixes (ecoinvent, 2006).

MJ eq./kwh kg CO2-eq./kwh

Base NORDEL

Hard coal NORDEL


7.8 0.18

11.4 0.95

8.2 0.04

81

the choice of electricity mix (Base NORDEL, hard coal NORDEL and Swedish electricity mix) had a large impact on the total GWP performance. Overall, taking the energy efficiency and GWP into account, DME showed the best performance of the fuel conversion scenarios studied. Acknowledgment

Sensitivity analyses showed that use of different allocation methods affected the LCA outcomes. An allocation method based on electricity equivalents, longer transport and a less renewable electricity mix led to larger PE input and worse environmental performance unless the Swedish electricity mix was used. The sensitivity analysis also showed that the choice of electricity mix greatly influences the results. In a decision-making context where GHG emissions are important, electricity mix can be more important than choice of biogas-based fuel. The assessment in this study was not a full LCA, as we only assessed the conversion of raw biogas to fuel and its use in city busses. This means that the results cannot be directly compared with those for fossil or other fuels. However, we can compare the results with the disaggregated default values in the Renewable Energy Directive [23] in which CBG from different substrates is estimated to emit 12–17 g CO2-eq. per MJ fuel for processing, transport and distribution. Although the Renewable Energy Directive assumes zero emissions for raw material supply, the numbers include emissions from the anaerobic digestion process, which makes a direct comparison difficult. However, in the present study the emissions calculated per MJ fuel was 17 (CBG), 22 (LBG), 24 (FTD), 12 (methanol) and 11 (DME) g CO2-eq. for process, transport and distribution, indicating that the CBG numbers are in the same range as in the Directive. In this study, the biogas production system was considered to be the same for all scenarios and was therefore not included. For future LCA studies, the scope can be extended by including the agricultural system regarding choice and supply of substrate and anaerobic digestion efficiency, utilizing the digestate as an organic fertilizer and CHP in comparison with transportation fuels. In addition, the LCA methodology can be further expanded, e.g. applying system expansion to include the use of co-products in other systems. There are many studies involving life cycle perspectives for assessing energy balances and environmental impact of biogas systems [50–54]. However, the wide variability in systems borders, scope, approach and state-of-the-art in end use technologies makes the comparison of the various results difficult [29]. As an example, Patterson et al. (2011) showed that CHP with 80% utilisation of surplus heat results in least overall environmental impact [54]. However, transportation fuel use (CBG) is more favourable when high level of heat utilisation cannot be achieved. Moreover, the small scale micro-channel GTL technologies assessed were modelled and more technical research and development is needed to verify the performance with biogas from anaerobic digestion in order to include economic assessment in future studies. Nevertheless, the results from this study show that converting biogas to liquid fuels in order to increase transportability could be beneficial regarding energy balance and GWP when compared with conventional upgrading and compression of biomethane. 5. Conclusions The CBG, methanol and DME-scenarios resulted in the highest specific fuel yields (5.2, 5.0 and 5.0 MJ primary energy/Nm3 biogas, respectively) among the scenarios studied, while FTD resulted in the lowest specific fuel yield (2.3 MJ PE/Nm3 biogas). However, when increasing the fuel distribution distance from 100 km to 1000 km, the DME, methanol and LBG scenarios showed the best specific fuel yield performance. The methanol, FTD and DME-scenarios resulted in a total GWP (including fuel conversion from upgrading to combustion in bus engines) of 219, 244 and 253 g CO2-eq./Nm3 biogas, respectively, which was approximately half of the GWP from the LBG and CBG-scenarios. However,

The authors would like to thank the Swedish Farmers' Foundation for Agricultural Research (SLF) for funding the study. Elham Ahmadi Moghaddam would like to thank Amir Sattari for comments that greatly improved the manuscript. Appendix A. Operating parameters for unit operations Unit

Temperature, inlet/outlet [°C]

Pressure, inlet/outlet [bar(a)]

Aspen model

Methane reformer Water–gas shift CO2/water removal Compression (0.72 isentropic efficiency used for all cases) Methanol DME Fischer–Tropsch

750/850 350/520 70/30 70 (5 stages with intercooling)

9/8 8/7.7 7.5/7.5

Gibbs free energy Equilibrium Separation block/flash Multistage compressor

125/260 300/300 250/250

100/92 20/18 27

Steam for power generation Turbine isentropic efficiency 0.9 District heating

70/500

90/90

500/39

90/0.07

Equilibrium Equilibrium Plug-flow reactor with power law reactions Steam generated down to 150 °C Compressor/turbine

-

-

Interval 150-70 °C

References [1] P.S. Nigam, A. Singh, Production of liquid biofuels from renewable resources, Progress in Energy and Combustion Science 37 (2011) 52–68. [2] F. Bauer, C. Hulteberg, T. Persson, D. Tamm, Biogas upgrading—review of commercial technologies, SGC Rapport2013. [3] N. Johansson, Production of liquid biogas, LBG, with cryogenic and conventional upgrading technologyOnline http://4energy2.eu/downloads/0424nina-j-ex-jobbimes-raport-opwerk-methoden.pdf (viewed 27 November 2013). [4] M.E. Dry, The Fischer–Tropsch process: 1950–2000, Catalysis Today (2002) 227–241. [5] S.R. Deshmukh, A.L.Y. Tonkovich, K.T. Jarosch, L. Schrader, S.P. Fitzgerald, D.R. Kilanowski, J.J. Lerou, T.J. Mazanec, Scale-up of microchannel reactors for Fischer– Tropsch synthesis, Industrial & Engineering Chemistry Research 49 (2010) 10883–10888. [6] S. de Llobet, J.L. Pinilla, M.J. Lazaro, R. Moliner, I. Suelves, Catalytic decomposition of biogas to produce H2-rich fuel gas and carbon nanofibers. Parametric study and characterization, International Journal of Hydrogen Energy 37 (2012) 7067e7076. [7] K. Engvall, H. Kusar, K. Sjöström, L.J. Pettersson, Upgrading of raw gas from biomass and waste gasification: challenges and opportunities, Topics in Catalysis 54 (2011) 949–959. [8] O.P.R. van Vliet, A.P.C. Faaij, W.C. Turkenburg, Fischer–Tropsch diesel production in a well-to-wheel perspective: a carbon, energy flow and cost analysis, Energy Conversion and Management 50 (2009) 855–876. [9] S. Gill, A. Tsolakis, K. Dearn, J. Rodríguez-Fernández, Combustion characteristics and emissions of Fischer–Tropsch diesel fuels in IC engines, Progress in Energy and Combustion Science 37 (2011) 503–523. [10] S. Liu, E.R. Cuty Clemente, T. Hu, Y. Wei, Study of spark ignition engine fueled with methanol/gasoline fuel blends, Applied Thermal Engineering 27 (2007) 1904–1910. [11] F.J. Keil, Methanol-to-hydrocarbons: process technology, Microporous and Mesoporous Materials 29 (1999) 49–66. [12] A. Demirbas, Biodiesel production from vegetable oils via catalytic and non-catalytic supercritical methanol transesterification methods, Progress in Energy and Combustion Science 31 (2005) 466–487. [13] R.F. Probstein, R.E. Hicks, Synthetic Fuels, McGraw-Hill, New York, 1982. [14] P. Fagerlund, B. Ramne, EffShip Project Summary and Conclusions, WP9 Final Reporting and Dissemination, 2013. [15] M. Yamada, High-quality transportation fuels, Energy & Fuels 17 (2003) 797–798. [16] T.A. Semelsberger, R.L. Borup, H.L. Greene, Dimethyl ether (DME) as an alternative fuel, Journal of Power Sources 156 (2006) 497–511. [17] M.G. Babu, K. Subramanian, Alternative Transportation Fuels: Utilisation in Combustion Engines, CRC Press, 2013.

82


[18] S. LeViness, A.L. Tonkovich, K. Jarosch, S. Fitzgerald, B. Yang, J. McDaniel, Improved Fischer–Ropsch Economics Enabled by Microchannel Technology, Velocys, 2011. (Online (http://www.oxfordcatalysts.com/press/wp/wp110224_microchannel_FT_ white_paper_24Feb11.pdf) viewed 20 October 2013). [19] L. Tonkovich, T. Mazanec, K. Jarosch, S. Fitzgerald, B. Yang, R. Taha, D. Kilanowski, J. Lerou, J. McDaniel, T. Dritz, Improved Fischer–Tropsch Economics Enabled by Microchannel TechnologyOnline http://www.adktroutguide.com/files/Microchannel_ FT_White_Paper_Sep08.pdf (viewed 12 November 2013). [20] J. Hu, Y. Wang, C. Cao, D.C. Elliott, D.J. Stevens, J.F. White, Conversion of biomass syngas to DME using a microchannel reactor, Industrial & Engineering Chemistry Research 44 (2005) 1722–1727. [21] H. Baumann, A.M. Tillman, The Hitch Hiker's Guide to LCA. An Orientation in Life Cycle Assessment Methodology and Application, External organization, 2004. [22] A.M. Tillman, Significance of decision-making for LCA methodology, Environmental Impact Assessment Review 20 (2000) 113–123. [23] EU, Directive 2009/28/EC of the European Parliament and of the Council of 23 April 2009 on the promotion of the use of energy from renewable sources and amending and subsequently repealing Directives 2001/77/EC and 2003/30/EC. COD(2008) 0016, 2009. [24] EU, Directive 2009/30/EC of the European Parliament and of the Council 23 April 2009 amending Directive 98/70/EC as regards the specification of petrol, diesel and gas-oil and introducing a mechanism to monitor and reduce greenhouse gas emissions and amending Council Directive 1999/32/EC as regards the specification of fuel used by inland waterway vessels and repealing Directive 93/12/EEC, 2009. [25] ISO, Environmental Management—Life Cycle Assessment—Principles and Framework (ISO 14040:2006), European Committee for Standardization, Brussels, Belgium, 2006. (I). [26] ISO, Environmental Management—Life Cycle Assessment—Requirements and Guidelines (ISO 14044:2006), European Committee for Standardization, Brussels, Belgium, 2006. [27] A. Zamagni, J. Guinée, R. Heijungs, P. Masoni, A. Raggi, Lights and shadows in consequential LCA, The International Journal of Life Cycle Assessment 17 (2012) 904–918. [28] P. Artaxo, T. Berntsen, R. Betts, D.W. Fahey, J. Haywood, J. Lean, D.C. Lowe, G. Myhre, J. Nganga, R. Prinn, G. Raga, M. Schulz, R. Van Dorland, Changes in atmospheric constituents and in radiative forcing, Climate Change 2007: The Physical Science Basis, in: IPCC Fourth Assessment Report: Climate Change, 2007. [29] F. Cherubini, A.H. Stromman, Life cycle assessment of bioenergy systems: state of the art and future challenges, Bioresource Technology 102 (2011) 437–451. [30] L. Luo, E. van der Voet, G. Huppes, H.A.U. de Haes, Allocation issues in LCA methodology: a case study of corn stover-based fuel ethanol, The International Journal of Life Cycle Assessment 14 (2009) 529–539. [31] J. Benjaminsson, R. Nilsson, Distributionsformer för biogas och naturgas i Sverige, Svenska Gasföreningen, 2009. [32] L. Göthe, bGöthe_2013 _Metanutsläpp CBG SGC.pdf N in: Metanutsläpp i den svenska fordonsgaskedjan—En nulägesanalys Svenskt Gastekniskt Center AB, Sweden, 2013. [33] M. Broses, S. Sannes, Compressed Natural Gas—Compressed Biogas. An Overview of the Technology, Economics, Short Elliott Hendrickson, 2012. [34] C.N. Satterfield, Heterogeneous Catalysis in Industrial Practice, Krieger Publishing Company, 1996.

[35] R. Edwards, J. Larivé, V. Mahieu, P. Rouveirolles, Well-to-Wheels Analysis of Future Automotive Fuels and Powertrains in the European Context, Version 2b, 2006. [36] J.J. Marano, J.P. Ciferno, Life-cycle greenhouse-gas emissions inventory for Fischer– Tropsch fuels, Report Prepared for the US Department of Energy, Energy and Environmental Solution, LLC, Gaithersburg, MD, USA, 2001. [37] Å. Furnander, Life cycle assessment of dimethyl ether as a motor fuel, Institutionen för transportteknik Chalmers tekniska högskola, 1996. [38] G. Ertl, H. Knözinger, J. Weitkamp, Handbook of Heterogeneous Catalysis, VCH, 1997. [39] J. Lundgren, T. Ekbom, C. Hulteberg, M. Larsson, C.-E. Grip, L. Nilsson, P. Tunå, Methanol production from steel-work off-gases and biomass based synthesis gas, Applied Energy 112 (2013) 431–439. [40] J.B. Hansen, S.-E. Mikkelsen, DME as a Transportation Fuel, Prepared for the Danish Road Safety and Transport Agency and the Danish Environmental Protection Agency, Lyngby, Denmark, 2001. [41] M. Kytö, P. Rantanen, N.-O. Nylund, Gas bus technology and operational experiences in helsinki area0nline www.balticbiogasbus.eu/web/Upload/Use_of_biogas/Act_6_ 1/WP6%201%20FINAL.pdf2012 (viewed 28 march 2014). [42] United States, Federal Transit Administration, ARCADIS Geraghty & Miller, Inc, Transit Cooperative Research Program, Transit Development Corporation, National Research Council (U.S.). Transportation Research Board, Guideline for Evaluating Selecting Implementing Fuel Choices for Transit Bus Operations, National Academy Press, 1998, ISBN 9780309062732. [43] N.-O. Nylund, K. Koponen, Fuel and Technology Alternatives for Buses: Overall Energy Efficiency and Emission Performance, VTT publication, Finland, 2012. [44] P.E. Meyer, E.H. Green, J.J. Corbett, C. Mas, J.J. Winebrake, Total fuel-cycle analysis of heavy-duty vehicles using biofuels and natural gas-based alternative fuels, Journal of the Air & Waste Management Association 61 (2011) 285–294. [45] J.R. Benjamin, Probability, Statistics, and Decision for Civil Engineers, McGraw-Hill, New York, 1970. [46] P. Tunå, C. Hulteberg, J. Hansson, A. Åsblad, E. Andersson, Synergies from combined pulp and paper and fuel, production, Biomass and Bioenergy 40 (2012) 174–180. [47] E. Andersson, S. Harvey, T. Berntsson, Energy efficient upgrading of biofuel integrated with a pulp mill, Energy 31 (2006) 1384–1394. [48] G. Finnveden, M.Z. Hauschild, T. Ekvall, J. Guinée, R. Heijungs, S. Hellweg, A. Koehler, D. Pennington, S. Suh, Recent developments in life cycle assessment, Journal of Environmental Management 91 (2009) 1–21. [49] Ecoinvent, Swiss Centre for Life Cycle Inventories, 2006. [50] P. Börjesson, M. Berglund, Environmental systems analysis of biogas systems—part I: fuel-cycle emissions, Biomass and Bioenergy 30 (2006) 469–485. [51] P. Börjesson, M. Berglund, Environmental systems analysis of biogas systems—part II: the environmental impact of replacing various reference systems, Biomass and Bioenergy 31 (2007) 326–344. [52] M. Pöschl, Evaluation of energy efficiency of various biogas production and utilization pathways, Applied Energy 87 (2010) 3305–3321. [53] A. Pertl, P. Mostbauer, G. Obersteiner, Climate balance of biogas upgrading systems, Waste Management 30 (2010) 92–99. [54] T. Patterson, S. Esteves, R. Dinsdale, A. Guwy, Life cycle assessment of biogas infrastructure options on a regional scale, Bioresource Technology 102 (2011) 7313–7323.