Plan Lievense Recalculated

‘Plan Lievense’ was developed in 1981 to overcome the problem of fluctuations with electricity produced by wind power. The basic setup of plan Lievense is displayed in figure 1 where windmills turn wind energy into electrical energy. If there is overcapacity on the electricity network, the energy is stored in a water basin, which is a pumped hydro storage (PHS). This storage can be used when there is a shortage in energy. The basin would have preferably been placed in the Markermeer.

Figure 1: A schematic set-up of plan Lievense (images courtesy of TU Delft.)

It eventually didn’t set through because of several problems, mostly financial and ecological. When I was doing some pre-research about this question, I looked at the DOE Global Energy Storage Database (energystorageexchange.org) for large storage systems in the Netherlands. No system larger than 10MW was found, whereas countries like the UK and Canada have larger storages to compensate for sustainable energy fluctuations. A big storage solution like Plan Lievense might support a transition to a bigger sustainable energy share in the Netherlands.

That is why I chose this question of Ewoud. Given losses in energy transfer from pumped hydro storages (PHS’s) in Norway to the energy consumer in the Netherlands, it might be attractive storing energy in a closer water reservoir. On top of that, a plan like Lievense’s would reduce the reliance on other countries on stored energy. This leads to the question:

1. Question:

If today’s fluctuations in electricity produced by wind energy have to be handled by a Plan Lievense equivalent in the Markermeer, how much height difference has to be there to handle this? And would this be achievable?

2. Final answer

Given a volume of discharge of 40m3, a height difference of 638m for a 535 km2 basin is needed to significantly have power capacity to catch up all wind energy fluctuations. This basin would store about 1500 times more energy than the original systems of Plan Lievense. Given the unviability of Plan Lievense and the risks of building a 638m basin near the Netherlands, this question is not realistic at all.

3. Procedure used to obtain your answer

To start, I looked up wind directions and wind strength from the Trintelhaven, which is at the middle of the dike between the Markermeer and the IJsselmeer. Most wind comes from south-west, whereas some wind comes from the south. If wind mills were aligned with the south-west and south wind directions, the wind strengths these wind mills are facing are displayed in figure 2.

Wind distribution NL

Figure 2: The wind strength in Beaufort scale, measured in Trintelhaven, the Netherlands. The horizontal axis displays the month of the year, whereas the vertical axis displays the amount of days that a certain wind strength occurs.

When taking into account a typical wind turbine power curve, as shown in figure 8, there is a cut-in wind speed of about 4 m/s (Lydia et al. 2014). That is the minimum wind speed to get the turbine going. To prevent the turbine from breaking down due to high wind speeds, a maximum cut-out speed (about 25 m/s) has been build in, where after the wind turbine stops. At about 17 m/s the limit of the electrical generator is reached, so this is the optimum wind speed. A wind speed of 17 m/s gives about ten times more wind power than a wind speed of 7 m/s.

Wind Power Curve

Figure 8: A typical wind turbine power curve that displayed amount of wind energy that flows in (in m/s), against the power generated (in kW). The curve is about similar for most rotor sizes, only the numbers are different. Since this graph is only about the curve and not about the numbers, input values like rotor sizes, density of air, etc. are not given. Details can be found in the article of Lydia et al. (2014).

When looking at the wind data of figure 6 and 7, it at first sight looks positive since there is always a little wind around to let the turbine run. But when taking figure 8 and the cut-in speed, rated speed and cut-out speed into account, it can be concluded that only wind speeds of about 10 m/s let wind turbines significantly contribute to the net, while a wind speed of more than 10 m/s occurs about 15% of the time (a bit more in areas like the Markermeer). When that wind speed peaks occurs during times that little electrical energy is demanded, storage may be required.

Wind power in the Netherlands has an amount of 2.85GW, enough for 4.98% of the total percentage of produced energy (Reijerman, 2014). Since Dutch windmills can reach 0 as well and sometime be full operational, it may be assumed that on a peak 2.5GW needs to be stored.

The approximate proposed Markerwaard area is depicted in figure 9 and has a circumference of 100 kilometers (Berny, 2013). The area is about 534.62 km² (Daftlogic, Google Maps Area Calculator Tool).

Assuming a volume of discharge of 40m3/s (Montero, 2015)

P=(m*g*h)/t=2.5GW
t=(A*h)/discharge-speed
t=(534.62*h)/40

(9.81*5.3462E11*h*h)/t=2.5GW
(9.81*5.3462E11*h*h)/((5.3462E8*h)/40)=2.5GW

(9.81*5.3462E11*x*x)/((5.3462E8*x)/40)=2.5E9W

h = 637.5m

Little is known about the volume of discharge for large PHS projects. Although it can be seen in figure 10 that hydro system costs largely depend on the amount of power capacity. Table 4 shows several calculated volumes of discharge and their needed heights. Plan Lievense originally needed 23-meter height difference (LievenseCSO Infra B.V, 1982).

Volume of discharge (m3/s) Height (m)
40 637.5
1000 25
10000 2.5

Table 4: The volume of discharge needed compared to the height needed for delivering enough power.

Figure 10: The costs of a hydro system to build per kW (RenewablesFirst) 

The reason why the height difference of above calculations compared to the height of Plan Lievense is so high, is because the philosophy of Plan Lievense is different than the plan described in this question. This question calculates the possibility to store all peak windmill energy in a single basin in the Markermeer. Whereas, Plan Lievense’s philosophy was to build a decentralized, integrated windmill-storage system. Part of the Markermeer was only required and a height difference of 23 meter was enough. So since Plan Lievense was already not viable, a 1500 times larger project, as proposed in this question, with a height difference of 628m, would not be viable at all.

4. Parameters used and describe for each parameter how it was found

  • Frequency distribution of yearly wind speed in m/s in the Netherlands: LievenseCSO Infra B.V.
  • Typical windturbine power curve: Lydia et al. (2014)
  • Wind power share in Netherlands: 2.85GW (Reijerman, 2014)
  • Markerwaard area: 100 kilometer circumference (Berny, 2013), 534.62 km² (Daftlogic, Google Maps Area Calculator Tool)
  • Volume of discharge of 40m3/s (Montero, 2015)
  • Plan Lievense original height difference: 23 meter (LievenseCSO Infra B.V, 1982)
  • Plan Lievense original power delivery: 1600 MW (LievenseCSO Infra B.V., 1982)

5. Discussion, including results from other sources or other estimates

As already discussed, realizing a Lievense equivalent of 2.5 GW to capture wind energy fluctuations would result in an unrealistically large basin and could do much harm if it breaks. Not even forgetting, the immense visual barrier it would take and the ecological effect it would have.

Even if a height difference of several (>10) meters is realized, it is a risky project. Even then, a breakthrough of dikes could lead to a tsunami. An underground PHS could be a solution for preventing this. More research to underground PHS systems may be necessary.

A next question may compare a Lievense’s storage to a high voltage DC-connected PHS Norway storage and an underground PHS and give the best source or mix of sources for coping with energy fluctuations.

Another option is to store energy in batteries, which is already compared in the previous question. An example of a recent planned battery storage facility is the Leighton Buzzard battery facility. It is a 6MW battery storage with a cost of 18.2 million pounds (BBC, 2014). If you would like to upscale that to the size of the original Plan Lievense (1600MW), it would cost 4.8 billion pounds. Although, there has never been build such a large battery storage.

The approximated area of the Markermeer could differ a bit, because the area was received by pointing out the area in an online tool (Daftlogic) manually. The used area is depicted in figure 7 and could differ a few percentages in reality.

6. References and/or sources.

Berny, Plan Lievense, 2013, retrieved on 25 January 2016

Daftlogic, Google Maps Area Calculator Tool, retrieved on 25 January 2016 from: https://www.daftlogic.com/projects-google-maps-area-calculator-tool.htm .

Herodotus, By Io Herodotus – Own work, CC BY-SA 4.0, retrieved on 25 January 2016.

Lydia, M., Kumar, S. S., Selvakumar, A. I., & Kumar, G. E. P. (2014). A comprehensive review on wind turbine power curve modeling techniques. Renewable and Sustainable Energy Reviews, 30, 452-460.

Montero, Niemann & Wortberg, UNDERGROUND PUMPED-STORAGE HYDROELECTRICITY USING EXISTING COAL MINING INFRASTRUCTURE, 2015, E-proceedings of the 36th IAHR World Congress, The Hague, the Netherlands – congress / semi-scientific source

Reijerman, CBS: windenergie voor de eerste maal belangrijkste bron voor hernieuwbare stroom, 2014, retrieved on 25 January 2016 from: http://tweakers.net/nieuws/101544/cbs-windenergie-voor-de-eerste-maal-belangrijkste-bron-voor-hernieuwbare-stroom.html

RenewablesFirst, How much does a hydropower system cost to build?, retrieved on 26 January 2016 from: https://www.renewablesfirst.co.uk/hydropower/hydropower-learning-centre/how-much-do-hydropower-systems-cost-to-build/

Wassink, ‘Plan Lievense eigenlijk bedoeld voor kernenergie’, 2008, retrieved on 26 January 2016 from: http://delta.tudelft.nl/artikel/-plan-lievense-eigenlijk-bedoeld-voor-kernenergie/17639

Wisuki, Trintelhaven, 2012, retrieved on 22 February 2016, http://nl.wisuki.com/statistics/2260/trintelhaven?wi_d9=1&wi_d10=1&wi_d11=1&wi_d12=1&a_wi=4&wi_m=0&temp=monthly&rain=quantity

7. Reliability

If possible, scientific sources had priority as a data source. Although, since little is published about plan Lievense or equivalents, mostly societal sources were used to find tentative information.

Amsterdam CNG Waste Truck Biogas Plan Hypothetical

(Related to Chapter 20 – better transport) – In several cities Canadian waste trucks are powered by purified and compressed biogas (compressed natural gas = CNG) that is created from organic waste collected by the truck [6]. In this way a “circle” is created (a waste truck can drive on the gas that is created from the waste it collects). This circle could possibly also be implemented in the Netherlands, although Dutch households do not separate biodegradable waste very well. in 2013 on average 38% of the Dutch solid (household)waste still consisted of biodegradable waste [7]. If this amount was separated by households, lots of energy could be won. Leading to the question:

1. Question: How many garbage collection trucks could drive on biogas, derived from separated biodegradable solid household waste in the city of Amsterdam, and if there is a surplus in fuel, how many additional personal cars can drive on the biogas?

2. Final answer: Unsorted solid waste still contains 38% biodegradable waste. For solely the city of Amsterdam about 110 waste trucks could drive on this waste, enough for collecting waste for the whole city, plus for the driving of about 1,000 typical cars.

3. Procedure used to obtain your answer

In 2013 the city of Amsterdam generated about 423 kg waste per citizen per year [8]. In 2013 Amsterdam had 799,442 citizens [9]. In 2013 38% of the solid household waste in the Netherlands consisted of biodegradable waste, in Dutch called ‘gft-afval’ [7]. The percentage of biodegradable waste is not known in Amsterdam’s waste. Given the small amount of biodegradable waste collection bins in Amsterdam, it is assumed that even less citizens of Amsterdam separate their biodegradable waste. That is why the assumed percentage of solid household waste that consists of biodegradable waste is about 50% for Amsterdam. The city of Amsterdam does currently not separate biodegradable waste from the solid house hold waste.

That leads to an amount of organic biodegradable waste Amsterdam of 0.5 * 0.423 * 799442 = 169,081 metric tons per year.

Basic research on biogas from biodegradable household solid waste

When we look at the organic waste processing facility of Dutch company Orgaworld in Lelystad, this generates 100 m3 biogas per metric ton of organic waste. Another research confirms that for a current plant a tonne of biogas gives at minimum about 100 m3 biogas per metric ton organic waste [10].

Biogas is a mixture of 55%-60% methane (CH4), 40%-45% carbon dioxide (CO2) and some small amounts of other gases [10]. A compressed natural gas (CNG) fuel requires at least 89.14% methane [11]. Upgrading to vehicle fuel biomethane (BioCNG) and pressure standards (3600 PSI), the compression costs are 3% of the energy content of the upgraded biogas, as calculated by a Canadian biogas engineering firm, co-funded by the province of British-Columbia [12]. It is assumed that all 60% methane is filtered when upgrading the biogas to an 89.14% BioCNG fuel. So when upgrading 100m3 biogas, 67m3 (=60/0.8914) BioCNG is received.

The production cost of the BioCNG would break down to Canadian Dollar (CAD) 7.72/GJ for the biogas and CAD 6.76/GJ for the upgrading [12]. Although these are Canadian values it gives a sense that the upgrading costs from biogas to biomethane are substantial. A reason for upgrading is that you have a higher energetic gas that you can store for vehicles or deliver to the gas network. All gas volumes in this assignment are given in normal cubic meter (so m3 = Nm3).

To cross-reference these results with existing and planned plants, there are several biogas installations in the Netherlands, as described in table 1.

 

Installation Biodegradable waste /year Energy output/ year
Orgaworld Lelystad Biocel [13] 30,000 tons 3 million cubic meters biogas (4.2 GWh)
Orgaworld Surrey (planned), Canada [14] 80,000 tons plus 25,000 tons industrial waste 3 million cubic meters biogas
Uppsala Biogas, Sweden [15] 23,000 tons food waste plus 4,000 tons slaughter waste 4.6 million cubic meters biogas

Table 1: Several biogas project in and outside the Netherlands for giving a feeling about input waste capacities and output energy capacity.. Biogas in this table means a gas that consists of more than 55% methane. It is noteworthy that the Uppsala Biogas has the highest output yield. They are using a different anaerobic digestion method, namely a wet digestion method, whereas the Orgaworld biogas installations use a dry digestion method. The investment data of the plant of Orgaworld were obtained via Business in Vancouver and were CAD 68 million [16] and confirmed by a spokesmen of Orgaworld. Although the newer plants have similar biogas outputs, the newer plants have a larger compost and rest sources output. These residuals like compost are not taken into account since they are not relevant for energy purposes in this calculation.

Amsterdam’s biodegradable-waste-created CNG leads to a ceiled biomethane output of about 10 million m3 BioCNG, including a 80% well-to-pump [17] efficiency (152,173 metric ton biowaste * 67 m3 BioCNG/metric ton biowaste * 0.8 * 0.97 is in the order of magnitude of 10 million m3).

Madrid uses a total of 445 waste trucks to collect their solid household [18]. These 445 Iveco CNG refuse/waste trucks are fuelled by 10 million m3 BioCNG [19].

Amsterdam has no numbers about the amount of waste vehicles that it is using and what routes they are driving. That is why the amount of trucks needed to collect Madrid’s garbage is scaled back to the size of Amsterdam (in terms of population). The population of Amsterdam is about 4 times smaller than Madrid [20] and therefore the city of Amsterdam would need about 110 garbage trucks to collect waste.

Based on 110 garbage truck, this would lead to extra 7.5 million m3 pump-available BioCNG to fuel regular vehicles.

As MacKay (2008) estimates, a typical car (50km/day) uses 40kWh/d = 14600 kWh/y. The caloric value of methane is about 38 MJ/m3 [21], the 90% methane content in CNG needs to be taken into account. Taking into account the CNG tank-to-wheel efficiency of about 20% [17], leads to a remainder potential of: 7.5 million m3 biomethane * 0.2 * 38 * 0.9 = 51.3 TJ = 14.25 GWh. That is enough for about 1,000 typical cars.

4. Parameters used and describe for each parameter how it was found

  • Related biogas plants for giving a feeling of normal energy outputs: [13], [14] and [15]
  • Amsterdam amount of waste per citizen in 2013: 423 kg per year [8]
  • Amsterdam population in 2013: 799,442 citizens [9]
  • Percentage of biodegradable waste still found in Dutch rest waste: 38% [7]
  • Metric tonne biodegradable waste, biogas yield: minimum 100 m3 per metric ton waste [10] and [13]
  • CNG well-to-wheel efficiency: about 15% (Curran et al., 2014)
  • CNG well-to-pump efficiency: about 80% (Curran et al., 2014)
  • CNG tank-to-wheel efficiency: about 20% (Curran et al., 2014)
  • CNG methane (CH4) percentage required for CNG driving: 89.14% [11]
  • CNG purification energy and financial cost: 3% of the upgraded biogas energy content and $0.58/GJ [12].
  • amount of Iveco waste trucks than can drive on 10 million m3 CNG: 445 [18]
  • Madrid population ratio compared to Amsterdam: 4 times larger [20]
  • Energy usage of an average person wagon: 40 kWh/day (MacKay, 2008)
  • Production cost of BioCNG in Canada: Canadian Dollar (CAD) 7.72/GJ for the biogas and CAD 6.76/GJ for the upgrading. Although it are no Dutch values, it gives a feeling for the percentage of cost for upgrading biogas to BioCNG [12].

5. Discussion, including results from other sources or other estimates

Amsterdam is used in this question, since it is a city with a big waste truck fleet and a city that wants to reduce CO2 emissions by 40% by 2025, compared to 1990 [22]. Biogas through anaerobic digestion has been evaluated as one of the most energy-efficient and environmentally beneficial technology for bioenergy production [23] and will help achieving these reduction goals. It is not known what the exact biodegradable waste percentage is in Amsterdam, that is why the Dutch percentage of biodegradable waste is taken into account. It is assumed that this amount is a bit higher for Amsterdam households than for Dutch households, since Amsterdam has a smaller amount of biodegradable waste collection bins, than other cities that provide bins with built-in biodegradable waste compartments. Further, the amount of 100 m3 biogas with about 55% methane content per ton biodegradable waste as estimated by Orgaworld is a bit lower than values of several crops [24] (e.g. wheat grain can generate a lot more than 100 m3/t). Although, because of the blend of several crops and less energetic waste, the estimated amount is a bit lower, so an estimate of 100m3 biogas per ton waste (Orgaworld, n.d.) seems reasonable. Plus, by using the minimum possible yield, the risk of overestimating yield numbers is diminished.

Cost aspect: Although there seems to be opportunity for converting biodegradable waste to biomethane, it has high investment costs. Another barrier is the low natural gas and oil price (RVO, Hugo Schotman, 2015), although these prices are higher in the Netherlands than in Canada where a waste loop concept is already viable. Subsidies may be necessary though and for example a CO2-tax on regular gas may be an incentive for producing biomethane (Schotman, 2015). Given the environmentally friendliness of biomethane and the CO2-goals cities like Amsterdam have, such a tax may be an interesting incentive. Another ‘cost’-aspect is the aspect of pickup. This may lead to a diversification of the waste fleet, requiring investments in different waste trucks, since separated biodegradable waste has to be collected separately from the normal solid waste.

In order to make sure if the case is possible and economically viable I asked a spokeswoman of DMT (a biogas installation company) if the case of Amsterdam as described above is possible. Her response can be found in Appendix A and she confirms that the case might be possible. Although, in a biogas installation other sources as sewage sludge may be needed to have enough energy for creating enough biogas to fuel the waste truck fleet. There are already other municipalities in the Netherlands where waste trucks drive on biomethane received from biodegradable waste [25].

Amsterdam has no numbers about the amount of waste vehicles that it is using and what routes they are travelling. Whereas Madrid concrete describes how much biomethane is needed to power its entire waste truck fleet, as given by the Natural & Bio Gas Vehicle Association Europe (2009). That is why these numbers are scaled to the size of Amsterdam. Feature research might take concrete numbers of Amsterdam into account, an attempt to contact the statistics department of Amsterdam has been done, but no response with relevant numbers has been received.

Finally, MacKay (2008) estimates that a typical car driver uses 40 kWh/day. He does not distinguish between diesel and petrol, because they have almost similar calorific values (about 10% difference). The calorific values of methane compared to petrol per kg have also about 10% difference. Because the final remainder of cars is rounded to 1000, a 10% differences has no impact on the remainder cars the could drive on the biofuel.

6. References

[6] Canadian Biogas Association, “Closing the loop,” June 2015. [Online]. Available: http://www.biogasassociation.ca/bioExp/images/uploads/documents/2015/close_the_loop/RNG_Close_the_Loop_Primer_Web.pdf“`. [Accessed 11 February 2016].
[7] Rijkswaterstaat, “Samenstelling van het huishoudelijk restafval, sorteeranalyses 2014 – Gemiddelde driejaarlijkse samenstelling 2013,” November 2015. [Online]. Available: http://www.rwsleefomgeving.nl/onderwerpen/afval/publicaties/downloads/samenstelling-4/. [Accessed 11 February 2016].
[8] Centraal Bureau voor de Statistiek, “Huishoudelijk afval per gemeente per inwoner,” 12 February 2016. [Online]. Available: http://www.cbs.nl/nl-NL/menu/themas/natuur-milieu/cijfers/afval/gemeentelijk-afval-per-inwoner-sl.htm. [Accessed 12 February 2016].
[9] Gemeente Amsterdam, “Amsterdam in cijfers 2015,” 2015. [Online]. Available: https://www.ois.amsterdam.nl/assets/pdfs/2015%20jaarboek%20amsterdam%20in%20cijfers.pdf. [Accessed 11 February 2016].
[10] Canadian Biogas Association, “Farm to Fuel Developers’ Guide to Biomethane,” 2012. [Online]. Available: http://biogasassociation.ca/bioExp/images/uploads/documents/2012/reports/Developers_Guide_to_Biomethane.pdf. [Accessed 11 January 2016].
[11] K. Subramaniana, V. C. Mathada, V. Vijayb and P. Subbaraoc, “Comparative evaluation of emission and fuel economy of an automotive spark ignition vehicle fuelled with methane enriched biogas and CNG using chassis dynamometer,” Applied Energy, pp. 17-29, 2013.
[12] Electrigaz Technologies Inc., “Feasibility Study – Biogas upgrading and grid injection in the Fraser Valley, British Columbia,” 2008. [Online]. Available: http://www.lifesciencesbc.ca/files/PDF/feasibility_study_biogas.pdf. [Accessed 12 February 2016].
[13] Orgaworld, “Lelystad: Biocel anaerobic digestion and composting plant,” [Online]. Available: http://www.orgaworld.com/more-about-our-business/our-locations/lelystad-biocel. [Accessed 11 February 2016].
[14] P. Oostelbos, in Biogas Association Canada – Closing the Loop, Abbotsford, 2016.
[15] G. Rogstrand, “Municipal Biogas Production in Sweden,” in Canadian Biogas Association – Closing the loop, Abbotsford, 2016.
[16] A. Reid, “Surrey breaks ground on biofuel processing facility,” 2 March 2015. [Online]. Available: https://www.biv.com/article/2015/3/surrey-breaks-ground-biofuel-processing-facility/. [Accessed 13 February 2016].
[17] S. J. Curran, R. M. Wagner, R. L. Graves, M. Keller and J. B. J. Green, “Well-to-wheel analysis of direct and indirect use of natural gas in passenger vehicles,” Energy, pp. 194-203, 2014.
[18] NGVA Europe, “Case Study. February 2009 1 CNG trucks in urban garbage collection. The successful case of the FCC fleet in Madrid,” February 2009. [Online]. Available: http://www.marcogaz.org/egatec2011/PS2/PS2I_Lage_egatec2011.pdf. [Accessed 16 February 2016].
[19] NGVA Europe, “Madrid’s fleet operators heavily rely on heavy-duty NGVs,” 27 March 2014. [Online]. Available: http://www.ngva.eu/madrids-fleet-operators-heavily-rely-on-heavy-duty-ngvs.
[20] United Nations Statistics Division, “City population by sex, city and city type,” 2014. [Online]. Available: http://data.un.org/Data.aspx?d=POP&f=tableCode%3A240#POP. [Accessed 13 February 2016].
[21] M. A. Laughton and D. Warne, “Electrical Engineer’s Reference Book,” in Electrical Engineer’s Reference Book, Oxford, Elsevier Science, 2003, p. 27/14.
[22] Gemeente Amsterdam, “Amsterdam: a different energy – 2040 Energy Strategy,” February 2010. [Online]. Available: https://www.amsterdam.nl/gemeente/organisatie/ruimte-economie/ruimte-duurzaamheid/ruimte-duurzaamheid/making-amsterdam/publications/sustainability-0/2040_energy_strategy_0/. [Accessed 14 February 2016].
[23] H. Fehrenbach, J. Giegrich, G. Reinhardt, U. Sayer, M. Gretz, K. Lanje and J. Schmitz, “Kriterien einer nachhaltigen Bioenergienutzung im globalen Maßstab,” UBA-Forschungsbericht 206, p. 41 – 11, 2008.
[24] P. Weiland, “Biogas production: current state and perspectives,” Appl Microbiol Biotechnol, pp. 849-860, 2010.
[25] TankPro, “Vuilniswagens gaan rijden op bananenschillen,” 18 November 2011. [Online]. Available: http://www.tankpro.nl/brandstof/2011/11/18/vuilniswagens-haarlemmermeer-gaan-rijden-op-bananenschillen/. [Accessed 13 February 2016].
[26] U.S. Department of Energy, “Case Study – Compressed Natural Gas Refuse Fleets,” 2014. [Online]. Available: http://www.afdc.energy.gov/uploads/publication/casestudy_cng_refuse_feb2014.pdf. [Accessed 13 February 2016].
[27] D. J. MacKay, “Sustainable Energy – Without the hot air,” Cambridge, UIT, 2008, p. 29.

7. Reliability of sources

This question involves lots of statistical data, scientific data, and other numbers. For statistical data, official statistical data was attempted to use as much as possible. These official sources include data from municipalities, national statistical data, and data from more international sources, like the United Nations. Since the question is mostly on a municipality-level, the data of municipalities had priority. If no data was available, data on a higher level (national and international) was used. For biogas characteristics it was tried to use as much scientific sources with more than 10 citations and published in a related and in a renowned journal. If that was not possible, related biogas association numbers were used. And if that was not possible, related contacts were personally contacted to get the right numbers and the right vision on the feasibility of this case. Only in a small amount of cases, if these data were not found, other sources like news and company sites were used. These sites were only used for exemplary data and not for critical data like the well-to-wheel efficiency of a CNG vehicle.