Can offshore wind farms be profitable, once subsidies stop, by storing their generated electricity on existing oil and gas platforms?

(1)

1

Can offshore wind farms be profitable, once subsidies stop, by

storing their generated electricity on existing oil and gas

platforms?

MSc Thesis International Economics & Business

University of Groningen – Faculty of Economics and Business

Abstract

Many wind farms will be installed in the coming years in the Dutch sector of the North Sea. They have a license to operate for 15 years, during which they will be subsidized. After this period the wind farms have to be decommissioned. This study investigates whether it is profitable to operate these wind farms after the initial 15 year period without subsidies. A major topic of this study is the potential of temporarily storage of the generated electricity on existing oil and gas platforms. This might be done by so called hydrogen storage. The calculations are based on a wind farm that will be installed in the Dutch part in the North Sea, called Gemini Windpark. The electricity prices vary depending on demand. Germany has already more experience with wind generated electricity and their price history is used for estimating prices in 2031 for the Netherlands.

Key words: Offshore wind energy, Electricity prices, Hydrogen storage, Subsidy

Author: H.C. Pipping Student ID number: 1885022

Date: January 28, 2015

(2)

(3)

3

1. Introduction

The availability of electricity is taken for granted by the public. The majority of the electricity is generated by fossil fuels such as natural gas, oil and coal but is not forever available and cause also high emissions of CO2. Wind is a renewable energy source and can continuously be used to produce power. Wind power, moreover, has the lowest CO2 emissions of all

methods of generating electricity. There are many initiatives to reduce global warming and more use of wind energy will help (Bundesverband WindEnergie, 2014).

It is well known that electricity generated by wind farms offshore is expensive and their costs can only be recovered when they are subsidized. Current wind farm owners in the Netherlands receive a subsidy called “Stimulering Duurzame Energieproductie” (SDE+) / “Encouraging Sustainable Energy Production” for 15 years. This subsidy is to compensate for the difference between the cost price of e.g. electricity generated by wind (green power) and the yearly average price of conventionally generated electricity (grey power). After 15 years the license to produce wind on the North Sea expires and so does the SDE+ subsidy. As a result of the license expiry, the wind farm owner has to remove the whole wind farm installation, so called decommissioning. It is technically possible to operate the wind farm after the initial 15 years. One could ask the question, why would you remove good working material? An answer could be that without the SDE+ subsidy the wind farm is not profitable.

The profitability of wind farms could be larger if the electricity can be stored, in one way or another, when electricity prices are low and electricity can be re-generated and sold when prices are high. This requires that the electricity can be stored. Hydrogen storage on existing oil or gas platforms is a possibility. When the electricity prices are low an electrolyzer can use the electricity to split water into hydrogen and oxygen. This hydrogen can be stored until the electricity prices are high. When the electricity prices are high enough the hydrogen will be converted back into electricity through fuel cells and will be sold.

This study investigates whether electricity can be economically stored and if that generates enough extra revenues to run offshore wind farms without subsidies. This study will not deal with operating costs of the existing infrastructure.

(6)

6

Groningen. This wind farm will start generating electricity in 2016 and has to be decommissioned in 2031 (Gemini Windpark, 2014).

(7)

7

2. Problem statement and research questions

The motivation of this research is based on increasing penetration of renewable energy sources and dismantling of good working windmills, in particular, in the North Sea. As a result of increasing supply of renewable energy and inelasticity of the electricity demand an imbalance occurs. The imbalance between supply and demand causes an oversupply of electricity and offers, from an economic perspective, an opportunity for arbitrage by means of storage (Veyer, 2014). Arbitrage is also called time-shifting supply because electricity is purchased at one time and sold at a later time when prices are high. Wouters (2014) took a look at electricity prices between three countries to find out when arbitrage can offset the costs of an electrolyzer. It turned out that electric energy storage, onshore, is not feasible yet.

The recent studies by Veyer (2014) and Wouters (2014) are used in this research for offshore wind farms in the North Sea. Due to the Fukushima nuclear plants disaster in 2011, Germany speeded up the changes in its energy supply system, the so called “Energiewende”, which meant high investments in renewable energy including offshore wind energy (Nederlandse ambassade, 2014). Offshore wind energy developed rapidly in Europe since the Energiewende, but also due to state subsidies and European Union objectives. In the Netherlands, tax incentives have strongly encouraged investments in wind turbines. Considerable more offshore wind capacity is projected to be installed in the North Sea area in the coming years. Once installed the installations only have permission to produce for 15 years. After these 15 years, the whole installation has to be decommissioned which is quite expensive. The decommissioning has also some negative effects on the environment. Instead of removing the offshore wind farm after 15 years, it will be investigated whether it is feasible for the offshore wind farm owners to continue after this period. A possibility might be to store the energy, in the form of hydrogen at an existing platform, during off-peak periods and selling at high-peak periods. The study will be based on the case of the Gemini Windpark, which has to be decommissioned in 2031 while the wind turbines have a lifetime which is much longer than 15 years (Gemini Windpark, 2014). This gives the following main research question:

(8)

8

This research will focus on the following five sub-questions:

- Why investigate the case of hydrogen storage?

- What are the marginal costs of an offshore wind farm?

- What is the difference in the development of renewable energy between Germany and the Netherlands?

- What is the value of green decommissioning?

- What is the context of the Gemini Windpark?

1 Why investigate the case of hydrogen storage?

This research focuses on storage of the generated electricity from offshore wind farms. As the wind farms are offshore it is interesting to found out if the storage can take place on an existing offshore platform. The storage of energy in the form of hydrogen is a cheap and efficient way of storage and has the ability to time-shift the energy supply to better meet the demand (Hedegaard and Meibom, 2011). If hydrogen storage is feasible it might be used by all wind farm owners, not only after 15 years when the license and subsidy stops, but even earlier. This method would support the use of sustainable electricity.

2 What are the marginal costs of an offshore wind farm?

(9)

9

3 What is the difference in the development of renewable energy between Germany and the Netherlands?

Germany is one of the leading countries in Europe in installed offshore wind parks, and are ahead of the Netherlands. The wind capacity that Germany currently has installed, is what the Netherlands wants to achieve in the near future. The amount of electricity generated in either Germany or the Netherlands, at a certain moment, cannot be accommodated in that country. It is also possible that prices in the neighboring country are higher. Arbitrage between Germany and the Netherlands is assumed to be reasonable because they are neighboring countries and bordering the North Sea.

4 What is the value of green decommissioning?

This research tries to find out to what extent decommissioning can be postponed or prevented by using the North Sea's oil and gas production infrastructure for storage by means of chemical conversion (of intermittent renewable capacity). Wind energy is an intermittent renewable energy source, because the supply of electricity generated by wind can fluctuate over time.

5 What is the context of the Gemini Windpark?

(10)

10

3. Literature review

The results of the literature investigated for this research are presented in this chapter. In the first section the difference in wind energy between Germany and the Netherlands will be discussed. Further on, section 3.2 explains the subsidy arrangements in the Netherlands. Then section 3.3 will review the implications of wind farm removal and the possibilities for green decommissioning. The last section, section 3.4, will expand on chemical conversion and hydrogen storage.

3.1 Renewable energy in Germany and the Netherlands

3.1.1 Energy situation in Germany

Germany is ahead of the Netherlands in terms of renewable energy sources. Increasing supply of electricity from renewable energy sources and the inelasticity of the demand, caused a balancing problem in Germany (Wouters, 2014). Today, Germany is the forerunner country in the field of renewable energy.

The electricity prices in Germany are compared with the electricity prices in the Netherlands for the year 2013. The year 2013 is chosen to be the current year in this research since the German electricity prices over 2014 are not available yet. This research assumes that the difference in electricity prices between the Netherlands and Germany in 2013 will be corrected within 18 years. This means that the Netherlands and Germany are at the same level in terms of developing renewable energy sources in the year 2031.

3.1.2 Energy situation in The Netherlands

(11)

11

Table 1 shows that in 2013 4.5% of the Dutch energy end use is derived from renewable energy sources. The consumption of energy from renewable sources is done for almost half of the time in the form of electricity. It mainly involves electricity produced by windmills, waste incinerators and biofuels for transport and consumption of wood by households (CBS, 2013). Wind energy penetration will increase in the Dutch energy market to realize a 16% energy consumption deriving from renewable energy by the year 2023. Due to this increase in renewable energy production, energy storage facilities are desirable to meet the future market needs (Auer et al., 2012). Table 1 shows the current Dutch wind energy capacity and the ambitions for 2023. The Dutch government has no specific goals yet for the year 2031 which is the year that the license expires for the Gemini Windpark and so the right to receive SDE+ subsidy.

Table 1 Dutch current and future wind energy capacity.

Onshore wind energy Offshore wind energy Total wind energy % wind energy of renewable energy sources % of energy consumption coming from renewable energy 2013 2,000 MW ±1,000 MW 3.000 MW 19.5% 4.5% 2023 6,000 MW (2020) 4,450 MW 10,450 MW 37% * 16% Increase 4,000 MW 3,450 MW 7,450 MW 17.5% 11.5%

Source: (Rijksoverheid, 2014),(Nijpels ,2014), (CBS, 2013) and (Draijer, 2013). *Estimated by ECN (Nijpels, 2014)

The existing and planned Dutch offshore wind capacity together add up to about 1000 MW electricity in 2013. This is 1/3rd of the total MW wind energy capacity in the Netherlands. The total 3000 MW wind energy, as given in Table 1, is 19% of the total renewable energy sources in the Netherlands. The estimated percentage wind energy of renewable energy for 2023 amounts up to 37%. This means that all offshore wind farms together have to produce 4450 MW electricity. As part of the Nederlandse Wind Energie Associatie (NWEA) private companies are working on 40% cost reduction of offshore wind generation, as agreed in the

(12)

12

3.1.3 Comparison of the energy situation between Germany and the Netherlands

Wind energy is the main and cheapest form of renewable energy in Germany. In 2013 Germany had 24,200 wind turbines installed, mainly onshore, with a capacity of around 34,000 MW wind energy. 9% of the electricity consumption in Germany comes from wind energy while renewable energy accounts for approximately 41.5% out of wind energy (Bundesverband WindEnergie, 2014). As presented in Table 2, 12% of the electricity consumption in Germany is derived from renewable energy in 2013 which is 7.5% lower than the 4.5 %, in Table 1, in the Netherlands in the same year.

Table 2 German current and future wind energy capacity.

Onshore wind energy Offshore wind energy Total wind energy % wind energy of green power % of electricity consumption coming from renewable energy 2003 15,000 MW - 15,000 MW Unknown Unknown 2013 33,480 MW 520 MW 34,000 MW ±41.5% 12% Increase 18,450 MW 520 MW 19,000 MW Unknown Unknown

Source: (Bundesverband WindEnergie, 2014), (Nederlandse ambassade, 2014)

While onshore wind energy is well developed in Germany, the share in offshore wind energy is little. In 2009 the first offshore wind farm was realized and in 2013 the offshore wind capacity amounted up to 520 MW provided by 116 turbines (see Table 3; Corbetta et al., 2014). A capacity of 520 MW offshore wind makes Germany the fourth largest offshore wind producer of Europe after the UK, Denmark and Belgium. The Netherlands follows as the fifth largest offshore wind producer of Europe with an installed offshore wind capacity of 247 MW (Corbetta et al., 2014). The 4 farms in the Netherlands, stated in Table 3, consist of 2 farms that are installed with a total capacity of 247 MW and 2 farms that are under construction and will have a power capacity of around 750 MW together (Corbetta et al., 2014).

Table 3 Installed offshore wind capacity in Germany and the Netherlands in 2013

Country Germany The Netherlands

No. of farms 13 4

No. of turbines 116 124

Capacity installed (MW) 520 247

(13)

13

Following the nuclear disaster in Fukushima in 2011 Germany increased the transformation of the energy supply system, the so called Energiewende. As a result, Germany became the forerunner country in renewable energy, both in the production and in the development of new technologies (Nederlandse ambassade, 2014). The Energiewende resulted in an investment in wind power capacity production (Veyer, 2014). The high amount of investments in renewable energy caused balancing problems in Germany between increasing supply of power and inelastic power demand. Since Germany has great experience with rapidly growing renewable resources, Germany is more advanced in the development of power storage through electrolysis than the Netherlands. Section 3.4 will discuss power storage in greater detail.

The German electricity price declining trend continues after 2012 resulting in a 6 to 9 €/MWh lower German wholesale prices compared to the Netherlands (ECN, 2013). The main reasons why the electricity price differs between the Netherlands and Germany are, state subsidies by the German government and more renewable energy sources in Germany (Bessenyei, 2014).

(14)

14

Through connecting wind farms in the North Sea with each other, and directly connect them to the power grids of different countries cheaper energy can be realized for European consumers and the costs of wind energy can decline dramatically. This would make it possible for Dutch wind farm owners to not only shift power supply over time but also between countries. For example, if the electricity prices are high in Germany and low in the Netherlands, electricity could flow from the Netherlands to Germany through the connected power grid (ECN, 2013).

3.2 Subsidies in the Netherlands

The Stimulering Duurzame Energieproductie (SDE+) / Encouraging Sustainable Energy Production arrangement is an operating subsidy in the Netherlands. The renewable energy cost price is higher compared to energy resulting from fossil fuel, which makes renewable energy production not always cost-effective (RVO, 2014). The SDE+ compensates for the difference between the cost price of renewable electricity and the yearly average conventional electricity price. This compensation is intended to offset the so called unprofitable component of the investment costs. The producers receive this subsidy for 15 years, depending on the technology they use. The government only pays the SDE+ subsidy when the wind farm actually generates electricity and really supplies to the national grid (RVO, 2014). The height of the SDE+ subsidy is determined annually by the Minister of Economic Affairs (NWEA, 2014). To receive a SDE+ subsidy, the wind farm owner has to win the tender offer pronounced by the government. The government now chooses a spot in the North Sea where wind energy may be generated and pronounces a tender offer for this spot, which is different than in the past.

(15)

15

government taxes increase to pay the SDE+ subsidy. The inhabitants still pay the same amount of money for their electricity.

3.3 Green decommissioning

The offshore wind farm owner gets permission to produce electricity for a maximum of 15 years and receives SDE+ subsidy for that period; see paragraph 3.2. After these 15 years normally the wind farm owner has to remove the whole installation and leave the area as they have found before. The costs for this decommissioning differ per wind farm, but dismantling costs are often around 2 to 3% of the total costs (Climate Change Capital, 2004). In the case, for instance, of the Gemini Windpark the bank guarantee for decommissioning is €40,000,000 (Gemini Windpark, 2014).

Decommissioning can have a negative impact on the economy, environment, ecology and commercial activities. It depends on the contract of the wind farm owner whether the sub-sea cables must be removed. It will contain costs to remove the cables which can be avoided if the cables can remain. Next to that, the removal of the installation can temporarily disturb the marine life that was created over lifetime. Seabirds, for instance, are temporarily not able to use the windmills as feeding ground, especially when the decommissioning takes place during breeding time. Mammals are also disturbed by the noise generated by vessels used for dismantling the installation. Decommissioning, moreover, has a negative influence on the shipping industry, fishermen and tourism because they are not allowed to use the area during the disposal (Januário et al., 2007). All these damaging effects can be avoided when the license of a wind farm will be extended. The technology is nowadays so advanced that the windmill material can last even longer than 20 years. As long as the maintenance costs are less than the market power price, the wind farm can keep producing (Reitsma and Janssens, 2010).

(16)

16

political situation and electricity prices at the moment. Furthermore, the operation and maintenance (O&M) costs of the installation can exceed the estimated future returns (Climate Change Capital, 2004).

3.4 Storage of electricity by chemical conversion

Chemical conversion is a process, by means of electrolysis, where wind energy is used to convert water into hydrogen and oxygen. This technique is so called Power to Gas (PtG). When energy supply exceeds energy demand the surplus can be transported to the PtG installation where large scale energy storage can occur.

Power to Gas has several benefits compared to other energy storage techniques. First of all, PtG provides a connection between the power and gas grid, which has never happened before in the Netherlands. Next to that, gas transport is cheaper than power transport. Besides, it is an efficient method since a limited amount of energy is wasted and it has potential for large scale storage of renewable energy. Power cannot be stored in large amount while gasses and liquids can, section 3.4.4 will go into more detail about this. Last, this innovative technique creates sustainability for the chemical industry and releases possibilities in other sectors (Brandsma and Broere, 2014).

3.4.1 Conversion of power to gas process by means of electrolysis

In this research the Alkaline electrolysis technique is used. There are more existing electrolysis techniques but Alkaline is the commercially available one (DNV KEMA, 2013). This electrolyzer has a power capacity of 6MW and the investment costs are estimated at €1000 per kW (DNV KEMA, 2013). The chemical conversion process where water is split into oxygen and hydrogen, by use of electricity, is illustrated by the following chemical equation:

2𝐻2𝑂 → 2𝐻2 + 𝑂2 (3.1)

(17)

17

the hydrogen can be combined with carbon dioxide (CO2) to produce green gas (CH4). The

equation below shows the chemical reaction of hydrogen with carbon dioxide.

2𝐻2 + 𝐶𝑂2 → 𝐶𝐻4 + 𝑂2 (3.2)

Figure 1 shows a graphical representation of the electricity flow from wind to electricity, using electrolysis techniques.

Figure 1 Schematic illustration of electricity flow from wind to electricity by chemical conversion

3.4.2 Conversion of gas to power process by means of fuel cells

The conversion of the stored gas (hydrogen) to power (electricity) is realized by the reverse chemical process. This process is called exothermic reaction or Gas to Power (GtP) which consists of merging oxygen and hydrogen to recover electricity. The following chemical equation gives the exothermic reaction:

(18)

18

When hydrogen hits oxygen, it releases energy as heat and produces water (Koroneo et al., 2004). Fuel cells are used to recover hydrogen back into electricity (Veyer, 2014). Proton exchange membrane (PEM) fuel cells are suitable for mass production which can reduce the costs (Tsuchiya and Kobayashi, 2004).

3.4.3 Cost estimates of the conversion processes

The Alkaline electrolyzer has an assumed efficiency of 75%. So 75% of the inputted electricity remain after chemical conversion. This means that the required fuel cell capacity equals 75% of the electrolyzer’s installed capacity, which is 75% of 6 MW equals 4,5 MW (Veyer, 2014). This study assumes that the electrolyzer is employed 24 hours a day while the fuel cell is only employed when the hydrogen has to be converted back into electricity which happens when the electricity prices are high.

Taljan et al. (2008) assumed that the efficiency of PEM fuel cells are between 40 and 60%. Additionally, they concluded that fuel cell efficiency has a positive effect on profit as a result of electricity sales. The investment costs of PEM fuel cells are estimated at €1000 per kW (DNV KEMA, 2013).

The roundtrip efficiency in the existing chemical conversion installations amounts to approximately 40 percent which is uneconomical now. According to professor Michael Sterner of the Ostbayerische Technische Hochschule in Regensburg will the chemical conversion installation be economical feasible within ten years (Sterner, 2014). The focus in this study will be on the feasibility of PtG and GtP.

3.4.4 Hydrogen storage

(19)

19

efficiency (Sioshansi, 2008, p. 270). Interesting to know is that in 2010 there was already nearly 2 times more power capacity available than the Netherlands needed at a peak time (Reitsma and Janssens, 2010). The renewable energy sources will continue to increase in the near future while the electricity demand is still inelastic which predicts an even bigger oversupply than in 2010. This increasing oversupply gives a reason to create storage capacity.

Hydrogen can be used for other purposes than converting it back into electricity via gas engines or fuel cells. There are four other purposes where hydrogen can be used for:

i) as fuel in the automobile industry in the form of methanol,

ii) sold to chemical industry,

iii) in feed in the natural gas grid and

iv) converted with carbon dioxide into green gas (European Commission, 2013)

These alternatives will not be discussed further here.

(20)

20

4. Electricity data and analysis methods

4.1 Electricity data

The Dutch electricity prices are compared with those from Germany over the year 2013; see Table 4. The Dutch electricity data is received from the Amsterdam Power Exchange (APX) while the German electricity prices are obtained from the European Energy Exchange (EEX). Both are collected using the ThomsonReuters data stream (Wouters, 2014). The hourly spot price data of the APX and EEX is from the 1st of January 2007 until the 31st of December 2013 but the focus will be on the data of the year 2013. The data consists of 6264 observations per year per country, because ThomsonReuters data stream only collects data for weekdays. So, 24 hours a day multiplied by 5 days a week and 52,2 weeks a year equals 6264 hours a year. This means that a year counts 6264 hours instead of 8760 (365 days*24 hours).

The calculations in this study are applied to a Dutch offshore wind farm, called Gemini Windpark, that will be installed in the North Sea in 2015 and is supposed to start producing electricity in the year 2016 (Gemini Windpark, 2014). This means that several data is based on the provided information by the Gemini Windpark website. The remaining descriptive statistics are summarized in Table 4. The minimum electricity price in Germany is negative in 2013, as can be seen in Table 4, which can happen, for example, when there is a big storm whereby the wind turbines have to shut down. If this occurs the wind farm owners have to give money to the electricity consumers instead of receiving money.

Table 4 Descriptive statistics for Dutch and German electricity prices in 2013

(21)

21 4.2 Methodology

Data collected over intervals of time on one specific economic unit is called time-series data. Time-series data often trend together over time. The macroeconomic data in this research are reported in hourly terms and the economic unit is ‘electricity price’. Univariate time-series is a stochastic process where a single '𝛾' is related to values of itself in the past and current. Univariate time-series do not have any explanatory variables (Hill et al., 2011). The Dutch and German electricity prices will be plotted against time to examine whether the prices are stationary in the year 2013. The Augmented Dickey-Fuller (ADF) test is used to test for nonstationarity of the electricity prices of the Netherlands and Germany. The ADF test will compute the test statistics which will be compared with the critical value to find out whether the null hypotheses has to be rejected or not (Adkins and Hill, 2008). The null and alternative hypotheses are as follows:

H0: The electricity prices are nonstationary

H1: The electricity prices are stationary

Next to that, the Granger causality Wald test is used to test whether there is a causality between the Dutch and German electricity prices. The 'varstable' test is used to find out if the time series can be used to generate forecasts.

4.2.1 Assumed SDE+ subsidy in the Netherlands

(22)

22

that the Netherlands will catch up with Germany in these 18 years. The cost price of renewable energy in 2013 and the wind turbine capacity in this research are based on the Gemini Windpark.

4.2.2 Assumed gross profit without subsidy in the Netherlands in 2031

The profit calculations will be applied to the Gemini Windpark. The Gemini Windpark starts producing wind energy in 2016 and has a producing license for 15 years so the wind farm will be decommissioned in 2031. This research will estimate for the Gemini Windpark whether it is more profitable to make use of storage facilities or not for the year 2031. The Gemini Windpark assumes that the lifetime of their wind turbines is way longer than 15 years but they do not mention the number of years (Gemini Windpark, 2014). That is why this research only calculates the profit for the year 2031 which can be assumed as yearly profit but has to be adjusted for inflation after 2031. Most of the prices will come from the year 2013 that is why the inflation is calculated over 18 years.

A. Profit without storage

The profit without storage will be calculated by the following equation:

𝑃𝑟𝑜𝑓𝑖𝑡 𝑤𝑖𝑡ℎ𝑜𝑢𝑡 𝑠𝑡𝑜𝑟𝑎𝑔𝑒 = 𝑇𝑜𝑡. ℎ ∗ 𝑞̅ ∗ (𝑃̅ − 𝐶𝑃) + 𝐺𝐷 (4.1)

In this equation 𝑇𝑜𝑡. ℎ is the total number of hours per year that the wind turbines of the Gemini Windpark produce electricity, 𝑞̅ denotes the electricity production amount per megawatt hour of these wind turbines, 𝑃̅ denotes the average electricity price in €/MWh, 𝐶𝑃 is the cost price in €/MWh and 𝐺𝐷 is the green decommissioning bonus.

B. Profit with storage

(23)

23

𝑅𝑒𝑣𝑒𝑛𝑢𝑒 = (ℎ > 𝑋 ∗ 𝑞̅ ∗ 𝑃̅ > 𝑋) + [(ℎ < 𝑋 ∗ 0,375 ∗ 𝑞̅) ∗ 𝑃̅ > 𝑋] + 𝐺𝐷 (4.2) 𝑊𝑖𝑛𝑑 𝑒𝑛𝑒𝑟𝑔𝑦 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑜𝑛 𝑐𝑜𝑠𝑡𝑠 = 𝑇𝑜𝑡. ℎ ∗ 𝑞̅ ∗ 𝐶𝑃 (4.3) 𝐶ℎ𝑒𝑚𝑖𝑐𝑎𝑙 𝑐𝑜𝑛𝑣𝑒𝑟𝑠𝑖𝑜𝑛 𝑓𝑖𝑥𝑒𝑑 𝑐𝑜𝑠𝑡𝑠 = 𝐸𝐶 + 𝑂&𝑀 𝐸 + 𝐹𝐶𝐶 + 𝑂&𝑀 𝐹 (4.4) 𝐶ℎ𝑒𝑚𝑖𝑐𝑎𝑙 𝑐𝑜𝑛𝑣𝑒𝑟𝑠𝑖𝑜𝑛 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒 𝑐𝑜𝑠𝑡𝑠 = ℎ < 𝑋 ∗ 𝑞̅ ∗ (𝑊 + 0.75𝐻 + 0.75𝑂) (4.5)

Where 𝑋 is the storage border. The wind farm owner can decide to store electricity when the electricity price drops below a certain value. The storage border is the switching point between storage and sale. All the electricity that will be produced during an electricity price below the electricity price of the storage border will be stored. The electricity that will be produced while the electricity price is above the electricity price of the storage border will directly be sold to the power market. The storage border can be between the minimum price, 1 €/MWh, and the maximum electricity price, 130 €/MWh, that is assumed to occur in 2031.

The revenue with storage of equation (4.2) consists of three parts:

- The first part is the income out of direct electricity sale which will be calculated as the hours per year that the wind turbines of the Gemini Windpark produce electricity while the electricity price is above the storage border (ℎ > 𝑋) multiplied by the electricity production amount per megawatt hour of these wind turbines (𝑞̅) and the average electricity price above the storage border.

(24)

24

- The third element is green decommissioning. As a result of the license expiry, the wind farm owner has to remove the whole wind farm installation, so called decommissioning. If the wind farm owner gets a license renewal the wind farm does not have to be decommissioned yet so the wind farm owner can save the money it had put aside. The interest that the wind farm owner receivers over his saved money is so called green decommissioning bonus (𝐺𝐷).

The wind energy production costs of equation (4.3) will be calculated as:

- The total hours per year that the wind turbines of the Gemini Windpark produce electricity (𝑇𝑜𝑡. ℎ) multiplied by the electricity production amount per megawatt hour of these wind turbines (𝑞̅) and the cost price of producing offshore electricity (𝐶𝑃).

The costs of storage consist of equation (4.4) and (4.5):

- The chemical conversion fixed costs in equation (4.4) consist of the lease costs (𝐸𝐶, 𝐹𝐶𝐶) and the operation and maintenance costs (𝑂&𝑀 𝐸, 𝑂&𝑀 𝐹) of the electrolyzer and the fuel cell.

(25)

25

5. Empirical results

5.1 The relation between Dutch and German electricity prices

The Augmented Dickey-Fuller (ADF) test is used, as said in section 4.2, to test the stationarity of the Dutch and German electricity prices. The hourly electricity prices in the Netherlands and Germany from 2007 until 2012 are added to the dataset of 2013 to make the results more reliable.

Firstly, is tested how many lags are needed for the ADF test. The test results in Tables A.1 and A.2 in Appendix A show that the Akaike’s information criterion (AIC) and Schwarz’s Bayesian information criterion (SBIC) have the lowest value at 54 lags for the Netherlands and 53 lags for Germany. A lower AIC and SBIC means that the model is assumed to be more likely to be true. This means that an ADF test with 54 lags for the Netherlands and 53 lags for Germany gives the most reliable model.

Secondly, Figures 2 and 3 contain graphs of the observation of the electricity price plotted against time. The electricity price per megawatt hour (€/MWh) that appears in Figures 2 and 3 has a tendency to fluctuate around a constant mean without trending or wandering. This is the reason why the Augmented Dickey-Fuller (ADF) test is done for both time series including a constant but without a trend. Tables A.3 and A.4 in Appendix A gives the ADF test results of the Dutch and German electricity prices. The Dutch test statistics of -10.382, in Table A.3, fall within the rejection region of 1% critical value of -3.430. This means that the null hypothesis of nonstationary will be rejected and so the Dutch electricity prices are stationary. Moreover, the test statistics of the German electricity prices which are -10.272 do also fall within the rejection region of 1% critical value of -3.430 (see Table A.4 in Appendix A.) So, the null hypothesis for Germany will also be rejected which means that the German electricity prices are stationary. Two nonstationary time series that move together through time are so called cointegrated. Due to the time series (Dutch and German electricity prices) are both stationary, they are not "Integrated of order 1" or I(1) and so the time series are not cointegrated (Adkins and Hill, 2008).

(26)

26

Appendix A, appear to be significant. This means that the null hypothesis of no causality has to be rejected and so the residuals have a causal relationship. Moreover, Table A.6 in Appendix A illustrates an autocorrelation robust regression model of the Dutch electricity price as the dependent variable and the German electricity price as the explanatory variable. An autocorrelation robust regressions is generated instead of a normal regression because the variables are assumed to correlate. The coefficient of the German electricity prices turned out to be positive (0.831) and significant (t=57.31 > 1.96). As the two series are stationary these variables are related. This means that when the German electricity price increases with 1 €/MWh the Dutch electricity price increases as a result with 0.831 €/MWh. Table A.7 in Appendix A shows that it also happens the other way around. The explanatory variable, which is in this case the Dutch electricity prices, have a positive (0.755) and significant (t=16.55 > 1.96) relation with the dependent variable; German electricity prices. So, when the Dutch electricity price increases with 1 €/MWh the German electricity prices will increase with 0.755 €/MWh.

Finally, a ‘varstable’ test is done to test whether the time series (Dutch and German electricity prices) can be used to generate forecasts. All the calculated eigenvalues generated by ‘varstable’ are less than one which illustrates that the model is stable. As a result, the electricity prices will show no explosive behavior in the long-run. This proves that the electricity prices in this dataset can be used to forecast the electricity prices.

(27)

27

Figures 4 and 5 show the histograms of electricity prices per hour in 2013 respectively in the Netherlands and Germany. Both electricity prices display non-normal distributions, as can be seen by comparing the normal distribution curves with the histogram distributions. Note that there are more peaks around the mean and tails. Distributions with these properties are so called leptokurtic (Hill et al., 2011). Although, the average electricity price in Germany is 28.48% lower than that of the Netherlands in 2013, the standard deviation is bigger for Germany; see Table 4. The standard deviation for the Netherlands is 13.17868 €/MWh while the German standard deviation is 15.8972 €/MWh. This means that there is more variation in the electricity prices in Germany than in the Netherlands in 2013. The results of a two-sample t-test with unequal variances state that the differences between the Netherlands and Germany in 2013 is significantly higher than zero (see Table A.8 in Appendix A). This means that difference in average electricity price and standard deviation can be assumed to be true.

Figure 2 Electricity prices per megawatt hour in the Netherlands in 2013

0 50 1 0 0 1 5 0 2 0 0 € /MW h 01jan2013 00:00:56 31dec2013 23:00:40 Hour APXH Mean

(28)

28

Figure 3 Electricity prices per megawatt hour in Germany in 2013

Figure 4 Histogram of electricity prices in the Netherlands in 2013

-5 0 0 50 1 0 0 1 5 0 € /MW h 01jan2013 00:00:56 31dec2013 23:00:40 Hour EEXHR Mean

Hourly electricity prices in Germany in 2013

0 5 10 15 Pe rc e n ta g e 0 50 100 150 €/MWh

(29)

29

Figure 5 Histogram of electricity prices in Germany in 2013

5.2 Assumed SDE+ subsidy in the Netherlands in 2013 and 2031

The SDE+ contribution is calculated as the difference between the cost price of renewable energy minus the yearly average electricity price. Based on the Gemini Windpark, the cost price of renewable energy is estimated at 170 €/MWh for the year 2013 and 243 €/MWh for the year 2031 (Gemini Windpark, 2014). The cost price for 2031 is adjusted for inflation over 18 years. The yearly inflation in this research is assumed to be 2%. The average electricity price is calculated as 53.77 €/MWh for the year 2013 and 54.91 €/MWh for the year 2031. The electricity price is adjusted for inflation over 18 years and this price is lowered by 28.48%, based on the fact that German prices were 28.48% lower than the Dutch prices in 2013 (see section 5.1). This is based on the assumption that by that time the electricity prices in the Netherlands are also lower because of occasional over-supply by green energy. So, the SDE+ contribution for 2013 is 116.23 €/MWh and assumed to be 191.09 €/MWh for 2031, as can be seen in Table 5.

The wind turbine capacity is also based on the Gemini Windpark and amounts to 4 MW per hour per wind turbine. The Gemini Windpark counts 150 wind turbines, so the total wind

0 5 10 15 20 Pe rc e n ta g e -50 0 50 100 150 €/MWh

(30)

30

capacity is 600 MW per hour (Gemini Windpark, 2014). The maximum production per year is 600 MW wind capacity multiplied by the 6264 production hours per year equals 3,758,400 MWh per year (see Table 5).

(31)

31

Table 5 SDE+ subsidy in the year 2013 and 2031

2013 2031

Cost price renewable energy Average electricity price 2013 SDE+ contribution

170 €/𝑀𝑊ℎ 53.77 €/𝑀𝑊ℎ 116.23 €/𝑀𝑊ℎ−

Cost price renewable energy Average electricity price 2023 SDE+ contribution

243 €/𝑀𝑊ℎ 54.91 €/𝑀𝑊ℎ 191.09 €/𝑀𝑊ℎ−

Capacity wind turbine Production hours per year Maximum production per year

600 𝑀𝑊𝑒 6264 ℎ𝑜𝑢𝑟𝑠/𝑦𝑒𝑎𝑟 3,758,400 𝑀𝑊ℎ/𝑦𝑒𝑎𝑟×

Capacity wind turbine Production hours per year Maximum production per year

600 𝑀𝑊𝑒 6264 ℎ𝑜𝑢𝑟𝑠/𝑦𝑒𝑎𝑟 3,758,400 𝑀𝑊ℎ/𝑦𝑒𝑎𝑟×

Wind capacity per year SDE+ contribution SDE+ subsidy per year

3,758,400 𝑀𝑊ℎ/𝑦𝑒𝑎𝑟 116.23 €/𝑀𝑊ℎ € 𝟒𝟑𝟔, 𝟖𝟑𝟖, 𝟖𝟑𝟐 ×

Wind capacity per year SDE+ contribution SDE+ subsidy per year

(32)

32

5.3 Assumed parameters of the gross profit models for the Gemini Windpark in 2031

After 15 years the SDE+ subsidy ends. The profit for the wind farm owner will therefore be calculated without subsidy after 15 years of production. As said in section 5.2 the inflation rate is assumed to be 2% per year. The taxes imposed by the government in 2031 are not estimated and so are not taken into account. This means that the gross profit will be calculated instead of the net profit. Besides, platform facilities rental costs are not taken into account because it is outside the scope of this research. Finally, transport costs are assumed to be zero.

The Gemini Windpark

As said in section 5.2 the Gemini Windpark has 150 wind turbines with a capacity of 4 MW each and a total maximum production per year of 3,758,400 MWh. The actual generated electricity per year, though, is estimated by the Gemini Windpark at 2,6 TWh (equals 2,600,000 MWh) (Gemini Windpark, 2014). This means that the expected generated electricity per hour is 2,600,000 MWh divided by 6264 hours a year resulting in 415 MWh. So one wind turbine produces 2.77 MWh of its 4 MW available capacity.

Green decommissioning bonus

The Gemini Windpark has a bank guarantee of €40,000,000 to decommission the wind farm in 2031 that can be saved in case of green decommissioning (Gemini Windpark, 2014). This research supposes that a yearly 4% interest rate can be received for €40,000,000 which is €1,600,000 per year. So the green decommissioning bonus in 2031 is €1,600,000.

Electricity price

(33)

33

Wind energy production cost price

After 15 years the capital investment costs are amortized. In this research it is assumed that the remaining costs only involve operation and maintenance (O&M) costs after 15 years. The costs of offshore wind farms consist for 25% out of O&M costs but can be reduced with 1/3rd within 10 years (ECN, 2014). This research assumes that the O&M costs will reduce with another 1/3rd over the next 8 years. This means that the O&M costs is assumed to be 11.11% of the total production costs after 18 years. The cost price in this is research, as said, is based on the Gemini Windpark and their cost price is 170 €/MWh in 2013. So, with an assumed inflation of 2% per year, the cost price in 2031 amounts 26.71 €/MWh.

Electrolyzer and fuel cell lease costs

The wind farm owner needs an electrolyzer and fuel cell for chemical conversion. The investment costs are not included in the calculations because it is supposed that the electrolyzer and fuel cell can be leased instead of purchased in 2031. The investment costs are needed, though, to calculate the lease costs. The lease costs are assumed to be 10% of the investment costs. The investment costs of an Alkaline electrolyzer are estimated at 1000 €/kW (equals 1,000,000 €/MW) and has a capacity of 6 MW (Auer et al., 2012 and DNV KEMA, 2013). So the investment costs of an Alkaline electrolyzer are €1,000,000 multiplied by 6 MW equals €6,000,000. 10% of these costs are €600,000 per year.

The investment cost of a fuel cell are estimated at 1000 €/kW (equals 1,000,000 €/MW) and has a capacity of 2,4 MW (DNV KEMA, 2013). So the investment costs of a fuel cell are €1,000,000 multiplied by 4,5 MW equals €4,500,000. 10% of these costs are €450,000. With an inflation of 2% per year the electrolyzer lease costs are €856,948 and the fuel cell lease costs are €642,711 in 2031.

Electrolyzer and fuel cell O&M costs

(34)

34

Water price

Water and power are the fundamental inputs for electrolysis. To process one megawatt hour of power requires approximately 168 liters of water. The water price in the Northern Netherlands is estimated at €0.000642 per 1 liter in 2015 (Waterbedrijf Groningen, 2014). Adjusted for inflation of 2% the water price is assumed to be €0.000881 per 1 liter in 2031. So processing one megawatt hour of power the water input costs will be 0.15 €/MWh in 2031.

Hydrogen storage price

After electrolysis the hydrogen has to be stored until the electricity prices are higher than before. One megawatt hour of power can be converted into 210 Nm3 of hydrogen and one Nm3 hydrogen is equal to 0.0899 kg. Veyer (2014) estimated to price per kg of large scale storage at €1.45. So the hydrogen storage price is 27.37 €/MWh in 2013. This price adjusted for 2% inflation is 39.09 €/MWh in 2031.

Oxygen storage price

Oxygen is needed to convert the hydrogen back into power through a fuel cell. It is assumed that the oxygen that is processed through electrolysis can possibly be stored and converted back to power afterwards. The oxygen price is estimated at 0.25 €/ Nm3

in 2013 (Wouters, 2014). Adjusted for 2% inflation this price is assumed to be 0.36 €/ Nm3

2031. To produce one megawatt hour of electricity, approximately 105 Nm3 of oxygen is needed (Veyer, 2014). So the assumed oxygen storage price in megawatt hours is 37.49 €/MWh.

(35)

35

Table 6 Parameters of the profit without storage model

Table 7 Parameters of the profit with storage model

(36)

36

5.4 Assumed gross profit for the Gemini Windpark in 2031

The profit without storage and the profit with storage for the year 2031 are calculated with the respectively equations (4.1), (4.2), (4.3), (4.4) and (4.5). The calculations of the profit with and without storage are given in Tables B.1 and B.2 in Appendix B while the results are illustrated in Figure 6. The profit margin shown in Figure 7 is calculated as the profit with storage divided by the total revenues with storage. The calculations of the profit margin are shown in Table B.3 in Appendix B.

Figure 6 Gross profit in for the Gemini Windpark in the year 2031

-€ 150,000,000 -€ 100,000,000 -€ 50,000,000 € 0 € 50,000,000 € 100,000,000 € 1 € 20 € 40 € 60 € 80 € 100 € 120 € 130 P ro fit in E UR Storage border in €/MWh

Gross profit curve in 2031

(37)

37

Figure 7 Gross profit margin of storage facilities for the Gemini Windpark in 2031

Profit without storage

Figure 6 illustrates that the profit without storage assumes to have an average profit of around €75,000,000. This result shows that after 15 years, when the license to produce wind energy on the North Sea expires and therewith the SDE+ subsidy, the Gemini Windpark is assumed to be profitable in 2031. The profit can be explained by the fact that the estimated average electricity price is higher than the estimated cost price for 2031.

Profit with storage

The profit margin in Figure 7 illustrates that the profit margin is positive for storage borders below 50 €/MWh. This means that below a storage border of 50 €/MWh the total revenues are higher than the total costs when storage facilities are included. The profit with storage, though, is at every storage border lower than the profit without storage which can be seen in Figure 6. This means that at every storage border the additional revenues of storage are lower than the additional costs of storage and therewith Gemini Windpark is better off without storage than with storage. This profit wit storage will probably be even lower if the rental costs of the offshore platforms and the transportation costs are taken into account.

-160% -140% -120% -100% -80% -60% -40% -20% 0% 20% 40% 60% € 1 € 20 € 40 € 60 € 80 € 100 € 120 € 130 G ro ss pro fit m a rg in in perc ent a g e Storage border in €/MWh

Gross profit margin of storage in 2031

(38)

38 5.5 Sensitivity analysis

5.5.1 Inflation rate

The inflation rate is assumed to be 2% in this study but this section examines the gross profit sensitivity of 4% inflation. The gross profit curve with an inflation rate of 4% is given in Figure 8. Compared to Figure 6 in section 5.4 the gross profit increases but the difference between the gross profit without storage and with storage stays the same. This means that the gross profits without en with storage are sensitive to the inflation rate but the conclusion does not change. It is still not profitable for offshore wind farms, if they are not subsidized, to store their power on existing platforms in the North Sea even with an inflation rate of 4%.

Figure 8 Gross profit curve with 4% inflation

5.5.2 Without correction for German electricity prices

In this research it is assumed that the electricity prices in the Netherlands will catch up with those in Germany within 18 years after 2013. This means that the Dutch electricity prices, predicted for 2031, are lowered for 28.48%, so called German correction. This section investigates how sensitive the gross profit curves are to this German correction. The gross profit without storage and with storage for 2031 are calculated without the Germany correction to illustrate the impact. The electricity prices of 2013 are therefore only adjusted for 2% inflation over 18 years. Figure 9 illustrates that both gross profits, without and with

-€ 200,000,000 -€ 150,000,000 -€ 100,000,000 -€ 50,000,000 € 0 € 50,000,000 € 100,000,000 € 150,000,000 € 1 € 20 € 40 € 60 € 80 € 100 € 120 € 140 € 160 € 180 P ro fit in E UR Storage border in €/MWh

Gross profit curve in 2031 with 4% inflation rate

(39)

39

storage, shift upward compared to the German corrected profit curves in section 5.4 This means that the gross profit curves are sensitive to the German correction. The ratio between the gross profit without storage and with storage, though, stays the same which means that the gross profit without storage is always higher than the gross profit with storage and thus storage is still not a profitable option without the German correction.

Figure 9 Gross profit curve without German correction

5.5.3 Lower assumed investment costs of the electrolyzer and fuel cell

The investment costs of the electrolyzer and fuel cell are estimated at 1000 €/kW each and are adjusted for 2% inflation over 18 years. This section will investigate how sensitive the gross profit with storage is to the investment costs of the electrolzyer and fuel cell. Assume that the investment costs of both come down with 60% for those 18 years. This means that the investment costs in 2031 for the electrolyzer are expected to be €3,427,791 and for the fuel cell €2,570,843. The lease costs are still assumed to be 10% and the operation and maintenance (O&M) costs to be 4% of the investment costs. This results in yearly lease costs of €342,779 for the electrolyzer and €257,084 for the fuel cell and yearly O&M costs of €137,112 for the electrolyzer and €102,834 for the fuel cell. Figure 10 shows the new gross profit with storage curve combined with the original gross profit without storage curve. With lower investment costs the gross profit with storage is obviously higher than with the normal

-€ 150,000,000 -€ 100,000,000 -€ 50,000,000 € 0 € 50,000,000 € 100,000,000 € 150,000,000 € 1 € 20 € 40 € 60 € 80 € 100 € 120 € 140 € 160 € 180 P ro fit in E UR Storage border in €/MWh

Gross profit curve withouth German correction

(40)

40

investment costs. The gross profit with storage curve in Figure 10, though, is not drastically shifted upward compared to Figure 6 in section 5.4. This concludes that the gross profit with storage is not very sensitive to the investment costs. This suggests that, for the storage case, the costs of hydrogen and oxygen are the cause of a lower gross profit.

Figure 10 Gross profit curve with 60% lower investment costs of the electrolzyer and fuel cell

5.5.4 Efficiency rate

The assumed roundtrip efficiency of chemical conversion is 37.5%. This is first of all based on the efficiency rate of the Alkaline electrolyzer which is 75% and so 25% of the inputted power is wasted after electrolysis. This means 25% of the generated electricity cannot be sold and so negatively influences the revenue but also 25% less hydrogen has to be stored which save costs. Secondly, of the 75% remaining power will 50% remain after the reversed chemical conversion. This section examined how sensitive the gross profit with storage is to an increase in the efficiency rate of only the electrolyzer and to a similar increase in the efficiency rate of the electrolyzer and fuel cell.

Table 8 shows the sensitivity of gross profit for an increase of 10% and 20% of the efficiency rate of the electrolyzer while the fuel cell efficiency rate stays the same. A 10% efficiency

-€ 150,000,000 -€ 100,000,000 -€ 50,000,000 € 0 € 50,000,000 € 100,000,000 € 1 € 20 € 40 € 60 € 80 € 100 € 120 € 130 P ro fit in E UR Storage border in €/MWh

Gross profit curve in 2031 with 60% lower investment costs

of the electrolzyer and fuel cell

(41)

41

increase means 10% more than the assumed 75% which is 85% and so 20% more than 75% is 95%. The calculated gross profit per storage border per efficiency increase is shown in Table C.1 in Appendix C.

Table 8 shows that the gross profit is negatively influenced by the electrolyzer efficiency increase which means that the gross profit is lower when more power remain after electrolysis. This indicates that the hydrogen and oxygen storage costs apparently increase more than the revenue, when more power remain after chemical conversion.

Table 8 gross profit sensitivity to the electrolyzer efficiency increase

Percentage efficiency increase Efficiency rate electrolyzer Percentage efficiency increase Efficiency rate fuel cell

Roundtrip efficiency rate Sensitivity gross profit with storage 10% 85% 0% 50% 42.50% -9.87% 20% 95% 0% 50% 47.50% -19.74%

Table 9 illustrates the gross profit sensitivity to similar efficiency increase of electrolyzer and fuel cell. The gross profit sensitivity are calculated for the electrolyzer and fuel cell efficiency rate increase of 10% and 20% compared to the assumed efficiency rates of respectively 75% and 50%. The calculated gross profit with storage are stated in Table C.2 in Appendix C. Table 9 shows that the gross profit is higher by an increase in both efficiency rates. This is because the fuel cell efficiency rate only causes higher revenues and no extra costs while the elecrolyzer efficiency rate cause higher storage costs next to higher revenues.

Table 9 Gross profit sensitivity to similar efficiency increase of the electrolyzer and fuel cell

Percentage efficiency increase Efficiency rate electrolyzer Percentage efficiency increase Efficiency rate fuel cell

(42)

42

(43)

43

6. Limitations

Within this research there are several limitations. First of all, ThomsonReuters does not collect data of electricity prices for the weekends which means only electricity prices of weekdays are taken into account.

Besides, this research focuses on predictions for 2031 which can deviate from reality at that time. To forecast the political situation or the electricity prices in 2031, for instance, can be very difficult.

Furthermore, the calculations only include the operation and maintenance costs of the electrolyzer which possibly influenced the results. As said in section 3.4.4 this research assumes that the electrolyzer can be leased but if this is not possible yet in the year 2031 the investment costs should be taken into account to calculate the profit.

Another bias of the methodology is the sales price. The sales price always increases when the storage border increases because the sales price is the average price above the storage border. When the storage border increases, the average sales price and the number of hours below the storage border increase which has an influence on the results. In reality the sales price can be different than the sales price assumed in this study.

Moreover, this research did not focus on detailed information about fuel cell conditions like the required capacity, conversion speed and total package costs of the fuel cell. These conditions could affect the profitability of power storage.

Next to that, this study focused on the reversed chemical conversion which means that all the stored hydrogen will be converted back into power while it is also possible to convert the hydrogen into gas. The chemical conversion of hydrogen to gas does not need a fuel cell which makes the conversion of hydrogen to gas cheaper than hydrogen to power. So, the fuel cell costs makes the storage facilities more expensive in this research.

(44)

44

7. Conclusions and recommendations

7.1 Conclusions

This study focuses on the research question “Is it profitable for offshore wind farms, if they

are not subsidized, to store their power on existing platforms in the North Sea?”. The first

part of the conclusion is based on the literature review and the second part is predicated upon the methodology and results of this research. Firstly, the literature review made it clear that hydrogen storage has to be investigated in this research because hydrogen storage is cheap and can be used for several purposes such as the automobile industry or chemical industry. Hydrogen storage is possible on a more efficient, sufficient and larger scale than electricity storage. Secondly, if the hydrogen can be stored and the wind farm does not have to be removed after 15 years, the wind farm can have a second life. This second life is also known as green decommissioning. Green decommissioning can prevent quite a few negative impacts of the wind farm removal. The possibility, however, to get a license renewal in the future depends on the political situation and electricity prices in the future. The estimated electricity prices in the Netherlands for 2031 are adjusted for inflation and based on the electricity price history of Germany, because after comparing the Netherlands with Germany, Germany turned out to be ahead of the Netherlands in terms of renewable energy sources.

(45)

45 7.2 Recommendations

Further research can add to this study and several suggestions are mentioned here. As discussed in the limitations, only weekdays are investigated in this research while it would give a more complete analysis if the methodology of this study is applied on electricity price data of every day including the weekend.

Since the SDE+ arrangement is specific to the Netherlands further research can take a look at different countries because each country has its own subsidy arrangements.

Furthermore, storage is assumed to be possible on an oil and gas platform in the North Sea. Amortized oil and gas platforms could be a great option to place an electrolyzer on. It is very expensive to remove an oil platform and if this platform can be reused as a storage platform, the dismantling costs can be saved. Further research could include more details about storage facilities on oil and gas platform.

Besides, the existing fossil grids can be used to transport power and gas. Further research can investigate whether the North Sea oil and gas infrastructure can be used for storage and chemical conversion of intermittent renewable capacity.

Moreover, every country has its own electricity grid while it would be more efficient if these countries connect their grid to each other’s grid so power and gas can easily flow from one country to another. If this is possible, which is realistic because countries bordering the North Sea are already negotiating about this, it provides opportunities for further research about arbitrage between countries.

Next to that, this research explains that hydrogen can be converted back into power but can also be converted into gas. As said in the limitations, the conversion of hydrogen into gas is cheaper than hydrogen into power. So, further research can focus more on power to gas in combination with gas sales between countries.

(46)

46

References

Adkins, L. C, Hill, R.C., 2008. Using STATA for Principles of Econometrics. Third Edition, John Wiley & Sons Inc., New York, USA.

Auer. J., Keil, J., Stobbe, A., & Mayer, T., 2012. State-of-the-art electricity storage systems. Deutsche Bank, DB Research.

Bessenyei, K., 2014. An Empirical Analysis Of Price Convergence Within The EU Electricity Sector.University of Groningen.

BP Energy Outlook 2035., 2014. Accessed on 25/01/2015 via bp.com/energyoutlook

Brandsma, M., Broere, A. (2014, 31 oktober). Duurzaam gas uit wind en zon. Dagblad van

het Noorden, p. 18.

Bundesverband WindEnergie (BWE). 2014. Wind Moves Germany – Information on the energy

transition. Accessed on 20/11/14 via

http://www.windenergie.de/sites/default/files/download/publication/wind-moves-energiewende-wind-moves-germany/flyer_windbewegt_engl.pdf

CBS (Centraal Bureau voor de Statistiek). 2013. Hernieuwbare energie in Nederland 2013. Accesed on 07/01/15 via http://www.cbs.nl/NR/rdonlyres/00DEA034-8FBE-4EFF-B488-18FC4A9BC7BC/0/WebversiefHernieuwbareenergie.pdf

ClimateChange Capital. 2004. Offshore Renewable Energy Installation Decommissioning

Study. Accessed on 18/11/14 via

(47)

47

Corbetta, G., Pineda, I., Moccia, J., 2014. The European Wind Energy Association (EWEA) The European offshore wind industry – key trends and statistics 2013.

DNV KEMA, 2013. System analysis power to gas. Deliverable 1: Technology review.

Draijer, W., 2013. Energieakkoord voor duurzame groei. SER

ECN (Energie-Nederland en Netbeheer Nederland). 2013. Energietrends 2013 vier gevolgen van de groei van hernieuwbaar voor het energiesysteem.

ECN (Energie-Nederland en Netbeheer Nederland). 2014. Accessed on 02/12/14 via https://www.ecn.nl/nl/nieuws/item/nieuw-onderhoudssysteem-verlaagt-kosten-offshore-wind/

European Commission. 2013. The future role of challenges of Energy Storage. Accessed on 06/11/14 via http://ec.europa.eu/energy/infrastructure/doc/energy-storage/2013/energy_storage.pdf

Gemini Windpark. 2014. Accessed on 01/12/14 via http://geminiwindpark.nl/financiering/

Greiner, C., Korpas, M., Holen, A. 2007. A Norwegian case study on the production of hydrogen from wind power. International Journal of Hydrogen Energy 32, 1500-1507.

Hedegaard, K., Meibom, P., 2011. Wind power impacts and electricity storage – A time scale perspective. Renewable Energy, 37, 318-324.

(48)

48

Januário C., Semino, S., Bell, M., 2007. Offshore Windfarm Decommissioning: A proposal for guidelines to be included in the European Maritime Policy.

Keles, D., Möst, D., Fichtner, W., 2010. The development of the German energy market until 2030 – A critical survey of selected scenarios. Energy policy, 39, 812-825.

Koroneo, C., Dompros, A., Roumbas, G., Moussiopoulos, N., 2004. Life cycle assessment of hydrogen fuel production processes. International Journal of Hydrogen Energy, 29, 1443-1450.

Mulder, M., Schoonbeek, L., 2013. Decomposing changes in competition in the Dutch electricity market through the residual supply index. Energy Economics, 39, 100-107.

Nederlandse ambassade. 2014, Duurzame Energie in Duitsland. Accessed on 19/11/14 via http://duitsland.nlambassade.org/zaken-doen/energie

Nijpels, E. (2014). Energieakkoord voor duurzame groei – voortgangsrapportage 2014. Borgingscommissie Energieakkoord

NWEA (Nederlandse Wind Energie Associatie). 2014. Accessed on 13/11/14 via http://www.nwea.nl/windenergie-de-feiten

Ӧzdemir, Ӧ, Scheepers, M., Seebregts, A., 2008. Future electricity prices – Wholesale market prices in and exchanges between Northwest European electricity markets. ECN, 08-044.

(49)

49

Rijksdienst voor Ondernemend Nederland (RVO). 2014. Berekening Stimulering Duurzame Energieproductie. Accessed on 17/11/14 via http://www.rvo.nl/subsidies-regelingen/berekening-stimulering-duurzame-energieproductie

Rijksoverheid. 2014., Windenergie op zee. Accessed on 19/11/14 via http://www.rijksoverheid.nl/onderwerpen/duurzame-energie/windenergie/windenergie-op-zee

Sioshansi, R., Denholm, P., Jenkin, T., &Weiss, J., 2009. Estimating the value of electricity storage in PJM: Arbitrage and some welfare effects. Energy economics, 31(2), 269-277.

Sterner, M., 2014. Verwachting Duits onderzoek: “Power-to-gas over tien jaar rendabel” – accessed on 06/11/14 via http://www.energiepodium.nl/nieuws/item/verwachting-duits-onderzoek-power-to-gas-over-tien-jaar-rendabel

Taljan, G., Fowler, M., Cañizares, C. Verbič, G., 2008. Hydrogen storage for mixed wind-nuclear power plants in the context of a Hydrogen Economy. International Journal of Hydrogen Energy, 33, 4463-4475.

Tsuchiya, H., Kobayashi, O., 2004. Mass production cost of PEM fuel cell by learning curve.

International Journal of Hydrogen Energy, 29, 985-990.

Veyer, J., 2014. The value of electric energy storage including real options: The case of Eemsdelta. University of Groningen.

(50)

50

Wei, Y., Li, F., &Tomsovic, K., 2012. Measuring the volatility of wholesale electricity prices caused by wind power uncertainty with a correlation model. IET Renewable Power Generation, 6(5), 315-323.

Can offshore wind farms be profitable, once subsidies stop, by storing their generated electricity on existing oil and gas platforms?