• No results found

Hydrogen economy in the Netherlands

N/A
N/A
Protected

Academic year: 2021

Share "Hydrogen economy in the Netherlands"

Copied!
35
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

Hydrogen economy in the Netherlands

Modeling weather patterns, forecasting future renewable energy output

while using hydrogen to store overproduced energy

Thomas van Haastrecht 30th of May 2021

First supervisor: Dhr. dr. ir. J.H. (John) van Boxel Second supervisor: Dr. A. (Albert) Tietema

University of Amsterdam

(2)

Abstract

To limit climate change to less than 2.0°C (1.5°C if possible), as stated in the Paris agreement, a severe reduction of emissions is necessary. In order to do so, more renewable energy capacity is installed. But intermittent renewable energy sources (e.g. solar and wind) can’t supply energy on demand. Over generated electricity that would otherwise damage the grid, could be temporarily stored in the form of hydrogen and used when to renewable energy output is insufficient. The aim of this scientific paper is to solve the mismatch between supply and demand and reduce the amount of curtailed (reduced in efficiency) renewable energy in the Netherlands. A model is developed to simulate different scenario’s, by varying the efficiency of hydrogen and the installed capacity of renewable energy sources. This model is combined with historical and projected climate data to evaluate the different scenarios. In two of the three scenario’s a solely renewable energy grid is possible before 2050, thereby reaching the goals stated in the Paris agreement. In scenario 1, which uses the climate policy of the Netherlands to simulate the energy mix until 2050, a totally-renewable energy grid is possible by the year 2046. With a total installed capacity of 9.1 GW from onshore wind turbines, 24.5 GW of offshore wind turbines; and 41.4 GW of solar energy (PV panels). In this scenario the average annual hydrogen production from 2046-2050 will be 0.72 Mt of hydrogen. This will fulfill energy demand when solar and wind energy output isn’t sufficient. This system will emit zero carbon and would contribute towards a sustainable way of living.

Abstract to the general public written in Dutch

Opslag van hernieuwbare energie in waterstof

Hernieuwbare schone energie is volgens vele de toekomst. De fossiele brandstoffen raken op en de effecten van klimaatverandering moeten worden tegen gegaan. Maar hoe gaan we dit toepassen? Energie opwekken via windmolens en zonnepanelen gebeurt al op grote schaal. Maar het opwekken via deze bronnen is heel onvoorspelbaar. Daarbij gaat de vraag naar energie gedurende de dag ook alle kanten op. Zo is deze in de middag laag wanneer juist net zonne-energie kan worden opgewekt. Daarom moeten we die energie gaan opslaan op het moment dat er veel van wordt opgewekt. Dit kan bijvoorbeeld in de vorm van waterstof. Deze waterstof kan worden opgeslagen in lege gasvelden onder de Noordzee. Zo kunnen we nog genieten van onze duurzame energie ook als de zon even niet schijnt.

(3)

Table of content

Abstract... 2

Abstract to the general public written in Dutch ... 2

1. Introduction... 4

2. Methods ... 6

2.2 Data collection ... 6

2.3 Wind generated energy ... 6

2.4 Photo-voltaic generated energy... 7

2.5 Hydrogen storage and transport... 8

2.6 Energy mix and demand ... 9

2.7 Forecasting ... 10 2.8 Scenarios ... 10 3. Results ... 11 3.1 calibrating ... 11 3.2 Linear regression... 11 3.3 Scenario’s ... 11 4. Discussion ... 14 4.1 Weather forecasting ... 14

4.2 Assumptions/improvements of the model ... 14

4.3 Scenarios ... 15

5. Conclusion ... 17

6. References ... 18

Appendix A: Technical specifications wind turbine ... 21

Appendix B: Technical specification photo-voltaic module ... 21

(4)

1. Introduction

The production of energy is the backbone of our current civilization. During the 18th century,

agriculture and industry became more efficient due to the use of fossil fuels. Which started a reinforcing feedback between growing population and increasing energy demand that will continue in the 21th century. According to BP energy outlook (2018), the energy demand will grow another 33% by 2040 (as of 2018). As a result, the earth is being depleted from its fossil fuels reserves. Shafiee and Topal (2009) predict that by 2042 the world will be completely depleted from its oil and gas reserves. On top of that, greenhouse gasses (GHG) are emitted when fossil fuels are burned. When GHG, like carbon dioxide, are emitted into the atmosphere, they trap heat. Thereby, causing the global temperature to rise. As a result of the rising temperature, altering precipitation regimes are causing extreme weather events, desertification is making regions uninhabitable, rising sea level is increasing flood risk, etc. (IPPC, 2018; IPPC, 2019).

The world needs to transition away from non-renewable fossil fuels towards clean and renewable alternatives such as hydro, solar photo-voltaic (PV) and wind. As hydro power requires a significant declining slop, this source of power isn’t suited for the Netherlands. In 2019 in the Netherlands 8.6% of the energy production came from renewable sources. The world produced more than three times that with 27.3% (Murdock et al., 2020). Unlike conventional electricity generators, the output of intermittent renewable energy sources (IRES) (e.g. wind and PV) is highly unpredictable (See figure 1; Su et al., 2013; Enappsys, 2021). This causes grid instability as a result of the mismatch between IRES energy output and the energy demand (Zhan et al., 2016). Figure 2 demonstrates the potential of over generating electricity in California. This risk occurs when conventional dispatchable energy resources cannot lower their output any further to balance the high production by IRES (Denholm et al., 2015). In addition, many conventional energy resources have a slow-start process. In order to ramp up for the evening demand, while IRES output is decreasing, over generation is destined to occur. Figure 3 shows that negative prices for energy producers are appearing throughout Europe (Nispel, 2021). This is a consequence of overgeneration and will eventually cause electricity prices for consumers to rise (Fanone et al., 2013). To avoid over generation, IRES output is curtailed; the output of IRES is decreased by deliberately lowering their efficiency or completely disconnecting them from the grid. If curtailment is not sufficient, damage to the grid could occur (Denholm et al., 2015).

(5)

In order to balance energy supply-and-demand and avoid the aforementioned, IRES energy needs to be stored so it can be used when necessary (Zhang et al., 2016). Among other ways of storing energy (e.g. batteries, pumped-hydro), hydrogen is seen as a sustainable cost efficient compound to store enormous amounts of energy. Hydrogen is highly suitable because it is: nontoxic, tasteless, colorless and non-metallic. Furthermore, hydrogen is the most abundant and lightest element, yet the highest energy dense known element and therefor highly combustible (Züttel et al., 2010). When converted to energy, water is the only exhaust product (2 H2+O2 -> 2H2O) (Lubitz & Tumas, 2007). One kg of

hydrogen can store 143 MJ. More than double the energy density of natural gas (53.6 MJ) and triple that of diesel (45.4 MJ) (Mazloomi & Gomes, 2012), at an overall efficiency rate of 44% (Abe et al., 2019).

The aim of this research is to increase the efficiency of IRES by reducing curtailment and avoid negative energy prices for energy producers. On top of that this will stabilize the electricity grid in the Netherlands and unlock the potential of sustainable energy.

Thereby, answering the following research question:

How can energy from intermittent renewable energy sources be stored as hydrogen to balance differences in energy demand and supply in the Netherlands in the 21th century?

With the following sub questions:

- How will solar and wind energy production and energy demand change in the future?

- How much PV panels, wind turbines and hydrogen is needed to facilitate energy demand from solely renewable sources?

- How much hydrogen has to be stored to overcome seasonal differences in renewable energy output?

This research yielded a model to investigate the potential storage of IRES energy in the form of hydrogen. A schematic rendering of the model is shown in figure 4. The storage capacity can facilitate as a mediator to balance the supply and demand of energy in the Netherlands.

(6)

2. Methods

This research will mainly focus on intermittent renewable energy resources (IRES) (e.g. wind and PV) in the Netherlands. It will take a quantitative approach by combining weather data (KNMI), calculating the energy supply and fulfil the energy demand in the Netherlands. A model will be created in Rstudio to analyze potential interventions from hydrogen storage to counter over generation with only limited use of curtailment, thereby, decreasing energy loss and increasing stability and efficiency of the grid.

2.2 Data collection

In order to create the model and predict potential IRES output and energy demand in the Netherlands, the hourly data of windspeed and solar radiation is downloaded from the KNMI (2021), Installed power in the Netherlands from CBS (2021), climate policy of the Netherlands from Planbureau voor de Leefomgeving (Hammingh et al., 2020) and hourly energy demand from the Ennappsys (2021).

2.3 Wind generated energy

To forecast the onshore wind energy output in the Netherlands, the hourly mean wind velocity [FH] (in 0.1 m/s) is taken from the following weather stations; Lelystad, Stavoren, Geulhaven and Lauwersoog, as can be seen in figure 5. All have available data since 1991 (KNMI, 2021). These weather stations are the ones closest located to the 10 largest windfarms in the Netherlands (e.g. Windpark Noordoostpolder, Westermeerwind, Krammer).

To forecast the offshore wind energy output in the Netherlands, the hourly mean wind speed [FH] (in 0.1 m/s) is taken from the following weather stations; AWG-1, P11-B and Europlatform, as can be seen in figure 5. Data is available since 2006, 2009 and 1996 respectively (KNMI, 2021). Data is used from 2010 till 2020. These weather station are the closest to the current and future largest windfarms at sea (e.g. Germini, Borssele, IJmuiden ver).

The model will treat onshore and offshore wind separately, because wind speeds at sea are usually higher and more uniform (Esteban et al., 2011). The hourly mean from the onshore group of weather stations will be calculated in order to compute the total onshore wind energy generated in the Netherlands. The hourly mean from the offshore weather stations will be taken in order to calculate the total offshore wind energy production.

The model will use one type of wind turbine to calculate potential wind energy output; Siemens SWT 4.0-130 (Windpark Gemini, 2021: Technical specifications; Appendix A,). This model is used in the off shore Gemini wind park in the North sea near IJmuiden. In 2015 this was the new high end model and has a rated power output of 4.0 MW, nowadays wind turbines can reach up to 8.0 MW. Older models could go as low as 0.8 MW (CBS, 2021b). So this number is about average. In 2020 the installed onshore wind power in the Netherlands was 4100 MW (CBS, 2021b) which is equivalent to 1025 SWT

4.0-130 wind turbines in the model. And the installed offshore wind power was 2500 MW (CBS, 2021b)

which is equivalent to 625 SWT 4.0-130 wind turbines in the model.

In order to cobvert the wind from the weather station height of 10 meters (WMO, 2021), to the wind turbine hub height of 90 meters, the logarithmic wind profile (1) will be used (Kaimal & Finnigan, 1994).

𝑉

𝑧

=

𝑢∗

𝑘

𝑙𝑛(𝑧 𝑧

⁄ )

0 (1)

u is the wind velocity [m/s], u* is the friction velocity [m/s], k is the Von Kármán constant (0.4 ; Kaimal & Finnigan, 1994), z is the height of wind to explore [m] and z0 is the surface roughness [m]

(7)

Afterwards the mechanical power output (2) is derived from the wind speed at hub height using the following formula (Xia et al., 2012):

𝑃 = 0.5 𝜌 𝐶

𝑝

𝐴 𝑉

𝑧3 (2)

P is power output of the wind turbine generator [W], Cp is the power coefficient, ρ is the air density (1.240 kg/m3, because the Netherlands has a yearly average temperature of 10.5 °C (CBS, 2021a) and

the atmospheric pressure is 101310 Pa at a relative humidity of 70% (Jones, 1977)), A is the cross sectional rotor swept area (m2) and V is the wind speed (m/s).According to the Betz limit a wind

turbine can have a maximum power coefficient of 0.59 (Xia et al., 2012). This research will calibrate the Cp coefficient according to the wind turbine’s technical specifications.

The wind turbine is cut-in at a windspeed of 4 m/s and is cut-out at 30 m/s and the maximal power is generated at the rated wind speed of 12 m/s. Therefore al wind speeds below 4 m/s and above 30 m/s will produce zero power output (Windpark Gemini, 2021).

2.4 Photo-voltaic generated energy

To forecast the PV output in the Netherlands, hourly data of global radiation [Q] (in J/cm2) are used from the following weather stations; De Bilt , De Kooy , Eelde , Maastricht & Vlissingen, as can be seen in figure 5. All have available data since 1991 (KNMI, 2021). As four of these stations can make a rectangle which covers almost all of the Netherlands with De Bilt in the centre, data from these weather stations will give an estimate of the average global radiation in the Netherlands.

The model will use one type of solar panel to calculate potential PV output: Sunceco Poly-crystalline 156 × 156 mm 72pcs 300 Wp (Sunceco, 2017: Technical specifications; Appendix B). The total PV capacity in 2020 in the Netherland was 10213 MW (CBS, 2021b), which is equivalent to 34 million Sunceco solar panels.

To estimate the generated electricity by a photovoltaic system (3) the following formula (Yadav, 2017) is used:

(3) E is the generated energy [kWh], A is the total solar panel area [m2], r is the solar panel efficiency

(15.46%; Sunceco, 2017), H is the hourly solar radiation [W/m2] and PR performance ratio or coefficient

for losses (e.g. inverter losses, cable losses, range from 0.5 to 0.9, default value is 0.75). The performance ratio is calibrated according to historical data provided by the CBS (2021b) from 2019, 2019 & 2020

(8)

2.5 Hydrogen storage and transport

Hydrogen will be used to store the surplus of energy generated by PV and wind. Hydrogen is highly suitable because it is: nontoxic, tasteless, colorless and nonmetallic. Furthermore, hydrogen is the most abundant and lightest element, yet the highest energy dense known element and therefor highly combustible (Züttel et al., 2010). When converted to energy, water is the only exhaust product (2 H2+O2 -> 2H2O) (Lubitz & Tumas, 2007).

But because hydrogen has an extremely low density, it needs to be compressed or liquified in order to be stored efficiently which uses 20% to 40% of the energy content of hydrogen (Edwards et al., 2008). In this model hydrogen is stored in depleted gas fields and will use 30% of the energy content of hydrogen (Abe, et al. 2019).

Electrolysis is one of the most capable, well established techniques to produce hydrogen with an efficiency of 70%-80%. In this model we assume a efficiency of 75%. It uses water (H20) and the only by

product is oxygen (O2). Thereby, electrolysis can utilize DC power generated by IRES (Kumar &

Himabindu, 2019). Pipeline transportation losses are estimated to be around 3% (Bergerson & Lave, 2005). Via small local fuel cells units, hydrogen can be turned into electricity at an efficiency rate up to 85% (Dutton, 2002). The overall efficiency after converting to hydrogen, storing and converting back to electricity is 44%. Lauwersoog Stavoren Geulhaven Lelystad Euro Platform P11-B AWG-1 Maastricht De Bilt De Kooy Vlissingen Eelde

KNMI station to track onshore windspeed KNMI station to track offshore

windspeed

KNMI station to track solar radiation

Figure 5, Map of the Netherland with KNMI stations. These stations will provide hourly data of the onshore and offshore windspeed and solar radiation in the Netherlands.

(9)

2.6 Energy mix and demand

The real time energy mix and demand of one week in the Netherlands is visualized in figure 5 (Enappsys, 2021). The gaps in demand and overproduction can be explained by the European shared electricity net. But in this model, the energy grid of the Netherlands is seen as disconnected from the rest of Europe. To forecast future energy demand, hourly energy demand from the Netherlands is taken from 2016 to 2020 (Enappsys, 2021).

The model will have a fixed baseline (FB) of 4000 MW of conventional power generated by nuclear (500 MW), coal (3000 MW) and waste (500 MW). On top of that the model will have a variable baseline (VB) of conventional power generated by gas and greenhouse combined heat and power which also uses natural gas. Because gas powerplants don’t have a long starting process, their output can be quickly adjusted (Tennet, 2020). This output can vary from 2000 to 17500 MW (Enappsys, 2021), and is able facilitate the highest demand without IRES measured from 2016 to 2020 (Enappsys, 2021). These outputs are manageable by the fuel mix in the Netherlands according to Tennet (2020).

If the output of IRES in combination with the variable and fixed baseline energy production exceeds energy demand, this surplus will be converted to hydrogen and stored. The potential energy (process efficiency loss taken into account) that can be provided by this stored hydrogen in MW per hour is called Hs.

The variable baseline will change every hour by subtracting the fixed baseline output (4000 MW) & the IRES power output (variable) from the demand (variable). If this is less than the minimum output of variable baseline (<2000 MW) the output is 2000 MW and no hydrogen is used. If this is more than the minimum output of variable baseline (≥2000 MW), the output will be the provided by the following formula: VB=Demand-FB-IRES-Hs (Hydrogen (Hs) is used until VB =2000 MW).

Figure 6, fuel mix of the Netherlands 9 May to 16 May 2021. Consisting of greenhouse CHP (heating), coal, gas, solar, wind (onshore), wind (offshore) waste and nuclear (Enappsys, 2021).

Fuel mix of the Netherlands 9 May to 16 May

2021

(10)

2.7 Forecasting

In order to forecast the wind speed and solar radiation a simple linear regression with one variable will be used. The yearly sum over 30 years of data is used. The linear regression model will copy the same data but then yearly increase this to keep the variability and unpredictability of the weather implemented in the model. The offshore windspeed at the North sea is increasing with 0.23% each year according to Siegismund & Schrum (2001). The 11 years of data will be copied and yearly increase with 0.23% until 2050 is reached. The demand will decrease according to Hammingh et al. (2020) by 1% a year as energy savings increase in efficiency and are steadily implemented throughout the Netherlands.

2.8 Scenarios

According to Planbureau voor de Leefomgeving (Hammingh et al., 2020), the Netherlands has set out a climate policy for 2030 to increase the offshore installed capacity to 11 GW. Thereby an increase of onshore wind turbine installed power to 6 GW is set out in this policy. On top of that, the photo voltaic installed capacity will increase to 26 GW by 2030. In addition, the 3 GW capacity of coal will be zero by the year 2030. Furthermore, projections have been made that the energy demand will be decreasing by 1% each year, due to energy saving measurement (Hammingh et al., 2020). Scenario 1 is representing this climate policy and runs the same policy until 2050. The model inputs for all the scenarios can be found in table 1.

Scenario 2 uses the same values for the increasing installed capacity as scenario 1. But in this case the overall efficiency of hydrogen is 30% and the efficiency will increase with 1% each year.

Scenario 3 represents a 25% outperform of the installed capacity increase set out by the climate policy of the Netherlands. Thus increases in installed power are 13.75 GW, 7.5 GW and 32,5 GW by 2030 for offshore wind turbines, onshore wind turbines and PV panels.

Scenario 1 Scenario 2 Scenario 3

Installed power offshore [yearly] 848 MW 848 MW 1124 MW

Installed power onshore [yearly] 192 MW 192 MW 340 MW

Installed power PV [yearly] 1578.7 MW 1578.7 MW 2228.7 MW

Coal output to zero [year] 2030 2030 2030

Gas minimum output to zero [year] 2040 2040 2040

Demand increase/decrease [yearly] -1% -1% -1%

Hydrogen overall efficiency [start] 44% 30% 44%

Hydrogen overall efficiency increase/decrease [yearly]

2% 1% 2%

(11)

3. Results

In this section the results of the projected climate data is given and the outputs of each scenarios is presented.

3.1 calibrating

In order to calculate the power generated by the PV panels, the performance ratio (PR) of the installed PV panels had to be calibrated. By rewriting the PV power equation to solve PR, installed capacity (A), the total yearly power (E) combined with the amount of solar radiation could give the PR (PR = E/(A*r)). This was done for the years 2018, 2019 and 2020, which gave a PR of 0.695 on average.

To calibrate the Cp coefficient for the wind turbine model ‘SWT 4.0-130’ which reaches its peak power output of 4MWh at 12 m/s, the wind power equation has to be rewritten(𝐶𝑝= 𝑃/𝜌/ 𝐴/ 𝑉𝑧3∗ 2). The power coefficient is 0.281.

3.2 Linear regression

By calculating a linear regression over the data period 1991-2020, the future onshore windspeed and solar radiation can be projected until 2050. For both regressions the assumptions of linearity, normality & heteroscedastic are met.

The linear regression for onshore total yearly windspeed showed that there is a decreasing amount of total yearly windspeed of -0.34% (pval= 4.26e-5).

The linear regression for total yearly solar radiation concluded that there is a increasing amount of total yearly solar radiation of 0.37% (pval= 3.356e-5).

3.3 Scenario’s

In scenario 1 the total energy demand of the Netherlands is completely facilitated by renewable energy sources from the year 2046, as can be seen in figure 7. With the only exception being 2047 where 0.15 TWh needs to be provided by gas. But as the model has a buffer capacity of 2.4 TWh this could easily be dealt with. The total installed capacity is 9092 MW, 24548 MW and 41400 MW, for onshore wind turbines, offshore wind turbines and PV panels. The amount of solar energy from 2035 onwards is more than that of the onshore and offshore power output combined. The overall hydrogen efficiency is 73.6% in 2046. The yearly average amount of hydrogen that is used to supplement the IRES output from 2046 until 2050 is 23.87 TWh. Which is equivalent to 0.72 Mt of hydrogen (Abe, et al. 2019). In 2050 a total of 1.56 Mt of hydrogen is produced. In figure 8 is the energy mix of a 5 day period in the Netherlands displayed. The red line is the energy demand of the Netherland, which is in general lower during the morning and afternoon and is increasing in the evening. The brown color is representing the remaining fixed baseline without coal. During the middle of the day the solar energy (yellow) is peaking at around 2 o’clock. The output is of all renewable energy is highly fluctuating even though these are 5 days in a row. The purple area represents the harvested offshore wind. Even though the offshore installed capacity is about 2.5 times more than the onshore capacity, the output could tenfold, as can be seen on 27 until 29 May of 2030. All the areas above the demand line (268.3 GWh) represents overproduced energy, which is converted to hydrogen. After periods of overproduction there is a green area of hydrogen that is replacing the gas output (gray).

Scenario 2 has a lower starting overall efficiency of hydrogen and has 1% of efficiency increase each year. Nonetheless, gas energy production is brought down to 2.7 Twh in 2050 coming from 56.6 TWh in 2021, as can be seen in figure 9. Looking at the drop in gas use from 2049 (8.25 TWh) to 2050 (2.7 Twh), it is likely that shortly after 2050, gas won’t be needed at all. Moreover, the model has a

(12)

hydrogen buffer of 2.4 TWh which would bring gas use to just 0.3 TWh in 2050. The total installed capacity of onshore wind turbines, offshore wind turbines and PV panels, are 9860 MW, 27940 MW and 57592 MW respectively. In 2050 the overall efficiency of hydrogen will be 40% and a total of 0.59 Mt of hydrogen is produced. Even though this efficiency is still lower than the current theoretical efficiency of 44%, hydrogen is almost at the yearly production level that is required in scenario 3. The 5 day course of the energy demand can be seen in figure 10. This is fairly similar to figure 8 but with a significant smaller green area (hydrogen) since fewer hydrogen is generated. Nonetheless, the supply of energy is at some points more than double the demand. If there no hydrogen to deal with this amount of energy, it cause serious damage to the grid.

In Scenario 3, total energy demand of the Netherlands is completely facilitated by renewable energy sources from the year 2042 onwards, as can be seen in figure 11. The total installed capacity of onshore wind turbines, offshore wind turbines and PV panels, are 11580 MW, 27228 MW and 59263 MW respectively. The total hydrogen efficiency is 68%. The yearly average amount of hydrogen that is used to supplement the IRES output from 2042 until 2050 is 20.16 TWh. Which is equivalent to 0.60 Mt of hydrogen (Abe, et al. 2019).This is lower than in scenario 1, as there is more installed solar and wind capacity, less hydrogen is needed to supplement this output. From 2045 onward, the yearly amount of hydrogen needed has found an equilibrium. However, as the demand is still decreasing this amount slowly becomes less. Nonetheless, the produced amount of hydrogen is still increasing. In 2050 a total of 1.88 Mt of hydrogen is produced, which is more than triple the amount needed to fulfil the demand. In figure 12 is the energy mix of a 5 day period in the Netherlands displayed. As expected the peaks in power output are about 25% higher than in scenario 1 and 2. All the areas above the demand line (442.2 GWh) represents overproduced energy, which is converted to hydrogen. This is almost double the amount of hydrogen that is produced compared to scenario 1 and 2. After periods of over generation, the green area (hydrogen) that is replacing the gas output is larger than in scenario 1 and more than triple the amount in scenario 2.

0 10000 20000 30000 27 28 29 30 31 P o w e r [MW ] Day

Energy mix scenario 1

27 May- 31 May 2030

Nuclear & Waste Offshore wind Onshore wind PV Gas Hydrogen Demand 0 50 100 150 2021 2023 2025 2027 2029 2031 2033 2035 2037 2039 2041 2043 2045 2047 2049 P o w er [ TW ] Year

Energy mix scenario 1

Nuclear & Waste Coal Gas

Hydrogen Offshore wind Onshore wind PV

Figure 7, yearly sum of the forecasted energy mix of the Netherlands; scenario 1.

Figure 8, the hourly energy mix of a 5 day period in the Netherlands; scenario 1. From the 27th of May till 31st of May, 2030.

(13)

0 50 100 150 2021 2023 2025 2027 2029 2031 2033 2035 2037 2039 2041 2043 2045 2047 2049 P o w e r [T W ] Year

Energy mix scenario 2

Nuclear & waste Coal Gas

Hydrogen Offshore wind Onshore wind PV

Figure 9, yearly sum of the forecasted energy mix of the

Netherlands; scenario 2. Figure 10, the hourly energy mix of a 5 day period in the Netherlands; scenario 2. From the 27th of May till 31st of May, 2030.

Figure 11, yearly sum of the forecasted energy mix of

the Netherlands; scenario 3. Figure 12, the hourly energy mix of a 5 day period in the Netherlands; scenario 2. From the 27th of May till 31st of May, 2030. 0 10000 20000 30000 27 28 29 30 31 P o w e r [MW ] Day

Energy mix scenario 2

27 May- 31 May 2030

Nuclear & waste Offshore wind Onshore wind PV Gas Hydrogen Energy demand 0 50 100 150 200 2021 2023 2025 2027 2029 2031 2033 2035 2037 2039 2041 2043 2045 2047 2049 Po w er [T W ] Year

Energy mix scenario 3

Nuclear & waste Coal Gas

Hydrogen Offshore wind Onshore wind PV 0 10000 20000 30000 27 28 29 30 31 P o w er [ M W ] Day

Energy mix scenario 3

27 May- 31 May 2030

Nuclear & waste Offshore wind Onshore wind PV

Gas Hydrogen

(14)

4. Discussion

In this section the results will be analyzed. Thereby, certain assumptions of the model and their impact on the results will be discussed. Recommendations to improve the model and if further research is necessary are presented.

4.1 Weather forecasting

In the results, a significant linear increase in the annual sum of solar radiation in the Netherlands is calculated. This does not mean the sun gets brighter. One explanation could be that there are less clouds during the day, as they prevent solar radiation from reaching the earth surface (Li et al., 1995). But, in this case it is most likely that a reduction of aerosols in the atmosphere due to a decrease in anthropogenic emissions, also known as the brightening effect, is causing more solar radiation to hit the earth surface above the Netherlands (Beelen & Delden, 2012). The significant rise in solar radiation after 1980 is consistent with direct and indirect effects of aerosols. For instance by the decline of Sulphur dioxide, which has dropped from a yearly average of 20 µ gm−3 in 1980 to about 2 µ gm−3 in

2010 (Van Beelen & Delden, 2012). These particles will find a new equilibrium, it is therefore unlikely that the declining solar radiation trend will continue until 2050. Furthermore, the linear regression does not explain any increases or decreases in variance or weather extremities. The linearity only takes the annual sum of solar radiation into account. Thereby, the model uses data from the past 30 years with a yearly increase, to forecast the solar radiation until 2050. It could therefore be that on a clear sunny day the theoretical maximum amount of solar radiation already reaches the earth surface but as the linear regression increases even the maxima, the new maximum would be higher and theoretically impossible (Da Rosa & Ordonez, 2021).

Secondly, the result present a decrease in the annual sum of onshore windspeed. This phenomena can be explained by an increasing surface roughness (Wieringa & Rijkoort, 1983). As more buildings enter the Dutch landscape, the surface roughness will increase. Therefore, the windspeed that is measured at 10 m above the surface is decreasing (CBS, 2021c). As the linear regression only takes the annual sum of windspeed into account, nothing can be said about variance or weather extremities and what their impact could be on the amount of power that is generated.

At last, a study conducted in 2001 by Siegismund & Schrum, analyzed wind speed patterns at the North sea. This study took data from 1958 until 1997 and thus only covers 4 decades. Even though, this is 33% more than the 30 years of data used to determine the linear regression of solar radiation and onshore windspeed, this amount of data is still too small to detect long term patterns nor any causes for the trends.

4.2 Assumptions/improvements of the model

In the constructed model certain assumptions have been made to be able to make calculation. The list below covers these assumption and also recommends improvements to the model:

- The model uses only one type of wind turbine and photo voltaic panel. This makes it less complicated to use the formulas, as there is only one list of specification that have to be inserted in these formulas. Any innovations in efficiency or power output is left out. But because the model uses installed power as guideline and then calculates how much wind turbines or PV panels it takes to fulfill this installed power, these factors don’t play a role. Especially when considered that the model isn’t measuring ground space to put wind turbines and PV panels as a limiting factor.

- The model uses data from the climate policy of the Netherlands in the scenarios (Hammingh et al., 2020). This data consist of plans to expand the amount of onshore/offshore wind

(15)

turbines, PV panels and is forecasting a decrease in energy demand. But as these plans are set out until 2030, the model assumes the same progress will be made after 2030 until 2050. - The model assumes a certain amount of annual innovation considering the efficiency in

converting renewable energy to hydrogen, transporting the hydrogen, storing the hydrogen and converting the hydrogen back to electricity. This assumption is based on efficiency increases in other technologies like solar cells. Solar cells have experienced a major improvement in efficiency over the years. From 1983 until 2008 the silicon solar cell has seen an efficiency increase of 57%. In the same period the price of solar panels has dropped tremendously. This can be explained by the wide acceptance and enormous input of funding in this technique (Green, 2009). If hydrogen could prove work of concept on a large scale, the momentum for this technique could increase and the overall efficiency of hydrogen could await the same efficiency boost as solar cells. Because hydrogen has to be created, transported, stored and turned back to electricity, there are multiple stages to lose the energy. But for the same reason there are many stages to improve the efficiency.

- In order to calculate the output of PV panels the model uses a fairly simple formula. It only uses the solar radiation, efficiency and size of the panel for the calculation. In reality the solar panel will only reach peak power output at a perfect combination of panel temperature and the amount of solar radiation. Increasing panel temperature decreases PV power output (Dubey et al, 2013). This parabolic curve isn’t implemented in the formula and therefor the panel will almost always produce less power than is calculated with this formula. But in order to counter this phenomena to some extent, the performance ratio was calibrated using data of the annual PV output and installed capacity within that year.

- In addition, the windspeed is measured at a 10 meter height and via the logarithmic wind profile it is converted to the windspeed at a 90 meter height. This formula can strictly, only be used under neutral atmosphere. The logarithmic wind profile is a close approximation of the windspeed at 90 meter height but with some degree of error (Kaimal & Finnigan, 1994). - On top of that, when calculating the power output of the wind turbine at certain a windspeed

it would be best to know the power coefficient (Cp) curve for the specific wind turbine. The model uses a factor instead of a curve and is therefore approximating the power output linearly under 12 m/s (Sedaghat, et al., 2017).

- At last, after hydrogen is produced during stages of energy over generation, the model immediately uses all of its hydrogen to replace the gas output. After the hydrogen is used, gas will have to ramp up production to facilitate the energy demand. Therefore, the installed capacity of gas powerplants cannot be lowered over the years. Even one year before becoming completely reliant on renewable energy in scenario 1, the maximum gas capacity has to be 11 MW. It would be better to replace a certain amount of the gas that is used every hour and divide the usable amount of hydrogen over a longer period so that the standby gas capacity can slowly be brought down to zero.

4.3 Scenarios

In scenario 1 the Netherlands will no longer need gas to fulfil the energy demand in the year 2046 and in scenario 3 in the year 2046. This means that the targets set in the Paris agreement will be met for the year 2050 (Rijksoverheid, 2020). When comparing the total installed onshore capacity of 9.09 GW in scenario 1 , 11.58 GW in scenario 3 to Germany’s installed onshore capacity in 2020 of 54.94 GW (Komusanac et al., 2021), the amount of turbines that have to be added looks manageable.

The total offshore capacity of 24.55 GW in scenario 1 and 27.23 GW in scenario 3 is far greater than Germany’s current installed offshore capacity in 2020, which is 7.69 GW. But when looking at a scenario study of the future of offshore capacity in the north sea by ‘Planbureau voor leefomgeving’ (Mathijssen

(16)

et al., 2018) these amounts are still less than the third in Mathijssen et al. (2018) which involves a rapid development of 32.0 GW installed power by 2050.

In addition, the PV capacity has to be increased to 41.4 GW in scenario 1 and to 59.3 GW in scenario 3. Compare this to Germany who has already an installed PV capacity of 49 GW in 2019 (Fraunhofer, 2020) and the it can be concluded that this should be reachable by the Netherlands within the next 20/25 years. If all rooftop area of the Netherlands is covered with solar panels a capacity of 66 GW will be reached, so considering the trend to create solar farms, 41 GW or 59 GW of installed power is feasible (Hammingh et al., 2020).

In scenario 1 on average 0.72 Mt and in scenario 3 on average 0.6 Mt of hydrogen has to be produced each year. A maximum of 264 ton of hydrogen has to be stored in scenario 1 in 2046 and 259 ton of hydrogen in scenario 3 in 2042. This will take up around 2.9 million m3 (density of hydrogen under

atmospheric pressure 0.09 kg/ m3). The Netherlands has the capacity to store 93 billion m3 in in

depleted gas fields underneath the Netherlands and another 60 billion m3 beneath the Dutch North

sea (Van Gessel et al., 2018). This gives the potential to store 13.77 billion kilo of hydrogen which can produce 46 TWh of energy since the energy density of hydrogen is 33.33 KWh/kg hydrogen (Abe et al., 2019).

In scenario 1 the maximum amount of hydrogen that has to be produced is in one hour is 880 kg. in scenario 3 this amount is 782 kg. More research should be conducted whether this amount of production is feasible. There should also be more research to see if the conversion process is best to be done on site via large fuel cells or at various locations via smaller fuel cells. Since there is a energy loss in transporting energy trough cables especially over longer trajectories, large on site fuel cell and transporting hydrogen, which loses less energy while transporting, appears to be the better option (Abe et al., 2019).

The efficiency of hydrogen is the major uncertainty in this model as innovation can’t be properly forecasted. The 73.6% overall efficiency in scenario 1 and 68% in scenario 3 is reachable, especially when taken into account that the hydrogen doesn’t have to be stored under enormous pressure as there is plenty of storage capacity in the Netherlands. The pressurized storage process took almost 30% of the hydrogens energy, without the need for increased pressure this percentage could definitely be lower, but more research in this field needs to be conducted.

In scenario 2 the overall efficiency of hydrogen is 30% and the yearly increase of hydrogen efficiency is 1%. This scenario could be plausible as it could be challenging to produce this amount of hydrogen at a large scale using only renewable energy. Even though a gas free economy isn’t reached before 2050 it will surely be reached within 5 years as the hydrogen production is steadily increasing (0.60 Mt in 2050) and gas use is decreasing.

(17)

5. Conclusion

The fossil fuel reserves of the world are depleting and the emission burning the remaining fossil fuels to facilitate energy demand, are heavily contributing to climate change. Therefore, renewable energy alternatives are increasing in capacity. But, as these alternatives cannot supply energy on demand, a surplus is energy is generated at some point and on the other hand shortage of renewable energy output have to supplemented by conventional fossil energy sources. Hydrogen could play a major role in stabilizing the grid by using overproduced electricity to generate hydrogen. By modeling a fuel mix model, projecting the climate of the Netherlands until 2050 and simulating three different scenarios, it is fairly feasible that with the current climate policy a gas free and solely renewable energy infrastructure could be in place in the year 2046. With an installed offshore wind power capacity of 24.55 GW, onshore wind power capacity of 9.09 GW and a PV capacity of 41.40 GW. With this capacity an overall efficiency of hydrogen (converting renewable energy to hydrogen, transporting the hydrogen, storing the hydrogen and converting the hydrogen back to electricity) of 73.6% has to be met. This efficiency could be reached if the technique would realize proof of concept and funding is increasing. But even if the overall efficiency reaches only 40% , with an installed offshore wind power capacity, onshore wind power capacity and a PV capacity of 9.86 GW, 27.94 GW and 57.59 GW respectively, a solely renewable energy hydrogen economy will be reached shortly after the year 2050. Depleted gas fields can provide plenty of storage for hydrogen even if when hydrogen isn’t pressurized to increase its energy density per m3. Either way, it would make sense to turn to hydrogen

to store over generated energy at a large scale, since it could otherwise damage the grid and will be curtailed. Especially when over generating electricity is destined to occur when installed renewable capacity is increasing in the near future.

(18)

6. References

Abe, J. O., Popoola, A. P. I., Ajenifuja, E., & Popoola, O. M. (2019). Hydrogen energy, economy and storage: review and recommendation. International Journal of Hydrogen Energy, 44(29), 15072-15086. Bergerson, J. A., & Lave, L. B. (2005). Should we transport coal, gas, or electricity: Cost, efficiency, and environmental implications, Environmental Science & Technology 39: 5905-5910. DOI: 10.1021/es048981t

BP Energy Outlook (2018). BP Energy Outlook. BP plc: London, UK. Retrieved on 26 march 2021, via

https://www.bp.com/content/dam/bp/business-sites/en/global/corporate/pdfs/energy-economics/energy-outlook/bp-energy-outlook-2018.pdf

CBS (2021a). Gemiddelde temperatuur. Centraal bureau statistiek . Retrieved on 22 march, 2021, via:

https://www.cbs.nl/nl-nl/maatschappij/natuur-en-milieu/groene-groei/milieukwaliteit-van-het-leven/indicatoren/gemiddelde-temperatuur

CBS (2021b). Hernieuwbare elektriciteit; productie en vermogen. Centraal bureau statistiek. Retrieved

on 19 march 2021, via:

https://opendata.cbs.nl/statline/#/CBS/nl/dataset/82610NED/table?ts=1616147466174

CBS (2021c). Ruim 69 duizend nieuwbouwwoningen in 2020. Centraal bureau statistiek. Retrieved on 19 May 2021, via:://www.cbs.nl/nl-nl/nieuws/2021/04/ruim-69-duizend-nieuwbouwwoningen-in-2020 Da Rosa, A. V., & Ordonez, J. C. (2021). Fundamentals of renewable energy processes. Academic Press, London UK, San Diego US, 944 pp.

Denholm, P., O'Connell, M., Brinkman, G., & Jorgenson, J. (2015). Overgeneration from solar energy in

california. a field guide to the duck chart (No. NREL/TP-6A20-65023). National Renewable Energy

Lab.(NREL), Golden, CO (United States).

Dubey, S., Sarvaiya, J. N., & Seshadri, B. (2013). Temperature dependent photovoltaic (PV) efficiency and its effect on PV production in the world–a review. Energy Procedia, 33, 311-321.

Dutton, G. (2002). Hydrogen energy technology. Tyndall Centre for Climate Change Research, Norwich, 31.

Edwards, P. P., Kuznetsov, V. L., David, W. I., & Brandon, N. P. (2008). Hydrogen and fuel cells: towards a sustainable energy future. Energy policy, 36(12), 4356-4362.

Enappsys (2021). Webservice data. EnAppSys Ltd. Retrieved on 26 march 2021, via https://app.enappsys.com/#nl/elec/renewables

Esteban, M. D., Diez, J. J., López, J. S., & Negro, V. (2011). Why offshore wind energy?. Renewable

Energy, 36(2), 444-450.

Fanone, E., Gamba, A., & Prokopczuk, M. (2013). The case of negative day-ahead electricity prices. Energy Economics, 35, 22-34.

Fraunhofer (2020). PHOTOVOLTAICS REPORT 2020. Fraunhofer Institute for Solar Energy Systems, ISE. Retrieved on 17th of May, via https://www.ise.fraunhofer.de/en/publications/studies/photovoltaics-report.html

Green, M. A. (2009). The path to 25% silicon solar cell efficiency: History of silicon cell evolution. Progress in Photovoltaics: Research and Applications, 17(3), 183-189.

(19)

Hammingh, P., Menkveld, M., Daniëls, B., Koutstaal, P., Schure, K. & Hekkenberg M. (2020). Klimaat-en Energieverkenning (KEV) 2020. Planbureau voor de Leefomgeving (PBL). Retrieved on 17th of May, via https://www.pbl.nl/publicaties/klimaat-en-energieverkenning-2020

IPCC (2018). Global Warming of 1.5°C.An IPCC Special Report on the impacts of global warming of 1.5°C above pre-industrial levels and related global greenhouse gas emission pathways, in the context of strengthening the global response to the threat of climate change, sustainable development, and efforts to eradicate poverty [Masson-Delmotte, V., P. Zhai, H.-O. Pörtner, D. Roberts, J. Skea, P.R. Shukla, A. Pirani, W. Moufouma-Okia, C. Péan, R. Pidcock, S. Connors, J.B.R. Matthews, Y. Chen, X. Zhou, M.I. Gomis, E. Lonnoy, T. Maycock, M. Tignor, and T. Waterfield (eds.)]. In Press”.

IPCC (2019). Climate Change and Land: an IPCC special report on climate change, desertification, land

degradation, sustainable land management, food security, and greenhouse gas fluxes in terrestrial ecosystems [Masson-Delmotte, V., P. Zhai, H.-O. Pörtner, D. Roberts, J. Skea, P.R. Shukla, A. Pirani, W.

Moufouma-Okia, C. Péan, R. Pidcock, S. Connors, J.B.R. Matthews, Y. Chen, X. Zhou, M.I. Gomis, E. Lonnoy, T. Maycock, M. Tignor, and T. Waterfield (eds.)]. In Press”.

Jones, F. E. (1977). The air density equation and the transfer of the mass unit. National Bureau of Standards.

Kaimal, J. C., & Finnigan, J. J. (1994). Atmospheric boundary layer flows: their structure and

measurement. Oxford university press.

KNMI(2021). Uurgegevens van het weer in Nederland. Koninklijk Nederlands meteorologisch intistuut. Retrieved on 18 march 2021, via: https://knmi.nl/nederland-nu/klimatologie/uurgegevens

Komusanac, U., Brindley, G., Fraile, D., & Ramirez, L., (2021). Wind energy in Europe 2020. WindEurope

Business Intelligence

Kumar, S. S., & Himabindu, V. (2019). Hydrogen production by PEM water electrolysis–A review. Materials Science for Energy Technologies, 2(3), 442-454.

Li, Z., Barker, H. W., & Moreau, L. (1995). The variable effect of clouds on atmospheric absorption of solar radiation. Nature, 376(6540), 486-490.

Lubitz, W., & Tumas, W. (2007). Hydrogen: an overview. Chemical reviews, 107(10), 3900-3903.

Matthijsen, J., Dammers, E. & Elzenga, H. (2018). The future of the North Sea The North Sea in 2030 and 2050: a scenario study.Planbureau voor de Leefomgeving (PBL). Retrieved on 17th of May, via https://www.pbl.nl/sites/default/files/downloads/pbl-2018-the-future-of-the-north-sea-3193.pdf Mazloomi, K., & Gomes, C. (2012). Hydrogen as an energy carrier: Prospects and challenges. Renewable

and Sustainable Energy Reviews, 16(5), 3024-3033.

Murdock, H. E., Gibb, D., André, T., Sawin, J. L., Brown, A., Appavou, F., ... & Mastny, L. (2020). Renewables 2020-Global status report. Retrieved on 26 march 2021, via https://www.ren21.net/wp-content/uploads/2019/05/gsr_2020_full_report_en.pdf

Nispel, A. (2021). Negative prices sweep across Europe. EnnAppSys Energy insights. Retrieved on March 15, 2021, via https://www.enappsys.com/negative-prices-sweep-across-europe/

Rijksoverheid (2020). Klimaatplan 2021-2030. Rijksoverheid. Retrieved on 18th of May 2021,

(20)

Sedaghat, A., Hassanzadeh, A., Jamali, J., Mostafaeipour, A., & Chen, W. H. (2017). Determination of rated wind speed for maximum annual energy production of variable speed wind turbines. Applied

energy, 205, 781-789.

Shafiee, S., & Topal, E. (2009). When will fossil fuel reserves be diminished? Energy policy, 37(1), 181-189.

Siegismund, F., & Schrum, C. (2001). Decadal changes in the wind forcing over the North Sea. Climate

Research, 18(1-2), 39-45.

Su, W., Wang, J., & Roh, J. (2013). Stochastic energy scheduling in microgrids with intermittent renewable energy resources. IEEE Transactions on Smart grid, 5(4), 1876-1883.

Sunceco (2017). 300 W – 320 W Poly-crystalline Solar Module. Sunceco. Retrieved on 19 march 2021, via http://sunceco.com/wp-content/uploads/2017/01/SEP300-320.pdf

TenneT (2020). Rapport Monitoring Leveringszekerheid. TenneT TSO B.V. Retrieved on 26 march 2021, via

https://www.tennet.eu/fileadmin/user_upload/Company/Publications/Technical_Publications/Dutch/Ra pport_Monitoring_Leveringszekerheid_20JAN2021.pdf

van Beelen, A. J., & van Delden, A. J. (2012). Cleaner air brings better views, more sunshine and warmer summer days in the Netherlands. Weather, 67(1), 21-25.

Van Gessel, S.F., Breunese, J., Juez Larré, J, Huijskes, T.D., & Remmelts, G. (2018). Ondergrondse Opslag in Nederland. TNO.

Wieringa, J., & Rijkoort, P. J. (1983). Windklimaat van Nederland. Staatsuitgeverij Den Haag.

Windpark Gemini (2021). The Gemini Offshore Wind Park. Retrieved on 18 march 2021, via https://www.geminiwindpark.nl/

WMO(2021). CHAPTER 5. MEASUREMENT OF SURFACE WIND. World meteorological organization. Retrieved on 18 march 2021, via: https://library.wmo.int/doc_num.php?explnum_id=3177

Xia, Y., Ahmed, K. H., & Williams, B. W. (2012). Wind turbine power coefficient analysis of a new maximum power point tracking technique. IEEE transactions on industrial electronics, 60(3), 1122-1132. Yadav, S. (2017). Calculations of solar energy output. Academia. Retrieved on 26 march 2021, via https://www.academia.edu/9005661/CALCULATIONS_OF_SOLAR_ENERGY_OUTPUT

Zhang, F., Zhao, P., Niu, M., & Maddy, J. (2016). The survey of key technologies in hydrogen energy storage. International Journal of Hydrogen Energy, 41(33), 14535-14552.

Züttel, A., Remhof, A., Borgschulte, A., & Friedrichs, O. (2010). Hydrogen: the future energy carrier. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering

(21)

Appendix A: Technical specifications wind turbine

Manufactures Siemens

Model SWT 4.0 -130

Rated power 4 MW

Cut-in wind speed 4 m/s

Rated wind speed 12 m/s

Cut-out wind speed 30 m/s

Rotor diameter 130 meter

Rotor swept area 13.274 m2

Hub Height 90 meter

Blade length 63 meter

Source: Windpark Gemini (2021)

Appendix B: Technical specification photo-voltaic module

Source: Sunceco (2017)

Manufactures Sunceco

Model 300W solar module

Cell dimensions 156x156mm

Total cells 72

Cell efficiency 17,46%

Module efficiency (used in formula) 15,46%

Operating temperature -40 °C to 85 °C Temperature Coefficient of Pmax (>25°C) - 0.45 % / °C

(22)

Appendix C: Rstudio model

All used data and the model is uploaded to fig share under Thomas van Haastrecht, Hydrogen economy in the Netherland Bsc thesis.

#Hydrogen economy in the Netherlands

# This model will calculates the potential future overproduction of renewable energy in the # Netherlands. The model stores this otherwise curtailed energy in the form of hydrogen, so that it # can be used at a later time.

# The model uses data from the past 30 years from the KNMI to forecast future Solar radiation and # onshore windspeed. Thereby, data from the past 10 years of offshore windspeed is used and # Siegismund, F., & Schrum, C. (2001) provides inside on trends in this offshore windspeed. # The Demand in the Netherland is from the past 5 years retrieved from ENappsys, where there is # no significant trend noticeable.

#Loading data Library (readxl)

YYMDHH_2021_2050 <- read_excel("Future Planet Studies/Thesis/Modelling/Clean/YYMDHH 2021-2050.xlsx")

Demand15_20 <- read_excel("Future Planet Studies/Thesis/Modelling/Clean/Demand15-20.xlsx") Hydrogen_potential<- read_excel("Future Planet

Studies/Thesis/Modelling/Clean/Hydrogen_potential.xlsx")

Windspeed_On <- read_excel("Future Planet Studies/Thesis/Modelling/Clean/Windspeed On.xlsx") MonthsHH_Onshore <- read_excel("Future Planet Studies/Thesis/Modelling/Clean/MonthsHH.xlsx") Windspeed_Off <- read_excel("Future Planet Studies/Thesis/Modelling/Clean/Windspeed Off.xlsx") YYMDHH_2021_2053 <- read_excel("Future Planet Studies/Thesis/Modelling/Clean/YYMDHH 2021-2053.xlsx")

Solar_radiation_ <- read_excel("Future Planet Studies/Thesis/Modelling/Clean/Solar radiation .xlsx") MonthsHH_solar<- MonthsHH_Onshore Gas_output<-Hydrogen_potential Gas_hyd<-Hydrogen_potential Hyd_stored<-Hydrogen_potential Hydrogen_consumed<-Hydrogen_potential YYMDHH_2021_2050onW <- YYMDHH_2021_2050

(23)

YYMDHH_2021_2050PV <- YYMDHH_2021_2050

#Wind Variables

#Technical specification wind turbine (Windpark Gemini, 2021)

Sweptarrea=13274 #crossectional roto swept area [m^2]

Windspeedmaxp= 12 #Windspeed when maximum power output can be generated [m/s]

Maxp=4e6 #Maximum power output of the wind turbine [W]

Hub_height=90 #Hub height win turbine [m]

lowerboundwindturbine= 4 #Minimum windspeed for the turbine to start rotating [m/s]

upperboundwindturbine= 30 #Maximum windspeed, at this windspeed the turbine is shut off [m/s]

#Installed power

WindturbinesNLsea= 2.5e9/Maxp #Amount of turbines, installed power 2.5 GW in 2020 (CBS, 2021)

WindturbinesNLland= 4.1e9/Maxp #Amount of turbines, installed power 4.5 GW in 2020 (CBS, 2021)

Maxtotalpsea= WindturbinesNLsea*Maxp/1e6 #Installed power [MW]

Maxtotalpland= WindturbinesNLland*Maxp/1e6 #Installed power [MW]

land_inc=round((6000-Maxtotalpland)/(Maxp/1e6)/10, digits = 0) #YEARLY. Plans to increase installed power to 6 GW by 2030 (Abels-van Overveld et al., 2020)

sea_inc= round((11000-Maxtotalpsea)/(Maxp/1e6)/10, digits = 0) #YEARLY. Plans to increase installed power to 11 GW by 2030 (Abels-van Overveld et al., 2020)

Inc_wind_year =2050 #Untill which year will the installed power increase

#Equations input

Air_density=1.24 #kg/m^3

surface_rough_land= 0.15 #Surface roughness on land [m]

surface_rough_sea= 0.0002 #Surface roughness on sea [m]

Von_Karman_constant= 0.4 #Von Kármán constant (Kaimal & Finnigan, 1994)

(24)

#calculating constants for converting windspeed measured at 10 meter to 90 meter for land and sea convert90land=Von_Karman_constant*1/log(Height_measurement/surface_rough_land)/Von_Karman_c onstant*log((Hub_height/surface_rough_land)) convert90sea= Von_Karman_constant*1/log(Height_measurement/surface_rough_sea)/Von_Karman_constant*log((Hu b_height/surface_rough_land))

#Calculating the power coefficient of the windturbine

# Maxp is reached at Windspeedmaxp, Cp can be calculated when rewriting the function: # P=0.5*Air_density*Cp*Sweptarrea*windspeed^3 (Xia, Ahmed & Williams, 2012)

Cp=Maxp/Air_density/Windspeedmaxp^3/Sweptarrea*2

#calculate annual offshore windspeed increase (Siegismund & Schrum, 2001)

Vwind1958=7.4 #Average windspeed at the north sea in 1958 [m/s] Vwind1997=8.1 #Average windspeed at the north sea in 1997 [m/s]

Yearly_offwindspeed_inc=(Vwind1997/Vwind1958)^(1/40) #Yearly increase in windspeed at the North sea

#PV Variables

#specifications solar panel (Sunceco, 2017)

Effiency_PV= 0.1546 #Solar panel efficiency

A=1.75 #Total solar panel area [m^2]

maxpsolar= 300 #Maximum power output [w]

#Installed power

InstalledpowerNL= 1.0213e10 #Installed power in 2020 [W] (CBS, 2021)

Solarpanels= InstalledpowerNL/maxpsolar #Amount of solar panels

panel_inc=round((2.6e10-InstalledpowerNL)/10/maxpsolar, digits = 0) #YEARLY. Plans to increase installed power to 26 GW by 2030 (Abels-van Overveld et al., 2020)

(25)

degrees30= 15.51/13.78 #KNMI horizontal surface solar radiation data to a 30 degrees slope angle

Wm=10000/3600 #changing the unit of measurement from J/cm^2 to W/m^2

#Calibrating the performance ratio (e.g. inverter losses, cable losses, range from 0.5 to 0.9) # Data is gathered from CBS; Total power output and installed power (data from

2018,2019,2020) (CBS, 2021)

Solar2020 <- Solar_radiation_$averagewm30[Solar_radiation_$Year == '2020'] #Taking the average solar radition [W/m^2] of De Bilt , De Kooy , Eelde , Maastricht & Vlissingen of 2020

Solarpanels2020= InstalledpowerNL/maxpsolar #Amount of solar panels

production2020CBS = 8127 #Total PV power output in 2020 [GW]

# PR can be calculated when rewriting the function: # E=PR*A*Solarpanels*Effiency_PV*Solar radiation PR2020= production2020CBS/sum(A*Solarpanels2020*Solar2020*Effiency_PV/1e9) #Same for 2019 solar2019 <- Solar_radiation_$averagewm30[Solar_radiation_$Year == '2019'] Solarpanels2019= 6874*1000000/maxpsolar production2019CBS = 5700 PR2019= production2019CBS/sum(A*Solarpanels2019*solar2019*Effiency_PV/1e9) #Same for 2018 solar2018 <- Solar_radiation_$averagewm30[Solar_radiation_$Year == '2018'] Solarpanels2018= 4609*1000000/maxpsolar production2018CBS = 3700 PR2018= production2018CBS/sum(A*Solarpanels2018*solar2018*Effiency_PV/1e9)

#Taking the average PR

PR=(PR2020+PR2019+PR2018)/3

(26)

#Energy mix

FB=4000 #Fixed baseload nuclear, waste & coal [MW]

VBmin=2000 #Minimum variable baseload gas [MW]

VBmax=17500 #Maximum variable baseload gas [MW]

storage=0 #starting value of hydrogen in storage

Buffer=2.4e6 #hydrogen buffer before starting implementing hydrogen in energy mix [MW]

efficiency_hydrogen=0.44 #efficiency converting to hydrogen, storing and converting to electricity

FB_decrease=300 #Yearly decrease in FB output [MW]

VB_decrease=100 #Yearly decrease in VBmin output [MW]

Demand_increase=0.99 #Yearly increase in demand

hyd_eff_increase = 1.03 #Yearly increase in hydrogen efficiency

#Onshore windspeed model

# Taking the average from the 4 stations (Lelystad, Stavoren, Geulhaven and Lauwersoog)

# and changing the unit of measurement from 0.1 to m/s

#I HAVE ALREADY DONE THIS IN THE EXCELL FILE

#Windspeed_On$Averagems=(Windspeed_On[5]+Windspeed_On[6]+Windspeed_On[7]+Windspeed_On[ 8])/4/10

#Determine linear regression of decrease in onshore windspeed

b=vector() vary=1990

future=c(2021:2050) Years<- c(1991:2020)

#Talking the sum of the hourly windspeed of the average taken from 4 weather station #(Lelystad, Stavoren, Geulhaven and Lauwersoog) for each year of data.

while (vary < 2020){ vary=vary+1

(27)

X=sum(X) b=c(b,X) }

#calculating the Yearly decrease of windspeed coefficient

mod=lm(b~Years) forecast=mod$coefficients[2]*future+mod$coefficients[1] Yearly_decrease_wind=forecast[2]/forecast[1] #plot(mod,1) #Heteroscedesity #hist(b) #normality #shapiro.test(mod$residuals) #normality #plot(b) #linearity

#Forecasting onshore windspeed untill 2050 by repeatingt the 30 years of data with the decrease

#factor ^30 and converting the windspeed to 90 meters

YYMDHH_2021_2050onW[5]= Windspeed_On$Averagems*Yearly_decrease_wind^30*convert90land

#[m/s]

names(YYMDHH_2021_2050onW)[5]<- "Forecast"

#Implementing wind turbine boundaries

YYMDHH_2021_2050onW$Forecast[YYMDHH_2021_2050onW$Forecast<lowerboundwindturbine|YYM DHH_2021_2050onW$Forecast>upperboundwindturbine]<-0

#Implenting a potential wind turbine increase, calculating the total power generated and making

#sure there is no more power generated than the Maxtotalpland

while(vary<2050){

if (vary<Inc_wind_year){WindturbinesNLland=WindturbinesNLland+land_inc Maxtotalpland=WindturbinesNLland*Maxp/1e6}

(28)

YYMDHH_2021_2050onW$Forecast[YYMDHH_2021_2050onW$Year == vary]=YYMDHH_2021_2050onW$Forecast[YYMDHH_2021_2050onW$Year == vary]^3*0.5*Air_density*Sweptarrea*Cp/1e6*WindturbinesNLland #[MW]

YYMDHH_2021_2050onW$Forecast[YYMDHH_2021_2050onW$Year == vary & YYMDHH_2021_2050onW$Forecast>Maxtotalpland]<-Maxtotalpland #[MW]

}

#Offshore windspeed model

#Taking the average of the 3 weather stations(AWG-1, P11-B and Europlatform)

# I HAVE ALREADY DONE THIS IN THE EXCELL FILE

# Windspeed_Off$`Average windspeed`[8]=( Windspeed_Off[5]+ Windspeed_Off[6]+ Windspeed_Off[7])/3

Offshore_windspeed2010_2020=Windspeed_Off$`Average windspeed`

#Forecasting onshore windspeed untill 2053 by repeatingt the 11 years of data 3 times with the #increase factor ^11, ^22, ^33, changing the windspeed to m/s and converting the windspeed to 90 meters

Offshore_windspeed2021_2053=c((Offshore_windspeed2010_2020*Yearly_offwindspeed_inc^11),(Offsh ore_windspeed2010_2020*Yearly_offwindspeed_inc^22),(Offshore_windspeed2010_2020*Yearly_offwin dspeed_inc^33))

YYMDHH_2021_2053$Windspeedms<- Offshore_windspeed2021_2053/10*convert90sea #[m/s]

#Implementing wind turbine boundaries

YYMDHH_2021_2053$Windspeedms[YYMDHH_2021_2053$Windspeedms<lowerboundwindturbine|YY MDHH_2021_2053$Windspeedms>upperboundwindturbine]<-0

#Implenting a potential wind turbine increase, calculating the total power generated and making

#sure there is no more power generated than the Maxtotalpland

vary=2020 varyear=1990 while(vary<2050){

(29)

if (vary<Inc_wind_year){ WindturbinesNLsea=WindturbinesNLsea+sea_inc Maxtotalpsea=WindturbinesNLsea*Maxp/1e6} vary=vary+1 varyear=varyear+1 YYMDHH_2021_2053$Energy_Off_wind[YYMDHH_2021_2053$Year == vary]=YYMDHH_2021_2053$Windspeedms[YYMDHH_2021_2053$Year == vary]^3*0.5*Air_density*Sweptarrea*Cp/1000000*WindturbinesNLsea #[MW]

YYMDHH_2021_2053$Energy_Off_wind[YYMDHH_2021_2053$Year == vary & YYMDHH_2021_2053$Energy_Off_wind > Maxtotalpsea]<-Maxtotalpsea #[MW]

}

#Solar Radiation Model

#Taking the average from the 5 stations( De Bilt , De Kooy , Eelde , Maastricht & Vlissingen), #changing the unit of measurement from J/cm^2 to W/m^2

#and calibrating from a horizontal surface to a 30 degrees slope angle

#I HAVE ALREADY DONE THIS IN THE EXCELL FILE

# a=0 # while (a<length(Solar_radiation_$Year)){ # a=a+1 # Solar_radiation_$averagewm30[a]=(Solar_radiation_[a,5]+Solar_radiation_[a,6]+Solar_radiation_[a,7]+S olar_radiation_[a,8]+Solar_radiation_[a,9])/5*wm*degrees30 # }

#Determine linear regression of yearly increase in solar radiation

b=vector() vary=1990

while (vary < 2020){ vary=vary+1

(30)

X=Solar_radiation_$averagewm30[Solar_radiation_$Year == vary] X=sum(X)

if (vary == 1992|vary == 1996|vary == 2000| vary == 2004|vary == 2008|vary == 2012|vary == 2016|vary == 2020){

X=X-sum(Solar_radiation_$averagewm30[Solar_radiation_$Year == vary & Solar_radiation_$Month == 2 & Solar_radiation_$Day == 29]) } b=c(b,X) } mod=lm(b~Years) forecast=mod$coefficients[2]*future+mod$coefficients[1] Yearly_increase_solar=forecast[2]/forecast[1] #plot(mod,1) #Heteroscedesity #hist(b) #normality #shapiro.test(mod$residuals) #normality #plot(b) #linearity

#calculating power generated from solar radiation

YYMDHH_2021_2050PV[5]=Solar_radiation_$averagewm30*Yearly_increase_solar^30*A*Solarpanels*Ef fiency_PV*PR/1e6 #[MW]

names(YYMDHH_2021_2050PV)[5]<- "Forecast"

#Implenting a potential PV increase and calculating the total power generated

x=0 vary=2020 varyear=1990 while(vary<2050){ if (vary<Inc_PV_year){Solarpanels=Solarpanels+panel_inc } vary=vary+1 varyear=varyear+1

(31)

YYMDHH_2021_2050PV$Forecast[YYMDHH_2021_2050PV$Year == vary]= Solar_radiation_$averagewm30[Solar_radiation_$Year==

varyear]*Yearly_increase_solar^30*A*Solarpanels*Effiency_PV*PR/1e6 #[MW]

}

#Energy mix & hydrogen model

#Repeating energy Demand data from 2016 to 2020 until 2050 is reached

Demand2021_2050<-rep(Demand15_20$`Energy Demand`[Demand15_20$Year == 2016 |Demand15_20$Year == 2017 |Demand15_20$Year == 2018 |Demand15_20$Year == 2019 |Demand15_20$Year == 2020],6)

YYMDHH_2021_2050$DemandMw=Demand2021_2050 #[MW]

#Imlementing yearly demand increase or decrease

vary=2020 u=0 while (vary<2050) { vary=vary+1 u=u+1 YYMDHH_2021_2050$DemandMw[YYMDHH_2021_2050$Year == vary]<- YYMDHH_2021_2050$DemandMw[YYMDHH_2021_2050$Year == vary]*Demand_increase^u }

#Calculating the potential overproduction, gas use when implenting and without implementing hydrogen

#and tracking the total hydrogen in storage

u=0 #loading empty vectors and variables that are used in the loop

w=3 vary=2020 varm=0 varyear=2 VBmin=2000

(32)

month=vector() dem=vector() off= vector() VBoutput=vector() Gas_hydrogen=vector() amount_storage=vector() Hydrogen_use=vector() Fullyear=vector() Demand_sum_year=vector() Offshore_sum_year=vector() Onshore_sum_year=vector() Solar_sum_year=vector() Gas_sum_year=vector() FB_sum_year=vector() Gas_hydrogen_sum_year=vector() hydrogen_produced_sum_year=vector() Hydrogen_used=vector() hydrogen_used=vector() while (vary<2050) { vary=vary+1 varyear=varyear+1 w=w+1

Dem= YYMDHH_2021_2050$DemandMw[YYMDHH_2021_2050$Year== vary] #Extracting 1 year of demand

off= YYMDHH_2021_2053$Energy_Off_wind[YYMDHH_2021_2053$Year== vary] #Extracting 1 year of offshore wind power output

on= YYMDHH_2021_2050onW$Forecast[YYMDHH_2021_2050onW$Year== vary] #Extracting 1 year of onshore wind power output

pv= YYMDHH_2021_2050PV$Forecast[YYMDHH_2021_2050PV$Year== vary] #Extracting 1 year of PV power output

FB=FB-FB_decrease #Decreasing coal output

(33)

Offshore_sum_year=c(Offshore_sum_year,sum(off)) #record sum of offshore wind power output

Onshore_sum_year=c(Onshore_sum_year,sum(on)) #record sum of onshore wind power output

Solar_sum_year=c(Solar_sum_year,sum(pv)) #record sum of PV power output

if (FB<1001){ #Only nuclear and waste output remains

FB=1000 }

VBmin=VBmin-VB_decrease #Decreasing minimum gas output, less overproduction occurs when output of gas is returned to zero

if (VBmin<1){ #No negative gas use is possible

VBmin=0 }

efficiency_hydrogen=efficiency_hydrogen*hyd_eff_increase #Increasing efficiency of hydrogen

VBoutput=vector() amount_storage=vector() Gas_hydrogen=vector() hydrogen_used=vector() z=0

Diff=Dem-(off+pv+on+FB) #Calculating the difference in supply and demand

while (z<length(Diff)){ #For every hour the gas use and hydrogen prodcution is calculated

z=z+1

if (Diff[z]<=VBmin){ #If there is overproduction

Diff[z]=Diff[z]-VBmin #Only VBmin is added to the supply

VBoutput=c(VBoutput,VBmin) #tracking gas output

}else { VBoutput=c(VBoutput,Diff[z]) #Else VBoutput is the remaining shortage

Diff[z]=0} #Difference is brought down to 0, no overproduction

storage=(storage+Diff[z]*-1*efficiency_hydrogen) #The overproduced energy is transformed to hydrogen and stored

Referenties

GERELATEERDE DOCUMENTEN

The international competitive position of energy-intensive industry in the Netherlands does not currently allow for the national increase in the carbon price that would be required

Hydrogen (H 2 ) is a key element in the Dutch energy transition, considered a sources of flexibility to balance the variable renewable energy sources, facilitating its integration

Electrochemical oxidation of dissolved sulphur dioxide at platinum and gold electrodes.. Determination of sulfur dioxide in solution by anodic voltammetry and by

 Backup lead time;  Backup order threshold;  Demand variation;  Chemical flexibility;  Venting level;..  Number of hours for the simulation

A high restriction of the flow of energy to and through the electrolyser results in consequences for the hydrogen output volume as well as for the hydrogen output relative to

For even larger liquid storage capacity is advantageous because of the low investment costs per GJ of stored hydrogen, the overall cost of liquefaction installation,

In the short-term case, a simulation model represents a supply chain configuration where household and mobility are relying on hydrogen supply through tanks transported

This research is different from existing studies because this study is about self-sufficient houses running on a hybrid heat pump, which is provided with energy from the