L OCAL E - GRID MANAGEMENT

(1)

T

ECHNOLOGY AND OPERATIONS MANAGEMENT

MS

C

T

HESIS

U

NIVERSITY OF

G

RONINGEN

,

F

ACULTY OF

E

CONOMICS AND

B

USINESS

Author: J. (Jelle) Keizer S3539962

Supervisor: Dr. M.J. (Martin) Land

Second supervisor: Prof. dr. ir. J.C. (Hans) Wortmann

Date: 24.06.2019

L

OCAL E

-

GRID MANAGEMENT

(2)

2 Acknowledgement:

(3)

3 ABSTRACT

Purpose:

The purpose of this study is to fill a gap in literature by analysing an innovative congestion management system, that uses electrolysis to relieve strain, caused by the increase in decentralized electricity production, from the electricity grid. By doing this from an operation management point-of-view, it is possible to zoom out of the technicalities and focus on analysing the bigger picture, based on the operations effects of different system configuration options.

Methodology:

This study uses a model-based quantitative research approach in order develop to a model for system configuration analysis, based on real-life data. In this model, hourly solar production patterns, electricity demand, hydrogen demand, and electricity distribution grid constraints are used as fixed data inputs. Next to that, three simulation variables are used, i.e., location of the electrolyser, priorities and handling of the electricity flows and solar park electricity output, to develop multiple scenarios. These scenarios are used to analyse the different possible configuration options from an operations point-of-view, in order to gain valuable insights and support real-life decision making.

Findings:

The findings of this study show that, the location of the electrolyser in such a system is vital for relieving strain of the electricity grid. Next to that, it is shown that, a mixture electricity feedback and hydrogen production is needed to optimise the utilisation and efficiency of the electrolyser, together with the electricity load flow of the system. Finally, it is shown that the electrolyser capacity should not be determined by the highest peak in available electricity for hydrogen production. Instead of that, choosing the electrolyser capacity should go hand-in-hand with determining the optimal mix between electricity feedback and hydrogen production.

Implications:

(4)

4 CONTENTS ABSTRACT ... 3 1 INTRODUCTION ... 6 2 THEORETICAL BACKGROUND ... 8 2.1. Congestion management ... 8 2.2. Balancing mechanisms ... 9

2.3. Hybrid renewable energy systems ... 9

2.4. System configuration ... 10

2.5. Viability of a local distribution grid supported by an electrolyser ... 10

2.6. Combining the literature ... 11

3 METHODOLOGY ... 12

3.1. Situation overview ... 12

3.2. The system ... 12

3.3. The simulation model ... 13

3.3.1. Justification ... 16

3.4. System details ... 16

3.5. Model inputs ... 17

3.5.1. Electricity demand ... 17

3.5.2. Hydrogen demand ... 20

3.5.3. Solar production patterns ... 20

3.5.4. Cable capacities ... 21

3.5.5. System parameters ... 21

3.6. Scenarios ... 22

3.7. Model rationale per scenario group ... 23

3.8. Summary of assumptions ... 24

4 FINDINGS ... 25

4.1. Supply and demand matching ... 25

4.2. Electricity flows in the system ... 26

4.3. Hydrogen buffer dimensions ... 33

4.4. Electrolyser dimensions... 37

4.5. Extension on the findings ... 42

5 DISCUSSION ... 44

5.1. Implications of main findings ... 44

(5)

5

5 CONCLUSIONS ... 47

REFERENCES ... 49

APPENDICES ... 52

Appendix A: Calculations total electricity demand ... 53

Appendix B: Comparison electricity flow analysis for all scenario groups ... 55

Appendix C: Comparison load flow analysis for all scenario groups ... 57

Appendix D: Comparison buffer dimensions for all scenario groups ... 58

Appendix E: Comparison electrolyser dimensions for all scenario groups ... 59

(6)

6

1 | INTRODUCTION

The energy sector is changing in a rapid pace and global demand for energy is rising due to technological advancement, rapid growth in industries, and an increase in household electricity demand (Goel and Sharma, 2017; International Energy Agency, 2018). Because of this growth in demand, and the depletion of fossil energy sources, engineers found a solution in alternative energy sources (e.g. wind and solar). These alternative sustainable energy sources come with many challenges (Denholm and Hand, 2011). For example, sustainable energy sources are often planned in rural locations because of the available space. However, these rural locations often lack the electricity distribution network capacity that is needed to handle the electricity supply of these sustainable energy sources (Crabtree et al., 2010). Because of this, the system operators of these local electricity distribution grids must be able to optimally utilise, and stabilise the currently available electricity transmission networks in order to achieve maximum efficiency with the current assets (Mohd et al., 2008). There are multiple system operators involved in the electricity network. The first, the transmission system operator (TSO) is the generic operator of an open-access, high voltage, electricity transmission system (further referred upon as the main grid) and responsible for the transmission of electricity. The second system operator is the distribution system operator (DSO) which is the generic operator of an open-access, medium and low voltage, electricity distribution network. The collective name for these two types of system operators are independent system operators (ISOs). An important tool for ISOs, to optimally utilise and stabilise the currently available electricity network, is congestion management. By using this tool, ISOs can take actions to avoid or relieve congestion, i.e., manage the electricity flows in such a way that the constraints of the electricity network do not have to be violated (Kirby and Dyke, 2002; Christie, Wollenberg and Wangensteen, 2000). The importance of congestion management for ISOs is also stressed by Karthika & Paul (2015) who found that, due to the increase in electricity demand and transmission line outages there is an increased chance for occurring congestion. They state that congestion management systems are essential for a stable and reliable power system.

(7)

7

case, generated by new-built houses in a suburb that is powered by hydrogen instead of natural gas, and electricity.

The investigation into this innovative method of managing the electricity distribution network, concerning the use of a grid-integrated electrolyser, is found to be of theoretical relevance. In the existing literature, it has been shown that the increase in capacity and variability of the energy production portfolio, results in a need for innovative ways to manage the electricity distribution network and its congestion (Brunekreeft, Neuhoff and Newbery, 2005; Schermeyer, Vergara and Fichtner, 2018; Bennoua, Duigou and Qu, 2015). Additionally, the potential of hydrogen production and storage as an innovative method of managing the electricity distribution network, and its congestion, is recognized by Qadrdan et al. (2015). They also state, that electrolysers in power-to-gas systems should be operated at times and locations where they have the most impact in minimizing the total cost of operating the system. However, in their research they do not go into the specifics of those timings and locations. Therefore, questions concerning possible different system configurations (e.g. location of the electrolyser, priority rules and strategy of handling electricity flows, and size of sustainable energy source) remain unanswered.

(8)

8

This thesis will be structured as follows. In section two, a literature review is done where the relevant literature for this study is discussed. In section three, the methodology will be discussed in which, the research approach is discussed and the model is presented and defined. In the fourth section of this thesis, the most relevant findings will be reported, after which they are discussed upon in section five. In section six, the final conclusions will be drawn and relevant managerial insights are presented.

2

| THEORETICAL BACKGROUND

In this section of the thesis, a literature review, concerning the relevant topics surrounding this study, is presented. The literature review starts by reviewing the topic of congestion management after which, it is funnelled towards the identification of the research gap in relation to the sustainable development of electricity grids.

2.1. Congestion management

Congestion, as defined by Karthika and Paul (2015), is the condition of the electricity grid when it reaches or exceeds the transfer capability limit (TCL). The TCL of the transmission lines are, among other things, a voltage limit, thermal limit and stability limit. Congestion of an electricity grid will cause huge power losses, poor voltage regulation, and high temperature which can cause the grid to fail (Karthika and Paul, 2015). Congestion management methods that are currently applied by ISOs, can be split up in three different categories; technological methods (e.g. transformer tap changers), market based methods (e.g. pricing strategies) and non-market based methods (e.g. FCFS policies) (Yusoff, Zin and Khairuddin, 2017). Due to the rapid growth of deregulated energy markets, congestion management has become a crucial factor for ISOs in order to avoid and deal with (possible) congestion. New challenges in congestion management relate to the use of new technologies, and creation of efficient methods that enhance power system performance as fast as possible to relieve the electricity grid from congestion (Yusoff, Zin and Khairuddin, 2017). An example of such new technology, paired with many challenges, is the use of hydrogen production and storage to support the electricity grid.

According to Bennoua, Duigou, & Qu (2015) hydrogen technologies offer a credible way of storing electricity and therefore can be used for balancing services in the electricity grid. Gutiérrez-Martin

et al., (2009) show that, the use hydrogen technology as a balancing mechanism is feasible and

(9)

9

provide a useful function in conjunction with electrical systems, and found that hydrogen production could play an important role in resolving problems in electrical systems. The main challenges, as identified by Bennoua et al., (2015), relate to technical advances (e.g. electrolyser life and performance/efficiencies), availability of storage infrastructures, and electrical energy management (e.g. balancing mechanisms).

2.2. Balancing mechanisms

Regarding the aspect of balancing mechanisms in the electricity network, Bennoua et al. (2015) identified that most ISOs work with a balancing mechanism based on different levels of reserves. The primary reserves contains the energy reserve as determined by regulations and the secondary reserve, is established in case the primary reserve is not sufficient. The primary reserve is activated automatically, a few seconds after an imbalance occurs. The secondary reserve is activated automatically after the primary reserves are depleted. The final reserve is the tertiary reserve. This reserve can be activated manually by asking generators connected to the electricity grid to alter their production plan as fast as possible. This form of balancing mainly satisfies the imbalance when the demand is higher than the supply. However, balancing the grid goes two ways, and should therefore also create and store output when there is an excess of supply. This determination of when to store electricity from the grid could be approached as congestion management method. According to González & Sánchez (2013), the determination of timing hints the need for a control system in order to control the energy consumed by a hydrogen production system when there is an excess of electricity. Next to that they identify multiple technical solutions (e.g. electro pneumatic valves, a complex control system based on algorithms and a data acquisition system) to control a hydrogen production system that is part of a hydrogen energy buffer.

2.3. Hybrid renewable energy systems

(10)

10 2.4. System configuration

More research into the proper selection of system configuration and optimized operation processes is done by, Qadrdan et al., (2015). They studied how the introduction of power-to-gas could reduce wind curtailment by converting excess wind energy into hydrogen. In their model, the produced hydrogen is injected into the gas network, which has an injection constraint, in order to transport it. Qadrdan et al. (2015) also studied a scenario where there was no constraint on the hydrogen injection into the gas network. By removing this constraint, the results became applicable for other (comparable) systems as well. The unlimited gas injection can be seen as an unlimited buffer, of which the capacity could be determined later. Qadrdan et al. (2015) concluded that by producing hydrogen from wind energy, the wind curtailment could be eliminated, resulting in an overall cost decrease for operating the network. Next to that, they stated that electrolysers in power-to-gas systems should be operated at times and locations where they have the most impact in minimizing the total cost of operating the combined system. In the end, they concluded that the use of hydrogen production could improve the optimal dispatch for electricity by using cheap electricity from other congested zones in the network and, that connecting a power-to-gas system to the electricity grid can lead to considerable reductions in operational costs.

2.5. Viability of a local distribution grid supported by an electrolyser

Other authors contributing to the idea of connecting an electrolyser to the electricity grid are Kiaee

et al. (2013). They found that an electrolyser connected to the electricity grid could help the

(11)

11

of sustainable energy sources in remote areas, without increasing grid capacity towards the central grid.

2.6. Combining the literature

Overall, previous research identified that, the development of new sustainable energy sources often leads to congestion near those energy sources (Schermeyer, Vergara and Fichtner, 2018). Additionally, the need for new technologies and systems to manage this congestion are described (Bennoua, Duigou and Qu, 2015; Brunekreeft et al., 2005), and the potential of hydrogen production and storage, as a method of managing the electricity grid and its congestion, is recognized (Bødal and Korpas, 2017; Qadrdan et al., 2015). However, previous research does not give a clear picture of the requirements regarding system configuration, in a system with a grid-integrated electrolyser. It is stated by Qadrdan et al. (2015) that, electrolysers in power-to-gas systems should be operated at times and locations where they have the most impact in minimizing the total cost of operating the local electricity network. However, the specifics of those timings and locations are not studied and therefore, questions concerning possible different system configurations remain. In conclusion, there is a lack of exact insights into the bigger picture of the possible system configurations.

(12)

12

3 | METHODOLOGY

This study analyses different system configuration options, in a congestion management system with a grid-integrated electrolyser, based on their operational effects. By doing this, it becomes possible to get a clear insights into the workings of such a congestion management system and provide decision support concerning possible system configurations. In this section of the thesis, the methodology used to perform these analyses is presented.

3.1. Situation overview

For this study, real-life data from a town, located in a rural province of the Netherlands, is used. In and near this town, multiple decentralized solar power plants are under development. The capacity of these solar power plants range between 1 MVA and 55 MVA (with the majority between 5 MVA and 20 MVA). The total capacity of solar power plants under development is around 185 MVA. To put that in perspective, the city under investigation uses around 50 MVA of installed capacity. Because of that, a lot of electricity is fed back into the local electricity distribution network. However, the electricity distribution network of this town, is not able to handle the increase in unstable electricity generation. Because of that, expansion of the electricity distribution network is recognized as needed, in order to comply with the growing decentralized sustainable energy sources. Next to that, this town has committed to a project of developing (80) new-built houses that are solely powered by electricity, and hydrogen instead of natural gas. Because of this, there is a local hydrogen demand, which is subject for up-scaling if the project is a success.

In conclusion, this town faces the problems of congestion, caused by the development of decentralized sustainable energy sources, and is actively involved in the energy transition, by building a suburb powered by hydrogen instead of natural gas. These two aspects fit together perfectly with the goal of this study. Because of this, this case is considered as an ideal case to investigate how a grid-integrated electrolyser can prevent congestion in the electricity distribution grid, while fulfilling local hydrogen demand to create a closed system of hydrogen production, storage and usage.

3.2. The system

(13)

13

that all demand is satisfied from the main grid, and that the distributed electricity is equal to the electricity demand at that specific time. The second state, is defined by a small amount of electricity input from the solar park, which is not enough to satisfy all demand. In this state, the so called “mixed input” state, the input from the main grid is determined by the difference between the electricity input from the solar power plant and the electricity demand at that time. In this state the electricity distributed through the network remains equal to the demand at that time, although the input originates from two different energy sources. The third state is the most interesting state for this study. This is the state where the electricity supply from the solar park exceeds the electricity demand. The way of handling of this surplus of electricity is determined by different scenarios, described section 3.6 of the report.

FIGURE 1- GENERAL OVERVIEW OF THE SYSTEM UNDER INVESTIGATION

3.3. The simulation model

In the simulation model, as depicted in figure 2, five different types of model inputs are converted into model outputs, through three simulation variables. These inputs are defined as the hourly electricity demand within the modelled system in kWh, the hourly hydrogen demand per house in kWh, the hourly solar production pattern as a fraction of the total yearly production, the cable capacities in amperes within the modelled system, and the relevant system parameters (e.g. electrolyser efficiency). The actual data of these inputs are presented in section 3.5 of this thesis.

(14)

14

discussed in section 3.6 of this thesis. Per scenario, the electricity flows and the operational effects in the system are monitored. Based on this, the influence of the different simulation variables is observed and the model outputs are derived.

(15)

15

FIGURE 2- OVERVIEW OF THE SIMULATION MODEL

TABLE 1- MODEL OUTPUTS AND INDICATORS USED FOR ANALYSIS

OUTPUTS PERFORMANCE INDICATORS

SUPPLY & DEMAND MATCHING 1. On-site energy matching (OEM) 2. On-site energy fraction (OEF) 3. Electricity supply main grid (GWh) 4. Grid dependency

ELECTRICITY DISTRIBUTION NETWORK LOAD FLOW

1. Average system load flow

2. Percentage of time that the system load flow is above 75% 3. Percentage of time that the system load flow is above 50%

ELECTRICITY FEEDBACK TO THE MAIN GRID 1. Outgoing flow of electricity (GWh)

2. Percentage of electricity feedback relative to total electricity production of the solar park

ELECTRICITY LOSSES DUE TO CONGESTION 1. Lost electricity (GWh)

2. Percentage of electricity losses relative to total electricity production of the solar park

ELECTRICITY AVAILABLE FOR HYDROGEN PRODUCTION

1. Electricity available for hydrogen production (GWh)

2. Percentage of available electricity for hydrogen production relative to the total electricity production of the solar park

HYDROGEN BUFFER DIMENSIONS 1. Capacity (GWh, kg, m3₎ 2. Starting inventory (GWh, kg, m3₎ 3. Ability to supply 80 houses with hydrogen 4. Optimal amount of houses supplied

(16)

16 1.3.1. Justification

In order to analyse the operational effects of different system configuration options within the system as described in previous sections. The system under investigation either has to be monitored or modelled. Since the system is not real yet, the only viable option is to model the system. Because a model like this requires many calculations due to its large amounts of input data, i.e., (hourly) solar production patterns, electricity demand, hydrogen demand and, electricity distribution grid constraints, a quantitative spreadsheet model is developed and used. The use of model-based quantitative research is classified as a knowledge generation approach (Meredith et

al., 1989; Bertrand and Fransoo, 2002). It is based on the rationale that an objective model can be

built, which explains the behaviour of real-life operational processes and capture decision making problems faced in real-life operations (Christer et al., 2016). It seems appropriate to use this type of research because the goal of this research is to give insight into a real-life operational system and the decision making problems surrounding that system. Because the modelling approach is based on causal relationships between variables, it is possible to predict the future state of the system and obtain useful (managerial) insights concerning the implementation of the system and the related decision making issues.

3.4. System details

(17)

17

FIGURE 3- DETAILED SYSTEM OVERVIEW 3.5. Model inputs

This section of the thesis is dedicated to presenting the different model inputs that are used..

3.5.1. Electricity demand

Unfortunately, there is no real-life demand data available for the modelled electricity distribution network. However, it is known that there are 1356 households, specified per junction in table 2, within the system, and that the maximum amount of electricity distribution measured in the system in 2018 has a total of 1840 kWh. By using a electricity demand pattern for an average household (E1C), as depicted in figure 4 (Vereniging Nederlandse Energie Data Uitwisseling, 2019), it is estimated that a total of 7.3 GWh is distributed through the modelled system in 20181_{. Based on the}

average electricity use of a household in the Netherlands (3370 kWh/y) (Centraal Bureau voor Statestiek, 2016) , it is estimated that, of this 7.3 GWh, 4.6 GWh is used by the nine junctions within the system and the other 2.7 GWh only flows through the system and is used externally. The total demand pattern of the system is shown in figure 5,6 and 7.

(18)

18

TABLE 2- OVERVIEW OF HOUSEHOLDS PER JUNCTION, SHARE IN ELECTRICITY DEMAND AND YEARLY ELECTRICITY DEMAND (MWH)

FIGURE 4- DEMAND FRACTION PROFILE OF AN AVERAGE HOUSEHOLD (E1C)(NEDU,2019)

JUNCTION HOUSEHOLDS SHARE IN DEMAND YEARLY DEMAND (MWH)

(19)

19

FIGURE 5–ELECTRICITY DEMAND PROFILE OF THE MODELLED SYSTEM – YEAR OVERVIEW

FIGURE 6– ELECTRICITY DEMAND PROFILE OF THE MODELLED SYSTEM - ZOOMED IN ON JANUARY,FEBRUARY, AND MARCH

FIGURE 7– ELECTRICITY DEMAND PROFILE OF THE MODELLED SYSTEM - ZOOMED IN TO 8 DAYS IN NOVEMBER AND DECEMBER

500 700 900 1100 1300 1500 1700 1900 S ys te m e le ctr ici ty d e m an d ( kW h ) Time (months) 400 600 800 1000 1200 1400 1600 1800 2000 S ys te m e le ctr ici ty d e m an d ( kW h ) Time (months) 0 200 400 600 800 1000 1200 1400 1600 1800 2000 S ys te m e le ctr ici ty d e m an d ( kW h )

(20)

20 3.5.2. Hydrogen demand

Because this study’s aim is to analyse how hydrogen production, used to satisfy local hydrogen demand, can influence the electricity distribution grid and its congestion, it is important to know how much hydrogen demand there is and at which times. As described in section 3.1, the hydrogen demand concerns the hourly heat-demand of 80 new-built houses in the hydrogen suburb. The hydrogen demand pattern for these 80 houses is depicted in figure 8. The total heat demand for one year sums up to 1.55 GWh for 80 houses. (19.4 MWh for one house).

FIGURE 8- HYDROGEN DEMAND PROFILE FOR 80 HOUSES (RENDO N.V.,2019)

3.5.3. Solar production patterns

In order to determine the electricity output of the solar park, the solar production pattern of 2017 is used. This pattern exists out of hourly fractions (with a total sum of 1 over a year) of the total electricity generation and can therefore be applied to differently sized solar power plants. The solar production pattern, depicted in figure 9, shows the hourly percentage of the maximum possible electricity generation (in the 5th_{hour of January the 1}st_{, the solar park produces 5% of its maximum}

generation). There are three sizes of solar parks being used in this study.

1) A solar power plant with a 6.1 GWh per year output with a maximum output of 500 amperes at 10 kV,

2) A solar power plant with a 12.2 GWh per year output with a maximum output of 1000 amperes at 10 kV

(21)

21

By combining the solar production pattern with the maximum electricity output of the solar parks, the exact electricity output in amperes per hour (for one year) is determined for all three solar parks.

FIGURE 9- SOLAR PRODUCTION PATTERN AS PERCENTAGE OF MAXIMUM ELECTRICITY OUTPUT (2017)(RENDO N.V.,2019)

3.5.4. Cable capacities

For all cables the capacity in amperes is retrieved from the real-life system. These capacities can be seen in table 3. In this study, the hard cap in amperes is used as the maximum amount of amperes that can be distributed through that cable. Meaning that the possibility to exceed the cap at a cost of increased energy losses cap is disregarded.

CABLE CAPACITY K-1 224 K-2 240 K-3 240 K-4 240 K-5 240 K-6 240 K-7 240 K-8 240 K-9 240 K-10 240

TABLE 3- CABLE CAPACITY PER CABLE IN THE MODELLED SYSTEM (RENDO N.V.,2019)

3.5.5. System parameters

Finally, some system parameters are needed to be able to perform calculations. The first system parameter is the electrolyser efficiency. This efficiency is assumed to be 65%. Another system parameter is the voltage difference on which the local electricity distribution grid is working. In reality, this voltage difference has minor fluctuations around 10 kV all the time, in this study however, the voltage difference is fixed to 10 kV in order to make systematic calculations possible.

(22)

22 3.6. Scenarios

In this study, the effect of different system configuration options on operational aspects is studied. In order to do this, multiple scenarios are developed to represent these different system configuration options. The variable inputs, and their possible values, used in these scenarios are presented in table 4 below.

TABLE 4- VARIABLE INPUTS AND THEIR POSSIBLE VALUES

Based on these values there are 12 possible scenarios in a full factorial design. These scenarios are presented in table 5. Within the model, there is one priority dominant in all scenarios. This priority makes sure that, electricity demand always gets satisfied first , either by the electricity output from the solar power plant or the main grid, before anything else happens in the model. Any excess electricity is dealt with in different ways for different scenarios resulting in different electricity flows. These flows are caused by two variables in particular, the electrolyser location and the priority rules and strategy of handling electricity flows. By combining all possible combinations between these two variables, 4 so-called main groups are recognised. The third variable, solar park electricity output, analyses the outcomes of these main groups for different solar park electricity outputs.

TABLE 5- OVERVIEW OF SCENARIOS

VARIABLE POSSIBLE VALUES

ELECTROLYSER LOCATION 1. Close to the solar park 2. Far away from the solar park

PRIORITY RULES & STRATEGY OF HANDLING ELECTRICITY FLOWS

1. Prioritise feedback to the main grid 2. Prioritise hydrogen production

SOLAR PARK ELECTRICITY OUTPUT 1. 6.1 GWh per year 2. 12.2 GWh per year 3. 24.4 GWh per year

GROUP SCENARIO ELECTROLYSER LOCATION PRIORITY SOLAR PARK SIZE

1 1 Far away (J1) Hydrogen production 6.1 GWh per year

2 Far away (J1) Hydrogen production 12.2 GWh per year 3 Far away (J1) Hydrogen production 24.4 GWh per year

2 4 Far away (J1) Feedback to main grid 6.1 GWh per year

5 Far away (J1) Feedback to main grid 12.2 GWh per year 6 Far away (J1) Feedback to main grid 24.4 GWh per year

3 7 Close by (J7) Hydrogen production 6.1 GWh per year

8 Close by (J7) Hydrogen production 12.2 GWh per year 9 Close by (J7) Hydrogen production 24.4 GWh per year

4 10 Close by (J7) Feedback to main grid 6.1 GWh per year

(23)

23 3.7. Model rationale per scenario group

3.7.1. Group one – scenarios 1 to 3

For all scenarios in scenario group one, the electrolyser is located far away from the solar park, while the priority lies on hydrogen production. This means that, for hydrogen to be produced, the electricity from the solar park is distributed through the system, i.e., from the solar park to the location of the electrolyser (J1). Because of this, it can occur that not all produced electricity can be distributed to the electrolyser, due to cable capacities. Therefore, electricity that is produced, but cannot be distributed due to cable capacities, is discarded and lost at the solar park location. The electricity that does reach the electrolyser, is transformed into hydrogen and subsequently transported to the hydrogen buffer.

3.7.2. Group two – scenarios 4 to 6

For all scenarios in scenario group two, the electrolyser is located far away from the solar park, and the priority lies on feeding back electricity to the main grid. This means that, the electricity input from the solar park into the system is maximized until both cables K-6 and K-7 reach their maximum capacity. All electricity that cannot be distributed from the solar park into the system due to cable capacities is discarded and lost at the solar park. Of all electricity that reaches location J1, most is fed-back through cable K-1. Only after the cable capacity of cable K-1 is reached, the remainder of electricity at location J1 is transformed into hydrogen.

3.7.3. Group three – scenarios 7 to 9

For all scenarios in scenario group three, the electrolyser is located close by the solar park, and the priority lies on hydrogen production. This means that, the electricity that is fed into the electricity distribution network, is always equal to the demand at that time. All produced electricity in the solar park that cannot be used for demand satisfaction, is directly transformed into hydrogen at the solar park location.

3.7.4. Group four – scenarios 10 to 12

(24)

24

3.8. Summary of assumptions

During this research, assumptions are made in order to make the study possible and manageable. These assumptions are listed below:

1) The entire electricity distribution network is operating at a stable voltage difference of exactly

10 kV.

2) There is a power factor of 1, this means that VA is assimilated with watts and therefore 1 MVA

is equal to 1 MW.

3) The energy loss due to distribution is neglectable and therefore not taken into account. 4) The electrolyser operates at 65% efficiency whereas 100 kWh can be transformed into 65 kWh

in hydrogen 300 bar (30Mpa).

5) There are only 3x 25 Ampere connections in the system where the price of electricity higher

during the daytimes (07:00 – 23:00) and cheaper during the night time (23:00 – 07:00).

6) All electricity users within the modelled system are average households, or the total electricity

demand in the system averages out to be in line with the demand pattern of system containing only average households.

7) At the at the highest demand point in 2018 according the E1C demand pattern (NEDU, 2019),

the maximum amount of electricity measured (1840 kWh) (Rendo N.V., 2019), is distributed through the system.

8) 1 kg of hydrogen is assumed to constraint 33.333 kWh of energy, meaning that hydrogen

(25)

25

4 | FINDINGS

This chapter contains the findings of this study and is structured according to the outputs defined in table 1. First the matching between supply and demand is presented. After that, an analysis of the electricity flows is made per scenario. Third, the required buffer dimensions per scenario group are presented, Fourth, the electrolyser dimensions per scenario group are shown and finally, an extension on the required electrolyser dimensions is presented.

Before presenting the findings. There are a few main conclusion that can be drawn from the different simulations results and the model analysis.

1) Hydrogen production close to the source of electricity production is preferred since that way,

no energy needs to be lost, and the grid can be relieved of potential strain even before the electricity reaches the grid.

2) To relieve strain in the electricity grid and fulfil local hydrogen demand, the local hydrogen

demand should be on par with the amount of electricity that needs to be subtracted from the electricity grid.

3) A trade-off is found between the amount of electricity feedback and the hydrogen production.

By optimising this ratio, the electrolyser needed for hydrogen production can be a lot smaller than the biggest peak in available electricity for hydrogen production.

4.1. Supply and demand matching

(26)

26

TABLE 6– SUPPLY AND DEMAND MATCHING PERFORMANCE INDICATORS 4.2. Electricity flows in the system

This section of the results is dedicated to analysing the electricity flows within the system per scenario group. These electricity flows contain the outputs; electricity distribution network load flow, electricity feedback to the main grid, electricity losses due to congestion, and electricity available for hydrogen production.

4.2.1. Scenario group one

Within scenario group one, it is noticed that by doubling the electricity output from 6.10 GWh/y (scenario 1) to 12.20 GWh/y (scenario 2), the electricity available for hydrogen production also increases by more than doubles (218%) (table 7). At the same time, the amount of lost electricity increases from nearly zero in scenario 1, to 1.60 GWh in scenario 2. By comparing scenario 2 and 3 it is noticed that, despite the doubling of the electricity output from 12.20 GWh/y to 24.40 GWh per year, the electricity available for hydrogen production only increases by 46%. This indicates that, the cable capacities are reached more often, if the electricity output increases. Because of that, not all electricity can be distributed through the system, resulting in an increase load flow (figure 10,11 and table 8) and an increase in electricity losses due to congestion (figure 12 and 13).

TABLE 7-ELECTRICITY FLOW PERFORMANCE INDICATORS (GROUP 1)

ELECTRICITY PRODUCED (SOLAR) 6.1GWH/Y 12.2GWH/Y 24.4GWH/Y

ELECTRICITY USED FOR DEMAND 2.63 GWh 3 .04 GWh 3.32 GWh

GRID INPUT 4.67 GWh 4.25 GWh 3.98 GWh

ON-SITE ENERGY MATCHING (OEM) 43% 25% 14%

ON-SITE ENERGY FRACTION (OEF) 36% 42% 45%

GRID DEPENDENCY 64% 58% 55%

ELECTRICITY FLOWS |SCENARIO GROUP ONE SCENARIO

1 2 3

TOTAL ELECTRICITY PRODUCTION (BY SOLAR PARK) 6.10 GWh 12.20 GWh 24.40 GWh

LOST ELECTRICITY 0.01 GWh 1.60 GWh 10.08 GWh

PERCENTAGE OF LOST ELECTRICITY RELATIVE TO THE TOTAL ELECTRICITY PRODUCTION

0% 13% 41%

ELECTRICITY INPUT FROM THE SOLAR PARK 6.09 GWh 10.06 GWh 14.32 GWh PERCENTAGE OF ELECTRICITY INPUT RELATIVE TO TOTAL ELECTRICITY

PRODUCTION

100% 87% 59%

ELECTRICITY USED FOR DEMAND SATISFACTION (OEM) 2.63 GWh 3.05 GWh 3.32 GWh

PERCENTAGE OF ELECTRICITY USED FOR DEMAND SATISFACTION RELATIVE TO TOTAL ELECTRICITY PRODUCTION

43% 25% 14%

ELECTRICITY AVAILABLE FOR HYDROGEN PRODUCTION 3.46 GWh 7.55 GWh 11.00 GWh

PERCENTAGE OF ELECTRICITY AVAILABLE FOR HYDROGEN PRODUCTION RELATIVE TO TOTAL ELECTRICITY PRODUCTION

57% 62% 45%

ELECTRICITY FED-BACK INTO THE MAIN GRID 0 0 0

(27)

27

TABLE 8–SYSTEM LOAD FLOW PERFORMANCE INDICATORS (GROUP 1)

FIGURE 10- AVERAGE LOAD FLOW OF THE SYSTEM – YEAR OVERVIEW (GROUP 1)

FIGURE 11- AVERAGE LOAD FLOW OF THE SYSTEM - ZOOMED IN ON 10 DAYS IN JUNE (GROUP 1)

FIGURE 12- LOST ELECTRICITY IN THE SYSTEM - YEAR OVERVIEW (GROUP 1)

LOAD FLOW SCENARIO

1 2 3

AVERAGE SYSTEM LOAD FLOW 17% 26% 33%

LOAD FLOW ABOVE 75%(PERCENTAGE OF TIME) 1% 13% 24%

LOAD FLOW ABOVE 50%(PERCENTAGE OF TIME 6% 19% 29%

0% 20% 40% 60% 80% 100% Ave rag e load flow in th e s ys te m Time (months)

Scenario 3 Scenario 2 Scenario 1

0% 20% 40% 60% 80% 100% Ave rag e load flow in th e s ys te m Time (days)

0 2.000 4.000 6.000 8.000 10.000 12.000 14.000 16.000 El e ctr ici ty lo st (k W h ) Time (months)

(28)

28

FIGURE 13- LOST ELECTRICITY IN THE SYSTEM - ZOOMED IN ON 10 DAYS IN JUNE (GROUP 1) 4.2.2. Scenario group two

Within scenario group two, it is noticed that the same amount of electricity is distributed through the system to junction seven (table 10). This results in the same amount of electricity lost due to congestion as in scenario group one (figure 12 and 13) The load flow is found to be almost similar to scenario group one. The only difference is that in this scenario group, the electricity is fed-back via cable K-1 while in scenario group one, cable K-1 is only used for the required grid input to satisfy demand. This results in a slightly higher average load flow of the system for scenario group two (table 8 and 9) In scenario group two, the doubling of the electricity output of the solar park only increases the fed-back electricity with 67% (from 6.10 GWh/y to 12.20 GWh/y) and 33% (from 12.20 GWh/y to 24.40 GWh/y). This shows that the electricity feedback is heavily capped by the cable capacity of cable K-1 (figure 14 and 15). By these doublings in electricity output of the solar park, the amount of electricity available for hydrogen production does increase a bit. However, it is still noticed that the electricity available for hydrogen is heavily capped by the cable capacities of cables K-6 and K-7.

TABLE 9-SYSTEM LOAD FLOW PERFORMANCE INDICATORS (GROUP 2)

0 2.000 4.000 6.000 8.000 10.000 12.000 14.000 16.000 El e ctr ici ty los t ( kW h ) Time (days)

4 5 6

(29)

29

TABLE 10- ELECTRICITY FLOW PERFORMANCE INDICATORS (GROUP 2)

FIGURE 14-ELECTRICITY FEEDBACK TO MAIN GRID – YEAR OVERVIEW (GROUP 2)

FIGURE 15-ELECTRICITY FEEDBACK TO MAIN GRID - ZOOMED IN ON 10 DAYS IN JUNE (GROUP 2)

ELECTRICITY FLOWS SCENARIO

4 5 6

TOTAL ELECTRICITY PRODUCTION (BY SOLAR PARK) 6.10 GWh 12.20 GWh 24.40 GWh LOST ELECTRICITY 0.01 GWh 1.60 GWh 10.07 GWh

0% 13% 41%

PRODUCTION

100% 87% 59%

43% 25% 14%

6% 19% 17%

ELECTRICITY FED-BACK INTO THE MAIN GRID 3.11 GWh 5.20 GWh 6.91 GWh

51% 43% 28% 0 500 1.000 1.500 2.000 2.500 El e ctri ci ty f e e dbac k to ma in gri d (k W h ) Time (months)

0 500 1.000 1.500 2.000 2.500 El e ctri ci ty f e e dbac k to ma in gri d (k W h ) Time (days)

(30)

30 4.2.3. Scenario group three

Within scenario group three, it is noticed that there are no electricity losses (table 11). This is because the electrolyser is located at the location of the solar park, and that all electricity that is not used to satisfy demand, stays behind at the solar park and is available for hydrogen production. Compared to scenario groups one and two, the amount of electricity available for hydrogen production is therefore significantly higher (table 11). Because the electricity used for hydrogen production is not distributed through the system, the system load flow is remarkable better compared to other scenario groups (figure 16,17 and table 12).

TABLE 11- ELECTRICITY FLOWS PERFORMANCE INDICATORS (GROUP 3)

TABLE 12-SYSTEM LOAD FLOW PERFORMANCE INDICATORS (GROUP 3)

7 8 9

LOST ELECTRICITY 0 0 0

0% 0% 0%

PRODUCTION

100% 100% 100%

43% 25% 14%

57% 75% 86%

ELECTRICITY FED-BACK INTO THE MAIN GRID 0 0 0

0% 0% 0%

7 8 9

(31)

31

FIGURE 16-AVERAGE LOAD FLOW OF THE SYSTEM – YEAR OVERVIEW (GROUP 3)

FIGURE 17–AVERAGE LOAD FLOW OF THE SYSTEM – ZOOMED IN TO 10 DAYS IN JUNE (GROUP 3)

4.2.4. Scenario group four

Within scenario group four, it is noticed that there is both electricity feedback the main grid, as well as electricity available for hydrogen production (table 13). Because of this, in scenario 10, almost all electricity is fed-back, leaving little to nothing for the production of hydrogen. However, in the other two scenarios there is more electricity available for hydrogen production. This is because a bigger electricity output from the solar park, results in higher peaks in production. These peaks cannot all be distributed through the system, which makes them unfit for feedback to the main grid, but available for hydrogen production. Within this scenario, cable K-1 is clearly the bottleneck

0% 5% 10% 15% 20% 25% 30% ave rag e load flow in th e s ys te m Time (months)

0% 2% 4% 6% 8% 10% 12% 14% 16% AV e rag e load flow in th e s ys te m Time (days)

(32)

32

for the electricity feedback. Because of this, the average system load flow reaches its maximum at around 65% (figure 18 and 19). This is explained by the fact that after cable K-1 reaches its capacity, there is no more electricity inserted into the system, while the rest of the system is still able to handle more electricity flows.

TABLE 13- ELECTRICITY FLOW PERFORMANCE INDICATORS (GROUP 4)

TABLE 14–SYSTEM LOAD FLOW PERFORMANCE INDICATORS (GROUP 4)

FIGURE 18-AVERAGE LOAD FLOW OF THE SYSTEM – YEAR OVERVIEW (GROUP 4)

10 11 12

LOST ELECTRICITY 0 0 0

0% 0% 0%

ELECTRICITY INPUT FROM THE SOLAR PARK 6.10 GWh 12.20 GWh 24.40 GWh

PERCENTAGE OF ELECTRICITY INPUT RELATIVE TO TOTAL ELECTRICITY PRODUCTION

100% 100% 100%

43% 25% 14%

6% ^33% 58%

ELECTRICITY FED-BACK INTO THE MAIN GRID 3.07 GWh 5.18 GWh 6.89 GWh

50% 42% 28%

10 11 12

LOAD FLOW ABOVE 50%(PERCENTAGE OF TIME 9% 21% 31%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% ave rag e load flow in th e s ys te m Time (months)

(33)

33

FIGURE 19– AVERAGE LOAD FLOW OF THE SYSTEM - ZOOMED IN ON 10 DAYS IN JUNE (GROUP 4)

4.3. Hydrogen buffer dimensions

This section of the results is dedicated to analysing the hydrogen buffer behaviour and dimensions per scenario group.

Within scenario group one, it is noticed that with an appropriate starting inventory, it is possible to supply 80 houses with hydrogen. This is because, the system can maintain a positive buffer inventory as shown in figure 20. The optimal buffer dimensions for this scenario group are calculated and shown in table 15. It is noticed that, the required buffer capacity decreases, if the amount of houses supplied with hydrogen increases. This is explained by the fact that the buffer inventory keeps increasing if there is not enough demand to empty the buffer. Subsequently, it is noticed that, the optimal case needs a higher amount of required starting inventory, than the case with 80 houses. This is explained by the deficit that occurs in the first months of the year. In the optimal case, there is more demand than in the case with 80 houses. Because of that, there is also a higher deficit in the first months. Subsequently, a higher starting inventory is required if the amount of houses in the demand pool grow (figure 21).

TABLE 15- HYDROGEN BUFFER PERFORMANCE INDICATORS (GROUP 1)

0% 10% 20% 30% 40% 50% 60% 70% AV e rag e load flow in th e s ys te m Time (days)

BUFFER DIMENSIONS SCENARIO

1 OPTIMAL 2 OPTIMAL 3 OPTIMAL

(34)

34

FIGURE 20– HYDROGEN BUFFER BEHAVIOUR WITH 80 HOUSES (GROUP 1)

FIGURE 21–HYDROGEN BUFFER BEHAVIOUR IN THE OPTIMAL SETTING (GROUP 1)

4.3.2. Scenario group two

Within scenario group two, the fact that there is a lower amount of available electricity for hydrogen production, due to priority on feedback to the main grid, resonates into to buffer behaviour as shown in figure 22. Because there is barely any electricity available for hydrogen production (table 16) in scenario 4, the hydrogen demand of 80 houses cannot be fulfilled. The other two scenarios do have sufficient electricity available for hydrogen production to supply 80 houses with hydrogen (79). This is explained by the fact that the system can distribute more electricity to junction seven, than cable k-1 can feed back into the main grid. However, since the system is capped by the cable capacities, up-scaling the hydrogen production in this setup is difficult. -1 0 1 2 3 4 5 6 7 B u ff e r inve ntor y (G W h) Time (months)

0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0 4,5 5,0 Buf fe r in ve n tor y ( G W h ) Time (months)

(35)

35

TABLE 16-HYDROGEN BUFFER PERFORMANCE INDICATORS (GROUP 2)

FIGURE 22-HYDROGEN BUFFER BEHAVIOUR WITH 80 HOUSES (GROUP 2)

4.3.3. Scenario group three

Within scenario group three, There is significantly more electricity available for hydrogen production and this resonates in the buffer behaviour and dimensions. It is noticed that, the demand of 80 houses can easily be satisfied by all scenarios in this group (figure 23). Next to that, it is noticed that doubling the electricity output from the solar park from 6.10 GWh/y to 12.20 GWh/y, the amount of houses that can be supplied more than doubles (265%) (table 17). If the electricity output from the solar park is doubled again from 12.20 GWh/y to 24.40 GWh/y, the amount of houses that can be supplied with hydrogen still increases with 230%. This indicates that in this scenario group, up-scaling the solar park electricity output to supply more hydrogen is viable.

AMOUNT OF HOUSES 80 12 80 79 80 137 CAPACITY GW 1.32 GW 0.20 GW 1.14 GW 1.15 GW 2.04 GW 1.77 GW KG OF HYDROGEN 39.533 kg 6.011 kg 34.345 kg 34.459 kg 51.270 kg 53.141 kg CUBIC METRE 988 m3 _{150 m}3 _{859 m}3 _{861 m}3 _{1.532 m}3 _{1.329 m}3 STARTING INVENTORY GW 1.32 GW 0.11 GW 0.64 GW 0.63 GW 0.50 GW 0.97 GW KG OF HYDROGEN 39.541 kg 3.340 kg 19.156 kg 18.870 kg 14.940 kg 29.100 kg -1,5 -1,0 -0,5 0,0 0,5 1,0 1,5 2,0 Hydr og e n b uf fe r in ve n tor y ( G W h ) Time (months)

(36)

36

TABLE 17-HYDROGEN BUFFER PERFORMANCE INDICATORS (GROUP 3)

FIGURE 23–HYDROGEN BUFFER BEHAVIOUR WITH 80 HOUSES (GROUP 3) 4.3.4. Scenario group four

Within scenario group four, it is noticed that combination of electricity feedback and hydrogen production only works for the scenarios with a higher electricity output (12.20 GWh/y and 24.40 GWh/y) (figure 24). This is because, when there is a lower electricity output from the solar park (6.10 GWh/y), the majority of it can be fed-back and barely anything remains for the production of hydrogen (table 18). For the solar parks with a higher electricity output (12.20 GWh/y and 24.40 GWh/y) the peaks in production cannot be distributed through the system due to cable capacities and therefore more electricity becomes available for the production of hydrogen.

TABLE 18–HYDROGEN BUFFER PERFORMANCE INDICATORS (GROUP 4)

AMOUNT OF HOUSES 80 116 80 307 80 707 CAPACITY GW 1.77 GWh 1.66 GWh 5.22 GWh 4.14 GWh 12.55 GWh 9.19 GWh KG OF HYDROGEN 53.975 kg 49.708 kg 156.704 kg 124.190 kg 376.490 kg 275.627 kg CUBIC METRE 1.349 m3 _{1.243 m}3 _{3.918 m}3 _{3.105 m}3 _{9.412 m}3 _{6.891 m}3 STARTING INVENTORY GW 0.06 GWh 0.09 GWh 0.04 GWh 2.29 GWh 0.01 GWh 5.07 GWh KG OF HYDROGEN 18.054 kg 27.470 kg 12.696 kg 68.641 kg 4.399 kg 152.127 kg -2 0 2 4 6 8 10 12 14 Hydr og e n b uf fe r in ve n tor y ( G W h ) Time (months)

(37)

37

FIGURE 24–HYDROGEN BUFFER BEHAVIOUR WITH 80 HOUSES (GROUP 4)

4.4. Electrolyser dimensions

This section of the results is dedicated to analysing the electrolyser dimensions per scenario group. The electrolyser dimensions are determined based on the assumption that all electricity available for hydrogen production is captured by the electrolyser.

Within scenario group one it is noticed that, the required electrolyser capacity for all three scenarios is almost the same. Even though, the total amount of hydrogen produced shows an increase whenever the electricity output of the solar park becomes higher (table 19). The increase in produced hydrogen between scenario one and two is 118%, whereas the increase between scenario two and three is only 46%. This indicates that, if the solar park output increases, the amount of electricity available for hydrogen production becomes more constant. However, due to the cable capacities, the system is not able to handle large peaks in production and therefore the peak inflow of the electrolyser remain fairly stable, while the peaks in production are lost as shown before in figure 12 and 13. This is also noticed by looking at the electrolyser utilisation as shown in figure 25 and 26. These two figures show that the electrolyser, with an inflow capacity of the highest encountered inflow peak, has to work closer to the maximum level more often if the solar park output connected to the system increases. Next to that, it is noticed that the electrolyser utilisation lies between 45% and 56% in this scenario group. This means that in these scenarios, the electrolyser is active between 11 and 13.5 hours per day on average throughout the year.

-2 0 2 4 6 8 10 Hydr og e n b uf fe r in ve n tor y ( G W h ) Time (months)

(38)

38

TABLE 19–ELECTROLYSER DIMENSION PERFORMANCE INDICATORS (GROUP 1)

FIGURE 25- ELECTROLYSER UTILISATION - YEAR OVERVIEW (GROUP 1)

FIGURE 26- ELECTROLYSER UTILISATION - ZOOMED IN ON 10 DAYS IN JUNE (GROUP 1)

4.4.2. Scenario group two

Within scenario group two, it is noticed that the required electrolyser capacity lies relatively close to each other for the same reason as in scenario group one. However, compared to scenario group one, the electrolyser utilisation as seen in figure 27, 28 and table 20, lies significantly lower than scenario group one. This is a logical consequence of prioritising electricity feedback over hydrogen production. Because the electrolyser only has to deal with the peaks that cannot be fed back into the main grid, the utilisation becomes lower. In comparison to scenario group one, the electrolyser

ELECTROLYSER DIMENSIONS SCENARIO

1 2 3

CAPACITY

PEAK INFLOW (KW) 4.139 kW 4.284 kW 4.460 kW

PEAK OUTFLOW (KW) 2.690 kW 2.785 kW 2.899 kW

PRODUCTION

TOTAL HYDROGEN PRODUCTION (GW) 2.25 GW 4.91 GW 7.15 GW

TOTAL HYDROGEN PRODUCTION (KG) 67.594KG 147.374 KG 214.629 KG

UTILISATION

ELECTROLYSER UTILISATION OVER 1 YEAR 45% 51% 56%

0% 20% 40% 60% 80% 100% El e ctr olys e r uti lis at ion Time (months)

(39)

39

is only active between 1.5 (scenario 4) and 7 hours (scenario 6) per day on average throughout the year.

TABLE 20- ELECTROLYSER DIMENSION PERFORMANCE INDICATORS (GROUP 2)

FIGURE 27-ELECTROLYSER UTILISATION - YEAR OVERVIEW (GROUP 2)

FIGURE 28-ELECTROLYSER UTILISATION - ZOOMED IN ON 10 DAYS IN JUNE (GROUP 2)

4 5 6

CAPACITY

PRODUCTION

TOTAL HYDROGEN PRODUCTION (GW) 0.23 GW 1.53 GW 2.66 GW

TOTAL HYDROGEN PRODUCTION (KG) 6.985 kg 45.906 kg 79.804 kg

UTILISATION

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% El e ctr olys e r uti lis at ion Time (months)

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% El e ctr olys e r uti lis at ion Time (days)

(40)

40 4.4.3. Scenario group three

Within scenario group three, the increase in electricity available for hydrogen production, compared to scenario group one and two, directly impact the electrolyser dimensions. As shown in table 21, the required capacity of the electrolyser is significantly higher than in scenario groups one and two. This is a logical consequence of relocating the electrolyser to the location of the solar park. By doing this, all peaks in electricity production can be captured by the electrolyser, and therefore the required electrolyser capacity is as high as the highest peak in the available electricity for hydrogen production. Concerning the utilisation of the electrolyser, in scenario group three the electrolyser is quite intensively used. Just like in scenario group 1, the electrolyser is active between 11 and 13.5 hours per day on average throughout the year. Noticed in figure 29 and 30 below, is that for all three scenarios in this group, the working times of the electrolyser are more or less similar to each other. The big difference lies in the amount of electricity that is processed by the electrolyser.

TABLE 21-ELECTROLYSER DIMENSIONS PERFORMANCE INDICATORS (GROUP 3)

FIGURE 29-ELECTROLYSER UTILISATION – YEAR OVERVIEW (GROUP 3)

7 8 9

CAPACITY

PRODUCTION

TOTAL HYDROGEN PRODUCTION (GWH) 2.25 GWh 5.95 GWh 13.70 GWh

TOTAL HYDROGEN PRODUCTION (KG) 67.612 KG 178.520 kg 411.115 kg

UTILISATION

(41)

41

FIGURE 30-ELECTROLYSER UTILISATION – ZOOMED IN ON 10 DAYS IN JUNE (GROUP 3) 4.4.4. Scenario group four

Within scenario group four, it is noticed the electrolyser capacity in scenario 10 lies fairly low (table 22). This is because most of the produced electricity in that scenario can be fed-back into the main grid. However, the required electrolyser capacity in scenario 11 and 12 lies remarkably higher. This is explained by the location of the electrolyser which makes it possible to capture all peaks in the electricity production rather than losing a large portion of it. Since the large peaks in production cannot be distributed through the system, they have to be captured by the electrolyser. Concerning the electrolyser utilisation as show in figure 31 and 32, it is noticed that the utilisation lies a fair bit lower than in scenario group three. This is caused by the combination of providing feedback to the main grid and the hydrogen production. If there is more electricity supply than demand, the priority is providing feedback to the main grid. This process does not activate the electrolyser. It only get activated when cable K-1 is congested, and therefore the electrolyser utilisation lies lower than in scenario group three.

Table 22 - ELECTROLYSER DIMENSIONS PERFORMANCE INDICATORS (GROUP 4)

10 11 12

CAPACITY

PRODUCTION

TOTAL HYDROGEN PRODUCTION (GWH) 0.25 GWh 2.58 GWh 9.23 GWh

TOTAL HYDROGEN PRODUCTION (KG) 7.662 KG 77.522 kg 276.797 kg

UTILISATION

(42)

42

FIGURE 31-ELECTROLYSER UTILISATION – YEAR OVERVIEW (GROUP 4)

FIGURE 32-ELECTROLYSER UTILISATION – ZOOMED IN TO 10 DAYS IN JUNE (GROUP 4)

4.5. Extension on the findings

The most promising scenario groups, that lie in line with the goal of the system, i.e., reduce electricity distribution grid load flow while fulfilling local hydrogen demand, are found to be scenario group three and four. Scenario group three shows that by locating the electrolyser close to the solar park, unnecessary distribution of electricity can be avoided and local hydrogen supply can be fulfilled. Scenario group four shows that it is possible to look at the combination of electricity feedback and hydrogen production as a trade-off. This trade-off can be put to use in the system configuration choices when implementing such a system. By taking a deeper look into the required electrolyser dimensions of the system. The required electrolyser capacity determined in this study

(43)

43

is equal to the highest peak in electricity available for hydrogen production. Since the electricity output from a solar plant is defined by its variability, i.e., peaks and troughs in electricity supply, this resulted in required electrolyser capacities up to 19.2 MW (scenario nine). However, by taking a closer look at the required electrolyser capacity in these scenarios, it is noticed that there is a trade-off to be found between choosing the capacity of the electrolyser and the amount of hydrogen produced. As shown in figure 33 below, in scenario nine, it is possible to use an electrolyser with a capacity of 60% of the highest peak in available electricity for hydrogen production (11.5 MW instead of 19,2 MW) and still capture 94% of the available electricity for hydrogen production (19.82 GWh instead of 21.08 GWh) resulting in 12.88 GWh in hydrogen instead of 13.70 GWh in hydrogen (6% loss). This pattern is noticed in all scenarios of this study. By choosing an electrolyser capacity of 50% of the highest peak, 87% to 88% of the electricity available for hydrogen can be captured, depending on the scenario. By choosing an electrolyser capacity of 60% of the highest peak, this value lies between 92% and 94% of the available electricity for hydrogen production, depending on the scenario.

FIGURE 33-ELECTROLYSER CAPACITY VS. CAPTURED ELECTRICITY AVAILABLE FOR HYDROGEN PRODUCTION |SCENARIO 9

This trade-off can be used in the determination of system configuration options when implementing a such a system. It would be possible to use an electrolyser smaller than the highest peak in the available electricity for hydrogen production and when the peak exceeds the electrolyser capacity, the remainder of the peak can be fed-back via the system until the cable capacity is reached. This way, an unnecessary investment in an electrolyser can be avoided while the system still reduces the electricity distribution grid load flow, while fulfilling hydrogen demand.

(44)

44

5 | DISCUSSION

The goal of this study was to explore different configuration options in a congestion management system using hydrogen production. This is done from an operations management point of view in order to explain the behaviour of this system and capture relevant decision options faced in implementing such a system in real-life. In this section of the thesis, the implications of the findings are elaborated on, after which the shortcomings of this study are discussed.

5.1. Implications of main findings

According to Gutiérrez-Martin et al., (2009), there should be a sufficient amount of surplus energy that reaches the electrolyser in order to render it cost-effective. In order to achieve this, this study shows that, a location close to the source of electricity, i.e., the solar park, is better than an electrolyser located elsewhere in the system. Whenever an electrolyser is located elsewhere in the system than where the electricity production takes place, the electricity needed for hydrogen production has to be distributed there. This results in high electricity grid load flow and congestion, Next to that, an electrolyser located far away from the solar park disables the ability to up-scale the production of hydrogen. This is because the amount of electricity available for hydrogen production would be capped by the distribution capacities of the electricity grid. In conclusion, in order to reach the goal of this congestion management system, i.e., lower the electricity grid load flow, and ensure a sufficient amount of surplus energy to reach the electrolyser, the electrolyser in this system should be located close to the solar park.

Continuing on the up-scaling of hydrogen production, it is noticed that, whenever upscaling is possible (e.g. scenario group three), the solar park with an output of 24.40 GWh/y results in hydrogen buffer requirement of 9.19 GWh for 707 houses (13 MWh per house supplied with hydrogen). The solar park with an electricity output of 6.10 GWh/y has a required hydrogen buffer of 1.66 GWh for 116 houses (14.3 MWh per house supplied with hydrogen). For the other scenario group (group four), in which the up-scaling of hydrogen production is also possible, a sort-like difference is noticed between the required hydrogen buffer for the small solar park and the big solar park. This indicates that, by locating the hydrogen production at a larger decentralized solar park, the required buffer capacity, in comparison with the amount of houses supplied with hydrogen, is lower.