Master’s Thesis TOM-DD Newcastle

(1)

Reducing grid balancing problems – use of hydrogen to match

electricity supply and demand

Master’s Thesis TOM-DD Newcastle

(EBM028A30.2019-2020)

University of Groningen

Faculty of Economics and Business

Newcastle University Business school at Newcastle university

MSc Technology and Operations Management and

(2)

Abstract:

Hydrogen is seen as one of the potential solutions to reduce congestion on the local electricity grid caused by the fluctuating supply of renewable energy sources (RES). However, hydrogen has its challenges and matching supply with demand is one of them. Research has focused on the use of hydrogen at either the demand-side or supply-side to optimize the usage of renewable energy. Hitherto, most literature neglects to see hydrogen production as an integral part of the entire energy system. For this study a simulation model has been designed for a rural town in the Netherlands, incorporating renewable electricity production, electrolysis, battery storage and different ways for fulfilling residential heating by means of hydrogen. Aiming to find the best match between sustainable energy supply and demand for residential heating. The findings in this study indicate that different ways of fulfilling residential heating causes trade-offs between the match of supply and demand. The trade-offs occur between the absolute required hydrogen buffer and electrolyser capacities on the one hand and relative required hydrogen buffer and curtailed energy on the other. This thesis contributes to the existing knowledge base by providing insights into choices in using hydrogen for fulfilling residential heating and their implications on decisions that have to be made on the supply-side.

Keywords:Hydrogen, Hybrid heat pumps (HHP), Renewable energy sources (RES), congestion, supply-side flexibility, demand-side management, storage, battery

Acknowledgments: I want to thank N.V. Rendo for allowing me to do an interesting research project

(3)

Reducing grid balancing problems – use of hydrogen to match electricity supply

and demand

1. Introduction

The fossil energy system is a key contributor to greenhouse gasses (GHG), mainly due to the CO2 it releases into the atmosphere (Höök and Tang, 2013). To reduce GHG emissions, many European governments have implemented policies to replace the fossil energy system with a system that is based on renewable energy sources (RES), such as wind power or solar energy (Union, 2009). They aim to achieve a 32% share of RES of the gross final consumption by 2030 (EEA, 2019). However, one of the difficulties faced in the energy transition from a fossil energy system to a RES system is the fluctuating electricity supply from RES due to weather-driven supply, limited dispatchability, and natural intermittency of these sources, which creates problems on the electricity grid when there is a significant supply (Bellocchi et al., 2019). Moreover, RES are predominantly located in rural areas, where the electricity infrastructure tends to have insufficient capacity to handle these peak production periods (Crabtree et al., 2011). Greater penetration of RES has resulted in increased levels of curtailment in recent years (Bird et al., 2016). One way to deal with the peak supply from solar or wind parks is to transform electricity into hydrogen. The production and storage of hydrogen can be beneficial for reducing congestion on the local grids (Korpas, 2017).

Green hydrogen can be produced by connecting off-grid power sources to an electrolyser (Bert, 2012). However, the electrolyser can only be loaded during the production time of the RES, and thus it cannot be evenly loaded. Due to the high costs of investing in electrolysers, it is generally concluded that intensive use of the electrolyser is necessary to ensure the investment is beneficial (Jørgensen and Ropenus, 2008; Rahil et al., 2016). A battery buffer can help increase the utilization of the electrolyser (Walker et al., 2016). Research has already shown that a supplementary battery leads to higher utilization of electrolysers and a better match of hydrogen production with hydrogen demand (Wermenbol, 2020). Nevertheless, for large-scale hydrogen production to occur, the costs of hydrogen storage and production need to be lowered (Kopp et al., 2017). This research analyses the opportunity to combine three different flexibility techniques – (1) storage, (2) demand-side management (DSM), and (3) supply-side flexibility – to both reduce the electricity curtailment and the additional costs. It does so by examining how these techniques can be used for fulfilling residential heating in a rural town in the Netherlands.

(6)

6 This research elaborates on the work of Wermenbol (2020), but, in contrast to his study, it also considers the demand side. The heating demand is to be fulfilled through heat pumps and hydrogen boilers and not solely by hydrogen. Demand-side management can provide system flexibility as it provides the option of switching from hydrogen to electricity and vice versa, therefore reducing the congestion on the grid when necessary but still keeping the costs of the production of hydrogen low (Grünewald et al., 2011).

This study’s contribution of also analysing the demand side is interesting for academics because the current hydrogen literature, which is based on mathematical modelling, lacks research which sees the hydrogen supply system as an integral part of the entire energy system. Thus, including the existing electricity grid, different electricity generation technologies, and hydrogen as the final product in the energy system design is an important move towards this integration (Welder et al., 2018). This broader perspective may offer more suitable solutions for reducing the costs of production and hydrogen storage (Li et al., 2019). Demand-side management practices should be complemented by a supply perspective (Beal et al., 2016). The expectation is that examining the demand side will influence, amongst others, the electrolyser capacity, the needed hydrogen storage at the supply-side, and the electricity demand.

To come to the above-described insights, the following research question is formulated: How can the best match between sustainable energy supply and demand for residential heating be realized in cases in which the supply of electricity results from RES while the demand is fulfilled by a combination of hydrogen and electricity?

(7)

7

2. Theoretical Background

The theoretical background focuses on the currently available literature on load mismatches caused by RES. It discusses hydrogen and batteries as energy storage and the role they can play in reducing the curtailment of electricity. Afterward, the section elaborates on the existing mismatches on the demand side of the grid and the way DSM can play a role in providing flexibility. Finally, it outlines an option for how to reduce the peaks on the electricity grid.

2.1 Load mismatches from a supply side perspective

Multiple studies have investigated the load mismatch between generated and consumed electricity produced by RES (for reviews see Orioli and Gangi, 2014, Yang et al., 2012). These studies have shown that both the supply and demand sides contribute to this mismatch.

(8)

8 carrier for helping decarbonize the energy systems and in oversupply situations, electrolysis can help prevent the curtailment from the generated electricity of the RES (Mulder, 2019). The next section now turn to hydrogen as energy carrier.

2.1.2 Hydrogen

The role of energy storage in the electricity grid has received renewed interest, and RES is the main driver behind this renewed interest (Denholm et al., 2010). As mentioned, hydrogen lends itself perfectly to being an energy carrier that can lower the electricity curtailment that occurs due to RES (Troncoso and Newborough, 2010). There are different kinds of production methods for hydrogen. The production of hydrogen considered in this paper is that of “green” hydrogen, which can be generated from RES in the process of electrolysis (Abdalla et al., 2018). An advantage of green hydrogen is that it does not emit CO2. A major disadvantage of the use of green hydrogen is the low round-trip efficiency of hydrogen when it is reconverted into electricity by fuel cells, which is about 30% (Pellow et al., 2015). Due to this low round trip efficiency, this study does not consider reconverting hydrogen back to electricity. In their paper, Dodds et al. (2015) mention that hydrogen comes into its own if it is used for space heating, water heating, and gas cooking. Thus, it might be possible to replace gas central heating with hydrogen central heating while providing the same service level to households (Dodds et al., 2015) with only minor adjustments to the gas infrastructure (Van Wijk, 2019). According to the available literature, hydrogen is a promising energy carrier that could replace the current gas central heating system in the Netherlands.

Nevertheless, if the electrolyser is only used when there are supply peaks, it would not be charged evenly. This would result in high investment costs for the electrolyser but low operating hours and probably high storage costs for the irregularly produced hydrogen. Currently, the main drawback of the production of hydrogen is the costs that are involved with the production and storage of hydrogen (Van Leeuwen and Mulder, 2018). A possible solution for this problem is to store electricity as an inventory buffer from RES to keep the electrolyser running when there is no wind or solar power available. An inventory buffer would thus extend the operating hours of the electrolyser. A promising way to extend the operating hours is to connect an electrolyser with battery storage, whereby the electricity can be consumed at a later stage (Tebibel, 2017). This process is discussed in the next section.

2.1.3 Battery storage

(9)

9 are generally feasible for short-term storage. There are many different battery technologies available, and each type has its advantages and disadvantages in terms of energy and power, efficiency, density, and costs (Lund et al., 2015). There is a lot of research on the different types of batteries and their advantages and disadvantages (for reviews, see Divya and Østegaard, 2009; Dell et al., 2001 and Lund et al., 2015). Nevertheless, due to the technological complexity of using different types of batteries in the model, this is considered outside the scope of this paper. The lithium-ion (Lio-on) battery is considered the most appropriate one for the buffer as it can help overcome the problem of the intermittency of RES and restrict the variability on the grid (Divya and Østegaard, 2009). While these batteries are easily portable and have high efficiency, they are really expensive for large-scale power storage (Sawle et al., 2018). Wermenbol (2020) has found that if PV and wind energy were combined, their complementary generation pattern could generate a more stable electricity supply to the electrolyser than either PV or wind energy alone. The use of such an energy mix would also reduce the necessary capacity of the electrolyser and battery.

2.2 Load mismatches from a demand side perspective

The electricity demand in the Netherlands is higher in winter than in summer (NEDU, 2020). In addition, the demand for natural gas is significantly higher in winter than in summer (NEDU, 2020). The higher demand for both gas and electricity in winter increases the mismatch with the produced electricity and hydrogen at the time if no storage options are available (Denholm and Hand, 2011). Increasing flexibility on the demand side (Lund et al., 2015) and storage capability are necessary to prevent such mismatches (Weitemeyer et al., 2015).

2.2.1 Flexibility

In the literature, a possible solution to avoid or reduce mismatches on the grid is system flexibility. System flexibility can be described as “the general characteristic of the ability of the aggregated set of generators to respond to the variation and uncertainty in net load” (Delholm and Hand, 2011). Lund et al. (2015) state that system flexibility is necessary to balance the supply and demand mismatches. Van den Burg et al. (2015) add that the integration of system flexibility in the energy system requires techniques that can reduce or absorb fluctuations in demand, supply, or both.

(10)

10 first approach, the grid ancillary services, are considered outside the scope of this study because grid ancillary services are not necessarily bound to RES power use, and this study aims to provide a solution in which houses are supplied as much as possible by RES.

2.2.2. Demand-side management (DSM)

Demand-side management can be defined as the planning, implementation, and monitoring of utility activities that are designed to influence customer use of electricity (Gellings, 1985). Demand-side management tries to find a response to produce desired changes in the load shapes of a power distribution system (Logenthiran et al., 2012). Gellings (1985) have identified six techniques for DSM to change these load shapes: load shifting, valley filling, peak clipping, strategic conservation, strategic load growth, and flexible load shape, which are shown in Figure 1.

Figure 1. Types of load shapes (Gellings, 1985)

Several studies have recognized that of all the demand-side resources, space heating and cooling have a high potential for load shifting (Braun, 2003; Houwing et al., 2010). Logenthiran et al. (2012) describe load shifting as taking advantage of time independence and shifting the loads from peak to off-peak time. To be successful with load shifting, it is important to find energy storage during peak load periods, which will have a minimum impact on the consumer’s comfort (Barzin et al., 2015). A promising technology that delivers load shape changes is the hybrid heat pump (HHP), discussed in the next section (Arteconi et al., 2013).

2.3 Combining heat pumps with hydrogen boilers

(11)

11 main reasons why it is difficult to replace these gas boilers. The first is that gas boilers are a strong incumbent technology, and people perceive them as safe, cheap, and easy to control (DECC, 2013). The second reason is that the infrastructure and markets for these gaseous heating fuels already exist (Dodds et al., 2015). Nevertheless, to become CO2 neutral it is important to identify ways to replace these natural gasses. A promising way to replace natural gasses which also considers the difficulties outlined is using an electric heat pump. Heat pumps could have an enormous role in this energy transition, as they are known for their low CO2 emission in the residential sector and their efficiency (Fischer and Madani, 2017). However, some drawbacks of electric heat pumps for

residential heating are that they increase the high peaks of demand in the electricity grid (Love et al., 2017), the losses of the already-existing gas infrastructures, and the high costs of the electric heat pump.

A potential solution to these issues is the HHP. In contrast to an electric heat pump, an HHP can switch from electricity to gas and therefore generates the potential to shift a percentage of the load from the heat-pump sub-system to the boiler sub-system in response to grid necessity (Stafford, 2017). Nevertheless, most studies so far have focused on a boiler-subsystem using natural gasses. However, natural gasses still emit CO2. Hydrogen boilers are promising, where hydrogen can be transported through the already-existing infrastructure of gas pipelines, can relieve the high peaks in the electricity grid when necessary, and is CO2 neutral. Furthermore, using an HHP instead of replacing natural gasses with hydrogen for residential heating is a promising way to lower the investment costs of hydrogen production: The energy from RES at oversupply moments can be used to convert the electricity into hydrogen, and so this energy does not have to be curtailed or exported to the main grid. For this to occur, it is important that there is a good balance between the produced and consumed electricity and hydrogen, where the performance indicators of reliability, costs, curtailment, and share of renewable energy sources are of crucial importance.

(12)

12 decisions in the system design will aid the decision-making process of an integrated hydrogen supply chain network.

(13)

13

3. Methodology

3.1 Justification of method

This research analyses the configurations of a hypothetical energy system containing production, storage, and consumption of energy. To answer the research question, a simulation-based quantitative study is conducted. The use of simulation makes it possible to compare alternative system designs, and to predict system performance (Robinson, 2004). This is appropriate as study sets to find out how to realize the best match between hydrogen supply and demand, and thus various cases should be compared to determine this. The simulation model is built in Microsoft Excel, where the simulation requires many calculations and a large amount of data.

3.2 Case description

The simulation is based on the available real-life data from a town in the northern part of the Netherlands. In a suburb of this town, an initiative for a local hydrogen economy is planned, with the aim of avoiding the use of natural gasses. The suburb will encompass 100 newly built stand-alone houses, and the residential heating for these houses will be provided by hydrogen and/or electricity. Next to these newly built houses are 400 existing houses that are connected to the electricity grid and included in this study. In line with a local hydrogen economy, the town will be supplied by solar parks and windmills.The strong expansion of RES generation is associated with increasing grid congestion, and many of the affected local electricity networks cannot handle the peak supply of the RES, resulting in curtailment (Schermeyer et al., 2018)

(14)

14

PV/wind

generation

Electrolyser

Local

electrcity

demand

Battery

Hydrogen

storage

Residential heating

+ DHW demand 100

newly built houses

2 Hydrogen

demand

Local grid

1 Main grid

4 Hydrogen/gas grid

2

1 Supply side

_{Demand side}

3

Figure 2. Simulation overview

3.3 The system

(15)

15 3.3.1 Supply side

For the third flow, when electricity generation exceeds local electricity demand, it is important to have priority rules that determine what to do with the surplus of electricity. The main purpose of these supply-side priority rules is to reduce the amount of electricity which would otherwise be curtailed or exported to the main grid. The priority rules in this flow are (1) supply local electricity demand; (2) produce hydrogen through electrolysis; (3) temporary electricity storage in batteries, in which this electricity will be fed to the electrolyser at moments when there is no oversupply; (4) export electricity to the main grid; (5) electricity will be curtailed. The priority rules at the supply side are indicated in Figure 2, numbered 1 to 4 on the left side of the figure.

3.3.2 Demand side

As described in the previous sections, two energy carriers are available, namely hydrogen and electricity. The energy demand of the houses in the simulation study can therefore be subdivided into hydrogen and electricity demand. The local electricity demand in this study is that of the 500 houses that are connected to the same electricity grid. The residential heating in this study is only that of the suburb, that is, the 100 newly built houses. The residential heating for the newly built houses will be provided through different combinations of hydrogen and electricity. On the demand side, the local electricity demand is fulfilled first. The priority rules for how to provide the residential heating demand are the simulation variables. For the 400 existing houses, it is assumed that the heating demand will always be fulfilled, and therefore the heating demand for these 400 houses is excluded in the model. The use of an HHP can reduce the hydrogen demand for the 100 newly built houses by switching from the boiler sub-system to the heat pump.

3.4. Data collection

(16)

16 Heat demand (100 newly built houses)

Inputs

Simulation

variables

Outputs

Local electricity demand

Heat demand

(100 newly built

houses)

Local electricity

demand

PV production

pattern

Wind production

pattern

Electrolyser

capacity

Battery

capacity

Residential

heating supply

PV/Wind mix

Electricity balance

Hydrogen

production and

consumption

Hydrogen buffer

Electrolyser and

battery

utalization

Allocation

hydrogen and

electricity

Figure 3. Overview of the simulation model

3.4.1 Data input

- Heat demand

(17)

17 Figure 4. Heat demand and domestic hot water profile of the 100 newly built houses over the year

- Local electricity demand

For this study, 500 houses of which 100 houses are newly built are connected to the electricity grid. The fractions of the electricity demand are gathered from a Dutch organization for energy data (NEDU, 2017), see figure 5. The average electricity consumption for one newly built house was 3,370 kWh per household per year. For one of the existing houses the average electricity demand was 2,660 kWh per household per year. This difference is because the newly built houses are larger than the already-existing houses. The fractions of the electricity demand was multiplied by the average electricity consumption per household per year and the number of houses connected to the grid. The electricity demand for the 400 existing houses remains the same regardless of how the residential heating is fulfilled. While the electricity demand of the 100 newly built houses varies depending on how space heating is fulfilled (see Section 3.6).

0 50 100 150 200 250 Av er age hour ly he at de ma nd (kWh) Time (hours)

Average hourly heat demand and DHW profile of 100 newly built

houses

(18)

18 Figure 5. Electricity demand fraction profile of an average household (NEDU, 2017)

- Photovoltaics (PV) production pattern

The solar production pattern is gathered from photovoltaic geographical information system (PVGIS;2017), where the PVGIS tool shows a complete data set of hourly PV power for a specific location. The peak power of the wind and PV park is calculated such that on a yearly basis the energy production always exceeds the expected combined demand for electricity and heat by 20%. The 20% extra production is chosen based on interviews with people who work in the energy company. The employees in the energy company are looking for a reasonable surplus that can be affordably generated in a short time in their plant and will produce enough energy despite the losses of the electrolyser efficiency. Different ways of fulfilling the residential heating are simulated, so the peak power will also be different depending on how the heating demand is fulfilled (see Section 3.6). Figure 6 shows the fraction of the maximum output that can be realized over the year. The figure shows that the output of PV is at its highest in the summer.

0 0,00002 0,00004 0,00006 0,00008 0,0001 0,00012 0,00014 0,00016 Fr ac ti on of e le ct ri ct y de ma nd pe r hou r Time (hours)

Demand fraction profile of an average household

Fractions of average hourly electricity demand

(19)

19 Figure 6. Photovoltaic generation profile (hourly output for one year)

- Wind production pattern

The wind speed hourly pattern over the year is gathered from the Royal Netherlands Meteorological Institute (KNMI, 2017). The different wind speeds correspond to the measured wind speeds from an adjacent weather station in the town. To determine the wind speed at a height of 100 m, the wind power law is applied (Peterson and Hennessey, 1978). The formula of this law is

𝑢 = 𝑢𝑟( 𝑧 𝑧𝑟 )

𝑎

where u is the wind speed in meters per second at height z in meters; ur is the known wind speed at

reference height; zr is the wind speed at the reference height; and a is an empirically derived coefficient

(Ibid.). The value of a is established at 0.143, where this is the average value for neutral stability conditions. The fraction of the maximum output that can be realized over the year is shown in figure 7. As can be seen, the fraction of the maximum output is very unpredictable, although the peaks seem to occur more frequently in the winter months than in the summer months. The calculation of the watt peak power of the wind park will be calculated in the same way as that of the solar park.

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 Fr ac ti on of ma x out pu t re al iz ed Time (hours)

Production profile of solar park

(20)

20 Figure 7. Wind generation profile (hourly output for one year)

3.5 Outputs

Table 1 shows the outputs of the simulation using the corresponding performance indicators. The first output, the electricity balance, measures the match between the supply of the RES and the local demand. Cao et al. (2013) discuss two commonly used basic matching indices for this match: the ‘on-site energy fraction’ (OEF) and ‘on-‘on-site energy matching’ (OEM). The OEF indicates the proportion of demand load covered by the on-site generated electricity, and the OEM indicates the proportion of supplied electricity that is used (Ibid). The grid dependency indicates the fraction of electricity that has to be imported from the main grid to fulfil the local demand. The second output monitors the amount of hydrogen consumed and produced each hour of the year. Based on the amount of hydrogen production, the percentage of the local electricity that is used for this production can be calculated. Furthermore, the effects of the simulation variables on the capacity of the hydrogen buffer are mapped. The bigger the required hydrogen buffer, the higher the costs, even though this provides a possible solution for the congestion problems. Therefore, is the required hydrogen buffer as a fraction of the total hydrogen production another performance indicator. The electrolyser and battery are

analysed using the operating hours and the percentage of used capacity. Lastly, the percentage of space heating demand covered by the heat pump and hydrogen boiler is measured, to monitor the effects of different ways of fulfilling heating demand.

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 Fr ac ti on of m ax out pu t re al ize d Time (hours)

Wind profile pattern

(21)

21

Outputs Performance indicators

Electricity balance - Grid dependency (%) - OEF (%)

- OEM (%)

- Electricity imported from the main grid (GWh) - Electricity exported to the main grid (GWh) Hydrogen production and

consumption

- Amount of hydrogen produced (kWh; kg) - Amount of hydrogen consumed (kWh; kg)

- Percentage of local electricity used for hydrogen production(%) Hydrogen buffer - Required capacity of the hydrogen buffer (kWh; kg)

- Required starting inventory (kg)

- Required capacity for the hydrogen buffer as a percentage of the total production (%)

Electrolyser utilization - Operating hours of the electrolyser in one year (hours) - Percentage of used capacity (%)

Battery utilization - Operating hours of the battery in one year (hours) - Percentage of used capacity (%)

Allocation hydrogen and electricity for fulfilling heating demand

- Percentage of space heating covered by heat pump (%)

- Percentage of total heat (including DHW) covered by heat pump (%) - Percentage of space heating covered by hydrogen boiler (%)

- Percentage of total heat (including DHW) covered by hydrogen boiler (%) Table 1. Overview of outputs with corresponding performance indicators

3.6. Simulation variables

- Allocation of hydrogen and electricity for residential heating

The simulation fulfilled the heat demand in different ways and affected the electricity and hydrogen demand in the system. First, the heat demand was simulated when residential heating was fulfilled by only the hydrogen boiler. Secondly, the heat demand was provided by means of a heat pump and a hydrogen boiler (the HHP), and the last scenario combined the use of an HHP with a DSM policy. The use of the heat pump for providing residential heating also creates a changing electricity demand for the 100 newly built houses compared to when only a hydrogen boiler is used. The DHW demand of the houses is always fulfilled by the hydrogen boiler because the heat pump is less efficient when it must provide DHW.

(22)

22 newly built houses are isolated well and the outdoor temperature ranges from -10°C to 35°C. Figure 9 shows the COP profile of the heat pump for different outside temperatures.

Figure 9. Coefficient of performance value for different outside temperatures derived from Haller et al., 2014

- Standard HHP rules

The temperature cut-off (Tcut-off) is defined as the outdoor air temperature where the HHP switches from the use of the heat pump to the use of the hydrogen boiler to provide residential heating (Dongellini et al., 2017). There is considerable disagreement in the literature about suitable values for this Tcut-off point. Since it will have a serious impact on the hydrogen demand, two Tcut-off points are simulated to measure the effects of the chosen Tcut-off points.

The first Tcut-off point is set at 0°C. This threshold corresponds with a COP value of 3.44. In Figure 10, the COP profile matching with the outside temperature for each hour of the year is shown. The Tcut-off point at 0°C implies that when the COP value drops below 3.44, the HHP uses the hydrogen boiler to fulfil the heat demand instead of the heat pump. Research has found the Tcut-off point at 0°C is the one that is the most economically viable (Bargella et al., 2016).

The second Tcut-off point is at , which is the temperature at which, according to Vuillecard et al. (2011) and D’Ettore (2019), the advantages arising from operating the heat pump do not outweigh the COP degradation due to the higher supply temperatures needed.

0 1 2 3 4 5 6 7 -8 -6,7 -5,4 -4,1 -2,8 -1,5 -0,2 1, 1 2,4 3,7 5 6,3 7,6 8,9 ₁₀,2 ₁₁,5 ₁₂,8 ₁₄,1 ₁₅,4 ₁₆,7 18 ₁₉,3 ₂₀,6 ₂₁,9 ₂₃,2 ₂₄,5 ₂₅,8 ,1₂₇ ₂₈,4 ₂₉,7 31 ₃₂,3 ₃₃,6

C

O

P

Temprature (

°C

)

(23)

23 Figure 10. Coefficient of performance values corresponding with outside temperature derived from Haller et al., 2014.

- The HHP with a DSM policy

The DSM policy is formulated in the following way: as soon as there is an overproduction of electricity, instead of using the hydrogen boiler the heat pump is used, even if the temperature is below the Tcut-off point. In this case, the HHP will use the heat pump instead of the hydrogen boiler to fulfil the heating demand. This policy decreases the hydrogen demand by making use of the surplus electricity to provide residential heating when possible.

- Electrolyser capacity

The efficiency of the electrolyser ranges between 65% and 80%, according to Barbir (2005), while Carmo et al. (2013) assume that the efficiency of the electrolyser ranges between 67% and 85%. In this study, the efficiency of the electrolyser is assumed to be 70%. The capacity of the electrolyser was minimized in half of the scenarios, such that enough hydrogen was produced to fulfil the hydrogen demand. In the other scenarios, the capacity of the electrolyser was restricted to 130 kW (residential heating fulfilled by hydrogen boiler), 100 kW (HHP having a Tcut-off point at 5°C), and 50 kW (having a Tcut-off point at 0°C). For the restricted capacities of the electrolyser, the battery size was minimized to fulfil the hydrogen demand.

The restricted capacities of the electrolyser depend on the hydrogen demand of the houses, which in turn depends on how residential heating is provided. If the restricted capacities were the same for the different ways of fulfilling residential heating, the effects of the battery in the different scenarios would

-10 0 10 20 30 40 50 60 0 1 2 3 4 5 6 7

Temp

ra

tur

e

in

°C

C

O

P

va

lue

Time (hours)

COP values over the year

(24)

24 not be measurable due to the enormous differences in capacities of the different ways of providing residential heating.

- Battery

The battery used in this research is a li-ion battery since this battery can reach an efficiency of nearly 100% (Chen et al., 2009; Lund et al., 2015), which this research assumed it does. The capacity of the battery was minimized for the restricted electrolyser capacities such that the hydrogen demand could be fulfilled.

- Allocation of wind versus PV power

The allocation of wind and solar power will range from 100% PV and 0% wind to 100% wind and 0% PV. The different combinations of RES influence when electricity will be available.

3.7 Overview of experimental variables

(25)

(26)

26

4. Results

The results of the simulation model are displayed and analysed in this chapter. In each section, the relevant settings are marked as a coloured box in Figure 11.

4.1 Base case: 50% PV and 50% wind

4.1.1 Residential heating provided by hydrogen boiler

(27)

27 100%PV-0% Wind 75%PV-25% Wind 50%PV-50% Wind 25%PV-75%Wind

PV-Wind

_capacity

Battery

Residential

_heating

0%PV-100% Wind

Minimized

100 kW

0 Minimized

Only

hydrogen

H

P

Electrolyser

capacity

Tcut-off point 0° C Tcut-off point 5° C

H

P

& DSM Tcut-off point 0° C Tcut-off point 5° C

130 kW

50 kW

Figure 13. Overview of the addition of a battery (Scenario 2)

(28)

28 Table 2. Results when the residential heating are supplied by hydrogen and the addition of a battery. 4.1.1.2 Hydrogen production

Details on the hydrogen buffer and production are presented in Table 3. Both Scenarios 1 and 2 use 0.84 GWh of electricity from the RES to produce hydrogen, which is 31% of the produced electricity from the RES that would otherwise be curtailed or exported to the main grid. The hydrogen

production and the hydrogen demand for each month are shown in Figure 14. For six months of the year, namely the first four months and the last two months, the hydrogen demand exceeds

production and heat demand is at its highest. The difference between hydrogen production and demand is especially high in the first month due to the seasonal mismatch.

Table 3. Hydrogen production and buffer details

Scenario 1 Scenario 2 Difference (%)

Capacities

Peak capacity solar park (kWp) 2,350 2,350 0.00%

Peak capacity wind park (kWp) 1,331 1,331 0.00%

Electrolyser size (kWh) 247 130 -47.37%

Battery size (kWh) 0 1,295

Electricty balance

Total electricity demand (GWh) 1,41 1,41 0,00%

OEF (%) 61.53% 61.53% 0.00%

OEM (%) 32.07% 32.07% 0.00%

Grid dependency (%) 38.47% 38.47% 0.00%

Electricity directly fed to electrolyser (GWh) 0.839 0.498 -40.62% Electricity fed from battery to electrolyser (GWh) 0 0.339

Electrolyser dimensions

Electrolyser utilization (amount of hours) 4,549 6,879 51.22% Electrolyser utilization (% of hours) 51.93% 78.53% 51.22% Electrolyser utilization (% of capacity overall) 38.79% 73.57% 89.66% Electrolyser utilization (% of capacity during operating hours) 74.69% 93.68% 25.43% Amount of hydrogen produced (kg) 17,624 17,593 -0.18%

Battery dimension

Battery utilization (amount of hours) 0 6,334 Battery utilization (% of hours) 0.00% 72.31% Battery utilization (% of capacity) 0.00% 47.79%

Hydrogen boiler

Scenario 1 Scenario 2 Difference (%)

Hydrogen production

Electricity used for hydrogen production (GWh) 0,84 0,84 -0.18% Percentage of (local) electricity used for hydrogen production (%) 30.96% 30.91% -0.16%

Hydrogen buffer

Required total capacity (kg) 5,811 6,012 3.46%

Required starting inventory (kg) 3,603 3,728 3.45%

(29)

29 Figure 14. Hydrogen production and demand in each month

The produced hydrogen in the absence of a battery is higher in the months when the average supply of the wind and/or solar park is quite high and vice versa (Appendix A). Thus, the addition of a battery balances the hydrogen demand and supply over the year in periods in which the production of RES is relatively low. The required capacity of the hydrogen buffer is equal to the maximum amount of hydrogen which must be stored in any given year, ensuring that the inventory that is built up is sufficient to meet the hydrogen demand throughout the year. In Scenario 1, the required hydrogen buffer is 5,811 kg, which is 33% of the total hydrogen production. In Scenario 2, the

required hydrogen buffer increases slightly to 6,012 kg, which is an increase of 3.46% from Scenario 1 and is 34% of the total hydrogen production. Due to the complementary character of PV and wind energy, there is a stable supply to the electrolyser and the effect from the battery on the required hydrogen capacity is negligible. To replace natural gasses with hydrogen for residential heating it seems necessary to import hydrogen or find cheaper ways of storing large amounts of hydrogen to become economically feasible. Such a large-scale energy storage method is found in salt caverns (Ozarslan, 2012). However, large-scale storage in these salt caverns is inaccessible yet. Therefore, the next section studies how the required hydrogen buffer can be reduced and the match between supply and demand can be improved using different ways of providing residential heating.

0 500 1000 1500 2000 2500 3000 3500 4000

H

yd

o

gen

p

ro

d

u

ctio

n

(k

g)

Time (months)

Hydrogen production vs demand

(30)

30

4.1.2 Residential heating provided by hybrid heat pump (HHP)

100%PV-0% Wind 75%PV-25% Wind 50%PV-50% Wind 25%PV-75%Wind

PV-Wind

_capacity

Battery

Residential

_heating

0%PV-100% Wind

Minimized

100 kW

0 Minimized

Only

hydrogen

H

P

Electrolyser

capacity

Tcut-off point 0° C Tcut-off point 5° C

H

P

130 kW

50 kW

Figure 14. Overview of residential heating provided by HHP

The previous section presented the hydrogen production and demand in cases in which the residential heating is provided by a hydrogen boiler. This section elaborates on the effect of providing residential heating with an HHP (see Figure 14) when compared to a hydrogen boiler. As aforementioned, two Tcut-off points were used to show the effects of the HHP. Scenario 3’s Tcut-off point is 0°C and Scenario 4’s is 5°C. The effects of the placement of the battery near the electrolyser have been left out of this section because the results of the effects of the battery (e.g., lower electrolyser capacity, increase of operating hours, etc.) compared to having no battery and

minimising the electrolyser are the same as described in the previous section. The DHW demand is constant every day and is equal to 876 kWh (26 kg) hydrogen a year for one newly built house.

4.1.2.1 Electricity balance

(31)

31 The peak power from both solar and wind park is recalculated resulting in a peak capacity of the solar park of 1,830 kWp and a peak capacity of the wind park of 1,036 kWp, such that yearly energy production exceeds the expected combined demand for electricity and heat by 20%. This decrease in the peak capacity of both parks can be explained using the efficiency of the heat pump. The

proportion of electricity that therefore is used and not exported to the main grid or curtailed increases by almost 27% (OEM), mainly due to the changed peak capacities of the wind and solar parks. In contrast, the OEF decreases by almost 8%, which can be explained due to the 7% increase of electricity demand compared to Scenario 1, in turn explained by the “extra” electricity use of the heat pump. Because less hydrogen needs to be produced over the year, the electrolyser will have lower operating hours when residential heating is provided by the HHP.

A possible way to increase the operating hours of the electrolyser is placing the battery near the electrolyser, as shown in Scenario 2. When comparing Scenarios 3 and 4, the operating hours increase from 3,892 hours to 4,227 hours per year, which is an increase of 8%. Because the

temperature is below or equal to 5°C in more hours (23% of the hours) of the year than for scenario 3, the hydrogen boiler covers 54% of the space heating and the operating hours increase. Thus, an increase in the Tcut-off point increases the operating hours of the electrolyser and the electrolyser’s capacity to fulfil the hydrogen demand.

Table 4. Results of using an HHP for two Tcut-off points 4.1.2.2. Hydrogen production

Details about the hydrogen buffer and production are presented in Table 5. For the Tcut-off point of 0°C (Scenario 3), the required capacity for the hydrogen buffer is 2,122 kg, which is 41% of the yearly

Difference (%) Difference (%)

Capacities

Peak capacity solar park (kWp) 1,830 -22.13% 2,055 -12.55%

Peak capacity wind park (kWp) 1,036 -22.16% 1,165 -12.47%

Electrolyser size (kW) 71 -73.99% 150 -39.27%

Total electricity demand (GWh) 1.51 7.09% 1.46 3.55%

OEF (%) 56.85% -7.61% 59.56% -3.21% OEM (%) 40.79% 27.18% 36.76% 14.61% Grid dependency (%) 43.15% 12.18% 40.44% 5.13% Electricity import (GWh) 0.65 16.83% 0.59 8.85% Electricity export (GWh) 1.00 0.01% 0.99 -1.64% Electrolyser use

Electrolyser utilization (hours) 3,892 -14.44% 4,227 -7.08%

Electrolyser utilization (% of hours) 44.43% -14.44% 48.25% -7.08% Electrolyser utilization (% of capacity overall) 39.80% 2.61% 39.09% 0.77% Electrolyser utilization (% of capacity during operating hours) 89.58% 19.93% 81.00% 8.45% Amount of hydrogen produced (kg) 5,198 -70.51% 10,785 -38.80% HHP distribution

Percentage of space heating covered by heat pump (%) 83.12% 45.55% Percentage of total heat covered included DHW by heat pump (%) 70.70% 38.75% Percentage of space heating covered by hydrogen boiler (%) 16.88% 54.45% Percentage of total heat covered included DHW by hydrogen boiler (%) 29.30% 61.25%

Hybrid heat pump (HHP)

(32)

32 hydrogen production that must be stored. The electricity that is used for hydrogen production is 0.25 GWh, which is a decrease of 70% as compared to Scenario 1. Thus, despite the reduced required hydrogen buffer, the use of an HHP decreases the percentage of electricity which used for the production of hydrogen when compared to Scenario 1. Figure 15 shows the hydrogen production and demand for each month when residential heating is provided by an HHP for the Tcut-off point of 0°C. The new hydrogen production pattern seems in general to have a better fit with the hydrogen demand pattern, except for the first month where there is a substantial difference between the hydrogen production and the hydrogen demand.

Therefore, it seems necessary to import hydrogen at least the first month to reduce the required seasonal hydrogen buffer and hence the percentage of yearly hydrogen production which should be stored. Appendix B visualizes the hydrogen demand and production for scenario 4. In Scenario 4 (Tcut-off point at 5°C), 22% of the local electricity is used for hydrogen production which will be otherwise curtailed or exported to the main grid. However, the required hydrogen buffer capacity is 5,031kg, which is 47% of the yearly hydrogen production that must be stored. This percentage is a high figure which indicates that a trade-off exists in the choice of the Tcut-off points (Scenarios 3 and 4) between the required hydrogen buffer capacity and the percentage of electricity that can be used for hydrogen production.

Table 5. Hydrogen buffer and hydrogen production details for HHP

Electricity used for hydrogen production (GWh) 0.25 -70.24% 0.51 -39.29% Percentage of (local) electricity used for hydrogen production (%) 11.73% -62.11% 21.76% -29.7%

Hydrogen buffer

Required total capacity (kg) 2,122 -38.80% 5,031 -13.42% Required starting inventory (kg) 2,005 -44.34% 3,579 -0.67% Required total capacity as percentage of total production (%) 40.83% 23.83% 46.64% 41.48%

(33)

33 Figure 15. Hydrogen demand and production for each month with HHP and Tcut-off of 0°C.

4.1.3 Residential heating provided by HHP in combination with DSM policy

100%PV-0% Wind 75%PV-25% Wind 50%PV-50% Wind 25%PV-75%Wind

PV-Wind

_capacityBattery Residential _heating

0%PV-100% Wind

Minimized

100 kW

0 Minimized

Only

hydrogen

H

P

Electrolyser capacity Tcut-off point 0° C Tcut-off point 5° C

H

P

130 kW

50 kW

Figure 16. Overview of residential heating provided by HHP in combination with DSM

This section elaborates on the effect of DSM in combination with the HHP (see Figure 16). As soon as there is an overproduction of electricity, the usage of a hydrogen boiler is replaced by that of a heat pump, despite a low COP value. The aim of DSM in this case is to lower the hydrogen demand in the

0 200 400 600 800 1000 1200 1400 1600 1800

H

yd

ro

gen

p

ro

d

u

ctio

n

(k

g)

Time (months)

Hydrogen production vs demand

Electrolyser capacity 71kW

(34)

34 winter months to ensure a better fit between hydrogen demand and production. Table 6 present the results of the HHP combined with DSM, where Scenario 5 has a Tcut-off point of 0°C, Scenario 6 has a Tcut-off point of 5°C, and the percentage difference indicates the difference between the outputs of the previous section when there is no DSM policy.

4.1.3.1 Electricity balance

When DSM is used, the space heating which is covered by the hydrogen boiler decreases by 11% for Scenario 5 and 16% for Scenario 6. As a result, the hydrogen that has to be produced decreases. The hydrogen demand can therefore be fulfilled with a smaller electrolyser capacity. For Scenario 5, in 59 out of 487 (12%) of the cases the COP is below the threshold but the heat pump replaces the

hydrogen boiler use due to an overproduction of electricity. For Scenario 6, this figure is 17% (343 out of 2018). Therefore, the effects of DSM are more noticeable for a higher Tcut-off point, where DSM can be used more frequently as there are more possibilities to apply it.

Table 6. Results of using HHP in combination with DSM policy

4.1.3.2 Hydrogen production

Figure 17 displays the hydrogen production and hydrogen demand for Scenario 5 in each month when the DSM policy is used. At first glance, the figure is quite similar to Figure 15. However, Figure 18 shows that the hydrogen demand in the first two months decreases when Scenario 3 is compared to Scenario 5. These are exactly the months where the hydrogen demand exceeds the production in Figure 15. Therefore, the required hydrogen buffer decreases by 10% for scenario 5 and by 13% for scenario 6 compared to when no DSM policy is used, as can be seen in Table 7. The percentage of the local electricity which is used for hydrogen production decreases when this DSM policy is applied

Capacities

Peak capacity solar park (kWp) 1,830 0.00% 2,055 0.00%

Peak capacity wind park (kWp) 1,036 0.00% 1,165 0.00%

Electrolyser size (kW) 67 -5.63% 129 -14.00%

Total electricity demand (GWh) 1.51 0,19% 1.48 0,82%

OEF (%) 56.93% 0.14% 59.89% 0.55% OEM (%) 40.93% 0.33% 37.27% 1.39% Grid dependency (%) 43.07% -0.19% 40.11% -0.82% Electricity import (GWh) 0.65 0.00% 0.59 0.00% Electricity export (GWh) 1.01 0.98% 1.04 5.05% Electrolyser use

Electrolyser utilization (hours) 3,892 0.00% 4,227 0.00%

Electrolyser utilization (% of hours) 44.43% 0.00% 48.25% 0.00%

Electrolyser utilization (% of capacity overall) 40.03% 0.58% 39.98% 0.89% Electrolyser utilization (% of capacity during operating hours) 90.09% 0.59% 0.83 1.85%

Amount of hydrogen produced (kg) 4,933 -5.10% 9,487 -12.04%

HHP distribution

Percentage of space heating covered by heat pump (%) 84.98% 2.24% 54.22% 19.03% Percentage of total heat covered included DHW by heat pump (%) 72.28% 2.23% 46.12% 19.02% Percentage of space heating covered by hydrogen boiler (%) 15.02% -11.02% 45.78% -15.92% Percentage of total heat covered included DHW by hydrogen boiler (%) 27.72% -5.39% 53.88% -12.03%

Influence of DSM policy

(35)

35 because the generated electricity in the overproduction will now also be used locally to provide residential heating using the heat pump. Demand-side management does not solve the problem of the seasonal mismatch on its own, where the main problem is still in the first month. Nevertheless, this DSM policy reduces the demand for hydrogen in the winter slightly and ensures a better match between the supply and demand of hydrogen in these months. The effects of DSM are more visible in Scenario 6 (Appendix C), where the DSM policy is used more as the heat pump can replace the hydrogen boiler for more cases, which provides a better match. Although, the problem of the seasonal mismatch is also for Scenario 6 clearly visible in the first month.

Table 7. Hydrogen buffer and hydrogen production details for HHP in combination with DSM policy

Figure 17. Hydrogen demand and production for each month with HHP and DSM (Scenario 5)

Electricity used for hydrogen production (GWh) 0.23 -5.10% 0.45 -12.01% Percentage of (local) electricity used for hydrogen production (%) 11.10% -5.10% 19.10% -12.01% Hydrogen buffer

Required total capacity (kg) 1,901 -10.43% 4,139 -13.42%

Required starting inventory (kg) 1,786 -10.94% 3,004 -15.13%

Required total capacity as percentage of total production (%) 38.54% -5.61% 43.63% -6.46%

Influence of DSM policy

Scenario 5 Scenario 6 0 200 400 600 800 1000 1200 1400 1600 1800

H

yd

ro

gen

p

ro

d

u

ctio

n

(k

g)

Time (months

)

Hydrogen production vs demand

(36)

36 Figure 18. Hydrogen demand in each month of Scenario 3 compared to Scenario 5.

4.2 Influence of PV and Wind combination

As of this point, each section assumed that the wind to PV ratio is 50-50. The following section elaborates on the effect that different combinations of PV and wind have on the outputs.

4.2.1 Electricity balance

Table 8 presents the OEF, OEM, and grid dependency for the different electricity generation scenarios. The same pattern that exists for the different ways of fulfilling the residential heating exists here, where the OEF reaches its lowest point for the scenarios in which the electricity is generated by 100% PV. The highest OEF is reached for the scenarios that have a combination of 75% wind and 25% PV, and when the electricity is generated by 100% wind, the OEF decreases. A similar pattern can be found for the OEM, where its lowest point is reached when electricity is generated by 100% PV, and the highest value of OEM is reached for a combination of 75% wind and 25% PV. The OEM decreases when the electricity is generated by 100% wind, which results in having to curtail or feedback to the grid a larger amount of electricity. In short, the performance indicators measures indicate that the best match between the supply of RES and the local electricity demand is the 75% wind and 25% PV combination.

Table 8 also presents the electrolyser capacities and battery capacities under different electricity generation scenarios and for different ways of fulfilling the heating demand. The electrolyser

0 200 400 600 800 1000 1200 1400 1600 1800 1-1-2017 1-2-2017 1-3-2017 1-4-2017 1-5-2017 1-6-2017 1-7-2017 1-8-2017 1-9-2017 1-10-2017 1-11-2017 1-12-2017 Hydrogen demand (kg) Ti me ( mon ths )

Effect on hydrogen demand with the use of DSM

(37)

37 capacities are at their highest when the energy mix is powered by 100% PV. The electrolyser

capacities are minimized when electricity is generated from 25% PV and 75% wind. These results align with the results of the electricity balance, where the highest OEF and OEM values are achieved at 75% wind and 25% PV. The electrolyser capacity increases for the 50% PV and 50 % wind

combination and when using 100% wind, although the differences are minimal.

(38)

38 Table 8. Electricity balance for different energy mixes and different scenarios of fulfilling residential heating

4.2.2 Effect of different energy mix on hydrogen production

Table 9 presents the details on the hydrogen buffer and the hydrogen production for the different energy combinations. The required hydrogen buffer decreases for every scenario in which PV substitutes wind when providing residential heating. The required hydrogen buffer always achieves its lowest point when the electricity is generated by 100% wind. Also, the used capacity as a percentage of yearly available capacity is at its lowest for 100% wind.

The inclusion of a battery results not always in a decrease of the required hydrogen buffer, depending on how the residential heating and energy mix is arranged. The battery has especially value when the electricity is generated by 100% PV, where the required hydrogen buffer for this energy mix for all the different ways of fulfilling residential heating decreases compared to the scenarios when no battery is used. A battery can store electricity and can use the electricity at a later stage for hydrogen production. For the scenarios when electricity is generated by 100% PV there are

Solar (%) 100% 100% 75% 75% 50% 50% 25% 25% 0% 0%

Wind (%) 0% 0% 25% 25% 50% 50% 75% 75% 100% 100%

Only hydrogen boiler

Electrolyser size (kWh) 365 130 307 130 247 130 233 130 256 130 Battery size (kWh) 0 1,896 0 1,478 0 1,295 0 1,760 0 4,836 OEF (%) 38.40% 38.40% 56.04% 56.04% 61.53% 61.53% 62.54% 62.54% 55.13% 55.13% OEM (%) 20.02% 20.02% 29.21% 29.21% 32.07% 32.07% 32.60% 32.60% 28.74% 28.74% Grid dependency (%) 61.60% 61.60% 43.96% 43.96% 38.47% 38.47% 37.46% 37.46% 44.87% 44.87% HHP with Tcut-off of 0°C Electrolyser size (kW) 99 50 84 50 71 50 67 50 72 50 Battery size (kWh) 0 375 0 272 0 193 0 174 0 321 OEF (%) 35.53% 35.53% 50.71% 50.71% 56.85% 56.85% 58.49% 58.49% 52.58% 52.58% OEM (%) 25.49% 25.49% 36.38% 36.38% 40.79% 40.79% 41.98% 41.98% 37.74% 37.74% Grid dependency (%) 64.47% 64.47% 49.29% 49.29% 43.15% 43.15% 41.51% 41.51% 47.42% 47.42%

HHP with DSM policy and Tcut-off of 0°C

Electrolyser size (kW) 96 50 82 50 67 50 61 50 64 50 Battery size (kWh) 0 351 0 253 0 150 0 110 0 166 OEF (%) 35.60% 35.60% 50.75% 50.75% 56.93% 56.93% 58.59% 58.59% 52.75% 52.75% OEM (%) 25.57% 25.57% 36.44% 36.44% 40.93% 40.93% 42.15% 42.15% 37.99% 37.99% Grid dependency (%) 64.40% 64.40% 49.25% 49.25% 43.07% 43.07% 41.41% 41.41% 47.25% 47.25% HHP with Tcut-off of 5°C Electrolyser size (kW) 217 100 182 100 150 100 141 100 155 100 Battery size (kWh) 0 866 0 654 0 480 0 485 0 1,039 OEF (%) 37.27% 37.27% 53.73% 53.73% 59.56% 59.56% 60.77% 60.77% 54.10% 54.10% OEM (%) 23.00% 23.00% 33.16% 33.16% 36.76% 36.76% 37.50% 37.50% 33.39% 33.39% Grid dependency (%) 62.73% 62.73% 46.27% 46.27% 40.44% 40.44% 39.23% 39.23% 45.90% 45.90% Required total capacity (kg) 6488 6012 5767 5489 5031 4992 4675 4839 4146 4486

(39)

39 short periods of electricity surpluses in the winter months. The possibility to store electricity during the scare oversupply moments benefits the hydrogen production in these winter months. When PV is substituted with wind the effects of the battery on the required hydrogen buffer for the different ways of providing residential heating diminish, where electricity generated from wind energy provides longer periods of oversupply in the winter months. A larger electrolyser capacity is able to get more benefit from this longer oversupply periods compared to a smaller electrolyser capacity with the addition of a battery. Therefore, the required hydrogen buffer does not benefit from a battery when a large part of the electricity is generated from windmills.

When applying the DSM policy, hydrogen demand is substituted with electricity demand when possible. The use of the HHP to fulfil residential heating in combination with DSM will change the electricity demand for each energy mix because the different generation patterns of PV and wind will generate overproduction at different times. Therefore, the times in which the hydrogen boiler can be replaced by the heat pumpdue to overproduction differ. The effects of the DSM policy are more visible when PV is substituted by wind energy; in wind energy, the moments of overproduction can better match hydrogen demand. Therefore, the DSM policy can be used more often to replace the hydrogen boiler with the heat pump. Figure 19 illustrates the number of hours that the heat pump replaces the hydrogen boiler for the different energy combinations and the DSM policy that can be applied.

(40)

40 Table 9. Required hydrogen buffer and electricity used for hydrogen production.

Solar (%) 100% 100% 75% 75% 50% 50% 25% 25% 0% 0%

Wind (%) 0% 0% 25% 25% 50% 50% 75% 75% 100% 100%

Only hydrogen boiler

Electrolyser size (kWh) 365 130 307 130 247 130 233 130 256 130

Battery size (kWh) 0 1,896 0 1,478 0 1,295 0 1,760 0 4,836

Required capacity of hydrogen buffer (kg) 8,654 7,570 7,261 6,745 5,811 6,012 5,078 5,760 4,119 4,833 Required capacity as percentage of total production (%) 49.15% 43.03% 41.24% 38.34% 32.97% 34.17% 28.80% 32.74% 23.38% 27.48% Electricity used for hydrogen production (GWh) 0.84 0.84 0.84 0.84 0.84 0.84 0.84 0.84 0.84 0.84 Percentage of electricity used for h2 production (%) 30.93% 30.90% 30.93% 30.90% 30.96% 30.91% 30.98% 30.90% 30.95% 30.90%

HHP with Tcut-off of 0°C

Electrolyser size (kW) 99 50 84 50 71 50 67 50 72 50

Battery size (kWh) 0 375 0 272 0 193 0 174 0 321

Battery size (kWh) 0 351 0 253 0 150 0 110 0 166

HHP with Tcut-off of 5°C

Battery size (kWh) 0 866 0 654 0 480 0 485 0 1,039

Battery size (kWh) 0 773 0 527 0 263 0 112 0 183

(41)

41 Figure 19. Hours that the DSM policy can be applied for different energy combinations.

0 100 200 300 400 500 600 100%PV 75%PV-25%wind50%PV-50%wind 25%PV-75% wind 100%wind

H

o

u

rs

wh

er

e

th

e

DS

M

p

o

lc

iy

ca

n

b

e

ap

p

lied

Energy mix

Effective cases where DSM can be applied for different

energy combinations

(42)

42

5. Discussion and conclusion

The rise of RES brings many challenges, one of which is the fluctuating supply of these sustainable sources. This fluctuating supply leads to problems on the grid, and, during oversupply, electricity grids become congested. One way to deal with the oversupply of solar or wind parks is to transform electricity into hydrogen. However, the unlimited production of hydrogen during oversupply

moments has its barriers, including that the electrolyser capacity and hydrogen storage to handle these oversupply moments are very expensive (Van Leeuwen and Mulder, 2018; Kopp et al., 2017). This study analysed how the hydrogen produced by the electrolyser can fulfil the need for residential heating to create a hydrogen supply system as an integral part of the entire energy system. This study aimed to give more specific insights into how hydrogen production can be integrated into the entire energy system, by taking both the supply and demand sides into account. This can create the best match between sustainable energy supply and demand for residential heating in a rural area. The integration of hydrogen production in the entire energy system produced several new insights.

This study has found that the necessary electrolyser capacity can significantly decrease when the heat demand is fulfilled by an HHP instead of a hydrogen boiler. The use of an HHP for fulfilling residential heating requires lower absolute hydrogen storage, where part of the hydrogen demand is replaced by electricity. However, the percentage of hydrogen which has to be stored from the total production increases, which seems to worsen the issues of seasonal mismatch. Nevertheless, except for the first month, the hydrogen demand compared with the hydrogen production seems to have a better fit when an HHP is used. Therefore, the HHP can bring the local hydrogen production and demand closer together if a solution can be found for the difference between supply and demand in the first month. Such a solution could be importing hydrogen in the first month, and further research can investigate the effects of importing hydrogen on the system.

(43)

43 The best match between the supply of RES and the local electricity demand is the 75% wind and 25% PV combination, which confirms Wermenbol’s study (2020). Previous studies have found that the complementarity generation pattern of PV and wind will result in lower energy storage requirements (Wermenbol, 2020; Heide at al., 2011). This study partially confirms these findings, where the lowest electrolyser and battery capacities are indeed achieved when a combination of PV and wind produce electricity. However, the lowest required hydrogen buffer is achieved when the electricity is supplied by 100% wind. All the simulations for fulfilling residential heating achieve the lowest required hydrogen buffer when this occurs. The effects of the DSM policy are most visible for the 100% wind supply as wind energy enables the moments of overproduction to better match the hydrogen demand. Therefore, the different energy combinations provide a trade-off in which one must decide what is most important: the electricity balance and the capacities of the battery and electrolyser, or the required hydrogen capacity.

This study shows that integrating hydrogen production in the entire energy system is important when making decisions that realize the best match between energy supply and demand. Decisions on how to fulfil residential heating can significantly reduce the demand for hydrogen and, hence, the required hydrogen buffer and electrolyser capacity. However, these different ways of providing residential heating have also an impact on several other performance indicators such as the

electricity balance, the relative required hydrogen storage, the percentage of electricity which is used for the production of hydrogen, and so on. As long as large-scale hydrogen storage is not available for solving the problems of the seasonal mismatch, it seems that there is not necessarily one best match between supply and demand in this research. Trade-offs between the electrolyser capacity, absolute required hydrogen storage on the one hand, and relative required hydrogen storage and curtailed energy on the other have to be considered.

5.1 Limitations and further studies

One of the limitations of this study is that the simulation is only done for 2017. All the model inputs are based on this year, although the patterns of these inputs will be somewhat similar in other years. The external validity of this study could be strengthened by simulating data for multiple years.

(44)

44 older houses. Further research could investigate the effects of the different ways of providing

residential heating discussed in this study when applied to older houses. The expectation is that this will have implications for the hydrogen demand, the electrolyser capacity and the required hydrogen buffer.

Furthermore, the peak powers of the wind and PV parks are calculated such that yearly energy production always exceeds the expected combined demand for electricity and heat by 20%. This will produce enough energy to have moments of oversupply and to not produce too much electricity which would have to be curtailed or exported to the main grid. A disadvantage of this assumption is that the peak powers of the solar or wind park change for the different ways of providing residential heating, which makes it difficult to compare the electricity balances. Future research could

Master’s Thesis TOM-DD Newcastle

Reducing grid balancing problems – use of hydrogen to match

electricity supply and demand