Hydrogen fuel stations as a network of buffers: resolving
the storage requirements in the hydrogen supply chain
Author
Lars Dieterman (S2940817/190615404)
l.dieterman@student.rug.nl Groningen, December 2020
Master Thesis (Dual Award)
MSc Technology & Operations Management (University of Groningen) MSc Operations & Supply Chain Management (Newcastle University)
Supervision University
Dr. Land (University of Groningen) Dr. Ying Yang (Newcastle University)
Supervision Field of Study
Mr. Zwetsloot MSc
Abstract
2
Acknowledgements
3
Table of Contents
1
Introduction ... 4
2
Theoretical background ... 7
2.1 The infrastructure ... 7
2.2 Supply and demand ... 8
2.3 Buffering ... 8
2.4 Distributed storage networks ... 9
2.5 Microgrids ... 9
2.6 Fuel stations to store hydrogen ... 10
3
Methodology ... 10
3.1 Experimental design ... 11 3.2 System input ... 14 3.3 Scenarios ... 16 3.4 Performance indicators ... 17 3.5 Sensitivity analysis ... 184
Results ... 19
4.1 Short-term case ... 21 4.2 Long-term case ... 24 4.3 Sensititvity analysis ... 295
Discussion ... 30
6
Conclusion ... 33
References ... 35
Appendices ... 39
Appendix A, simulation description ... 39
Appendix B, choice of experimental variables ... 41
Appendix C, choice of scenario values ... 42
Appendix D, base settings... 44
Appendix E, absolute values performance indicators ... 44
4
1 Introduction
Energy, in any form or shape, is one of the most valuable elements to mankind. Ever since the industrial revolution, the world has become more and more aware of the fact that the possessor of the source of energy is in possession of power as well. Transitions have been made in the past from one energy source to the other, causing shifts in power (Hancock and Vivoda, 2014). Currently, with the acknowledgement of being in a transition from fossil fuels to sustainable energy, governments, businesses and scientist are on the lookout for the next leap forward in terms of energy generation. However, although in the past the source of energy provided tangible energy resources such as wood, coal or oil, renewable energy sources are intangible (Crabtree et al., 2004). Wind or solar power for example have to be converted into something that can be saved and distributed. The dilemma surrounding energy becomes two sided, how to collect energy and what form to store it in. An answer to the second question seems negligible at first sight, but is more crucial than the first. In a future scenario, different energy sources can co-exist in harmony to provide all energy needed as long as the generated energy is converted into the same form or shape. Without doing so, energy will be lost by iteration of energy conversion or will not be transferable from source to user at all. Therefore, countries are trying to get a grip on the coordination of renewable energy sources while stimulating them as well.. For example, the Netherlands have recently, and quite drastically, implemented a new climate agreement in order to reduce emissions almost by half of what it was at the moment of signing the agreement by the year of 2030 (Ministry of Economic Affairs and Climate, 2019). In order to obtain this reduction, a fundamental change should be made from fossil fuels to renewable sources together with aligning the characteristics of converted renewable energy.
5 is to store generated energy in the form of hydrogen. Hydrogen has had a long history of being seen as the successor of petroleum and natural gas since electricity can be formed into liquid hydrogen or gaseous hydrogen, similar to petroleum and natural gas (Awad and Veziroglu, 1984). According to Andrews and Shabani (2012), hydrogen was thought of the one and only replacement for fossil fuels in a newly proposed hydrogen economy. Even though this concept is not generally supported, hydrogen often plays a major role in the possible adaptation of renewable energy (Balat, 2008). Nevertheless, there still are many barriers for the full implementation of hydrogen in the supply chain of renewable energy as a solution to the mismatch between energy supply and demand. Production costs and the availability of technologies have proven to be the main obstacles for the adaptation of hydrogen (Lemus and Martinez Duart, 2010).
Another challenge for the so-called hydrogen economy is its own match between supply and demand. Storing hydrogen is difficult and expensive, even though six different methods for storage have been distinguished (Züttel, 2004). This paper will mainly focus on one of those methods, storage in gaseous form under high pressure. Storage is needed to overcome the mismatch between supply and demand. According to Mulder et al. (2019), currently the only two proven methods for hydrogen storage are high-pressure storage for small scale storage and salt caverns for large scale storage. Although large scale storage would be feasible in some countries, the geographical limitations of salt caverns create an unattainable solution in other areas without any salt caverns (Michalski et al., 2017). Moreover, storing in salt caverns is a very costly operation due to the fact that the quality of stored hydrogen from salt caverns is insufficient for fuel cells, which makes the already expensive hydrogen even more expensive (Tarkowski, 2019).
6 from a commercial perspective (Brandsma, 2020). A future where hydrogen fuel stations are part of the infrastructure becomes not only more likely, but closer as well. Although one small buffer at a single hydrogen fuel station cannot cope with high variability, a larger network of small buffers may. The key element to such a network in practice is synergy. Such a solution is not necessarily cheaper. In order to make it more cost-efficient all elements in the hydrogen supply chain should be coordinated properly to make it function accordingly. Therefore, this paper proposes a novelty solution in the form of a network of hydrogen fuel stations to solve the mismatch in supply and demand of hydrogen with its main research question being:
How can distribution of storage across networks of hydrogen fuel stations resolve the storage requirements in the hydrogen supply chain?
7 innovative solutions with in practice experience makes this state-of-the-art research an interesting addition to the current literature.
2 Theoretical background
Research on the integration of hydrogen storage in the supply chain itself to be able to become self-sufficient is not exclusive to present literature. As early as 1997, research on the integration of a solar-hydrogen system for households was performed (Ulleberg and Mørner, 1997). According to Marbán and Valdés-Solís (2007), the hydrogen economy is a very promising solution for the energy problem, but the path to it is decisive for the success or failure. It is not without reason that hydrogen is a long-studied subject and that in the present time the hydrogen economy is far from in place. However, the potential of hydrogen has made national governments willing to invest their time and effort into hydrogen for a near future (Ministry of Economic Affairs and Climate, 2019).
2.1 The infrastructure
8 movement of hydrogen by trucks, while on the other hand a more distant future often involves the use of pipeline networks (Newman, 2020).
2.2 Supply and demand
The reason for an unbalanced supply and demand comes from the fact that hydrogen is generated by renewable energy sources that are dependent on natural influences (Jensen et al., 2007). Consequently, there are two effects on both the supply and demand side that augments the total effect. First, weather dependent renewable energy sources such as solar panels and wind mills deliver very volatile and sometimes unpredictable supply profiles (van der Wiel et al., 2019). Secondly, although industry demand and hydrogen demand for transportation is assumed to be relatively stable throughout the year, peak demand of households is very weather dependent as well and fluctuates heavily during the day (Wermenbol, 2020). For example, in colder climates peak demand is during winter since households require more heating (Ishaq et al., 2018). On the other hand, it should be acknowledged that in other parts of the world there is peak demand during summer due to the need for air conditioning. Generally, the imbalance of supply and demand for hydrogen is acknowledged and balancing is required (Menanteau et al., 2011). Aforesaid mismatch between supply and demand can be tackled at both the demand side and the supply side. Literature on demand side management mostly involves smart homes which assume to have some sort of energy storage capabilities to counterbalance the disparity (Tascikaraoglu et al., 2014). However, this only solves the problem for imbalance of household supply and demand, thus neglecting mobility and industry. Having recognized the need for a buffer for balancing supply and demand of energy, a fitting improvement is to create a buffer that is all-inclusive. Therefore, a more comprehensive solution is necessary to overcome complete imbalance.
2.3 Buffering
9 technical point of view. Storage seems to be seen as a burden, additional to, instead of integrated in, the hydrogen supply chain.
2.4 Distributed storage networks
Supply chain network design studies are mostly concerned about the (al)location of distribution centers for certain goods to meet demand under uncertainty (Shu et al., 2005). Such a typical supply chain network for goods is presented in figure 1 (Govindan et al., 2017) with a flow from suppliers to customers, but also intra-layer flows from one intermediate layer facility to another. The hydrogen supply chain can be put into a similar perspective as well with the electrolyzer as an upper layer facility and users of the fuel station as customers. However, the intra-layer flow is what is important for the proposed network of fuel stations. Intra-layer flow creates better optimization options within the supply chain in order to cope with uncertainties since the collective buffer is larger than a single buffer (Aghezzaf, 2005). Furthermore, the network aspect creates less need for capacity of one single buffer as long as the total capacity of all buffers, functioning as a network, can cope with the supply and demand uncertainties.
Figure 1: Supply chain network (Govindan et al., 2017).
2.5 Microgrids
10 own renewable energy providers through for example solar panels, their own electrolyzer and their own small capacity buffering facility. Moreover, it has been proven that a car can be used as a hydrogen buffer for a whole household (Cao and Alanne, 2015). A car is much more inefficient than a complete household when it comes to the use of hydrogen, making its onboard hydrogen tank a suitable buffer to compensate for demand and supply fluctuations. However, this is based on the assumption that each household is capable of generating enough energy in the first place. Although this may hold for rural areas, more urban environments (with a lot of high-rise) might struggle. Yet, the smart solution of using a car not only as a user of hydrogen, but also as a buffer of hydrogen does create incentives for other smart solutions as well.
2.6 Fuel stations to store hydrogen
As a part of the hydrogen supply chain, fuel stations already partly fulfill a storage purpose to be able to supply hydrogen continuously. In the same fashion as with cars in a microgrid, hydrogen fuel stations may function as a storage facility while being a user of hydrogen as well. However, the literature on using hydrogen fuel stations as a buffer for not only itself is practically non-existent. Moreover, literature on the topic of hydrogen fuel stations is limited in every way. Previous studies have shown that different storage methods are feasible in different situations. For example, liquid hydrogen storage is more feasible if the demand at the fuel station is relatively high, and gaseous storage is more feasible if demand is relatively low (Kurtz et al., 2019). The novelty of this subject makes it exciting and challenging at the same time. Combining the literature discussed so far, it can be concluded that the hydrogen supply chain is in need of a sufficient buffer to cope with demand and supply fluctuations. Network solutions and smart buffers with dual purpose create an incentive for a network of hydrogen fuel stations to be analyzed in detail. This paper will therefore contribute to the current literature of the hydrogen supply chain by researching the possible integration of the hydrogen fuel station into the hydrogen supply chain not only at the demand side, but as a distributed storage buffer in between supply and demand.
3 Methodology
In this section, the method of answering the research question is discussed.
11 This research will be performed based on a setting in the province of Groningen in the Netherlands in collaboration with Groningen Seaports. Even though many investments in hydrogen have been made in this area, there are not yet enough fuel stations in place to be able to quantify the data. Therefore, this research constitutes of a simulation of hypothetical scenarios. According to Karlsson (2016), a simulation-based quantitative study is the correct approach when dealing with non-existent operational processes. Furthermore, simulation research mainly focuses on operational processes that deal with variability, interconnectedness and complexity (Robinson, 2004). This applies to the proposed research where the overall model complies to these three key points and even more so the proposed network of buffers in itself is affected by variability, interconnected and more complex than a single buffer. The most obvious reason however remains that the current system does not exist and thus the correct approach is to make use of simulation (Robinson, 2004). The model is an hourly simulation of one full year, consisting of 365 days and 8760 hours.
This research will be split up into two main scenarios, a short-term case and a long-term case. In the short-term case, the simulation will be based on current supply chain configurations. Only the industry sector is currently making use of hydrogen together with one hydrogen fuel station. Both of these are directly connected to the source of hydrogen through a pipeline. Groningen Seaports is planning to build a distribution point for other sectors that are not connected by a pipeline to the hydrogen network. The distribution of hydrogen for all these other sectors comes from one filling station that can fill hydrogen tanks which are transportable by trucks over the road. On the other hand, in the long-term case it is assumed that all sectors in the supply chain are linked to a hydrogen supply network. Groningen Seaports has developed over the last few years a so-called ‘smart’ hydrogen pipeline that is made out of plastic. This pipeline is thinner and cheaper compared to conventional hydrogen pipelines. This strengthens the applicability of the setting in the long-term case, and ties the theoretical and practical value of this research. In both cases, the data on hydrogen demand and supply comes from Groningen Seaports, either by making use of historical data, or by making assumptions based on in-field experts from the firm itself. Therefore, the cases represent the practical value of a theoretical research question.
3.1 Experimental design
12 distinctive users, industry, mobility and households. However, the infrastructure between supply and demand side are different for both cases.
Short-term case
The similarities between the long-term case and the short-term case are that both have similar supply and demand, but the route from supply to demand is different. In figure 2, the simulated system is depicted. The main difference is that there is no pipeline network in place. In the short-term case, a renewable energy source (1) generates energy which is then converted by an electrolyzer (2) into hydrogen. Hydrogen is directly supplied through a pipeline to industry (3) and the filling station (4), capable of filling the large hydrogen tanks (5). These tanks can be transported to other users of hydrogen within the system, or exported. The integration of hydrogen fuel stations serves a triple purpose in the form of a so-called hub (6) in this setting. Assuming that households do not have access to the hydrogen network as well, hydrogen fuel stations will be used as a central storage facility for a neighborhood (7), while also capable of supplying hydrogen to mobility (7) and forming a buffer for demand and supply fluctuations. Although it can be used as a buffer for both households and mobility, it is assumed that it cannot for industry. Shortages for industry have to be imported externally (8).
Figure 2: Simulation model excluding pipeline network
13
Long-term case
In figure 3, the simulated system for the long-term case is depicted. A renewable energy source (1) generates energy which is then converted by an electrolyzer (2) into hydrogen. The hydrogen is supplied to households (3), industry (3) and a network of hydrogen fuel stations (4) that supplies hydrogen to mobility (7). Both household and industry, and hydrogen fuel stations have their own demand patterns. If hydrogen supply is relatively low and cannot satisfy all demand, households and industry have priority over fuel stations. This can be further explained by the working mechanisms of the proposed configuration. The hydrogen generated by the electrolyzer is used to supply industry and households. If supply exceeds demand, excess supply is directed to the network of hydrogen fuel stations (H2excess). There is one main hydrogen fuel station that can supply hydrogen to hydrogen tanks (5). This is done according to a certain demand pattern and those tanks are used for generators or ships for example. If all buffering capacity in the network of fuel stations is used and all demand for tanks is fulfilled, then additional tanks are filled and exported (H2export) to the hydrogen market (6). However, when there is a lack of hydrogen supply for industry and households, their demand is covered by using the hydrogen that is stored in the network of hydrogen fuel stations (H2short). This is prioritized over the filling of hydrogen tanks. If thereafter there is still a shortage, the hydrogen needs to be imported (H2import). By using a network of hydrogen fuel stations in such a manner it can be seen as a buffer that serves as a storage facility when there is excessive supply, and serves as an additional supplier when there is excessive demand.
14
3.2 System input
All data used in this simulation model is deterministic. No stochastic data will be incorporated. The reason for this is to guard the balance of the simulation model. If stochastic data would be included, buffers would increase to infinity if average supply and demand are equal. Even though in real life data is stochastic, this would not benefit the outcome of simulation model due to unbalanced results. The data used in this model is a combination of data from previous research at Groningen Seaports, assumptions made based on theory or in-practice experience or historic data figures that were not available at previous research stages. In the remainder of this section, the individual elements of the supply chain are evaluated and presented. A more detailed description of the simulation mechanics can be found in appendix A.
Hydrogen supply
For the supply side, a previous study’s findings for the NortH2 project are used (Nijnens, 2020). In that study, the NortH2 project is simulated with various settings for output of the electrolyzer. The study is based on the concept of a future windfarm of 10GW, which should be in place in 2040. However, due to this study’s more current setting, the most conservative supply profile is chosen in order to represent a more reasonable total supply. In this conservative supply profile, the 10GW windfarm is responsible for the supply of renewable energy, coupled with an electrolyzer capacity of 34% of the total windfarm capacity. This represents a total yearly hydrogen supply of 474 million kilograms. Since the total hydrogen supply used in this simulation is significantly smaller, the hourly supply pattern is converted into percentages of the total. Each hour represents a certain percentage of total supply. By doing so, the total amount of hydrogen supply can be altered without interfering with the supply profile.
Hydrogen demand
15 regulated more efficiently. Currently, at the only hydrogen fuel station in use by Groningen Seaports, five commercial busses each day make use of it with a very consistent demand pattern. Therefore, an on/off demand profile will be used. During the day, the demand profile shows a constant positive level for 12 hours, and at night the demand is equal to zero for 12 hours, creating the on/off setting. Finally, households will have detailed demand profiles based on historic natural gas profiles. It is expected that hydrogen demand by households will be similar to their current demand for natural gas (Lacko et al., 2014). This is due to the fact that hydrogen is used for heating purposes, similar to the usage of natural gas by households and is therefore assumed to take over that role.
Filling station (no pipeline network)
For the short-term case, where there is no network, the newly built filling station will operate like a regular hydrogen fuel station apart from the fact that it will only fill hydrogen tanks of 350kg. Mobility or any other elements cannot make direct use of the filling station. The filling station itself does not have any buffer capacity and is assumed to fill tanks in a continuous process. It can do so with a speed of 7.5kg of hydrogen per minute. However, to compensate for the operational time it takes to disconnect a full tank and connect an empty tank, it is assumed that the station is capable of filling exactly one tank of 350kg per hour.
Fuel station hub (no pipeline network)
16
Fuel station network
In the long-term case, the main difference for the network of fuel stations is that they are no longer supplied by tanks that are transported by trucks, but make use of a pipeline network. All households are also assumed to make use of a network and are no longer linked to stand-alone hydrogen fuel station hubs. Since the hydrogen fuel station is directly connected to the network, the hydrogen is supplied with a constant flow, instead of in batches similar to the truck transport for the hubs. Consequently, a re-ordering policy is not required. The hydrogen flow is determined by oversupply and shortages from the other sectors. In total, there are 32 fuel stations, each with their own buffer.
3.3 Scenarios
The different configurations of the experimental variables are determined by which of the two cases is being treated, the amount of hydrogen supply in total over the year, the ratio of either individual demand profiles or supply and demand ratio in total, and the buffer size per fuel station. In absolute values, both cases are different for all settings. For example, the expected amount of hydrogen supply in the short-term setting differs greatly from the expected amount of hydrogen supply in the long-term setting. In table 1 and 2, the experimental variables are given. In table 3, the scenarios are depicted in more detail of which there are 12 in total. These variables and the values of it are chosen based on theory, recommendations from in-field experts or based on model behavior. The choice of variables is explained in detail in appendix B. Furthermore, the reasoning behind each value is explained in appendix C.
Table 1: Short-term experimental variables Table 2: Long-term experimental variables
Short-term case Long-term case
Experimental variables Variation Experimental variables Variation
Total supply Expected Total supply Expected
High High
Demand ratios* 60%, 30%, 10% Supply-demand ratio 100%
50%, 25%, 25% 150%
Buffer size Low Buffer size Low
High High
17
Table 3: Scenarios Scenario Case Total hydrogen
supply
Demand ratios* Supply-demand ratio Buffer size 1 Short 14 000 000 kg 60%, 15%, 5% 100% 350kg 2 Short 14 000 000 kg 60%, 15%, 5% 100% 1050kg 3 Short 14 000 000 kg 50%, 12.5%, 12.5% 100% 350kg 4 Short 14 000 000 kg 50%, 12.5%, 12.5% 100% 1050kg 5 Short 20 000 000 kg 60%, 15%, 5% 100% 350kg 6 Short 20 000 000 kg 60%, 15%, 5% 100% 1050kg 7 Short 20 000 000 kg 50%, 12.5%, 12.5% 100% 350kg 8 Short 20 000 000 kg 50%, 12.5%, 12.5% 100% 1050kg 9 Long 28 000 000 kg 60%, 24%, 16% 100% 1050kg 10 Long 28 000 000 kg 60%, 24%, 16% 100% 3500kg 11 Long 28 000 000 kg 60%, 24%, 16% 150% 1050kg 12 Long 28 000 000 kg 60%, 24%, 16% 150% 3500kg 13 Long 40 000 000 kg 60%, 24%, 16% 100% 1050kg 14 Long 40 000 000 kg 60%, 24%, 16% 100% 3500kg 15 Long 40 000 000 kg 60%, 24%, 16% 150% 1050kg 16 Long 40 000 000 kg 60%, 24%, 16% 150% 3500kg
*Percentage of demand ratio for industry, mobility and households respectively
3.4 Performance indicators
18
Table 4: Performance indicators Performance indicator Case Explanation
Hours buffer empty Both The total hours that the buffer is empty and unusable to compensate for a shortage in supply. Mobility exclusion Both The total hours that mobility is excluded from using the hydrogen fuel station. Hours buffer full Long The total hours that the buffer is full and unusable to compensate for a surplus in supply. H2short Long The total yearly shortage of hydrogen supply,
compensated by the buffer.
H2excess Long The total yearly surplus of hydrogen supply, compensated by the buffer. H2import Long The total yearly shortage of hydrogen supply remaining after buffering. H2export Long The total yearly surplus of hydrogen supply remaining after buffering. Tank-orders Short The total yearly amount of tank-orders.
3.5 Sensitivity analysis
19 perhaps not future proof, assumptions. So, the mobility demand will be used to test for sensitivity. A change is made from the on/off setting, to a constant hydrogen demand from mobility. The total demand does not change, which results in the demand becoming more evenly spread over the hours.
4 Results
The results of the simulation models of both cases will be presented in this section. First, some general results will be presented such as the average supply pattern. After that, the cases will be presented separately, each with their own results. The key take-away points from both cases will conclude this section.
A commonality for both cases is the supply profile. Although the total supply varies between cases and scenarios, the average supply profile remains similar. For each hour, according to the profile, a certain amount of hydrogen will be supplied. To show what the seasonal profile looks like, the monthly average hydrogen supply in 24 hours of scenario 1 is given in figure 4. This shows that during the summer months supply is relatively low compared to winter months. This creates the trough in the middle of the graph starting in April and ending in September. Dependent on the demand profile, this could cause a shortage during summer and a surplus during winter.
20 Another commonality is the demand profile. For the supply side, a weekly figure would not give as good as an insight as an average yearly supply profile. However, for demand, the same cannot be said. Whereas the supply side is not influenced by human behavior, the demand side certainly is. Household demand is directly related to human activity and patterns throughout a day, and mobility as well. In figure 5, a detailed analysis of a hypothetical demand profile in a week is given where the share of total demand for each sector is equal. The industry demand pattern is the only demand sector that is not related to human activity and remains constant over time. The mobility demand pattern fluctuates between constant positive values during the day and being non-existent at night, while the household pattern follows a constantly changing pattern. The combination of the three demand patterns creates a heavily fluctuating demand pattern within 24 hour intervals, in this setting ranging from below 300kg of hydrogen per hour to over 1000kg of hydrogen. A clear day and night pattern can be recognized in the total demand profile of one week. To be able to compare it to the supply side, additionally in figure 6 the supply profile of a random week from the simulation is given.
Figure 5: Weekly Hydrogen Demand Profile (in kg per hour)
Figure 6: Weekly Hydrogen Supply profile (in kg per hour) 0 200 400 600 800 1000 1200 0 24 48 72 96 120 144 H yd ro ge n (k g p er h o u r) Hours
Industry Mobility Households Total
21 The main difference between supply and demand side profiles in this research is that one is affected by humans and the other is not. The combination of these creates a regular mismatch. In the simulation model, both short and long term cases will represent a possible solution to mitigate the imbalance between hourly supply and demand. Without any interference, it is impossible to make a hydrogen supply chain work due to consistent mismatch.
4.1 Short-term case
In the short-term case, a simulation model represents a supply chain configuration where household and mobility are relying on hydrogen supply through tanks transported by trucks to hydrogen fuel station hubs, and the industry sector is directly linked to the source through a pipeline. Before looking at the results of the simulation, the input gives some important considerations. The only element of the demand side directly connected to the hydrogen network is industry. However, since demand from industry is assumed to be constant over time, this does not alter the supply profile significantly. In figure 7, the supply profile of scenario 1 is given before and after industry demand, with total supply at 14 million kg of hydrogen a year. The filling station used to fill hydrogen tanks of 350kg will use the leftover supply after industry demand. In this setting, there is either enough hydrogen supply in one hour to fill a tank, or there is no supply at all.
Figure 7: Yearly supply profile 0 250 500 750 1000 1250 1500 1750 2000 2250 H yd ro ge n (k g p er h o u r) Months
22 Following the visualization of the supply side, it is clear that the filling station will either be able to supply hydrogen at a certain hour, or it cannot. If the filling station could supply hydrogen at all times, the total demand of households and mobility would be met regardless of the buffer size. In table 5, the output of the different scenarios for the short-term case is presented. Here, all grey shaded rows represent a scenario with a larger buffer size of 1050kg. The larger buffer size results in less hours empty and less hours of mobility exclusion compared to the smaller buffer size of 350kg, independent of the other experimental variables. Furthermore, the higher relative demand of households creates more uncertainty in the model, which results in more tank-orders and more hours of empty buffer and mobility exclusion. When comparing the results of scenarios 1, 2, 3 and 4 to 5, 6, 7 and 8 respectively, it can be seen that those are exactly equal. A higher total hydrogen supply seems to have no effect at all. This can be explained by the fact that the supply bottleneck in this case is not the total supply itself, but the filling station. Apparently, the total supply from the first four scenarios causes the filling station to either be able to supply one full tank of hydrogen per hour, or non. The same is true for the total supply from the last four scenarios from scenario 5 onwards.
Table 5: Simulation output short-term case (scenarios 1-8) Total supply Buffer
size
Demand ratios Buffer
empty Mobility exclusion Tank orders 1 14 000 000 kg 350kg 60%, 15%, 5% 297 403 226 2 14 000 000 kg 1050kg 60%, 15%, 5% 2 3 250 3 14 000 000 kg 350kg 50%, 12.5%, 12.5% 435 411 274 4 14 000 000 kg 1050kg 50%, 12.5%, 12.5% 14 10 310 5 20 000 000 kg 350kg 60%, 15%, 5% 297 403 226 6 20 000 000 kg 1050kg 60%, 15%, 5% 2 3 250 7 20 000 000 kg 350kg 50%, 12.5%, 12.5% 435 411 274 8 20 000 000 kg 1050kg 50%, 12.5%, 12.5% 14 10 310
23 where it does reach a hydrogen level of zero, the larger buffer is clearly able to consistently supply mobility and households with hydrogen throughout the year.
Figure 8: Buffer Behavior Comparison (Scenario 1 and 2)
Part of the fact that the lower buffer size reaches zero more often with the smaller buffer size is due to the re-ordering point. However, the re-ordering point for the smaller buffer size could not be higher than 1, as is the current setting. The reason for this is that hourly demand from households and mobility combined fluctuates between 1.4 in the lowest demand scenario and 27.7 for the highest demand scenario. If, for example, a re-ordering point of 5 is chosen, and the demand in the following hour equals 1.4, it would not be possible to add the full hydrogen tank of 350kg to the buffer. The buffer would overflow, meaning that the chosen re-ordering point is logical to be able to add the full 350kg of hydrogen supply to the buffer.
Although the short-term case shows evidence for matching supply and demand through the newly proposed system, this is overshadowed by the fact that a complete sector on the demand side is ignored when it comes to solving its shortage internally. An assumption is made that any shortages of the industry sector are compensated externally. In practice, this might be by the industry having its own buffer, or by importing their hydrogen elsewhere. Nevertheless, this is outside of the scope of this simulation. Therefore, the industry’s own supply and demand mismatch is not further regarded. Furthermore, currently all industry that uses hydrogen provided by Groningen Seaports is already connected to a hydrogen pipeline system due to said different demand characteristics compared to household and mobility. Moving industry
24 away from such a directly linked network to a transport through trucks by road would imply a step back in convenience.
The simulation model of the short-term case presents evidence for a working system of incorporating hydrogen supply in a nearby future setting that could overcome supply and demand mismatch if a large enough buffer is used. The total amount of tank orders in all scenarios remain below an average level of less than one per day, which is operationally feasible. This means that there is still enough supply left to cater other users of the filling station as well, for example for export of hydrogen to sectors left out of the scope of the simulation model. However, the actual solution to the industry sector’s shortages are outside of the scope of the short-term case, which leaves a supply and demand mismatch for one sector completely ignored. The results of this simulation present a good case for the integration of hydrogen fuel stations as buffer for the hydrogen supply chain, but an inclusion of all demand sectors should prove whether or not a similar statement can be made considering the long-term scenario.
4.2 Long-term case
25
Figure 9: Initial Supply and Demand Mismatch (Scenario 9 and 10)
From the yearly supply and demand mismatch it follows that the buffer within the network will have to deal with many fluctuations. The results of the simulation for the scenarios 9 to 12 are presented in table 6. The demand ratios for all scenarios are equal at 60%, 24% and 16% for industry, mobility and households respectively. Furthermore, the total supply in all scenarios is equal as well at 28 million kg of hydrogen per year. From these results, a comparison can be made between the different buffer sizes and the supply ratios. In appendix E, all results from all scenarios are presented in absolute values. In some cases, there is a difference between the total shortage of a certain scenario (H2import) and the total surplus (H2export), even if the ratio between supply and demand is 100%. This can be explained by the fact that the buffer at certain moments is either empty or full. This means that a time when it is full, it cannot decrease the surplus anymore in that hour. If later on in time, the buffer is empty in another hour, it cannot decrease the shortage. However, since the shortage does not equal the surplus at every hour, a difference could occur due to the buffer being full or empty at certain moments in time. Considering the implications of the higher supply-demand ratio in scenarios 11 and 12, it can be seen that the buffer continuously struggles to cope with the oversupply, both in the smaller buffer setting as in the larger buffer setting. There is a relatively small difference between the hours full between scenario 11 and 12. On the other hand, the larger buffer size in this case is able to compensate 90% of shortage, which is the highest of all. From this, it can be concluded that the if oversupply is deliberate and can be exported for other purposes, the network of buffers of 3500kg are a suitable storage solution for remaining shortages.
26
Table 6: Simulation output long-term case (scenarios 9-12)
9 10 11 12
Supply-demand ratio 100% 100% 150% 150%
Buffer size 1050kg 3500kg 1050kg 3500kg
Buffer empty (in hours) 1347 764 882 218
Buffer full (in hours) 2840 1607 5241 4698
Effect on surplus (H2export) -47% -70% -17% -26%
Effect on shortage (H2import) -47% -69% -60% -90%
Comparing the buffer behavior of scenario 9 and 10, it becomes evident that the larger buffer size is more capable of coping with the unbalance in supply and demand. In figure 10, the buffer behavior of scenario 9 is represented. Almost half of the time throughout the year the buffer is either empty or full. Consequently, the purpose of the buffer, to compensate for supply and demand mismatches, cannot be fulfilled and makes it useless during those periods. It can be concluded that the relative buffer size in scenario 9 is too small to make it operationally feasible. Not only does it have a relatively small impact on supply and demand mismatch, it also would create problems functioning as a fuel station as well. Mobility trying to make use of it would face no supply during 1347 hours (which is more than 56 days) per year in total.
27
Figure 11: Buffer Behavior (Scenario 10)
From the results in table 6, it seems that a bigger buffer has a positive effect on the capability to deal with supply and demand fluctuations. In figure 11, the buffer behavior of scenario 10 is presented. Compared to figure 10, the hydrogen level of the buffer is more often usable for excess supply or excess demand. Where the smaller buffer size seems to be insufficient operationally, the larger buffer size on the other hand does seem to be effective and usable in practice. Nevertheless, the larger buffer size of 3500kg is not capable of compensating for all supply and demand mismatches. Exporting hydrogen surplus, and importing hydrogen shortage is still necessary, even with the larger buffer. An even bigger buffer size might overcome this problem, but a buffer size of 3500kg is a reasonable maximum size of a buffer at a single hydrogen fuel station. If it could have been proven that a smaller buffer size than 3500kg would be sufficicient storage to not have any remaining shortage and surplus, then that would be the preferred option. However, since the maximum size of 3500kg is still not able to do so, this is the preferred solution at the moment.
28
Figure 12: Supply and Demand Mismatch after buffering (Scenario 10)
Considering the scenarios with higher supply, the results can be found in table 7. This time, the common setting is a demand ratio of 60%, 24% and 16% for industry, mobility and households respectively, along with a higher volume of total supply at 40 million kilograms of hydrogen per year. In comparison with the results from the scenarios with a lower supply level, it can be seen that the increased volume puts more stress on the buffer performance, resulting in lower effects for all comparative cases. Consistently, the effect on either surplus or shortage is lower than in the case with regular supply levels. All other observations are similar to scenarios 9 to 12. In figure 13, a comparison is made of the effect of the buffers on H2export between scenarios where only the total supply volume differs. From this, it can be concluded that the higher supply has a negative effect on the capability of matching supply and demand.
Table 7: Simulation output long-term case (scenarios 13-16)
13 14 15 16
Supply-demand ratio 100% 100% 150% 150%
Buffer size 1050kg 3500kg 1050kg 3500kg
Buffer empty (in hours) 1532 936 1098 393
Buffer full (in hours) 3250 1961 5430 4857
Effect on surplus (H2export) -39% -63% -14% -24%
Effect on shortage (H2import) -40% -63% -50% -82%
29
Figure 13: Comparison of effect on total surplus between similar scenarios with different total supply volume
The results of the simulation model of the long-term case provide evidence for an all-inclusive buffer that creates synergy between different sectors of the supply chain. In an equal supply-demand setting, approximately 70% of all supply-demand and supply mismatch can be overcome by incorporation of a network of buffers at hydrogen fuel stations directly linked to the supply chain network of hydrogen. Compared to the short-term case, the results have proven to have an effect on each and every sector. However, where in the short-term case the hub could provide complete cover of demand for two sectors apart from 2 hours per year, the buffer at the long-term case suffers from too high fluctuation volumes to be able to do so.
4.3 Sensititvity analysis
To test the sensitivity of the simulation, scenario 9 and 10 have also been simulated with a different supply profile from the study by Nijnens (2020) and by altering the demand profile. These scenarios are selected based on the fact that they are representative scenarios for the long-term case with both the small and large buffer size included. First, a comparison is made with a supply profile that has less supply limitations as a result of a higher capacity electrolyzer. In all scenarios, the supply profile stays consistent. However, it is expected that a setting with less supply limitations is much more capricious than the original setting with stronger supply limitations. In table 8, the results of the sensitivity analysis are presented. As expected, the higher level of uncertainty in the supply level is causing more stress on the buffer which results in more hours full or empty for example. However, similar to the original setting, the effect of a higher buffer size has a constant effect on output in both settings. Where the greater buffer size in the original supply setting has a constant effect of approximately -43% on output, the
-80% -70% -60% -50% -40% -30% -20% -10% 0%
9 &13 10 & 14 11 & 15 12 & 16
Ef fe ct o n H 2e xp o rt Scenarios
30 greater buffer size with a different supply profile has a constant effect of approximately -35%. Therefore, the absolute results of all scenarios are senstitive to different supply profiles, but the delta of output between two scenarios remains constant.
Table 8: Sensitivity analysis supply
Scenario 9 Scenario 10 Delta 9-10
Supply limitations High Low High Low High Low
Buffer empty (in hours) 1347 2586 764 1682 -43% -35%
Buffer full (in hours) 2840 2682 1607 1747 -43% -35%
H2export (in tons) 108.0 190.5 60.8 123.5 -44% -35%
H2import (in tons) 108.0 190.2 62.0 124.4 -43% -35%
Furthermore, one of the assumptions in the model is the demand profile of mobility. To see how sensitive the outcome is to this profile, an alteration is made. In the simulation model, it is assumed that the mobility profile relies on an on-off setting. Meaning that there is only mobility demand during 12 hours of the day. However, in this sensitivity test this assumption is removed and mobility demand becomes constant throughout the full 24 hours of a day. In table 8, the results are presented. What can be found, is that the performance indicators related to matching supply and demand are not sensitive to a change in mobility demand while all being less than 1%. This indicates robustness of these findings. However, the buffer behavior itself is influenced. Mobility is excluded more often, which is a logical consequence due to a doubling in the amount of mobility demand hours while maintaing equal total demand. The total hours of emptiness at the buffer remains similar, but the total hours of a full buffer is reduced. This can be explained by the fact that the demand is spread over all hours, and thus reduces the amount of times that the buffer reaches its full state. In appendix F, the absolute values of the sensitivity analysis can be found.
Table 9: Sensitivity analysis mobility demand
Percentage change Percentage change
Scenario Buffer empty Buffer full Mob. excl.*
H2excess H2short H2export H2import
9 <1% -20% -57% <1% <1% <1% <1%
10 <1% -24% -55% <1% <1% <1% <1%
5 Discussion
31 new storage solution within the hydrogen supply chain to make a leap forward in terms of efficiency and likeliness of a hydrogen economy.
From the first part of this study, the findings in the short-term case indicate that the integration of a hydrogen fuel station in the form of a hub for the sectors households and mobility could overcome their supply and demand mismatch. A positive relationship between buffer size and the effect on matching supply and demand can be found. This relationship is expected and the behavior of the buffer does not surprise. Interestingly, the simulation model provided new insights considering the limited importance of total supply. It was shown that a change in total supply does not change the buffer behavior, as long as the minimum hourly supply at the moments that there is supply of hydrogen exceeds the volume of one tank of hydrogen. It can be concluded that supply is limited not by the supply profile, but by the operational speed of the filling station. Especially for the short-term case, previous research is limited. Therefore, the findings cannot be directly compared to preceding findings. Nevertheless, compared to previous literature on smart homes for example (Tascikaraoglu et al., 2014), in a similar fashion the benefits of small-scale buffering at the demand side are found. Due to a lack of data on operational costs of this configuration in practice, the results cannot be evaluated in terms of financial feasibility. However, if mobility starts making more and more use of hydrogen, hydrogen fuel stations are bound to exist. Not only using it for the mobility sector, but also including households similar to this study will be an efficient measure that is at least financially interesting to further investigate. A more concrete limitation of the short-term case studied in this research is the exclusion of the industry sector from making use of the storage solution at hydrogen hubs. Due to the operational inefficiency of a hydrogen buffer indirectly connected through transport by trucks and the high demand of industry, such a buffer is only usable by households and mobility. This does limit the findings considerably. With that taken in mind, the generalizability of the results from the short-term case have to be restrained to hydrogen supply chain configurations eliminating the industry sector’s shortages.
32 the filling station in the short-term case. As a result, the buffer performance is influenced by the volume of total supply. This is due to the fact that the buffer size becomes relatively small compared to the amount of supply and demand variability. For smaller buffer sizes in various scenarios, the buffer becomes unstable and it is constantly moving quickly from an empty state to a full state, or vice versa. Consequently, the basic operational purpose of a hydrogen fuel station, supplying hydrogen to mobility, becomes impossible. In light of previous research, the effect of the buffer does show similarities to existing literature. Where Cao and Alanne (2015) found a positive relationship between incorporating a car itself as a hydrogen buffer for a household and matching supply and demand, the same can be said about incorporating hydrogen fuel station in a more extensive supply chain configuration considering other sectors as well. As with the short-term case, the financial feasibility can only be confirmed if future research will be done on the efficiency gains over costs.
As a result of the sectorial conditions of the short-term case and the remaining storage requirements in the long-term case, a compromise remains. Either partial demand can be fully matched through buffering, or full demand can only be matched partially with supply. Regardless, in both cases storage requirements still exist after buffering. Considering other limitations, this study has been built on a mixture of historic data and assumptions. The assumptions are sufficient for the purpose of this research, but cannot create the same confidence as historic data. Moreover, data used for this study comes from one region in the Netherlands. Acknowledging the fact that the supply profile is specific for this case, along with the household demand profile which is based on Dutch households, there are regional limitations. These limitations make it harder to generalize the results, the greater the geographical distance becomes. Nonetheless, the results are uniformly applicable for regions that have similar climates due to the fact that most other constraints in the model are technological constraints that are not bounded by its regional descent.
33 the fact that this research was done in cooperation with inside experts on the topic of energy transition makes this a practically applicable study within the literature.
6 Conclusion
This study aimed to identify the effect of a novelty solution in the form of the integration of hydrogen fuel stations as a buffer within the hydrogen supply chain. Based on a quantitative analysis, it can be concluded that the integration of hydrogen fuel stations has a positive effect on matching supply and demand. The results indicate that a high enough buffer size could reduce the storage requirements significantly.
Summarizing, the two cases have each provided their answers to the main research question. What can be learnt from the study as a whole is the fact that two distinctive cases have proved the possible efficient integration of hydrogen fuel stations in their own respective supply chain configuration. However, in answering the research question both cases are limited to a certain extent. A trade-off exists between resolving the storage requirement for selected sectors, or partly resolving the storage requirements for all sectors. Therefore, the integration of a network of hydrogen fuel stations can reduce the storage requirements in the hydrogen supply chain significantly, but it cannot resolve all storage requirements.
34 In conclusion, this study’s main findings contribute to the current literature on hydrogen supply chains by proposing the integration of hydrogen fuel stations as buffers. The insights provided introduce a practical application to an element of the hydrogen supply chain that is bound to exist if hydrogen proves to become an important energy carrier. A gap in the literature left within the self-sufficient microgrids studies is addressed, where its applicability is evaluated on a macro level. The results confirm the expected positive effect on the ability to match all-inclusive supply and demand.
35
References
Aghezzaf, E., 2005. Capacity planning and warehouse location in supply chains with uncertain demands. Journal of the Operational Research Society, 56(4), pp.453-462. Agnolucci, P., Akgul, O., McDowall, W. and Papageorgiou, L., 2013. The importance of economies of scale, transport costs and demand patterns in optimising hydrogen fuelling infrastructure: An exploration with SHIPMod (Spatial hydrogen infrastructure planning model). International Journal of Hydrogen Energy, 38(26), pp.11189-11201.
Almansoori, A. and Shah, N., 2012. Design and operation of a stochastic hydrogen supply chain network under demand uncertainty. International Journal of Hydrogen Energy, 37(5), pp.3965-3977.
Andrews, J. and Shabani, B., 2012. Where does Hydrogen Fit in a Sustainable Energy Economy?. Procedia Engineering, 49, pp.15-25.
Awad, A. and Veziroglu, T., 1984. Hydrogen versus synthetic fossil fuels. International
Journal of Hydrogen Energy, 9(5), pp.355-366.
Balat, M., 2008. Potential importance of hydrogen as a future solution to environmental and transportation problems. International Journal of Hydrogen Energy, 33(15), pp.4013-4029.
Barthelemy, H., Weber, M. and Barbier, F., 2017. Hydrogen storage: Recent improvements and industrial perspectives. International Journal of Hydrogen Energy, 42(11), pp.7254-7262.
Brandsma, R., 2020. Holthausen en Hyzon bouwen in Winschoten vrachtwagens op waterstof. Fabriek levert honderden banen op. [online] Dvhn.nl. Available at:
https://www.dvhn.nl/economie/Holthausen-en-Hyzon-bouwen-in-Winschoten-fabriek-voor-vrachtwagens-op-waterstof-26205916.html
Cao, S. and Alanne, K., 2015. Technical feasibility of a hybrid on-site H2 and renewable energy system for a zero-energy building with a H2 vehicle. Applied Energy, 158, pp.568-583.
Crabtree, G.W., Dresselhaus, M.S. and Buchanan, M.V., 2004. The hydrogen economy.
Physics Today, 57(12), pp.39-44.
Dagdougui, H., 2012. Models, methods and approaches for the planning and design of the future hydrogen supply chain. International Journal of Hydrogen Energy, 37(6), pp.5318-5327.
Dunn, S., 2002. Hydrogen futures: toward a sustainable energy system. International Journal
36 Govindan, K., Fattahi, M. and Keyvanshokooh, E., 2017. Supply chain network design under uncertainty: A comprehensive review and future research directions. European Journal of
Operational Research, 263(1), pp.108-141.
Gray, E., Webb, C., Andrews, J., Shabani, B., Tsai, P. and Chan, S., 2011. Hydrogen storage for off-grid power supply. International Journal of Hydrogen Energy, 36(1), pp.654-663.
Ishaq, H., Dincer, I. and Naterer, G., 2018. Performance investigation of an integrated wind energy system for co-generation of power and hydrogen. International Journal of Hydrogen
Energy, 43(19), pp.9153-9164.
Jensen, S., Larsen, P. and Mogensen, M., 2007. Hydrogen and synthetic fuel production from renewable energy sources. International Journal of Hydrogen Energy, 32(15), pp.3253-3257.
Jepma, C. J., Spijker, E., and Hofman, E. 2019. The Dutch Hydrogen Economy in 2050.
Hancock, K.J. and Vivoda, V., 2014. International political economy: a field born of the OPEC crisis returns to its energy roots. Energy Research & Social Science, 1, pp.206-216.
Karlsson, C., 2016. Research Methods for Operations Management. In Research
Methods for Operations Management.
Kurtz, J., Sprik, S. and Bradley, T., 2019. Review of transportation hydrogen infrastructure performance and reliability. International Journal of Hydrogen Energy, 44(23), pp.12010-12023.
Kyriakarakos, G., Dounis, A., Rozakis, S., Arvanitis, K. and Papadakis, G., 2011.
Polygeneration microgrids: A viable solution in remote areas for supplying power, potable water and hydrogen as transportation fuel. Applied Energy, 88(12), pp.4517-4526.
Lacko R., Drobnic B., Sekavcnik M. and Mori M., 2014. Hydrogen energy system with renewables for isolated households: The optimal system design, numerical analysis and experimental evaluation. Energy and buildings, 80, pp.106–113.
Lemus, R. and Martínez Duart, J., 2010. Updated hydrogen production costs and parities for conventional and renewable technologies. International Journal of Hydrogen Energy, 35(9), pp.3929-3936.
Marbán, G. and Valdés-Solís, T., 2007. Towards the hydrogen economy?. International
37 Menanteau, P., Quéméré, M. M., Le Duigou, A., and Le Bastard, S. (2011). An
economic analysis of the production of hydrogen from wind-generated electricity for use in transport applications. Energy Policy, 39(5), pp.2957–2965.
Michalski, J., Bünger, U., Crotogino, F., Donadei, S., Schneider, G., Pregger, T., Cao, K. and Heide, D., 2017. Hydrogen generation by electrolysis and storage in salt caverns: Potentials, economics and systems aspects with regard to the German energy transition. International
Journal of Hydrogen Energy, 42(19), pp.13427-13443.
Ministry of Economic Affairs and Climate., 2019. Klimaatakkoord. [online] Available at:
https://www.klimaatakkoord.nl/binaries/klimaatakkoord/documenten/publicaties/2019/06/28/ klimaatakkoord/klimaatakkoord.pdf
Mulder, M., Perey, P. L., & Moraga , J. L., 2019. Outlook for a Dutch hydrogen market: economic conditions and scenarios. (CEER Policy Papers; No. 5). Groningen: Centre for Energy Economics Research, University of Groningen
Newman, N., 2020. Hydrogen Delivers New Future for Holland’s Pipeline Network. Pipeline
& gas journal, 247(9), pp. 26-29.
Nunes, P., Oliveira, F., Hamacher, S. and Almansoori, A., 2015. Design of a hydrogen supply chain with uncertainty. International Journal of Hydrogen Energy, 40(46), pp.16408-16418.
Ogden, J. and Nicholas, M., 2011. Analysis of a “cluster” strategy for introducing hydrogen vehicles in Southern California. Energy Policy, 39(4), pp.1923-1938.
Ophof, A., 2020. Drenthe Krijgt Zes Nieuwe Waterstof Tankstations. [online] RTV Drenthe. Available at: https://www.rtvdrenthe.nl/nieuws/156205/Drenthe-krijgt-zes-nieuwe-waterstof-tankstations
Robinson, S. 2004. Simulation : The Practice of Model Development and Use. Wiley.
Rooijers, F. 2017. Grid for the future. [online] Available at:
https://www.ce.nl/publicaties/2030/netvoor-de-toekomst
Samsatli, S., Staffell, I. and Samsatli, N., 2016. Optimal design and operation of integrated wind-hydrogen-electricity networks for decarbonising the domestic transport sector in Great Britain. International Journal of Hydrogen Energy, 41(1), pp.447-475.
Schaber, K., Steinke, F., Mühlich, P. and Hamacher, T., 2012. Parametric study of variable renewable energy integration in Europe: Advantages and costs of transmission grid
38 Shu, J., Teo, C. and Shen, Z., 2005. Stochastic Transportation-Inventory Network Design Problem. Operations Research, 53(1), pp.48-60.
Stiller, C., Bünger, U., Møller-Holst, S., Svensson, A., Espegren, K. and Nowak, M., 2010. Pathways to a hydrogen fuel infrastructure in Norway. International Journal of Hydrogen
Energy, 35(7), pp.2597-2601.
Tarkowski, R., 2019. Underground hydrogen storage: Characteristics and prospects. Renewable and Sustainable Energy Reviews, 105, pp.86-94.
Tascikaraoglu, A., Boynuegri, A. and Uzunoglu, M., 2014. A demand side management strategy based on forecasting of residential renewable sources: A smart home system in Turkey. Energy and Buildings, 80, pp.309-320.
Tomasgard A., Rømo F., Fodstad M. and Midthun K. 2007. Optimization Models for the Natural Gas Value Chain. In: Hasle G., Lie KA., Quak E. (eds) Geometric Modelling, Numerical Simulation, and Optimization. Springer, Berlin, Heidelberg.
Ulleberg, Ø. and Mørner, S., 1997. TRNSYS simulation models for solar-hydrogen systems. Solar Energy, 59(4-6), pp.271-279.
van der Wiel, K., Bloomfield, H., Lee, R., Stoop, L., Blackport, R., Screen, J. and Selten, F., 2019. The influence of weather regimes on European renewable energy production and demand. Environmental Research Letters, 14(9), p.094010.
Wermenbol, K., 2020. Green hydrogen through electrolysis: Deployment of supplementary batteries to smoothen electricity input for electrolyzers. MSc Thesis. University of Groningen.
Yilanci, A., Dincer, I. and Ozturk, H., 2009. A review on solar-hydrogen/fuel cell hybrid energy systems for stationary applications. Progress in Energy and Combustion Science, 35(3), pp.231-244.
39
Appendices
Appendix A, simulation description
Both cases are simulated in hourly steps for exactly one year, which results in 8760 hours in total. All input is hourly, as well as the flows and output. To create a complete understanding, the simulations for both cases are explained in detail in this section. It should be noted that everything depicted here, from movement of hydrogen to buffer behavior, is occurring in one single hour. A commonality for both cases is the supply side. Hourly supply is given by the percentage retrieved from the supply profile. Each hour represents a certain percentage of supply according to the supply profile. With the percentage coupled to the total supply, hourly supply is simulated.
41
Appendix B, choice of experimental variables
For both long and short-term case, there are three experimental variables and in total there are four due to one being not similar. These experimental variables have been either chosen based on theory, or in-field experience from experts on the topic from Groningen Seaports. In the table below, the choice of experimental variables is further explained.
Variable Based on Case Reasoning
Total supply Theory Both
Resulting from the ongoing energy transition we are currently faced with, the total share of renewable energy will increase over the upcoming years. This causes the total hydrogen supply to increase as well (Rooijers, 2017).
Buffer size Theory Both
Sufficient storage facilities are needed in future scenarios to handle hydrogen flows occurring from seasonal fluctuations (Jepma et al., 2019).
Demand ratios Practice Short
Due to zooming in on the demand of mobility and households in the short-term case and the uncertainty of future scenarios, the demand ratio is chosen as experimental variable.
Supply-demand
ratio Practice Long
42
Appendix C, choice of scenario values
For all scenarios, certain values represent the variation of experimental variables. All of these values are chosen either on recommendations by Groningen Seaports, or on certain performance indicators of the simulation model. Due to the fact that the values for both cases are different, they will be treated separately. In the tables below, the choice of values is further explained.
Variable Value Reasoning
Total supply 14 000 000kg The expected hydrogen supply in 2021.
20 000 000kg A possible hydrogen supply in the next 3 years.
Demand ratios
60%, 15%, 5% A recommendation by an in-field expert.
50%, 12.5%, 12.5% A recommendation by an in-field expert.
Buffer size
350kg The minimal buffer size required.
1050kg The first multiplication of 350kg that is able to supply over 99% of household and mobility demand.
43
Variable Value Reasoning
Total supply 28 000 000kg A possible hydrogen in supply in the next 10 years.
40 000 000kg A possible hydrogen supply in the next 10 years .
Supply-demand ratio
100% Model balance.
150% A recommendation by an in-field expert.
Buffer size
1050kg The large buffer size from the short-term case. 3500kg The largest multiplication of 350kg that would be
feasible at a single hydrogen fuel station according to in-field experts. Larger buffer sizes are technically possible, but this defeats the purpose of a network of small-scale storage facilities.
44
Appendix D, base settings
Scenarios Total surplus
(H2export) Total shortage (H2import) Maximum hourly surplus Maximum hourly shortage
9-10 202.5 ton 202.5 ton 2.2 ton 4.7 ton
11-12 613.5 ton 176.0 ton 4.4 ton 4.7 ton
13-14 289.3 ton 289.3 ton 3.2 ton 6.7 ton
15-16 876.5 ton 251.5 ton 6.2 ton 6.7 ton
Appendix E, absolute values performance indicators
Buffer behavior in hours Buffer effect in tons of hydrogen
Scenario Empty Full Mob.
excl.*
H2excess H2short H2export H2import
9 1347 2840 884 94.5 94.5 108.0 108.0 10 764 1607 507 141.7 140.5 60.8 62.0 11 882 5241 470 105.6 105.6 507.9 70.4 12 218 4698 129 159.2 158.0 454.3 18.0 13 1532 3250 1008 114.1 114.3 175.2 175.0 14 936 1961 602 183.0 182.0 106.3 107.3 15 1098 5430 591 125.2 125.4 751.3 126.1 16 393 4857 207 206.5 205.5 670.0 46.0 *Mobility exclusion
Appendix F, sensitivity analysis
Buffer behavior in hours Buffer effect in tons of hydrogen
Scenario Empty Full Mob.
excl.*
H2excess H2short H2export H2import
9 1346 3412 1388 94.7 94.9 108.2 108.0
10 762 1992 786 142.2 141.2 60.7 61.7
*Mobility exclusion
