How will the storage cost behave when the green energy supply share
increases in a green energy supply chain?
MSc thesis Supply Chain Management
Faculty of Economics and Business, University of Groningen
Supervisors: prof. dr. ir. J.C. Wortmann & phd. J.E. Fokkema
Acknowledgements
1
Abstract
Green energy brings a lot of fluctuations to the green energy supply chain because of its special characteristics, intermittence and volatility. Storage could ensure a more stable green energy supply as a buffer against many of these fluctuations. The extra expenditure, storage cost, is unavoidable if aiming to supply green energy and especially to supply a big percentage of green energy. This research provides insights regarding the relationship between the storage cost and the green energy supply share with which the current researches rarely involve. This paper also gives the influence of the storage with original inventory and the starting moments to store green energy on the storage cost and the greenness share of the whole supply chain. Finally, this research gives some advice for power companies and customers on decisions of greenness share from a perspective of storage cost.
2 Table of Content Abstract ... 1 Table of Content... 2 1. Introduction ... 3 2. Theoretical background... 5 2.1 Green Energy ... 5
2.2 Energy supply Chain ... 5
2.3 Energy Storage & Storage Cost ... 6
2.4 Green energy fluctuations ... 7
2.5 The percentage of green energy supply ... 8
3. Methodology ... 9
3.1 Research context ... 9
3.2 Model Description... 10
3.3 Data collection ... 12
3.4 Simulation and Experiments ... 12
4. Results ... 19
4.1 Experiment from 1st January ... 19
4.2 The experiment with original storage ... 21
4.3 Swift seasons to start storage ... 23
5. Discussions... 27
5.1 The situation when there is no storage facility ... 27
5.2 Relations among greenness share, capacity and the storage cost ... 27
5.3 The influence of original storage ... 29
5.4 The influence of starting seasons ... 30
6. Conclusions ... 31
Reference ... 33
3
1. Introduction
In order to perform green, an increasing number of customers are willing to pay for the green energy because of its low-carbon characteristics (Ozturk &Yuksel., 2016). This holds especially for IT companies who experience pressure from the general public and from NGOs (Bernath, M., 2009). Green energy here is defined as renewable energy (Huber et al., 2017; Oncel., 2017; Victor., 2014). Many power companies start to develop green energy business because the increasing green energy demand creates a potential market, despite current cost levels for green energy. However, it is still difficult to achieve a high percentage of green energy supply with relatively low cost and many providers will but cannot guarantee a100% green energy supply (Weckmann et al., 2017).
Infrastructures can ensure a stable and continuous energy supply in a green energy supply chain (Shafiei et al., 2017) by guaranteeing integrity and efficiency of all facilities. This means a better infrastructure can provide a larger greenness share. Thus, many solutions about infrastructures can help increase a high greenness share, such as a larger capacity transmission cable, a bigger storage facility and a better conversion. Among them, storage is probably a better one because the area of storage is more centralised than transmission cables and storage can more naturally and efficiently reduce fluctuations and balance the supply and demand (Anon., 2017). However, though energy customers and suppliers need storage to reach a high greenness share, it unavoidably increases the cost like storage investment cost. The reason is a larger greenness share needs a larger storage capacity (Shafiei et al., 2017) and the storage investment will grow as the storage capacity rise up (Scapino, et al., 2017). Therefore, in order to try to properly both keep a high greenness share and decrease the cost from an operational view, it is essential to know the relationship regarding these two issues that could influence the storage operations.
4
into this relationship. In order to achieve the goal, the research question below is developed:
How will the storage cost behave when the green energy supply share increases in a green energy supply chain?
Sub-questions also need to be answered.
How can the storage capacity setting influence the green energy supply?
Is it economically feasible for a power company to supply customers with 100% green energy from a cost perspective?
What is a proper share of green energy supply for a power company from a cost perspective?
5
2. Theoretical background
In this section, the theoretical background of this study is provided. Firstly, the definition of green energy is given. Then, the energy supply chain, the research background in this paper, in current literature is referred and discussed. Next, storage is introduced and related literature is reviewed. Moreover, some theory about the cost of storage is described, showing the need for more knowledge about the storage cost in the green energy supply chain. After that, the influence of the green energy supply and demand fluctuations is given. Finally, the greenness share is mentioned and the importance of investigating the greenness share in a green energy supply chain is discussed.
2.1 Green Energy
Green energy is unconsciously considered by many researchers as one concept opposite to traditional energy like fossils and similar to renewable energy (Huber et al., 2017; Cullen, 2017; Gaspar et al, 2017). Oncel (2017) states that the term ―green‖ covers the healthy integration of human with the nature for a flouris hing future and that green energy could be called an alternative for fossil fuels. Furthermore, Victor (2014) gave a more clear definition that green energy is a concept that refers to renewable and environmentally sound energy sources, but most importantly is how to effectively enforce supply and use of green energy. In this paper, green energy is defined as renewable energy and wind and solar powers are be involved.
2.2 Energy supply Chain
Storage is rarely considered and emphasised as an independent concept when researchers study the energy supply chain (Shafieiet al., 2017; Bas Wijnberg., 2016). Commonly, researchers always describe general concepts when they investigate the energy supply chain. For example, Shafiei and other researchers (2017) say that infrastructures guarantee a stable energy supply and demand in Figure 2.1. A more stable green energy supply and demand could contribute to a relatively larger percentage of green energy supply (Zhang et al., 2011). Therefore, a better infrastructure can result in a bigger greenness share. However, the infrastructure is the foundation or framework that supports a system or organization (Gibbs, Johnson & McCarthy., 2015). It includes so many facilities such as transmission systems, storage and productions sites. Which part of infrastructures and how they can guarantee a good energy supply are not introduced in Shafiei‘s article. A detailed component of infrastructure, however, will give power companies and customers a better understanding of which part of infrastructure they should care more about.
6
to the sizes of green energy supply and demand in the green energy supply chain. Except for the influence of storage on the common energy supply chain, the special influence of storage on the green energy supply chain especially is worth investigating (Donker et al., 2015; Zhou et al., 2015). This paper considers the special influence of storage on the green energy supply chain as the change of greenness share that storage causes because greenness share is a popular parameter that many power companies and customers want to know in a green energy supply chain (Cédric Philibert., 2015). Existing literature falls short in researching from a perspective of green energy supply percentage. Therefore, it is meaningful to do this research in this paper.
Energy
Supply Infrastructures
Energy demand
Energy Availability Availabile supply
Figure 2.1. Part of energy supply roles. (Shafiei et al., 2017)
2.3 Energy Storage & Storage Cost
Since the storage has been emphasised many times in this research, t he paper reviews existing literature and gives an overview of the role of storage in the energy supply chain. Generally, storage plays a role as an important part of infrastructures between suppliers and customers in supply chain (Styles, G., 2015) guaranteeing the energy supply can demand. Specifically, energy storage can reduce the frequency and cumulative duration of interruptions (Zhou et al., 2015), buffer the supply and demand fluctuations (Anon., 2017& Zhou, Mancarella & Mutale., 2015), and increase the flexibility of the energy supply chain (Bas Wijnberg., 2016). This paper will still support these opinions. However, the study in this paper will provide a new mathematical concept, greenness share, to reflex advantages that the storage could bring to the green energy supply and demand. From a quantitative perspective, this paper will prove that the storage could increase the efficiency of the green energy supply. This research perspective is different from existing literature but important because the percentage of green energy supply can well reflex whether more green energy is delivered to customers or not. In the existing theory, such new perspective can objectively show the role of the storage and make it more convincing that storage make the energy chain more green.
7
profit. However, in their article, the profit is the only standard to get the optimal size of capacity. This consideration is unilateral and insufficient. Although companies care a lot about the profit, the storage cost and the goal of the energy level they want to achieve are also important and should be considered (Crampeset al., 2010).. For example, the cost commonly can influence whether a project can start due to the budget. This paper probably does not provide an optimal size of storage capacity. However, it gives a proper capacity size with two different constraint conditions, the storage cost and greenness share. This gives the researches in the future more references.
Energy storage unavoidably brings the storage cost. About the storage cost, firstly, it has to been known that the storage cost commonly consists of the capital cost, which is fixed cost, and the cost of storing energy mainly because of loss of energy (RMI., 2017). The capital cost is the cost of storage capacity in this research and the cost of storing energy is zero and thus not considered because of the assumption of this simulation, which is explained next chapter. Furthermore, many articles model the storage cost as a sum of capital cost and variable cost (Giaouris et al., 2015; Cau, Cocco& Serra., 2012; Crampeset al., 2010).
Some literature researches the relationship between the storage cost and types of green energy. In Giaouris‘ research (2015), supplying different forms of green energy can cause different storage cost and wind power is possibly more cost-saving when there is a loss of energy storage and supply. Some literature (Cau, Cocco& Serra.,2012; Crampeset al., 2010) points the storage cost increases as the capacity size of storage rises up. This increasing trend is not linear because of other issues like avoidable loss of storage (Scapino et al., 2017). If buying the storage service from the third party that provides this service, the storage cost is linear to the capacity because the rule of payment from the storage service company relies on the capacity size bought from suppliers (Weckmann et al., 2017). The study in this paper is inspired by the researches above and extends the relationship above into a new rela tionship between the storage cost and greenness share with the storage capacity as a mediator. The relationship between the storage capacity cost and greenness share has never been researched before. However, it is a realistic problem to know a percentage of green energy that can be supplied by considering and balancing the cost like the storage cost. The gap of theory about this relationship can be coved by this paper.
2.4 Green energy fluctuations
8
supply and demand really rely on climates and temperatures. These two items change apparently in different seasons. However, existing literature above always focuses on the influence of fluctuations and rarely researches that the influence of such a context, different seasons. Some articles involve different seasons. For example, different green energy sources have different production and supply patterns in different seasons (Midilli et al., 2006). Solar power can be produced more in sunny days than in cloudy days and more in summer than in winter (Wang, Vilathgamuwa & Choi., 2008). Nevertheless, researching the impact on costs from a perspective of seasons is missing. This paper would build this relationship between seasons and costs and focus costs on the storage investment of green energy. As it is mentioned that the greenness share is an important parameter in green energy supply chain, linking seasons, storage cost and greenness share in this paper would contribute to a different theory to tell whether different seasons to store green energy has an impact on the storage cost and greenness share or not.
2.5 The percentage of green energy supply
9
3. Methodology
In this paper, an appropriate model on a green energy supply chain was developed. Since the data that can be collected comes from these two places, t his simulation model is based on a computer centre of University of Groningen as a customer and a green energy company DNV.GL as a consultant to coordinate the green energy supply. One advantage of this approach is that it can simulate the e nergy supply in a specific case. Because all data is from real institutions the whole process could be performed in a more real way. Another important strength of this approach is that cost and greenness shares can be tested quantitatively and efficiently because of a large number of data gained in the end of simulation. Generally, a graph that illustrates the curveon the storage cost and the green energy supply share should be received from the results in this simulation. The storage cost increases as the percent of green energy supply rises up.
3.1 Research context
The research is based on a project of the company DNV.GL. From the information of this project of this consulting company, it is known that the power companies traditionally supplied energy for their customers without opportunities for them to choose the energy sources. Now, the DNV company wants to provide customers with an energy platform that allows them to choose different energy source supply plans with different green energy share supplied by the power company and these plans have different impacts on the cost of the power company. This power company is in charge of supplying different energy to customers and the whole supply chain can be described in Figure 3.1.
Producers (wind, solar, fossils, biofuel
and others)
Market Place Suppliers
Customers (From 0% Greenness to 100%
Greenness) Data centers Others
Certified Measure
Storages
Fossils Green energy
Figure 3.1. All stakeholders in this energy supply chain
10
to customers based on the different percentage plans of green energy supply for customers. Finally, the customers, including houses, data centres and others, have selected different energy supply share policies and they receive their energy from suppliers.
In this research, the data centre in the University of Groningen is considered as a customer of green energy. The fluctuations of demand in the data centre exist because that the amount of computer consumptions is different in different time period. More specifically, from the beginning of office hours, the demand of energy starts to grow up. When more people come and use computers in the afternoon, the demand increases rapidly. It goes more stably during the night. However, the demand is smaller and more stable in the weekend and holidays because there are not many students or staffs come to use computers. Commonly, the data centre buys energy according to the demand. Therefore, it is a respond-to-demand procurement. Considering the prices of green and non- green energy and the length of period that data centre buy energy one time, demand will fluctuate again. Controlling the demand can be conducted by the data centre but it often happens at night. Moreover, the data centre is unwilling to control the demand.
This paper will focus on the final part of the supply chain shown in the red square in figure 3.1, which means from energy suppliers to customers with different green energy share requirements. The form of energy is centred on the electricity. High percentages of green energy supply (from 80% to 100%) will be mainly analyzed because customers prefer a larger greenness share to perform green. However, supplying large percentages of green energy is difficult to achieve now because green energy is intermittent and green energy supply has more fluctuations, needs more technical support and creates more costs. This makes it more interesting to investigate how cost can behave within a large percentage range.
3.2 Model Description
11
supply is overproduced too much while demand roles could take in the energy from public grids when the green energy supply is insufficient.
Demand Demand Demand Supply Supply Supply Storage
Connect to the grid Connect to the grid
Figure 3.2.Model of energy supply
About the energy supply in this model, only three energy sources are considered, solar power, wind power and fossils. The supply patterns of these two green energy sources are quite different. Solar is generally more produced in the midday and afternoon while wind is more stable in the period of whole day. Moreover, the green energy supply of solar and wind really depends on the weather if it is sunny and windy and the amount of green energy that is produced varies very largely every hour (Wang et al., 2008 & Meschede et al., 2017). These uncertainties above would result in the variability of green energy supply. If the fluctuations of green energy supply cannot be well controlled, it would be difficult to correctly and sufficiently meet the green requirements of customers every hour.
There are also many fluctuation factors in the demand side. The fluctuations are related to the customers‘ utility, special time, the purchasing strategy, and so on. Fluctuations are caused by the fact that every customer uses the energy at different time every day and the amount of energy consumed is different as well. Commonly, the purchasing strategy of customers including ‗respond to demand‘ and ‗respond to supply‘ is another fluctuation. Here, a respond to demand is reflected because the data centre from which data comes does buy the energy according to the ‗respond to demand‘ strategy. However, the data centre buys energy once for three years and the period that is analyzed is only one year. Therefore, the purchasing strategy is not considered as a fluctuation.
12
storage plays an important role in buffering the supply and demand fluctuations and increasing the flexibility of the energy supply chain( Anon. 2017).Moreover, every storage facility has a capacity that limits the size of storage.
The model aims to investigate the relationship between the storage cost and green energy shares, which are both related to these three concepts that have been introduced above. This kind of relationship is quite interesting and meaningful for a power company and customers because it can show how the storage cost will increase or decrease when a larger green energy is supplied. Meanwhile, supplying a largerandlarger percentage of green energy is now a new trend of market. The results of this model in this paper would help the power companies that provide customers with green energy to make decisions on how large percentage of green energy supply would be best choice for these companies when they refer to the storage cost as a standard.
3.3 Data collection
As mentioned, the data about green energy supply was received from the DNV.GL company and it includes the solar and wind power productions. Meanwhile, the data about green energy demand is collected from the data centre of University of Groningen including the energy consumptions.
Generally, the data that is collected from both sides is quite useful because the data can reflect the situation of the green energy market and the amount of data is big enough. More specifically, the energy form here is electricity. Moreover, these two sets of data are specified into a number per quarter of hour and it covers one-year (2016) data starting from 1st January 2016 to 31st December 2016. There are 25136 rows of data in total. The unit of green energy from the supplier part is MHW and the unit of energy demand from data centre is KWH. Moreover, according to the data received, the amount of energy supply is far larger than the amount of energy demand. Such a big difference of the unit and amount becomes a burden to this simulation and the demand and supply would be changed to a same level in order to make a further analysis by using a factor resulting from the merchant of the average of supply and the average of demand.
3.4 Simulation and Experiments
13
Table 3.1 all symbols in the simulation
Symbol Unit Meaning
i None the ordinal number of time unit of 15 min; i>0; integer Si kwh the amount of green energy supply at time unit i di kwh the amount of green energy demand at time unit i
Di kwh the amount of adjusted green energy demand at time unit i
S kwh the average of all Si d kwh the average of all di
Ii kwh the inventory of the green energy at time unit i C euro the cost of storage capacity
P None the greenness share
Grid kwh the total amount of energy used from public grids NSGE kwh the total amount of green energy that cannot be stored
TD kwh the total adjusted demand
Cap kwh the capacity size of storage facility
The simulation was conducted in excel. Generally, there are four parts of concepts, which are supply, demand, inventory and cost. However, in excel, more than four columns are created because it helps to analyze the process of the calculation more clearly. The real settings in excel can be shown below in figure 3.4.1 in Appendix. As it can be seen, there are 9 columns in total and time and capacity just show the basic values themselves. There is a column for green energy supply. Here, the supply is 1/1000 of the whole green energy supply in Netherlands because the data for the whole nation is too big, which may make results of experiments so large that they seem unbelievable. Because it is not just one power company that supply all the green energy in Netherlands, such downscaling can divide the whole market to different power companies in Netherlands. Downscaling into 1/1000 of the whole national green energy supply can make the original number level of the national green energy supply decrease from the unit of GWH to the unit of MWH. This adjusted unit is at a normal level of green energy that a common power company can operate. Moreover, this research stand from a power company who share 1/1000 of green energy market in Netherlands. This adjustment can also make the size of results of the storage capacity investment and the capacity size become 1/1000 of the total national investment and capacity size. However, the greenness share doesn't change because a power company who share a fixed percentage of the market share not only the fixed percentage of green energy supply but also the same percentage of green energy demand. This helps the results of experiments more practical for a power company in reality.
14
of magnitude of original supply and demand makes the analyze meaningless. However, the specific process on how to change the original demand into an estimated one will be explained later and the average column is used not only to show the averages of demand and supply data but to calculate the estimated demand. There is also a column for the inventory concept and the difference of supply and estimated demand is used to easily calculate the inventory. Capacity cost is shown to show the cost based on the supply-demand situation and capacity size settings. Moreover, a greenness share column is created to show the result of greenness of t his green energy supply chain this year. In addition, supply is represented by the data of green energy that is produced by the power company and demand is represented by the data of green energy that is consumed by the data centre. The data is a per-quarter dataset and the difference between the supply and demand per time unit is the amount of energy that could be stored. The time unit is 15 minutes and the unit of energy either demand and supply or other related items here is KWH.
In order to better explain how to conduct this simulation, a conceptual model is created, which is shown below in figure 3.3. Commonly, every simulation includes eleven steps. Firstly, the supply data should be input into excel as green energy supply. Since the energy unit of the original supply data is MWH, the energy unit is changed into the same unit of the demand from the data centre: KWH. Secondly, original demand data is taken in. However, there is a big difference between amounts of original green energy supply and demand. In other words, the data of demand from data centre is far smaller than the data of supply, which means di ≪ Si.It has to
become the same order of magnitude of supply and demand. The same order of magnitude makes it is meaningful to investigate the function of the storage for supply and demand side Otherwise, the current situation where supply can always meet such small demand does not need the storage. Therefore, an adjusted demand has to been created. A new variable Di is set, which represents the adjusted demand. Specific calculation processes have been shown below. The amount of green energy supplied by the supplier per quarter of hour is called as Si and the amount of green energy demanded by the data centre is called as di.Two factors, the average of supply data : S and the average of demand data:d , are used to calculate the estimated demandDi and make the supply and demand into the same order of magnitude.
Di = S
d∗ di;
15
Ii = Ii−1 + Si − Di;
After that, a capacity size should be set. Here, there is a rule to set capacity, which would be introduced in next section.
Then, the storage cost has to be computed. Here, the storage cost means the sum of the cost of storage capacity and the cost of storing energy. The cost of storing energy results from the loss of energy for each storage cycle. But the cost of storing energy in this research is zero because of the assumption, which is shown below, that there will not be a loss of energy in storage when energy is stored. Therefore, the storage cost in this paper equals the storage capacity cost. The parameter of unit cost is 40 euro/ kwh. Each kwh unit of green energy storage space will cost 40 euro. This parameter is referred from the Chen‘s research about a review of the electrical energy storage systems (2009). We selected the storage unit cost of one of the most popular energy storage systems called Metal- Air batteries that are suitable for long-term storage and cost lower. Within the range of the capacity cost of the Meta l-Air in that paper, we randomly selected a number (40 euro/kwh) as the unit cost of the storage capacity because the possibility of every cost price is equal and it is reasonable to use any number within this range in the research. Related assumptions are shown in last two paragraphs of this section. As it can be seen in the formula, the cost of storage capacity only relies on the capacity of the energy storage facility in this year.
Ci = 40 ∗ Cap(Ii);
Next, the amount of energy from the public grids is computed. First, only when the difference between demand and supply at the specific time unit is bigger than the current inventory level in this time unit can the energy from grid be delivered to customers in order to cover the loss of green energy. This means that this part of energy is non-green. We accumulate the amount of green energy each simulation and get the total amount of energy from grids that is consumed this year.
Finally, the greenness percent of this supply chain should be calculated. The greenness share in this research is set from a demand side because green energy demand is a key driver of the green energy business and the goal of power companies is try to meet the energy demand of customers. The non-greenness percent here equals the quotient between the amount of energy that is received from public grids and the total number of green energy supplied. The green energy share is represented by Pi. The specific formula is shown below.
16 3. Demand From Data centre (di) 4. Estimated Demand (Di) 5. Set Capacity Size (Cap) 7.Calculate the Capacity cost(C) 9. Calculate the Greenness (Pi) 6. Calculate Inventory(Ii) 10. Set new Capacity (Cap) 8. Calculate the energy from Grids(Grid) 2. Input Supply (Si) 1. Start Capacity >0 Yes 11. End No
Figure 3.3 The conceptual model of simulation
All assumptions in this simulation are given in next four paragraphs. Firstly, when the demand cannot be met because of the loss of green energy supplied, fossils are always available and sufficient to meet the part of demand that cannot be matched by the current green energy supply and inventory level. In this research, green energy is a major energy to support the whole energy supply chain and the traditional energy is rarely input into the supply chain except for the situation where demand is not met by the produced and stored green energy. However, it is well known that nowadays the amount of fossils that is produced is bigger than the amount of green energy. Therefore, fossils are always available when it is needed.
Secondly, the capacity of transmission is big enough and the transmission and storage facility does not lose any energy. In this research, we don't investigate the influence of the transmissions on the green energy supply chain. Supposing a big enough transmission setting can help to weaken the influence of the transmission and better focus on the influence of the storage on the whole green energy supply chain.
Thirdly, energy suppliers only buy the storage service. Commonly, there are two forms of energy storage business. The first one is that suppliers have their own storage facilities and they run their storage by themselves. The other is that suppliers rent a storage service. Considering two types of storage business would increase the difficulties of this experiment. Another energy storage business form, owning storage facilities by suppliers, could be involved in the further research. This research will focus on the second form and payments rely on the capacity of storage (Chen et al., 2009).
17
met in this green energy supply will be replaced with traditional energy from the public grids. These two ways are both apparently non green.
3.4.1 Performing Experime nts
Six experiments will be conducted. Generally, every experiment will continuously change the capacity of storage. In the first experiment, the experiment will be considered as a standard sample that is used to make a comparison with the other experiments. In the second experiment, the original storage is added to check the change. The rest of experiments will change the season to start the storage and every experiment will only change the uncertainty one time.
In the first experiment, the date to start storage was the first day of the year 2016: 1st January. The first simulation follows the steps of figure 3.4. Firstly, start the first four steps. The supply and demand has been input and the difference between supply and demand for each time unit was calculated. Secondly, start the fifth step to set the capacity which at first was set as a big enough capacity. Therefore, no matter how large amount of green energy was overproduced, it can be stored. Moreover, the second capacity is set according to the largest inventory in the first experiment when the capacity is infinite, which here we define big enough because the calculation of storage cost depends on a real number of the capacity size. An integer that is close to the largest inventory is set as the second capacity. Next, we decrease the capacity continuously until there is no capacity any more, which means the capacity is zero. Continue step 6 and 7. Inventory was accumulated and the cost of storage capacity was calculated. After that, the amount of energy from grids was computed and then greenness share was calculated. After step 8 and 9, according to the maximum of inventory of the first simulation, a capacity would be set for a new simulation. For example, the largest inventory of the first simulation was 1417.8 mwh and we put 1400 mwh as the second capacity. Then the second simulation started following the same steps just like what is shown in figure 3.4. There are 28 simulations in this experiment and from the first simulation to the last simulation. In each simulation, only the storage capacity was changed purposely, decreasing from infinite, 1400mwh, to zero mwh and the specific capacity setting situation is given in Table 3.3. After that, this experiment was finished. In these simulations of this paper should it be noticed again that we define that the word ‗infinite‘ means large enough instead of numberless. A real numberless capacity is meaningless in the simulation because the calculation of storage capacity cost relies on a real number of capacity size.
Table 3.3 the range of capacity settings.
18
The second experiment changed the original storage level and it still started the storage on 1st January. Here, we did not do all simulations with all different capac ities mention above and we just selected several special capacities to investigate the influence of the original storage on the cost and greenness share. These special capacities are the ones that can cause different percents of green energy supply from more than 70%, more than 80% to more than 90%. After selecting these capacities (infinite,106,105,104, 0), different original storage amounts were set for every
capacity. For the infinite capacity, because the maximum of inventory is around 1.42*106 kwh and we get the experience from the data from the first experiment that
the inventory in the first increase will reach to around half of the capacity and we started the original storage with half size of maximum inventory 8*105,which can make the inventory level become full in first increase, then 7*105, and set 105, later smaller number level 104 and 103respectively.For other capacities, the original
storage levels were also set based on the maximum of inventory. And the specific setting is in Table 3.4 below. And the simulation circle was still following the figure3.4.
The other four experiments changed the season to start the green energy storage. The months of spring are from March to May, the months of summer are from June to August, the months of autumn are from September to November and the months of winter are from December and February. And the simulation circle was still following the Table 3.4.
Table3.4 Original settings for storage facilities with different capacity
Types of Original storage settings
O ri gi na l S tor age Large enough Capacity
10e6 capacity 10e5 capacity 10e4 capacity No capacity
0 kwh 0 kwh 0 kwh 0 kwh 0 kwh
1*10e3 kwh 1*10e3 kwh 1*10e3 kwh 1*10e3 kwh 1*10e4 kwh 1*10e4 kwh 1*10e4 kwh 8*10e3 kwh 1*10e5 kwh 1*10e5 kwh 8*10e4 kwh
19
4. Results
The model results provide the insights into the relationship between the cost of storage capacity and greenness shares, the influence of original s torage level on the cost and the short-term difference of cost when starting the storage in different seasons. Depending on the capacity size, different green energy supply shares and the storage capacity costs are determined. Moreover, it should be mentioned that all experiments are based on the dataset of green energy production in the whole Netherlands. Therefore, all results are so large that it seems not convincing. In order to make the results look more practical and closer to the reality, all results in this chapter will be shown from a perspective of green energy supplier who shares 1/1000 of the green energy market in Netherlands. The reasons why doing this have been discussed in section 3.4.
4.1 Experime nt from 1st January
The first experiment starts on the 1st January of 2016. In all experiments, the total adjusted demand and supply are of the same size. There is no original storage. In the beginning, there is an amount of green energy is overproduced and the difference is 275.6 kwh. This makes the storage facility have an inventory in the beginning. In this experiment, when the capacity of storage facility is infinite, the maximum of inventory this year is 1417.8 mwh, which means that storage investment in this situation is around 56.7 million euro. The total amount of energy that has to be given from the public grids is 801.6 mwh. Therefore, according to the formula of greenness share, greenness share is about 97.6%. When there is a specific capacity for the storage facility, during the period of the year 2016, the storage facility with each capacity could always become full. In other words, the maximum of the inventory is the capacity. However, the amount of green energy that cannot be stored definitely increases. After this experiment, a graph of greenness share and the storage cost is created. It can be seen in the figure 4.1 below.
20 0 10 20 30 40 50 60 0.68 0.7 0.720.740.760.78 0.8 0.820.840.860.88 0.9 0.920.940.960.98 1 M ill ion Eur o Percentage
The relationship between greenness share and storage cost
Jan 1st
high level like 92%, if rising up the greenness share only by increasing the storage capacity, a very large capacity is required to increase such a small increase of greenness share. A large capacity means a large storage capacity cost.
Figure 4.1 the results of experiment one (1st Jan)
Another important thing to know is that a too small capacity of storage cannot almost influence the greenness share. Only when the capacity size of green energy storage is becoming relatively close to the order of magnitude of the amount of overproduced green energy, the influence might become obvious. As it is shown in the Table 4.1 below, before the storage capacity size of 1000 kwh, the greenness share is not almost different from the greenness share without storage. After the capacity size becomes bigger, like 5000kwh, the greenness share starts to increase largely. This is because when the capacity size of storage is too small, the amount of green energy that is overproduced is far bigger than the capacity and the capacity is full so earlier that green energy overproduced later cannot be stored. This makes the storage lose its function to keep storing the green energy for a long time. It also gives an advice for a power company that when this company wants to store energy, it is necessary to set a relatively big enough capacity of storage. Otherwise it does not well increase the greenness share and does increase a new cost of this company.
Table 4.1. the relationship between greenness share and small capacity sizes
21
This trend in Figure 4.1 below is a common trend, which can be proven by the trends in the figure 4.3.4 in Appendix. Different capacity sizes not only influence the cost of storage capacity investment but also have an impact on the amount of energy that is needed from grids to meet the energy demand. The change of grid energy amount further influences the greenness shares of the whole energy supply chain. Furthermore, from the whole trend of this curve, it is easy to see that when the capacity size is relatively small, every time the capacity of storage increases will results in a big decrease of the total amount of energy needed from grids, which causes a big growth of the greenness share. For instance, from the storage capacity of 103 to 105 kwh, the
amount of grid energy becomes almost half and the share increase from 73% to 86%. However, this type of influence becomes weaker as the capacity increases.
4.2 The experiment with original storage
The second experiment starts on the 1st January of 2016. But it starts with an original storage. Here, the original storage means that before suppliers will officially start to provide storage service for customers, suppliers firstly prepare some amount of green energy in the storage facility to better meet the demand in the future. In this experiment, the storage facility with four different capacities, which are respectively infinite (large enough), 106, 105 and104 kwh is set to start with original storage of
green energy. The reason why these capacities are selected to do the simulations in this experiment has been explained at section 3.4. All the important results can be shown in figure 4.2 below.
Firstly, small original inventory cannot make any difference on the storage capacity cost. It should be known that, for the influence of original inventory o n the storage capacity cost, this paper only could investigate the situation when the capacity is infinite because the assumed storage capacity in this situation can be changed and other finite capacities are fixed. This is set and explained in Chapter 3. From the first line graph, it can be seen that before the original storage is smaller than 104 kwh, the
22 0.96 0.97 0.98 0.99 1 1.01 55 65 75 85 P e rc e nt ag e M ill ion
Original storage level Kwh
1. Greenness share & storgage cost (Max Capacity with Original storage )
new storage cost Greenness share 0.96 0.962 0.964 0.966 0.968 0.97 30 40 50 60 P e rc e nt ag e M ill ion
Original storage level Kwh
2. Greenness share & storgage cost (10e6 Capacity with Original storage )
new storage cost Greenness share 0.861 0.8615 0.862 0.8625 0.863 3 4 5 6 0 1000 10000 80000 P e rc e nt ag e M ill ion
Original storage level Kwh
3. Greenness share & storgage cost (10e5 Capacity with Original storage )
new storage cost Greenness share 0.768 0.7685 0.769 0.3 0.4 0.5 0.6 0 1000 8000 P e rc e nt ag e M ill ion
Original storage level Kwh
4. Greenness share & storgage cost (10e4 Capacity with Original storage )
new storage cost
Greenness share
Figure 4.2 Results of Experiment 2
A proper original storage can truly help increase the greenness share of the green energy supply chain. Small original inventory cannot also have an apparent influence on the greenness share. As it can be seen in first line graph, before the original inventory is 104 kwh, the percentage of green energy supply almost does not change. When it is 105 kwh, the greenness share increases 0.3% with about 2.4 million euro of storage investment. After increasing the original inventory 8*105 kwh, the greenness share can achieve even 100%. Here, the concept of small is meaningful while comparing the number level of overproduced green energy. The reason why small influence of the small original storage is the same with the one mentioned in last paragraph.
An increasing trend of greenness share happens in other three experiments. For example, in second line graph, the greenness shares starts to increases and the increasing rate is rising until the original inventory is 105 kwh. After that, the
23
availability of original storage makes the influence of the storage on the greenness share last longer. This means a larger green energy demand can be met.
A larger storage capacity contributes to a bigger influence the original inventory on the greenness share. As long as the difference of capacity sizes is big enough, no matter how much the original inventory increases, the greenness share when the capacity size is small cannot surpass the greenness share when the capacity is big. When the capacity is infinite, the greenness share increases around 2.4 percentage points from 97.6% to 100%. When the capacity is 106 kwh, the greenness share
increases 0.118 percentage points. Moreover, when the capacity is 105 kwh and 104
kwh, the greenness share increases about 0.11 and 0.02 percentage points respectively. This is because the capacity is a foundation and precondition for the original storage to influence the greenness share. When the capacity size is not big enough, the capacity limits the storage and a large original inventory cannot make a further bigger influence on the greenness share.
4.3 Swift seasons to start storage
These four experiments respectively start storage in winter, spring, summer and autumn. There is no original storage for them. This also means that the inventory in beginning of four experiments is all zero. Generally, it can be seen in figure 4.3.3 in appendix that the relationships between greenness shares and the storage capacity costs for each of five experiments with no original inventory are similar. They all show a kind of trend that the cost of capacity increase as the greenness share grows up and when the share is above 90%, increasing greenness share is far more expensive. The reasons have been given in section 4.1.
However, it should be noticed that these results below in this section is practical for the short-term period of green energy supply. If the period last a many years, it could be seen that the influence of the season to start the storage is becoming weak and almost disappears in the end because the distribution of data of green energy supply and demand in such a long period might be more similar and the difference of which season to store green energy behaves slightly.
The small capacity size of storage limits the influence of the starting season to store green energy on the greenness share. The results of four experiments with the capacity of less than 6*103 kwh, which means less than 2.4 million euro investments, are described in the figure 4.3 below. Generally, it is clear that the results of greenness shares of these four experiments are quite close. When the capacity is less than 3*104
kwh (1.2 million euro), the percentages of all experiments are of the same size. When the capacity is less than 6*104 kwh, the greenness shares in spring and autumn are always the same. The greenness shares in summer and winter orderly start to slightly fall behind with the capacity of 3*104 kwh and 6*104kwh respectively. The reason
24
due to these relatively small capacity sizes. Every time unit the overproduced green energy can make the storage facility full, which causes the storage cannot store energy in the next few time units and further makes the amount of energy needed from grids slightly decrease. Even though the supply pattern is different with each other, the difference that should have happened in the storage doesn't happen.
Figure 4.3 the comparisons among these four experiments (less than 6*10e7)
Winter is the most potential and advantageous season to start the green energy storage by balancing the greenness share and the storage capacity cost. In figure 4.4, further comparisons in this section have shown that winter performs its advantage after the greenness share is 90%. For example, within the range from 73% to 90%, the difference of the greenness share between in winter experiment and in summer experiment is always really small (less than 0.01%). After 90% of greenness share, this difference starts to become bigger and bigger, increasing from 0.44% when the capacity is 6* 105kwh to 4.5% when the capacity is infinite (big enough). This happens because starting storage in winter is the only experiment that can have a positive inventory in the beginning of the storage. The positive inventory in the beginning can positively influence the greenness share like an original inventory does. However, summer experiment in the beginning of the simulation experiences a long period when the inventory level is zero and some energy from grids have to be needed. This influence might be one of the important reasons why the winter experiment could perform better all the time and achieve a larger greenness share in the end than other experiments. 66.00% 68.00% 70.00% 72.00% 74.00% 76.00% 78.00% 80.00% 82.00% 84.00% 86.00% pe rc e na ta ge million euro
The comparisons among these four experiments
25
Furthermore, the greenness shares in spring, summer and autumn experiments keep the same in the last few big capacities. O nly the greenness share in winter still increases. When the capacity is infinite, the winter one is the most advantageous because the share can reach 98.4%. After that, starting from autumn can arrive at 96.3%. Then, the greenness share in spring is also low (94.7%) following with the share in summer experiment (93.9%). A relatively big capacity gives winter more space to fulfil its advantages. The other three experiments face the bottleneck to increase the greenness share when the capacity is relatively big. This means the capacity is not an important constraint for the greenness share in this time of period. Supply patterns can truly influence the greenness share. Spring experiment and summer experiment has similar supply patterns and the biggest difference between two greenness shares of spring and summer experiment is only 0.87%. The basic reason why winter performs better than the other three seasons is also that the supply pattern in winter experiment is more beneficial.
Figure 4.4 the results of four experiments
Another important thing that the results of these four experiments can bring is the relationship between the storage capacity cost and the greenness share. These four experiments help to multiply check the relationship in section 4.1. Generally, the cost of storage capacity increases as the greenness share increases. More specifically, investing 1.2 million euro can make the green energy share increase to around 80%
0.65 0.66 0.67 0.68 0.690.7 0.71 0.72 0.73 0.74 0.75 0.76 0.77 0.78 0.790.8 0.81 0.82 0.83 0.84 0.85 0.86 0.87 0.88 0.890.9 0.91 0.92 0.93 0.94 0.95 0.96 0.97 0.98 0.991 0 20000 40000 60000 80000 100000 P e rc e nt ag e Million Euro
The cost and Percentage of five different Experiments
26
27
5. Discussions
The model in this paper is designed to get knowledge about the relationship between the greenness share and the storage capacity cost in a green energy supply. This section will reflect the reasons to the results of experiments in last section and give discussions and insights based on the relevant knowledge. Also, shortcomings of the model and simulations will be depicted and possible additions and improvements to the model will be proposed.
5.1 The situation whe n there is no storage facility
As the results show, when the capacity is zero, which means that there is no storage facility, the greenness energy still can achieve around 73%. This is because in most time, a majority demand can be met by the supply. Loss of supplying green energy 27% will be totally offset by the energy from grids. This is a good thing for a power company because it is easy to achieve a high share of green energy supply and this company does not need to pay for the cost of storage capacity. However, this happens in a realistic environment that has be told by the descriptions of assumptions in chapter 3, such as big enough transmission system, no time delay and other issues. This number of green energy share depends on the formula of greenness share in this paper. Moreover, supplying a high greenness share has a prerequisite, which is to ensure a sufficient and continuous green energy supply in most time. Here, green energy storage plays a role in achieving a stable and sufficient energy supply. In fact, a green energy supply chain has too many uncertainties and variables that both cause the loss of green energy and supply mistakes or conflict. This has a same influence as the insufficient inventory does. It will result in a lower percentage of green energy supply.
5.2 Relations among greenness share, capacity and the storage cost
A result on the relationship between green energy supply share and the cost of storage capacity has come out. The trend shows that the cost rise up slowly when the green energy supply share increases from 73% to 85%. After 85% of greenness share, the storage cost increases largely and after 90%, the increase rate of adding one percent of greenness share is even bigger. Here, the capacity size plays an important role like a mediator. In other words, the greenness share influences the capacity size and then capacity size directly and indirectly influences the cost.
28
because it has been supposed that the cost of storage investment, when setting a capacity that is bigger than the maximum of overproduced green energy during the whole year, is the same with the cost when the capacity is infinite. Furthermore, the amount of energy from grids does not change, which means that the greenness share keep the same big.
Additionally, not only does the greenness share influence the capacity size but also the capacity influences the greenness share. We need to take the amount of energy from grids into consideration, which is a factor of formula to calculate the greenness share and thus decides the size of share. When the capacity is big enough, all overproduced green energy can be stored in order to meet the loss of supply in the future. This makes it unnecessary to use energy from grids and thus it can cause a high percent of green energy supply. The reason why the greenness share is not 100% when the capacity is big enough is that though storing all green energy that is produced, this amount of inventory cannot meet the whole demand in each time. However, when the capacity is small, an amount of green energy that is overproduced and should be stored cannot be stored and the failure of inventory to meet a larger demand is offset by supplying energy from grids. This causes the growth of the amount of energy from grids. According to the formula of greenness share, the total demand does not change the total amount of energy used from the public grids reduces. Therefore, the greenness share will decrease.
Considering the capacity size and energy from grids, we can analyze some interesting changes in the trend. Firstly, when the capacity is very small (less than 103 kwh) the
greenness share does not or slightly go up. This is because that this kind of capacity does not change a lot about the amount of energy from grids and the amount of green energy that is not stored compared with no storage. Therefore, for a power company, if the total capacity of storage facilities is too small compared with the maximum of green energy that needs to be stored, this kind of investment for storage facilities is a kind of waste because it creates extra storage cost but does not make a contribution to increasing the green energy supply share. Moreover, in fact, there will be many other costs involved such as the building cost, the operation cost and the repairing cost (Timmons, Harris and Roach., 2014). Therefore, it is not smart for power companies to create a total small storage capacity.
29
energy supply want to continue to increase the greenness share, it is wise to increase the capacity to 2* 105kwh with only 4 million and even 3* 105 kwh with 8 million. As what has been said above, in order to increase greenness share, especially over 90%, the capacity size has to become bigger. If aiming to over 95% of green energy supply share, the company needs to set the total capacity infinite (big enough) or at least a bit bigger than the maximum of inventory now when the capacity is infinite. In conclusion, the kind of relations among these three concepts could help power companies with an overview about costs and shares. It could possibly provide a reference about the greenness share decision from a perspective of the cost of storage capacity.
5.3 The influence of original storage
Firstly, the original storage plays an important role in contributing to increasing the greenness share. But adding the inventory in the beginning possibly increases or does not change the green energy supply share and it depends on the amount of original inventory and the size of storage capacity. As there is a same trend of greenness share in the experiment that increasing original inventory will increase the share, it can be concluded that the original inventory can always achieve the goal to increase the share as long as this original inventory is proper. The reason of this trend is that the amount of energy used from the public grids decreases continuously. However, how much inventory in beginning should be added in the facility tota lly is not sure yet. The amount of original inventory is not the situation that the bigger the better. Too big amount will result in too large storage capacity investment cost but the greenness share can go up lightly, like the first line graph in figure 4.2, or too big amount of original inventory does not increase the greenness share, like the second line graph in figure 4.2.
Secondly, a big capacity size is a foundation that the original inventory can largely increase the greenness share. As it can be seen in figure 4.2, when the capacity is big, original inventory can increase 2.5% of the greenness share while when the capacity is small, original inventory can only increase 0.025%.This proves that the capacity size can limit the development of greenness share though the original inventory can help increase the share.
30
5.4 The influence of starting seasons
Different seasons to start to store the green energy do make a difference in a short term. Winter and autumn are good seasons to start the storage if a power company wants to start to store green energy. First of all, the influence of starting storage seasons is investigated for a power company that want a big percent of green energy supply ( more than 90%) because when it is less than 90%, starting storage seasons does almost not make any big difference. During the period between 90% and 96%, starting the storage in spring is the best choice because with the same storage, it results in a larger greenness share than the other three seasons. However, after 96%, starting the storage in winter is the best choice because only starting in winter can achieve a share of more than 96% and the storage capacity cost is smaller.
Additionally, for a power company, if there is not original inventory in the storage facilities, starting in spring, summer and autumn should not set a capacity size that is bigger than 1.1*106 kwh, because these three seasons experiment meet their own bottleneck and even though setting a bigger capacity, the greenness share is the same. This only wastes the investment. The reason is that when the capacity is larger than 1.1*106 kwh, the amount of energy from grids does not change any more. This also means that the capacity size this time is not an influential issue any more because increasing capacity cannot change the amount of energy used from grids. In this time, only starting storage in winter is still potential to increase greenness share.
31
6. Conclusions
Green energy brings a lot of fluctuations to the green energy supply chain because of their special characteristics. Storage could ensure a more stable green energy supply as a buffer against many of these fluctuations. The extra expenditure, cost of storage capacity, is unavoidable if aiming to supply green energy and especially to supply a big percentage of green energy. However, the existing literature lacks of the knowledge about the relationship between the storage capacity cost and greenness share. Therefore, it is meaningful to research the relationship between these two items. Next few paragraphs will respectively and orderly answer one main question and four sub-questions, give limitations and show the further research direction.
This research provides insights regarding the relationship between the storage capacity cost and the green energy supply share, at which the current researches fall short. The results of this model give a detailed trend of cost following the greenness share, to which the power company might refer when it is making the decisions on green energy supply share. This gives a kind of standard from the view of storage investment. The results are truly based on some assumptions that simplify and idealize the situation. However, they still can show a reasonable trend that when aiming to increase a higher percentage of green energy supply, the total storage capacity cost is bigger.
The paper also illustrates that the storage capacity has an impact on the green energy supply from a perspective of greenness share. The capacity of green energy storage can plays a role like a mediator between the storage cost and greenness share. It influences both two items. The capacity determines the amount of green energy that can be stored. A big storage capacity gives the energy supply chain an opportunity to achieve a higher percentage of green energy supply because more green energy can be stored to cover the loss of green energy production later.
Additionally, it is found that the original inventory will help to build a bigger greenness share with the condition of a proper big storage capacity. But when the capacity is small, the original inventory could only slightly increase the share. Furthermore, this simulation achieves a 100% greenness share from a single perspective, the storage capacity cost, when the capacity is infinite with proper inventory. However, this is based on many assumptions set in this paper. Therefore, relying on this paper to completely prove the feasibility of 100% of green energy supply is insufficient.
32
Finally, this paper provide another important finding that if deciding a season to start to store green energy without any original inventory, when the greenness share goal is less than 96%, autumn is a better option because it can achieve a higher greenness share with the same storage cost. When the goal is more than 96%, winter is a better choice because the other three seasons meet the bottleneck of increasing the share and only winter can continue this growth. This type of advantage is apparent in the period of one year. However, if the period covers many years, such advantage is becoming weak.
33
Reference
Al-Gwaiz, M., Chao, X., & Romeijn, H. (2016). Capacity expansion and cost efficiency improvement in the warehouse problem. Naval Research Logistics (nrl), 63(5), 367-373. Anon. (2017). Grid Electricity Storage: Size Matters. Spectrum. 54(1), 23–23.
Bas Wijnberg. (2016). Capturing flexibility in an inflexible market: energy system flexibility from a service perspective. 1-40.
Bernath, M. (2009). Green IT – sustainable profit through energy efficiency : Views, comments and opinion. Engineerit, 38, 12-13.
Cau, G., Cocco, D., & Serra, F. (2012). Energy and cost analysis of small-size integrated coal gasification and syngas storage power plants. Energy Conversion and Management, 56, 121-129.
Cédric Philibert. (2015). Three Reasons Why Renewable Energy Is So Important To The Power Industry. http://www.gepowerconversion.com/inspire/three-reasons-why-renewable-energy-so-important-power-industry
Chen, H., Cong, T., Yang, W., Tan, C., Li, Y. & Ding, Y. (2009). Progress in electrical energy storage system: A critical review. Progress in Natural Science, 19(3),291-312.
Crampes, C., & Moreaux, M. (2010). Pumped storage and cost saving. Energy Economics, 32(2), 325-333.
Cullen R. (2017). Evaluating renewable energy policies. Australian Journal Of Agricultural And Resource Economics, 61(1), 1-18.
David Elliott. (2016). Balancing green energy. International Journal of Ambient Energy, 37(5), 437-438.
Deng, Z., Xu, Y., Gu, W., & Fei, Z. (2017). Finite-time convergence robust control of battery energy storage system to mitigate wind power fluctuations. International Journal of Electrical Power and Energy Systems,91, 144-154
Donker, J., Huygen, A., Westerga, R., Weterings, R., & Bracht, M. Van. (2015). Naar een toekomstbestendig energiesysteem: Flexibiliteit met waarde. Delft.
Fukaya, Y., & Goto, M. (2017). Sustainable and safe energy supply with seawater uranium fueled hTGR and its economy. Annals of Nuclear Energy, 99(8), 19-27.
Gadd, H., & Werner, S. (2013). Heat load patterns in district heating substations. Applied Energy, 108, 176-183
