FACULTY OF ENGINEERING AND SUSTAINABLE DEVELOPMENT
Department of Building Engineering, Energy Systems and Sustainability Science
Energy storage and their combination with wind power compared to new nuclear power
in Sweden
A review and cost analysis Simon Englund-Karlsson
2020
Preface
This master thesis was written as a part of the Energy System programme, at University of Gävle.
To my family and friends, an enormous thank you for the support you have given me during the time for my studies. It would have been a lot harder without your support.
Thank you, my colleagues, who have made it possible for me to take the time off that I needed to complete my studies.
I would like to give a special thanks to my supervisor Magnus Mattsson, University of Gävle, that helped me to improve the quality of my thesis with his suggestions.
Simon Englund-Karlsson 2020-05-25
ii
Abstract
As intermittent renewable energy sources such as wind and solar power gradually increase around the world, older technologies such as nuclear power is phased out in Sweden and many other countries. It is then important to ensure that the total power need is secured, and that the power grid can remain stable. One way of man- aging intermittent renewables is by using energy storage.
The main goal of this thesis was to compare energy storage methods and their costs.
A secondary aim was to investigate how the cost of developing more renewable en- ergy sources, in combination with different energy storage methods, compares to erecting new nuclear power. This thesis was limited to three energy storage tech- nologies, namely pumped hydro storage (PHS), compressed air energy storage (CAES), and four battery storage technologies. They were combined with wind power in the cost analysis. The comparison was done by performing a literature re- view and economical calculations, which focused especially on levelized cost of stor- age (LCOS).
The results from the economic calculations indicated that PHS and CAES had lower LCOS than battery storage technologies. Similar results could be seen in the litera- ture review as well. When comparing levelized cost of energy (LCOE) nuclear power had the lowest, €0.03-0.12 kWh-1, followed by wind power in combination with PHS and CAES, both around €0.07-0.24 kWh-1. This result was maintained also at sensitivity analysis regarding the discount rate, which both nuclear power and PHS proved rather sensitive to.
Keywords: energy storage, nuclear power, wind power, pumped
hydro storage, compressed air energy storage, battery energy storage, levelized cost of energy, Sweden
iv
Nomenclature
Latin and Greek Symbols Description Unit
Ct Capital expenditures €
Mt Maintenance and opera-
tions expenditures
€
Ft Fuel expenditures €
r Discount rate %
Et Electricity generated kWh
n Lifetime Year
Eout Electricity discharged kWh
Costelec Price of electricity € kWh-1
ηs Efficiency of storage %
DoD Depth of discharge %
vi Abbreviation and acronyms
Letters Description
kW Kilowatt
MW Megawatt
GW Gigawatt
kWh Kilowatt Hour
MWh Megawatt Hour
GWh Gigawatt Hour
PHS Pumped hydro storage
Li-Ion Lithium-Ion
LCOE Levelized Cost of Electricity
LCOS Levelized Cost of Storage
CAES Compressed Air Energy Storage
BWR Boiler Water Reactor
PWR Pressure Water Reactor
CHP Combined Heat and Power
VRB Vanadium Redox Battery
NaS Natrium-Sulphur
UN United Nations
Table of contents
Preface ... i
Abstract ... iii
Nomenclature ... v
1 Introduction ... 1
1.1 Background ... 1
1.2 Aims and limitations ... 3
2 Method ... 4
3 Results ... 6
3.1 Literature review ... 6
3.1.1 Energy storage ... 6
3.1.2 Nuclear power ... 13
3.1.3 Renewable energy ... 15
3.2 Economic comparison ... 17
3.2.1 Levelized cost of energy and storage ... 17
3.2.2 Energy storage economic comparison ... 17
3.2.3 Wind power and energy storage, compared to nuclear power ... 20
4 Discussion ... 22
5 Conclusion ... 25
5.1 Study results ... 25
5.2 Outlook ... 25
5.3 Perspectives ... 26
References ... 27
x
1 Introduction
1.1 Background
As the world shifts towards a more sustainable future within the energy field, some old technologies are being replaced by new. In 2015, the members of the United Nations (UN) adopted the 2030 agenda for sustainable development [1]. In the 2030 agenda UN states 17 goals for sustainable development, one of them being afforda- ble and clean energy [1]. As a result, countries around the world started acting, in an attempt, to reach the goals set by the UN. The political parties in Sweden
reached an agreement; that Swedish electrical production should strive to be derived from 100 % renewable sources by 2040 [2]. While not a demand, still a goal. Nu- clear power, being neither a renewable nor fossil energy source, could remain a large contributor in Swedish electrical production. Although it faces large invest- ments in the future, and the decision has been handed over to the owners of the nu- clear power plants [2]. Figure 1 shows the Swedish electrical production from 1970 to 2017, in percentages for the main sources. Nuclear power has been the second largest source for electricity since the seventies, some years producing more than hydro power. Since early 2000 wind power has been growing steadily, seen in Fig- ure 1.
Figure 1 Swedish electrical production from 1970-2017 [3]
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
Hydro power Wind power Nuclear power CHP (industry) CHP (district heating) Other thermal power
2
If Sweden were to replace nuclear power with renewable energy sources, instead of maintaining some nuclear reactors, energy storage technologies could be beneficial to mitigate the intermittent nature of wind and solar power in the electrical grid [4], [5]. A downside of both wind and solar power is that both are intermittent produc- ers, resulting in problems when trying to accurately predict how much electricity will be produced from the different plants. Figure 2 shows the Swedish electrical production for 2019, where nuclear power was used as baseload, while hydro power acted as load-following. As can be seen in the power duration diagram, inter- mittency of wind was high, but could be managed due to the large amounts of hydro power in the Swedish power system. As previously mentioned, to mitigate these ef- fects energy storage could be used, especially if old nuclear power would be re- placed by intermittent renewables. Another solution could be constructing new nu- clear power.
Figure 2 Power duration diagram of the Swedish electrical production during 2019 [6]
However, constructing new nuclear power is not an easy topic, from neither a fi- nancial, scientific, or political viewpoint. There is presently an ongoing debate, which path would be most suitable for Sweden in the future, even though the goal set for 2040 is an entirely renewable power system [2]. In the following section the aims and limitations of this thesis are presented.
0 5 000 10 000 15 000 20 000 25 000 30 000
Power (MW)
Timeperiod
Nuclear power Hydro power Wind power Solar power CHP Gas and disel Other
1.2 Aims and limitations
The first objective of this thesis was to compare the cost of different energy storage methods, by calculating their levelized cost of storage. The second goal was to per- form a cost comparison between renewable energy sources in combination with en- ergy storage, with new nuclear power; in both cases replacing the old nuclear reac- tors.
As there are many different methods to store energy, this thesis was limited to: (1) pumped hydro storage, due to it being a mature and well-established energy storage technique; (2) compressed air energy storage, as it has potential for large scale appli- cation; (3) different types of battery energy storage technologies, as they are cur- rently being either used for some applications, or are emerging on the market.
Other energy storage technologies, such as flywheels and fuel cells, to mention a couple, have either a limited area of use, or have a rather low overall system effi- ciency in combination with high capital costs. As wind power has been growing, in Sweden, far more compared to solar power in recent years, it was chosen for calcu- lations on renewable energy sources. Later generations of fission reactors, from the beginning of the second millennia, were used for the nuclear power calculations.
Next section contains the methods used in this thesis.
4 2 Method
For the literature review, the search engines, Google Scholar, DiVA, and ScienceDirect were used. To establish that the articles had been peer-reviewed, UlrichsWeb was used. ScienceDirect was primarily used to find articles on the stud- ied subject. Google Scholar was helpful to find some of the material. DiVA was used the least, as it was difficult to find relevant reports for the study. Search terms that was used to find relevant articles for the different topics contained phrases and words e.g.: energy storage, electricity storage, nuclear power, Sweden, levelized cost of energy, levelized cost of storage, economics, cost, renewable energy, wind power and solar power.
Other sources used were the books: Elkraftssystem 1 and Energy Systems Engineer- ing. As well as various reports from authorities e.g. the Swedish Energy Agency, Swedish Radiation Safety Authority, and International Renewable Energy Agency. A full reference list is found at the end of the thesis.
For the economic comparison, data from the literature review was aggregated and recalculated to present day Euro (2020). The selected data for PHS, CAES, nuclear power (Gen 3), and wind power, had smaller uncertainties than battery technolo- gies when it came to capital cost, as they were more utilized. This is reflected in the input data with smaller, percentual, variations in prices. To compare the costs for the energy storage and electricity, levelized cost of storage (LCOS) and levelized cost of energy (LCOE) was calculated for the different energy storage methods and power sources.
Levelized cost of energy, LCOE
To calculate LCOE, for wind and nuclear power technologies, equation 1 was used [7]. By utilizing the systems lifetime costs, as well as its expected electrical produc- tion during its lifetime, the levelized cost of electricity for the system to break even can be calculated. Included in the total system cost: capital cost, maintenance and operations cost, as well as eventual fuel costs. Electricity produced during its life- time can be calculated by using the capacity factor (fraction of full load hours during a year) and installed power (the nominal power that the power plant produces).
𝐿𝐶𝑂𝐸 = 𝑠𝑢𝑚 𝑜𝑓 𝑐𝑜𝑠𝑡𝑠 𝑜𝑣𝑒𝑟 𝑙𝑖𝑓𝑒𝑡𝑖𝑚𝑒
𝑠𝑢𝑚 𝑜𝑓 𝑒𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑎𝑙 𝑒𝑛𝑒𝑟𝑔𝑦 𝑝𝑟𝑜𝑑𝑢𝑐𝑒𝑑 𝑜𝑣𝑒𝑟 𝑙𝑖𝑓𝑒𝑡𝑖𝑚𝑒=∑
𝐶𝑡+𝑀𝑡+𝐹𝑡 (1+𝑟)𝑡 𝑛𝑡=1
∑𝑛𝑡=1𝐸𝑡 (1) Where
Ct = Capital expenditures in year t Mt = Maintenance and operation expendi- tures in year t
Ft = Fuel expenditures in year t Et = Electricity generated in year t
r = Discount rate n = System lifetime
Levelized cost of storage, LCOS
The same principle used in equation 1 also applies for equation 2. LCOS [8]. The main difference being that electricity is discharged from the storage, and uses differ- ent parameters, instead of being produced. Capital costs, Cts, consists of power and storage related costs. Electricity out, Eout, contains the capacity of the storage, in kWh, cycles during a year, depth of discharge, DoD, and expected lifetime [8].
𝐿𝐶𝑂𝑆 = =∑
𝐶𝑡𝑠+𝑀𝑡𝑠+𝐹𝑡𝑠 (1+𝑟)𝑡 𝑛𝑡=1
∑𝑛𝑡=1𝐸𝑜𝑢𝑡 +𝐶𝑜𝑠𝑡𝑒𝑙𝑒𝑐
𝜂𝑠 (2)
Where
Cts = Capital expenditures in year t
Mts = Maintenance and operations cost for storage in year t Fts = Fuel expenditures in year t
Eout = Electricity out in year t r = Discount rate
Costelec = Price of electricity ηs = Efficiency of storage n = System lifetime
6 3 Results
3.1 Literature review
Previous research was studied in the areas of energy storage methods, nuclear power, and renewable energy sources. Recent studies, which were especially inter- esting, contained a more recent cost information, as well as possibilities and limita- tions that come with each technology.
3.1.1 Energy storage
There are several ways to store energy, for both short and long periods of time, un- til demand requires it. Which technology that is best suited is not straightforward as it depends on geographical, environmental, and other factors [5]. There have been several studies done on different energy storage methods from many parts of the world, as the growth of intermittent renewables continue. Some of the technologies that could be used for electrical energy storage are presented in the following sub- sections.
3.1.1.1 Pumped hydro storage, PHS
The most common energy storage technique, in 2020, was pumped hydro storage [5] with nearly a 99 % market share for electrical energy storage [9]. It has been used for a long time and can be considered a reliable technology [5]. It however comes with high capital costs and a long construction time, it is also not very flexible when it comes to down scaling if needed [5]. Sweden has two pumped hydro stor- ages, Letten (36 MW) and Kymmen (55 MW) [10]. When there is either, an over- production of electricity, or during hours with low demand, it is possible to pump water to a dam or reservoir, to store energy that can be used when there is a de- mand for electricity. Pumped hydro thus uses potential energy of water in order to generate electricity [5], [10]. It is at this time mostly used for peak reduction, and has an expected lifetime ranging from 50 to 100 years [10].
There have been several studies conducted, on how to utilize pumped hydro storage in different power systems. Percebois and Pommeret [11] investigated how the cost, in the French power system, would be affected by an increased amount of intermit- tent wind power, as 25 % of nuclear power plants has begun being either phased out or retired, and how pumped hydro energy storage could help mitigate the effects.
They concluded that there were several uncertainties that made it difficult to deter- mine the result. Indications were however that there was a need for both short-term and long-term energy storage in their power system, for France to manage the shift [11].
A Croatian study by Krajacic et.al [12], researched how financial tools could be used to increase profitability of pumped hydro storage, to make them more attractive in the market. Like previous studies, they found that pumped hydro storage can have beneficial effects in the power system but can be expensive. The cost of PHS depend on geographical location, and whether existing water reservoirs could be used for storage or not. They stated that, for Croatia to increase grid penetration (a percen- tual increase) of intermittent renewables, energy storage will likely be required, and to finance the system feed in tariffs and taxes could be used. An Australian study by McConnell et.al [13] investigated pumped hydro storage as a potential competitor to open cycle gas turbines for peak generation. They concluded that the two methods were both cost-efficient and could be used as peak generation capacity. They also found that the efficiency of the storage did not have a large impact on the storage cost. Electrical network costs were however not included in their study, and due to Australia’s extensive land area this would likely impact the cost. Sioshansi et.al [14]
compared using pumped hydro storage and compressed air energy storage, as load shifting devices, buying electricity when the cost is low and selling when it is high.
Their capital cost for PHS was set between $1500-2000 kW-1 and $750-1000 kW-1 for CAES, in 2008 dollar. While their maintenance and operation cost for CAES were set to $1 MWh-1, and with a 12 % discount rate. In their study, compressed air energy storage ended up being slightly more profitable than pumped hydro stor- age as capital cost was lower for CAES. There is, however, a need to consider fuel cost for CAES, as natural gas is used in the process to increase efficiency [14].
Other studies have used levelized cost of storage to compare different energy stor- age technologies. In Jülch [8] several different energy storage methods were com- pared, one of them was pumped hydro storage. Jülch collected cost data from litera- ture for older technologies, while she gathered market prices and data for newer technologies where it was possible, since there were great cost variations in litera- ture. Like previous studies, Jülch found that life expectancy for the pumped hydro storage was long, while pointing out that pumps and turbines likely needs to be re- placed after about half of the storage lifetime. Both short-term and long-term stor- age was analysed, short-term with 365 cycles per year, as it could be expected to be used daily, and long-term storage had a cycle once per year. In both the short and long-term perspective, pumped hydro had the lowest cost, with compressed air en- ergy storage closing in as long-term storage, due to storage capacity for CAES being less expensive. Battery energy storage methods were catching up as short-term stor-
8
More than 90 GW PHS were installed around the world, with new PHS facilities being constructed. It is important to consider geographical location to keep costs at a minimum [15]. Ekman and Jensen [16] explored the possibility to use pumped hy- dro storage or compressed air energy storage, to mitigate the effect of an increasing amount of intermittent wind power in the Danish power system. Due to Denmark having a flat landscape and not being suitable for regular PHS, they investigated un- derground PHS systems instead. Underground PHS use the same working principle as regular PHS [9], but has the lower reservoir underneath the surface, resulting in a more complex design. While underground PHS had higher capital cost than regular PHS systems, regular PHS $600-1000 kW-1 for power production and $1-20 kWh-1 for storage, and underground PHS $1200 kW-1 for power production and up to
$50 kWh-1 for storage, it could be a viable option for Denmark according to Ekman and Jensen [16]. They stated that CAES and underground PHS can utilize similar ge- ographical locations in Denmark. For storage applications to be profitable, however, there must be large variations in electrical prices [16]. As the power grid experi- ences an increase in wind penetration, energy storage will likely be required to keep the grid operational when fossil power sources are phased out [16]. In the following subsection compressed air energy storages is reviewed.
3.1.1.2 Compressed air energy storage, CAES
In a compressed air energy storage, air is compressed and stored in various types of reservoirs when there is an excess amount of electricity being produced [5]. When the compressed air is discharged, to produce electricity, it is heated to improve the expansion taking place in the turbine. During discharging, this can be achieved by combusting e.g. natural or biogas, and heating the discharged air [5], [17]. As previ- ously mentioned, CAES systems can be utilized in a similar way as PHS. Das et.al [18] studied how CAES could be used, as a short-term storage unit, with different levels of wind penetration in the power grid. Without CAES, there were larger fluctuations in the power supply, adding a CAES unit helped to mitigate and dampen the intermittent nature of wind power [11], [18]. Depending on the energy mix in the grid, Das et.al [18] discovered that the addition of a CAES could both have a positive and negative impact on CO2-emissions, that economic profitability de- pended on how the CAES operated and on the level of wind power penetration (percentage of wind in the power system). Lower wind penetration resulted in lower profitability. To get a better understanding on how an energy storage could be beneficial, both functionally and economically in a power grid, it is important to consider not only the storage on its own, but the entire system surrounding it [18].
In [8] compressed air energy storage ended up not being the best option, but it showed promise. Current CAES technology requires natural or biogas, to increase efficiency of the process, when discharging the air. This affects cost, as maintenance and operations had the highest impact on levelized cost of storage for this technique [8]. A newer technology, which is not yet commercially available, is adiabatic- CAES, that stores the heat from the compression process and uses it to reheat the air when discharging, to increase efficiency [8], instead of using fossil fuels.
A case study from Greece [20] studied several different energy storage technologies, and how well they could be integrated into electrical grids, with a high fraction of wind and solar photovoltaics, one of them was CAES. Capital cost, input electricity cost, fuel cost, and both fixed and variable maintenance costs, were included [20].
CAES was not among the technologies best suited, for the studied cases. The study also indicated that input electrical cost was an important factor for the economic profitability [20]. In 2007, Greenblatt et.al [21] studied which of gas turbines and CAES, would be most suited to complement wind power as baseload. Due to the large difference in capital cost between gas turbines and compressed air energy stor- age, wind as baseload in combination with gas turbines had the lowest cost in their
10
Sopher et.al [19] evaluated the feasibility of establishing a CAES in Gotland (Swe- den), where wind power produces a large portion of the electricity, around 40 % in 2013 for the island. Their results indicated that CAES could be a suitable energy storage in Gotland, at least from a geological perspective, but that further studies with focus on engineering and economics, were required.
When it comes to cost of compressed air energy storage, there are not exceptionally large variations in literature, compared to battery technologies, where the variation was substantially larger. As previously mentioned in [14], the cost of CAES was (in 2008 dollar) $750-1000 kW-1 with maintenance and operations cost at $1 MWh-1. A more recent review [22] found, that costs was $800-1000 kW-1 for power produc- tion and $50-150 kWh-1 for storage capacity. The coming section contains a review over battery energy storage technologies.
3.1.1.3 Battery storage
Utility scale batteries are only in the beginning of being implemented as a support function to the electrical grid, but will likely be an important part of the energy sys- tem as the technology matures [5]. There are different types of battery energy stor- age methods, such as; flow batteries, lithium-ion, lead-acid, and natrium-sulphur [5], [9], [10]. Batteries work by converting chemical energy into electrical [9].
There are presently different types of flow batteries being used and developed, one being the vanadium redox battery (VRB). These types of batteries have an efficiency of around 75-85 %, and use two liquid electrolytes, which can flow in and out of the cell through a membrane, to cause an electrochemical reaction. This allow for inde- pendent sizing of power conversion unit and storage [9], [15]. Some of the ad- vantages that a flow battery has; increased safety due to separation of the chemicals, higher depth of discharge compared to conventional batteries, and low requirement for maintenance, among others [9], [22].
One of the techniques that Ekman and Jensen [16] investigated, for enabling imple- mentation of a higher wind power penetration in the Danish power grid, was flow batteries. While ultimately not ending up using it, they saw future potential for it.
They stated that the cost of flow batteries, could range between $1100-4500 kW-1 for the power conversion unit, and $110-320 kWh-1 for energy storage tanks that contain the electrolyte (in 2010 dollar). In a more recent study [22] from 2014, prices for VRB was found to be in the ranges of $1200-2000 kW-1 for the power conversion unit, and $350-800 kWh-1 for the energy storage capacity.
Lithium-ion have seen large price drops in recent years, enabling a rapid growth of lithium-ion batteries [5]. While they are currently more common in new battery ve- hicles than in utility scale applications, there will likely be an increase of them in the utility field as well [5]. A lithium-ion battery is less harmful to the environment, when compared to lead-acid batteries, as the salt and lithium is less toxic and can be recycled [9]. Lithium-ion batteries have a large energy density, compared to other existing battery technologies, and have an efficiency of 80-90 % [10]. According to [22] the efficiency was 5 % higher than the 80-90% mentioned in [10], and had a lifetime of 10 to 15 years. Capital cost ranging from $400-1000 kW-1 for power conversion, and $500-1500 kWh-1 for energy storage capacity (in 2014 dollars).
Lead-acid are among the oldest batteries and is thus highly developed. It is a re- chargeable battery that has low capital costs tied to it [9]. However, as lead is a heavy metal that is toxic for both humans and the environment, there are better op- tions to use today [9]. Further, lead-acid batteries are not used in large-scale applica- tions due to its limited lifetime, 5 to 10 years depending on how they are used [16], compared to other energy storage technologies [9]. As regular batteries have a low amount of discharge cycles, ranging from hundreds to thousands, they are more suited for small to mid-scale energy storage application. However, according to [15]
lithium-ion batteries are likely even more suited than lead-acid, for small-scale en- ergy storage. In the previously mentioned Danish study [16], lead-acid batteries had a cost of $225 kW-1 for power conversion, and $300 kWh-1 for the storage capacity (in 2010 dollar). Castillo and Gayme [22] presented that power conversion costed around $300-800 kW-1 and energy storage capacity $150-500 kWh-1, in 2014.
When Kaldellis and Zalfirakis [20] performed an economic analysis in Greece, on which energy storage solution would be the optimal, their conclusion was that lead- acid batteries fell short to natrium-sulphur (NaS) batteries, which will be explored next.
12
Since early 1980’s, natrium-sulphur batteries have been developed [9]. The battery utilizes molten sodium and sulphur, and has a lifetime around 15 years with little need for maintenance during that time period [9]. A consequence of the high tem- perature it operates on, is that it can be dangerous for operators in case the battery would fail. Should air leak in to the battery it is highly probable that the sodium would explode when it comes in to contact with air [9]. In 2006 New York, Wala- walkar et.al [23] studied economics of different types of energy storage technolo- gies, for regulation and load-shifting purposes. They investigated whether gas tur- bines, flywheel, or a battery storage technique should be used for the stated pur- poses. In the end flywheels and batteries were selected due to being less bound by location. They found that flywheel were not suited for load-shifting, and chose na- trium-sulphur over lead-acid, due to the fact that the properties of natrium-sulphur were better and had already been used for peak shaving and load-levelling functions.
A similar result was found in [20], where NaS also proved superior to lead-acid.
Prices for natrium-sulphur is dependent on how it will be operated, in [23] capital cost were around $1125-2250 kW-1, with yearly maintenance and operation costs at $15-90 kW-1 (in 2007 dollar). While [22] showed similar capital cost at
$1000-2000 kW-1 for power conversion and $125-250 kWh-1 of storage capacity (in 2014 dollar). Following section will review literature containing important criteria’s that should be considered when studying different energy storage methods and their purpose.
3.1.1.4 Additional important factors to consider for energy storage
In previous subsections, some of the areas where energy storage is useful was pre- sented. In this subsection a more extensive summary over what energy storage could be utilized for can be found. Several studies [11], [12], [22] - [25], points at the im- portance of being able to utilize energy storage when the energy mix shifts; from a more predictable generation, provided by fossil and nuclear energy sources; to a more unpredictable generation, from both wind and solar power as they are harder to forecast, and thus are not particularly good at load-following regulation by itself.
In addition, it is important to know what supplementary service the energy storage is going to provide, e.g.; if it is going to be operating to support power quality, by maintaining voltage and frequency stability; as an energy arbitrage; supporting peak loads, by storing energy when there is an overproduction of electricity or during the off-peak hours; or as a load-following device, where it helps to maintain power bal- ance in the grid as it fluctuates [22]. In the coming section the state of nuclear power plants in Sweden, together with recent studies on current and future fission technol- ogies, is presented.
3.1.2 Nuclear power
In the 1950’s, the first nuclear reactors were being built in the world, and the tech- nology developed over the coming decades [26]. In 2017, nuclear power generated 10 % of the worlds gross electricity [27], and in Sweden 39.3 % of the net electric- ity was generated by nuclear power [3]. With the world moving more towards re- newable energy sources, and the interest in nuclear power is in decline [28], studies have been made on what impact shifting from both fossil and nuclear power, to re- newable, could have on power systems. Both technological and economical aspects have been studied around the world.
As depicted in Figure 2, nuclear power serves as baseload in the Swedish grid. Nu- clear power is very suitable for baseload electrical production, as it allows for the nuclear power plant to be operational most of the year [28]. According to [28], nu- clear power is not suited for load-following operation, due to the capacity factor not being kept at an optimum; thus reducing the total electricity produced over the nu- clear reactors lifetime. Although in [11], it is stated that France developed tech- niques, to use some of their nuclear reactors as load-following units due to demand.
When performing economic analysis on high capital cost investments, such as nu- clear power, it is important to carefully consider the discount rate, as it will affect the cost of electricity [28]. Both nuclear and renewable power production have lower CO2-emissions, compared to fossil-based power plants [28].
In [29] Kan et.al investigated how Swedish power production would be affected, by striving towards a low carbon power production, with and without nuclear power in it, and how interconnectivity between European countries could mitigate the in- termittency of larger fractions of renewables. Wind and hydro power ended up be- ing the largest sources of power; supported by biomass, biogas, and solar, if using existing interconnections [29]. Based on installed capacity, the Swedish power pro- duction could supply the electrical demand, however, if the European interconnec- tions would be increased; installed wind power should be increased to approxi- mately 30 GW, for more export options of electricity [29]. There could be some small remainder of nuclear power that could be useful for export purposes, accord- ing to [29]. When it came to energy storage technologies, barely any was used in [29]. As the high capital cost is an issue for investors when it comes to nuclear power, there could be a use for smaller standardised reactors. Smaller reactors could be manufactured in higher numbers, as the regulations for large nuclear reac-
14
When analysing economics of nuclear power plants, its capacity factor can weigh in heavily on the cost. Most reactors produced since the 1980’s have a capacity factor around 90 %, with some downtime for refuelling and maintenance [30]. Nuclear power plants have a capital cost around $4200 kW-1 (in 2007 dollar) and mainte- nance and operation cost were around 3 %. More recent studies have capital cost in the range of $4700-5700 kW-1 (adjusted for inflation to present value), and a similar maintenance and operations around 3 % of annualized capital cost, as the previous study [28], [29]. Construction time for nuclear power plants can be 4 to 6 years, not including all preparations that must be met before construction can begin. There are however plants that have gone over planned schedule, an example is the Olkiluoto 3 reactor in Finland [30]. Harding [30] states that most new nuclear reactors being constructed is in Asia, and that many reactors will retire around 2030. Fewer new nuclear reactors are being constructed in the world, than older reactors being de- commissioned. Which will further reduce the share of nuclear power, for electrical production, around the world.
While nuclear power could still have a future in the power system, it bears a stigma due to the three nuclear accidents that have occurred: Three Mile Isle, Chernobyl, and Fukushima. How to handle the expensive final storage of spent nuclear fuel and waste, so it can be safely stored to prevent future generations to suffer from its ef- fects, is another issue that follows nuclear power [28]. Following subsection pro- vides a brief overview of the active Swedish nuclear power plants and their reactors.
3.1.2.1 Current active nuclear reactors in Sweden
There are currently, in 2020, seven nuclear reactors in operation in Sweden [31], they consist of boiler water reactors (BWR), and pressure water reactors (PWR), of varying installed capacity as shown in Table 1. Ringhals 1 will be decommissioned in 2020, while the remaining reactors were estimated to operate until around 2040 [31]. In the beginning of 2020, Sweden had an installed capacity around 7.8 GW, and at the end of 2020, when Ringhals 1 is shutdown, 6.9 GW installed capacity will remain. Swedish nuclear power production ranged between 5.0 to 8.4 GW in 2019, before Ringhals 2 shutdown [6].
Table 1 Nuclear reactors in Sweden 2020 [32], [33] [34].
Name Forsmark
1 Forsmark
2 Forsmark
3 Oskarshamn
3 Ringhals
1 Ringhals
3 Ringhals 4
Type BWR BWR BWR BWR BWR PWR PWR
Installed
capacity 990
MW 1 118
MW 1 172
MW 1 400 MW 881
MW 1 063
MW 1 130
MW Starting
year
1980 1981 1985 1985 1976 1981 1983
The active Swedish BWR and PWR are light-water reactors and are second genera- tion nuclear reactors [35]. The following subsection will outline the concept of fourth generation nuclear power plants, which is not yet implemented.
3.1.2.2 Nuclear reactor: Generation IV
A newer concept that intends to eliminate the weaknesses from the previous genera- tion nuclear power, is generation IV [35]. By eliminating the risk for severe nuclear accidents, a reduction in the lifetime of radioactive waste by fuel reprocessing, and a process that prevents nuclear material from being weaponized, a safer system can be created, that is more economically viable compared to current energy sources [35].
Breeder reactors output more fissile material than is put in, this material can be used as fuel in new reactors [35]. Spent fuel from current nuclear reactors could be used as well in breeder reactors, thus reducing time needed to store the spent fuel in or- der for it to become harmless [35]. In the next section, the renewable energy sources hydro, wind and solar power are presented.
3.1.3 Renewable energy
Hydro power is the main contributor to electrical production in Sweden [3]. In 2019 the maximum power output from hydro was 12.9 GW, with a minimum out- put at 1.5 GW [6]. Wind power has increased steadily during the last decades. In 2017, 11 % of Sweden’s generated electricity came from wind power [3]. Figure 3 shows how Sweden has been developing the wind power capacity since 2000 to 2017; from 2010, an accelerated build rate can be seen.
0 2 000 4 000 6 000 8 000 10 000 12 000 14 000 16 000 18 000 20 000
2000 2002 2004 2006 2008 2010 2012 2014 2016
16
In a previously mentioned study, Kan et.al [29] analysed how interconnectivity be- tween nations in Europe, could help mitigate the effects of Swedish wind power.
Also, if it could be scaled up further and suggested a case where 30 GW wind power was installed, with a low share of solar power in the energy mix. Indications from their study were that it was possible to accomplish. They also considered weather scenarios and how it could affect power production from hydro for the past 20 years; however, no extreme weather scenario was included. Ireland has, like Swe- den and other countries, been building up their wind power plants, and have investi- gated how an increased wind penetration affects the power grid [36]. Tuohy and O’Malley [36] discovered that usage of energy storage had a beneficial effect, at higher penetrations of wind power in the power grid. They used pumped hydro storage in their analysis; for it to be economically sustainable, wind penetration had to be around 50 %. They also brought up technological issues in their study, the ma- jor one being that with a large amount of wind power in a power grid, stability problems with voltage and frequency will follow. Due to these issues, energy stor- age could be justifiable from a technological perspective to match power supply to demand [36]. Other studies that approached high wind penetration concluded that without energy storage, the energy losses from wind power would be rather large, when grid penetration is high [24], [37]. Solomon et.al [24] also pointed out that, large energy storage like PHS and CAES, could go underutilized in a system with high wind penetration due to higher short-term storage demands.
In [38] the effect of climate change on the electricity market in New Zealand was studied when there was a large share of hydro power in the power system, which was used for seasonal storage. Due to more extreme weather effects, and more un- predictable weather patterns; the inflow of water to the hydro storage could be af- fected negatively, which in turn leads to lower profits. To offset this effect, in- creased hydro storage could be used [38]. In 2007, capital cost of wind power was around $1700 kW-1 (in 2007 dollar) according to [30]. Since then, costs for part production and construction, as well as labour costs, have decreased and started to level out [28]. More recent capital costs used in literature was $1090 kW-1 for on- shore, and $2880 kW-1 for offshore plants [29]. Maintenance and operations cost ranges from 1.5 to 3.5 % of capital cost per year [29], [30]. Next section contains levelized cost of energy comparisons between different energy storage methods, as well as how wind power together with energy storage compares to nuclear power.
3.2 Economic comparison
Cost comparisons of energy storage methods, as well as how the combination of wind power and energy storage compare to nuclear power will be presented in the following subsections. From reviewed literature discount rate was pointed out as one of the parameters that could influence the result significantly, which is why it was chosen for the sensitivity analysis.
3.2.1 Levelized cost of energy and storage
When performing economic calculations on investments within the energy sector it is common to use the method LCOE; to allow for easy cost comparisons between different power plants and energy storage. Several studies [7], [39] - [41], have used LCOE and LCOS, for cost analysis on different types of systems. A couple being smaller household systems, with photovoltaics and batteries, and other being large scale grid tied systems. Discount rate has a large impact on total costs, especially for larger systems with high capital costs, and should thus be carefully considered [7], [28]. In [8] it is stated that, LCOS is a fast method to use when comparing different energy storage methods, but that it is sensitive to large variations of energy that is being charged in long-term storage.
3.2.2 Energy storage economic comparison
A high and low-cost scenario was created, based on data from some of the refer- ences mentioned previously, it was compiled, and adjusted for inflation to present currency value (2020). After that, LCOS and LCOE was calculated for both scenar- ios. Both scenarios used 1 GW installed power and 6-hour storage capacity, as a base. It was assumed that the energy storage would discharge one cycle per day, thus eliminating the need to consider self-discharge. For the battery storage, degradation was excluded to simplify the calculations, and leads to a lower LCOS.
18
In Table 2, input data is presented for the high cost scenario; power conversion and storage make up the capital cost.
Table 2 Input data for the high cost scenario
High cost scenario PHS CAES VRB Li-
ion Lead-
acid NaS Wind Nuclear [28]
[14]
[16]
[22]
[19]
[9]
[16] [9]
[22] [22]
[19] [22]
[23] [28]
[27] [28] [27]
Power Conversion (€ kW-1) 2100 800 1800 800 700 1800 12001 45002 Storage (€ kWh-1) 25 40 600 1300 500 200 - - Maintenance and Operation
(%/Year) 2.0 1.0 1.0 1.0 1.0 2.0 4.0 4.0
Discount Rate (%) 15 15 15 15 15 15 15 15
Lifetime (Years) 80 40 20 15 10 15 30 60
Efficiency (%) 75 60 75 80 70 75 - 33
Capacity factor (%) - - - - - - 40 90
Fuel (€ kWh-1 fuel) - 0.010 - - - - - 0.030
Depth of discharge (%) 95 55 95 75 45 60 - -
Cycles (Yearly) 365 365 365 365 365 365 - -
In the low-cost scenario (input data in Table 3) the capital cost, maintenance and op- erations cost, and fuel costs were lower. While efficiency and depth of discharge, DoD, were higher.
Table 3 Input data for the low-cost scenario
Low cost scenario PHS CAES VRB Li-
ion Lead-
acid NaS Wind Nuclear [28]
[14]
[16]
[22]
[19]
[9]
[16] [9]
[22] [22]
[19] [22]
[23] [28]
[27] [28] [27]
Power Conversion (€ kW-1) 1500 500 1000 300 300 1000 10003 40004
Storage (€ kWh-1) 1 20 300 400 200 100 - -
Maintenance and Operation
(%/Year) 0.5 0.3 0.5 0.5 0.5 1.0 2.0 2.0
Discount Rate (%) 5 5 5 5 5 5 5 5
Lifetime (Years) 50 20 10 10 5 10 20 40
Efficiency (%) 85 80 85 90 90 85 33
Capacity factor (%) - - - - - - 40 90
Fuel (€ kWh-1 fuel) - 0.005 - - - - - 0.003
Depth of discharge (%) 100 70 100 85 55 80 - -
Cycles (Yearly) 365 365 365 365 365 365 - -
1 Capital cost for wind power
2 Capital cost for nuclear power
3 Capital cost of wind power
4 Capital cost of nuclear power
In Figure 4, the different types of energy storage methods are compared. Pumped hydro storage (PHS) and compressed air energy storage (CAES) had the lowest LCOS of the compared energy storage methods, while the battery storages had higher LCOS. Lithium-ion (Li-ion) had the largest cost variation, and lead-acid had a similar LCOS as li-ion. Natrium-sulphur (NaS) and vanadium redox (VRB) had the lowest LCOS of the batteries in the comparison.
Figure 4 Comparison of levelized cost for energy storage methods, without cost of electricity. Bars indicating the range between high and low-cost scenarios
3.2.2.1 Sensitivity analysis
To check the sensitivity of a crucial input parameter, a sensitivity analysis was per- formed on discount rate for the different cases. For the high-cost scenario, lithium- ion batteries ended up having the highest LCOS, €0.86-0.90 kWh-1, while CAES had the lowest LCOS at 15 % discount rate, €0.09-0.15 kWh-1 (see Figure 5). A discount rate between 8 to 12 % were used for PHS and CAES in [14].
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
PHS CAES VRB Li-ion Lead-acid NaS
LCOS (€/kWh)
0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
LCOS (€/kWh) PHS
CAES VRB Li-ion Lead-acid
20
In the low-cost scenario (see Figure 6) lead-acid had the highest LCOS €0.31-0.35 kWh-1, and CAES €0.04-0.9 kWh-1 the lowest – like in the high cost scenario.
Figure 6 Sensitivity analysis low cost scenario LCOS, without cost of electricity
In the next section, the combined LCOS and LCOE for wind power is compared with LCOE for nuclear power followed by a sensitivity analysis on the discount rate.
3.2.3 Wind power and energy storage, compared to nuclear power Just as in the comparison of energy storage, a high and low-cost scenario was cre- ated for wind and nuclear power, LCOE was then calculated for both. Figure 7 shows that wind power, in combination with pumped hydro storage, and com- pressed air energy storage, both around €0.07-0.24 kWh-1, could almost provide electricity at a cost comparable to nuclear power €0.03-0.12 kWh-1. While battery technologies combined with wind power still have higher LCOE than nuclear power; natrium-sulphur and wind were the closest at €0.15-0.47 kWh-1.
Figure 7 LCOE of wind power and energy storage compared to nuclear power, with cost of electricity. Bars indicating the range between the high and low-cost scenarios
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40
5% 6% 7% 8% 9% 10% 11% 12% 13% 14% 15%
LCOS (€/kWh)
Discount Rate (%)
PHS CAES VRB Li-ion Lead-acid NaS
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
Wind + PHS
Wind + CAES
Wind + VRB
Wind + Li- ion
Wind + Lead-acid
Wind + NaS Nuclear
LCOE (€/kWh)
3.2.3.1 Sensitivity analysis
Presented in Figure 8 is the high-cost scenarios sensitivity analysis of the discount rate. Nuclear power had the lowest LCOE, at €0.06-0.12 kWh-1. With the combi- nation of lithium-ion and wind, being the most expensive, at €0.89-0.96 kWh-1. In [28] it was stated, that a 10 % discount rate would be appropriate in a deregulated electricity market, while a lower discount rate could be expected in a regulated market.
Figure 8 Sensitivity analysis high cost scenario combined LCOE, with cost of electricity
For the sensitivity analysis of the low-cost scenario; wind in combination with lead- acid batteries had the highest cost, and nuclear power the lowest. Most sensitive to the discount rate was wind power combined with PHS, and nuclear power, see Fig- ure 9.
0.00 0.20 0.40 0.60 0.80 1.00
5% 6% 7% 8% 9% 10% 11% 12% 13% 14% 15%
LCOE (€/kWh)
Discount Rate (%)
Wind + PHS Wind + CAES Wind + VRB Wind + Li-ion Wind + Lead-acid Wind + NaS Nuclear
0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45
LCOE (€/kWh) Wind + PHS
Wind + CAES Wind + VRB Wind + Li-ion Wind + Lead-acid Wind + NaS Nuclear
22 4 Discussion
Levelized cost of energy is a common method to use when comparing costs of en- ergy sources and energy storage technologies. Although somewhat modified for en- ergy storage, it was deemed as a suitable choice of method. Some of the assumptions that were taken affected the result. For battery technologies, including degradation would have increased the prices even higher. If the capacity factor for wind power had been lower, the LCOE would have been higher. A sensitivity analysis was then performed on one of the parameters – the discount rate – as it was deemed to have a high impact on the result of LCOE, according to the literature review. To perform a more extensive financial analysis, other methods should be used, and input parame- ters should be studied thoroughly – to establish more accurate results.
Data from several studies were integrated and used for input data, thus providing a wider cost perspective than what fewer studies could provide. A downside to using several studies, however, was that cost were presented in different manners in dif- ferent studies, which complicated evaluation. As the studies came from different years the cost data, in some of the studies, were adjusted for inflation to 2020 Euro value. Jülch adjusted for inflation in [8] this was, however, not common in the re- viewed literature. In the performed calculations more recent data was used over older, as for some of the technologies price have changed in recent years, especially for battery technologies.
From the comparison of energy storage methods, in Figure 4, pumped hydro stor- age (PHS) and compressed air energy storage (CAES) had the lowest costs. This was also the case in the reviewed literature. If CAES could become independent of fuel requirement to boost efficiency, both cost and CO2-emissions would be reduced, which would likely result in a lower LCOS than PHS.
