IN
DEGREE PROJECT TECHNOLOGY,
FIRST CYCLE, 15 CREDITS ,
STOCKHOLM SWEDEN 2020
Investigation of the siting process
of Swedish nuclear power plants
using GIS
EGIL MORAST
DANIEL WOLLBERG
KTH ROYAL INSTITUTE OF TECHNOLOGY
Abstract
In order to find the optimal placement for a nuclear power plant (NPP) many different factors have to be taken into consideration. Choosing the wrong location could result in higher construction costs and/or taking avoidable risks. GIS can be used to compare different spatial criteria efficiently. There are currently six nuclear reactors operating in Sweden, producing approximately 40% of the total electricity produced in the country every year. These reactors are divided over three nuclear power plants: Ringhals, Forsmark and Oskarshamn. The objective of this thesis is to evaluate the suitability of existing NPPs location in Sweden according to the current standards and guidelines and find good potential sites for a new hypothetical NPP in Sweden. The thesis is limited to investigate the
geographical aspects of siting an NPP, and therefore economic decisions and law restrictions will have less impact on the end result.
The methodology of the thesis is in line with International Atomic Energy Agency’s step by step siting procedure that includes a site survey in which candidate areas are derived using a set of exclusionary and avoidance criteria. This is followed by the comparison and ranking of these sites by implementing a suitability analysis in the site selection process to derive potential sites. The site survey process result in a map layer containing the candidate areas that fulfills the spectre of exclusionary and avoidance criteria set by the analyst. The site selection process includes weighting criteria compared to each other using the analytic hierarchy process which is based on the perceived importance of each criterion. The criteria are found in an initial study of available literature. Based on the gathered information, data is collected to execute the analysis in ArcMap. Both vector operations and raster calculations are made to produce the end result map layers.
Sammanfattning
För att hitta det optimala området för ett kärnkraftverk måste man ta hänsyn till flera kriterier. Fel val av område kan resultera i extra konstruktionskostnader och/eller onödiga säkerhetsrisker. GIS kan användas för att jämföra geografiska kriterier på ett effektivt sätt. Idag finns det sex stycken reaktorer i bruk i Sverige. Dessa producerar ungefär 40% av landets el. Reaktorerna är fördelade på tre stycken anläggningar: Ringhals, Forsmark och Oskarshamn. Syftet med detta arbete är att utvärdera
lämpligheten av placeringen av de existerande kärnkraftverken i Sverige i enlighet med nutida standarder och riktlinjer och att hitta bra potentiella områden för ett nytt hypotetiskt kärnkraftverk i Sverige. Arbetet är begränsat till att undersöka de geografiska aspekterna av nyförläggningsprocessen och därmed har ekonomiska beslut och lagrestriktioner mindre påverkan på slutresultatet.
Metodiken för arbetet är i linje med IAEAs steg för steg nyförläggningsprocedur som inkluderar en
site survey där candidate areas tas fram genom att använda en rad med kriterier för områden som antingen ska exkluderas eller undvikas. Dessa områden jämförs och rankas sedan genom att
implementera en suitability analysis i site selection processen för att framhäva potentiella områden. Site survey processen resulterar i ett kartlager med de kandidatområdena som uppfyller spektrumet av kriterier som analytikern har bestämt antingen ska exkluderas eller undvikas. Site selection processen inkluderar att vikta kriterier i förhållande till varandra med hjälp av den så kallade analytic hierarchy
process som baseras på den uppfattade vikten av varje kriterium. Kriterierna formuleras efter en initial litteraturstudie. Baserat på den samlade informationen hämtas data för att kunna utföra analysen i ArcMap. Både vektoroperationer och rasterberäkningar utförs för att producera de kartlager som sedan återfinns i slutresultatet.
Acknowledgements
Terms and Abbreviations
AHP - Analytic Hierarchy Process BWR - Boiling Water Reactor
GIS - Geographical Information System IAEA - International Atomic Energy Agency NPP - Nuclear Power plant
Table of content
Acknowledgements 3
Terms and Abbreviations 4
1. Introduction 7
1.1 Background 7
1.2 Research objectives 9
1.3 Limitations 9
2. Literature overview 10
2.1 GIS and Suitability analysis 10
2.2 Weighting method 10
2.3 Siting an NPP 11
2.4 Criteria for siting 12
2.4.1 Cooling water 12
2.4.2 Natural hazards 13
2.4.3 Power grid 13
2.4.4 Safety and emergency planning 13
2.5 Related work 14
3. Study area and data description 15
3.1 Study area 15 3.2 Data description 16 4. Methodology 18 4.1 Site survey 18 4.2 Site Selection 24 5. Results 31 5.1 Candidate Areas 31 5.2 Potential Sites 35
6. Discussion and Conclusion 38
7. Further studies 39
7.1 Fourth generation nuclear power 39
7.2 Small modular reactors 40
7.3 Cooling towers 40
Table of figures
Figure 1. Existing Nuclear Power Plants in Sweden. 7 Table 1. The Swedish nuclear power reactors type and output effect. 9 Table 2. Saaty’s fundamental scale. 11 Table 3. The different land cover available in the land cover layer. 16 Figure 2. Land cover visualized in vector format to the left and raster format to the right. 17 Figure 3. The mosaic layers combined into an elevation map. 17 Figure 4. The site survey process presented as a flow chart. 19 Figure 5. Map showing the buffer zones around airports and military facilities and
all the nature reserves. 20
Figure 6. Map showing where there are sufficient amount of available cooling water for
an NPP to operate. 21
Figure 7. All the buffer zones surrounding the towns. 22 Figure 8. Power lines in Sweden and the calculated buffer zone. 23 Figure 9. Railroads in Sweden and the calculated buffer zone. 23 Figure 10. Criteria coverage showing areas that fulfill the exclusionary criteria and/or
the avoidance criteria. 24 Figure 11. The site selection process presented as a flow chart. 25 Figure 12. The top left figure shows the significant buildings and antiquities, the figure on
the right shows the extent of important roads and railroad and the bottom figure shows the vector polygons used in the suitability analysis, nature reserves and
populated area. 26
Figure 13. Euclidean distance map to the left and slope map to the right. 27 Table 5. Breaking points for the different attribute layers in the different areas.
The highest value for the same layer is bolded. 27 Figure 14. Reclassified euclidean distance map to the left and reclassified slope map to
the right. 29
Table 6. Example of a comparison matrix. 29 Table 7. The three different weightings. 30 Figure 15. A closer look at the three candidate areas of interest, displayed in red. 31 Figure 16. A map showing the most important criteria when analyzing the existing NPP
in Forsmark. 32
Figure 17. A map showing the most important criteria when analyzing the existing NPP
in Oskarshamn. 33
Figure 18. A map showing the most important criteria when analyzing the existing NPP
in Forsmark. 34
1. Introduction
1.1 Background
There are currently six nuclear reactors operating in Sweden producing approximately 40% of the total electricity produced in the country every year. These reactors are divided over three nuclear power plants (NPPs): Ringhals, Forsmark and Oskarshamn, seen in Figure 1. Nuclear power is and have always been a controversial topic in Sweden and the public opinion about the subject have changed constantly during the last half century.
Figure 1. Existing Nuclear Power Plants in Sweden.
of dismantling was instead abandoned and the Swedish government voted yes to allow the construction of new nuclear power reactors, but only if they were built to replace old ones.
Parties in the Swedish government raised the issue of expanding the nuclear power industry with new reactors in 2016 and the taxation to operate nuclear reactors were removed (Lundberg, 2020). The same year the majority of the large Swedish parties signed an energy convention, agreeing to aim for “100 percent renewable energy” by the year 2040. The term renewable energy did not include nuclear power and some asked whether or not that meant that the last nuclear reactor had to be dismantled by this year, this was not the case and one party suggested if “100 percent renewable” should be changed to “100 percent fossil fuel free” in the agreement, this to include nuclear power.
In 2018 the issue was raised once more and the one party leader for one of the larger Swedish parties said that rather than shut down two reactors in Ringhals, Sweden should invest in more NPPs and reactors. Later in 2019 two parties left the energy convention because of the absence of investment plan of nuclear power (Björkman, 2019). The future of nuclear power in Sweden is yet undecided and a study from 2019 shows that only 11% of the people asked want to dismantle the existing nuclear reactors while 43% want to continue using nuclear power as an energy source, and if necessary, expand and build new reactors (Lindström, 2019).
Reviewing this events and the change in the political climate and public opinion in the last ten years it is possible that nuclear power is to play a greater part in Swedish energy production in the near future. In 2019 the total electricity production in Sweden was 165,4 TWh, an increase of 3,8% from the year 2018 according to the Swedish energy authority (Energimyndigheten, 2020). The need of electricity is increasing every year and in the next following 25 years it can be up to 60% more compared to now (Nylander, 2019). All together the six reactors at Forsmark, Oskarshamn and Ringhals produce the same amount of electricity as more than 2000 hydro plants in Sweden (Lindholm, 2019) and almost three times more than the 4200 wind power plants (Lindholm, 2020). In addition to this, nuclear power plants is dependent of wind or water and are able to produce electricity in a consistent and trustworthy rate. This raises the question if it is even possible to meet Sweden's future electricity demand without the involvement of nuclear power.
Even though the political climate have changed into a more positive view on nuclear power there are many legal restrictions that prevents the nuclear power industry to generate the momentum necessary to expand which requires determination and involves large economic investments. In a study from 2011 the company Elforsk compared more in depth the Swedish legal process of giving permission to build new nuclear reactors today to the process in the 1970’s. The major difference between now and then is the establishment of the environmental legal code, Miljöbalken in Swedish (Gåhlin, 2011). Miljöbalken introduces additional terms and restriction for expanding the nuclear industry, and as for now new reactors can only get permitted if they are to replace old reactors according to the 17 kap. 6a§. How suitable are the locations where the NPPs are operating today and how would finding more suitable sites affect the nuclear power debate in Sweden? There are many criteria a site need to fulfill to be considered a suitable location for an NPP and far from all sites do so. One can use Geographical Information System (GIS), to evaluate existing or site new NPPs efficiently by implementing
1.2 Research objectives
The objective of this thesis is to build a model using GIS and implement it to:
1. Evaluate the suitability of existing NPPs location in Sweden according to the current standards and guidelines based on available knowledge and information in the literature 2. Find good potential sites for hypothetical NPPs in Sweden
The objectives will be achieved by implementing GIS and by performing a suitability analysis with collected spatial data. Information about the criteria for siting an NPP will be be searched for and reviewed in a literature study part of the project. This newly acquired knowledge will then be put to practice following the general guidelines and process of siting an NPP combined with spatial data and GIS software.
1.3 Limitations
The process of site selection of an NPP is complex and includes taking aspects as law restrictions, economic decisions, technological advancement and policies in consideration, information with varying availability in the time frame for this project. Therefore this study will focus foremost on the geographical aspect of siting an NPP rather than the ones mentioned above. These aspects will still be discussed but not taken into consideration when selecting, comparing and evaluating sites. The sites that are considered as potential sites for an NPP are selected on the basis of the specifications of current reactors in Sweden with respect to size, type and output effect seen in Table 1. Two types of reactors are used in Sweden, boil water reactor (BWR) and pressurized water reactor (PWR). The operational area of a typical nuclear facility in the US with similar output effect is about 1 square mile according to energy.gov.
The data used to carry out the objectives of the thesis is free and provided from different authorities. For more in depth analysis, additional data may need to be purchased. It is important to know that the computations and use of data done for this project, including making a suitability analysis, can be done manually without the use of GIS. The main reason for using GIS is to be able to process great amounts of data at the same time and hence provide an easy way of visualizing the results, making the process much more efficient.
Table 1. The Swedish nuclear power reactors type and output effect.
Reactor Type Output effect
Forsmark 1 BWR 990 MW
Forsmark 2 BWR 1 118 MW
Forsmark 3 BWR 1 172 MW
Ringhals 4 PWR 1 130 MW Oskarshamn 3 BWR 1 400 MW
2. Literature overview
2.1 GIS and suitability analysis
GIS is used to collect, process and analyse data efficiently and make it easy to visualize and understand the data one is working with (ESRI, 2020a). GIS were first put in use in the 1960s and have continued to evolve and its uses today are unlimited in the digitalized society. Some fields where GIS is commonly used are network analysis, geographical mapping, forecasting, prevention and suitability analysis.
Suitability analysis is used to determine the most appropriate location for any facility or building by using geographic information and spatial data to aid the decision making process in a variety of hard decisions connected to spatial and geographic problems (Ryan, 2019).
The decision making process can be divided into four steps:
● “Problem definition”: The present state is described as well as the desired state and the difference between them.
● “Search for alternatives and selection criteria”: Potential solutions to the problem is explored and criteria are set up in order to evaluate the different alternatives.
● “Evaluation of alternatives”: The outcome of each alternative with its corresponding criteria is evaluated.
● “Selection of alternatives”: After the alternatives have been evaluated they are ranked from best to worst, where the best ones can be investigated further.
Instead of using the traditional method of overlaying maps manually, GIS is used to overlay digital maps (Jankowski 1995).
The analysis can be performed with both vector and raster data, but the latter is more widely used (Briney, 2014). By selecting and defining different criteria, suitability analysis can show how different candidate areas compare to each other depending on these criteria. All criteria are not binary but instead placed on a scale ranging from extremely beneficial to extremely unbeneficial and therefore the criteria layers have to be compared based on how much they are expected to affect the result. This is often referred to as weighting.
2.2 Weighting method
To determine the weights of each criteria there are plenty of methods to use. For this thesis the Analytic Hierarchy Process (AHP) , suggested by Thomas L. Saaty (2008). The method was
expected outcome and purpose of the decision. From here the decision hierarchy can be built. The next step is to construct the comparison matrices in which the different criteria are pairwise evaluated to each other. To be able to do this Saaty implemented a scale from 1 to 9 used to specify the ratio between two criteria seen in Table 2.
Table 2. Saaty’s fundamental scale.
Intensity of Importance Definition
1 Equal importance
2 Weak or slight importance 3 Moderate importance 4 Moderate plus importance 5 Strong importance
6 Strong plus importance 7 Very strong importance 8 Very, very strong importance 9 Extreme importance
2.3 Siting an NPP
There are guidelines and standards for siting an NPP provided by the International Atomic Energy Agency (IAEA), which is an organ under the United Nations with its headquarter located in Vienna, Austria. There are two major processes regarding the placement of an NPP: “Siting process” and “site evaluation process” (IAEA, 2015). The purpose of the siting process is to identify potential regions and compare different sites to each other to find the best one suitable for an NPP. The site evaluation process on other hand is initiated when the site is chosen and consist of site screening and monitoring on site. The transition between these two processes vary depending on the methodology used. This thesis focus mainly on the siting process. The siting process further consists of two stages: “site survey” and “site selection”. Where the first stage, site survey, involves the investigation of a large region and rejecting inappropriate sites and the latter stage involves judging and comparing the sites depending on their relative suitability based on different aspects.
of possible sites for an NPP. If the size of the candidate areas are too small or if there are too few candidate sites or the opposite, the avoidance criteria need to be refined to improve the quality of the candidates (AELB, 2011). The next step is to adjust map scale and investigate all the candidates areas more in depth, with more specific exclusionary and avoidance criteria possible on a smaller scale. The outcome of this procedure is potential sites. In the third step, considered the site selection stage, suitability criteria are used to compare and rank the relatively small number of potential sites to decide which site is best suited. The later steps are not included in the site survey or site selection stages and involves on site screening, geotechnical investigations and decision making and are therefore not regarded relevant to this thesis.
2.4 Criteria for siting
There are three identifiable types of criteria to be used in the siting process of an NPP: “exclusionary”, “avoidance” and “suitability criteria” (IAEA, 2012). The exclusionary criteria are used to determine if it is even possible to establish a nuclear power plant with the current level of technology and
knowledge, thus removing all sites without the basic requirements (IAEA, 2015). Avoidance criteria are criteria similar to the exclusion criteria but where current technology and knowledge can be implemented to prevent potential faults and hazards, thus increasing the cost of constructing the NPP unnecessarily much to consider the area to be a candidate area in the first selection. Avoidance criteria are flexible and used to regulate the diversity and number of potential sites (AELB, 2011). The last type of criteria is the suitability criteria which is used to compare sites and determine which sites that are more suitable than others. The process of scoring, weighting and ranking sites in regards to the established suitability criteria is iterative and subjective. In following subchapters the criteria found in literature are mentioned.
2.4.1 Cooling water
One of the, if not the, most important exclusionary criterion for locating a BWR or a PWR is the availability of cooling water. NPPs use large quantities of water to not only cool the reactor from overheating, but also to power the turbine, thus generating electricity (UCS, 2010). There are two types of cooling systems used for NPPs. Water consumption and withdrawal requirements are dependant on which cooler system that is put in use. The two systems are closed-cycle and
once-through. A 1 GWh PWR or BWR with a once-through cooling, similar to the current reactor in Sweden, has a requirement to be able to withdraw between 95 000 and 227 000 l/MW h from nearby water sources (Woods, 2016). The water source in this case needs to be large and reliable due to the withdrawal being equivalent to 26,6 - 62,5 m3 water per second. A closed-cycle cooling system uses a cooling tower to cool and condensate the water before reusing it in the reactor. Between 3 000 and 9800 l/MW h, equivalent to between 0,83 and 2,72 m3 per second. Since none of the existing NPPs in Sweden today uses cooling towers, that option will not be explored in this thesis.
though the possibility to withdraw cooling water from rivers has not been discussed at any greater lengths in Sweden, there is nothing that confirms that it would be impossible from a technical
perspective. Therefore that possibility will be taken into account for this study. This mainly applies to the major rivers in Sweden, since NPPs need a lot cooling water as stated earlier. Rivers in Sweden that have been identified with an average water flow of 200 m3/sor more are: Ljusnan, Kalixälven,
Dalälven, Torneälven, Umeälven, Indalsälven, Ångermanälven, Luleälven and Göta älv (SMHI, 2020). Smaller rivers have also been examined but something common for the average smaller rivers is that the water flow varies a lot throughout the year, sometimes coming close to 0 m3/s. For
example looking at daily measurements of water flow at a measuring station in Vanån during 2019, levels vary between 9,6 m3/sand 104 m3/s (SMHI, 2020).
So far only one commercial NPP in Sweden has been built in close proximity to a lake. The name of that plant is Ågestaverket, the first Swedish plant built to provide heat and electricity to the nearby households and not only for scientific research and study purposes (Nohrstedt, 2020). Ågestaverket was built adjacent to the lake Magelungen and operated between 1963 and 1974 until production was discontinued because of insufficient safety measures. Compared to reactors active today, Ågestaverket had a low net effect at only 68 MW for district heating and 12 MW for electricity. Since this study will focus mainly on larger sized NPPs, with a scale similar to the ones in use today, smaller lakes such as Magelungen will not be taken into consideration as a sufficient source of cooling water. The lakes in Sweden with a water flow needed to sustain the cooling of the nuclear core are Siljan,
Storsjön, Vänern and Mälaren. (SMHI, 2009; SMHI 2012a; SMHI 2012b; SMHI 2016) These lakes will be taken into consideration for this project.
2.4.2 Natural hazards
One of the largest threats to an NPP is natural hazards which is often dependant on the location. These include seismic activity and ground motion, volcanic activity, extreme weather, slope instability, erosion processes, flooding to name a few (IAEA, 2015). Sites that are more affected by one or more of these conditions are regarded as unsuitable.
2.4.3 Power grid
Proximity to the existing power grid for transportation of the electricity is important in order to decrease the cost of the construction of additional power lines and stations (IAEA, 2012). This is regarded as an avoidance criterion according to IAEA. Because of the high output from the NPP this criterion refers to power lines with 220kW voltage or more. Because of energy losses in the
transmission lines it is important that the NPP is located close to where there is demand for electricity such as cities or high density industrial areas (IAEA, 2012).
2.4.4 Safety and emergency planning
During normal operating procedure radiation emission is minimal and will not affect the siting process (Löwenhielm, 1997). However, a risk assessment must be made in case of nuclear meltdown and safety and emergency siting criteria is made with this assessment in mind.
the order has been issued. According to SSM, the Swedish Radiation Safety Authority, 5 km is enough to prevent death during an accident when an NPP’s security system fails in 90% of the occuring weather conditions.
According to the IAEA and AELB areas within 16 km of a major airport or within 8 km of hazardous facilities should be excluded. Examples of hazardous facilities are military bases, oil or gas wells, chemical facilities, dams and significant manufacturing facilities (AELB, 2011).
The risk of flooding, both natural or human induced, should be evaluated as well as the instability in ground surrounding lakes, coasts and rivers due to erosion or sedimentation (IAEA, 2012). The risk of flooding is deemed to be considered as an avoidance criteria since there are engineering solutions to minimize the risk. However it could potentially be difficult and/or costly to implement these solutions.
2.5 Related work
A similar analysis to this thesis has been made in the most populated nation in Africa, Nigeria. In Nigeria the demand of electricity is high due to the fast development rate. Eluyemi (2019) states the methodology in the study is in line with the IAEA safety standards, similar to this thesis and, the team managed to include almost every requirements made by the IAEA.
Because of Nigeria's geographical location a major part of the work circled around seismological activity and safety measures based on historical seismological events. Buffer zones are created around fractures, active fault lines and epicenters of earlier seismic activities. Other interesting buffer zones that were used in the analysis are:
● 20 km buffer zone from towns and villages. ● 5 or 30 km buffer zone from major rivers. ● 40 km buffer zone around state capitals. ● 40 km buffer zone from fault lines. ● 2 km buffer zone from roads.
The analysis was done in ArcMap and resulted in 12 candidate areas. It is discussed that these areas would be suitable for other sensitive facilities as well such as coal or gas power plants. Eluyemi states that other areas can be suitable for constructing an NPP, but the level of hazard would be higher and thus increasing the construction cost to reach the same level of safety.
In a research paper from 2007 Andrew Macintosh writes that half the population in Australia are not in favor of of nuclear power in other countries and that even two thirds of the population would oppose an NPP were they live (Macintosh, 2007). Australia has never had an operating NPP.
Macintosh continues by stating that if the government in Australia were to promote nuclear energy the siting of the NPP would be a large scale political issue. Therefore choosing these sites carefully are of priority. Macintosh identified four primary criteria:
● The NPP must be located close to relevant existing electricity infrastructure, such as high voltage power lines.
● The NPP is preferably located close to large centres of demand.
● Access to sufficient quantities of cooling water.
In addition to these four primary criteria Macintosh identified seven more secondary criteria that an appropriate site should fulfil:
● Having an appropriate buffer distance to populated areas. ● Having a good geological and seismological background. ● Low risk of extreme weather.
● Low security risks.
● Minimal impact on nearby important ecological areas. ● Minimal impact on nearby important heritage areas. ● Meet the economic and social demands.
The analysis was made by hand by listing all sites most likely to be suitable for an NPP and
comparing them to the criteria. Macintosh came to the conclusion that further research was needed to be sure these sites actually are suitable to site an NPP but also explained that there are for sure other sites that are potentially suitable as well.
In a research paper from Rosniza Idris (2012) investigate the possibility to site a nuclear power facility in Malaysia. In 2012 Malaysia relied heavily on coal, oil and natural gas but that source of energy was deemed to be unsustainable with the increasing electricity demand. Malaysia is considered to be qualified to meet the the electricity needs with nuclear power. An investigation of suitable locations for nuclear power plants issued. With an area of 2,271 km2, the town Raub in the federal state of Pahang was chosen as the study area. GIS and the AHP method were used to analyse the area. The important criteria taken into consideration in the decision making were:
● Residential area; Nearest residential area is at least 8 km away from the site. ● Topography; flat land or land with less slope than 20 degrees.
● Land ownership; Government land is prefered to private owned land.
● Land use; forest area or open land is prefered, it is not specified why in the work. ● Proximity to water source; site near a river or along the coastline to meet the demand of
cooling water.
● Population density; low population density is preferable.
● Electricity infrastructure; proximity to appropriate existing electricity infrastructure.
No information about how the criteria are weighted is given. The GIS analysis resulted in five suitable sites which was verified by visiting the sites.
3. Study area and data description
3.1 Study area
location for nuclear power because of the solid bedrock and the relative small amount of recorded tectonic and seismic activities. This promotes both the safety of the NPP when operating but also the possibility to safely store the nuclear waste on site or in dedicated facilities
3.2 Data description
Most of the data used in the study is collected from Lantmäteriet using Swedish University of Agricultural Sciences (SLU), geodata extraction tool (SLU, 2020). A land cover map for each candidate area is downloaded in vector format, seen in Figure 2, as polygon shapes through SLU’s geodata extraction tool contained in a dataset named “Fastighetskartan Markdata vector”. The file is then converted to a raster layer, seen to the right in Figure 2, with the resolution 10 meters using the
polygon to raster tool in ArcGIS. The resolution is chosen as a compromise between high detail level and reasonable processing speed for further analysis. The similar land covers are combined into one category, as seen in Table 3, by using the reclass tool in ArcGIS:
● Coniferous forest and deciduous forest are reclassified into forest. This because the type of trees do not affect the outcome of the analysis.
● High buildings, low buildings and squares are generalized as residential areas, as people live and socialise in these areas.
● Orchards and farmland are combined into farmland.
Table 3. The different land cover available in the land cover layer.
Abbreviation Land cover Reclassified as
Figure 2. Land cover visualized in vector format to the left and raster format to the right.
An elevation map for each candidate area, provided by Lantmäteriet in 2 m grid format, is downloaded using SLU’s geodata extraction tool. The elevation is expressed in the Swedish height system RH 2000 The dataset consisted of smaller individual map pieces which are combined using the mosaic to
new raster tool in ArcGIS and the combined map’s resolution is set to 10 meters. The result can be seen in Figure 3 below.
Figure 3. The mosaic layers combined into an elevation map.
Datasets with the relevant infrastructure is downloaded using SLU’s geodata extraction tool.
with information about the extent of swedish nature reserves, heritage monuments and protected bird areas are downloaded using SLU’s geodata extraction tool. The locations of the existing NPPs and airports are found on google maps and converted to a format that could be imported to ArcMap. This is done manually by the authors.
4. Methodology
The GIS software used to carry out this project is ArcMap 10.7.1 provided by Esri. ArcMap provides the user with a user friendly graphical interface and a good variety of easy to use tools to proceed with the suitability analysis.
4.1 Site survey
Sweden has an approximate size of 450 000 km2which in the targeted resolution, 10 x 10 meters, would result in 4,5 billion pixels. Processing that large amount of data many times requires great computer capacity and therefore the solution is to first find the candidate areas in a site survey by following IAEA’s proposed methodology. The first step is to find the candidate areas on a regional scale and then choosing promising candidate areas to derive the potential sites from. Finding candidate areas is done by applying the set of exclusionary and avoidance criteria on the region of interest by using a mask. This operation results in all areas outside the mask being rejected. The site survey process is presented as a flow chart in Figure 4.
First the mask layer is created, containing all the areas where it is possible to site an NPP considering the exclusionary criteria. To create the mask four geoprocessing tools in ArcMap are frequently used:
buffer, intersect and erase.
● The buffer tool creates circular buffer zones surrounding the selected objects or features with a given diameter, and the output is a polygon layer.
● The erase tool removes the features in one layer from another layer.
Figure 4. The site survey process presented as a flow chart.
Starting off with a polygon of Sweden's surface, the exclusionary criteria are applied creating a mask: ● An NPP has to be built on land and therefore all water bodies such as rivers, lakes and oceans
are removed using the erase tool.
● A 16 km buffer zone to larger airports is appropriate following the guidelines provided by IAEA, see section 2.4.6. For safety measures this is applied to all commercial airports in Sweden, seen in Figure 5.
● A 30 kilometer buffer zone to military practice facilities, ranges and storing facilities is issued following the the guidelines provided by IAEA, seen in Figure 5.
● An NPP may not be sited in areas classified as nature reserves, bio-sensitive regions and other important ecological areas. Neither should an NPP be located where it may have a negative impact on important heritage sites, seen in Figure 5.
Figure 6. Map showing where there is a sufficient amount of available cooling water for an NPP to operate.
The avoidance criteria are applied similarly to the exclusionary criteria but these are not strict and are allowed to be tweaked. The reason for this is to regulate the amount of candidate areas to be further examined, which makes this part of the process more interactive. It will be repeated until the number of candidate areas satisfy the decision maker. The avoidance criteria used are:
● Buffer zones from towns with a population of less than 20 000 people are set to 5 km, seen in Figure 7. Following the recommendations from the Swedish Radiation Safety Authority, it needs to be possible to evacuate all inhabitants within the first 4 hours of an eventual disaster. ● Buffer zones from towns with a population between 20 000 and 80 000 people are set to 10
km, seen in Figure 7.
● Buffer zones from towns with a population of more than 80 000 people are set to 25 km, seen in Figure 7.
● The maximal distance to any railroad is set to 15 km, seen in Figure 9. At first this was set to 30 km but that later resultatet in too many candidate areas to choose from and thus it was changed to 15 km to eliminate redundant candidate areas.
Figure 8. Power lines in Sweden and the calculated buffer zone.
Figure 10. Criteria coverage showing areas that fulfill the exclusionary criteria and/or the avoidance criteria.
The avoidance criteria and the exclusionary criteria map layers seen in Figure 10, are combined using the intersect tool, resulting in a map layer with areas that fulfill all the criteria combined. From this map layer 3 - 4 areas are selected for further investigation. The result from the combination of the two criteria layers is represented in Figure 15 found in section 5.1 in the result chapter.
4.2 Site Selection
In the site selection stage elevation, slope, flood risk and proximity to different features is taken in consideration. The site selection process is presented as a flow chart in Figure 11.
Figure 11. The site selection process presented as a flow chart.
The data used to perform the suitability analysis is:
● A layer containing all land cover. Different land covers entails different engineering costs. ● A layer is created with all the major roads that can affect the suitability of the site for safety
reasons, seen in Figure 12.
● A layer with railways, seen in Figure 12. Proximity to railroads favor transportation of material and waste material as well as long distance commuters.
● A layer with preserved antiquites, seen in Figure 12. There are several laws concerning how these sites should be treated and not interrupted.
● A layer with nature reserves, seen in Figure 12.
● A layer with the elevation map, used to calculate slope and flood risk.
Figure 12. The top left figure shows the significant buildings and antiquities, the figure on the right shows the extent of important roads and railroad and the bottom figure shows the vector polygons
used in the suitability analysis, nature reserves and populated area.
Flood risk assessment is applied using a mask. According to a rapport issued by Länsstyrelsen (2015) essential buildings should be located at least 2,7 m above the mean sea level around Mälaren and Sweden's coast to the Baltic sea expressed in the height system RH 2000. This guideline is also applied to the area around Vänern but instead of using mean sea level, the comparison is made using 2,7 m above the mean water level of the lake, approximately 46 meters above sea level. This mask is derived from the elevation map using the greater than tool in ArcMap and setting the target to 2,7 and 46 m respectively.
Map layers with the calculated euclidean distance from each respective feature is created using the
● Populated areas ● Roads ● Significant buildings ● Railroads ● Antiquities/Heritage grounds ● Nature reserves
For some features, the generated output map layer does not cover the whole study area. In that case the
cost distance tool is used instead with a raster layer with each cell assigned the value 1 as cost raster input.
From the elevation map a slope map layer is created using the slope tool to use in the weighting, seen to the right in Figure 13. Flat ground is to prefer when choosing site for an NPP.
Figure 13. Euclidean distance map to the left and slope map to the right.
To be able to compare the candidate areas all the attribute map layers are standardized by using linear transformation. Linear transformation is used because it is easy to apply and understand. To do the linear transformation the breaking points are needed which are the lowest and highest value in every specific layer, seen in Table 5. The lowest overall value for each layer is 0.
Table 5. Breaking points for the different attribute layers in the different areas. The highest value for the same layer is bolded.
Vanneberga Sjötorp Mariefred Forsmark Ringhals Oskarshamn
buildings Proximity to antiquites 5000 m 5000 m 5000 m 5000 m 5000 m 5000 m Proximity to nature reserves 11 810 m 11 342 m 16 359 m 10 713 m 11 439 m 12 191 m Slope 52,936
degrees 45,354 degrees 68,283 degrees 50,232 degrees 74,845 degrees 64,551 degrees
Proximity to railroad
17 503 m 15 533 m 21 177 m No Railroad
12 938 m No Railroad
The interval for each map layer for all areas are divided into 15 smaller intervals with equal length.These intervals are assigned a standardized score. The standardized score for proximity to roads, populated areas, proximity to significant buildings, proximity to antiquites and proximity to nature reserves are calculated using the following formula:
0 0
S = 5 − X−Xmin
Xmax−Xmin · 5
The standardized score for the slope and proximity to railroads are calculated using the following formula:
0
S = X−Xmin
Xmax−Xmin · 5
These calculations are made in Matlab and the result is rounded to the nearest integer. S is the standardized score for a given interval. X is the high end breaking point for the given interval. Xminis the lowest overall value for the given map layer. Xmaxis the highest overall value for the given map layer, the bolded values in Table 5. The scale factor 50 is used to differentiate values in the map layer close to each other when standardized. For example, if the highest value is 100 and the breaking points for two intervals, 1 & 2 are 10 and 14 respectively. Using a scale factor of 10 would result in both having a standardized score of 1. If a scale factor of 50 is used instead, interval 1 would have a standardized score of 5 and interval 2 a standardized score of 7.
Figure 14. Reclassified euclidean distance map to the left and reclassified slope map to the right.
The challenging part is to construct the comparison matrix, an example seen in Table 6. Depending on the decision maker’s knowledge in the subject and intention, the weighting can look different between decision makers. In this case, because of the lack of expertise, the ratio between tte different criteria is based on the perceived importance of every criteria in the literature study.
Table 6. Example of a comparison matrix.
The weighted value is calculated by first normalizing every element in the matrix by dividing them by the sum of each column. The result is then summarized for each row and divided by the number of criteria which in this case is 8. The result is a weighting percentage that says how influential the criteria should be in the suitability map. The sum of all the weighted layers should be equal to 1 and are combined using the weighted sum tool in ArcMap.
Three different weightings are used in the analysis, seen in Table 7. In weighting 1 all layers are equally influential. In weighting 2, proximity to populated areas and other buildings are taken more into consideration. In weighting 3 the potential impact on antiquities/cultural heritage sites and nature reserves are taken more in consideration.
Table 7. The three different weightings.
Roads Populated Areas Significant Buildings Antiquities Nature Reserves
5. Results
5.1 Candidate Areas
Figure 15 below shows the result when applying the intersection tool to the exclusionary criteria layer and the avoidance criteria layer. The red areas are candidate areas. Three candidate areas of interest, Sjötorp, Vanneberga and Mariefred have been selected as seen in the figure.
Figure 15. A closer look at the three candidate areas of interest, displayed in red.
Figure 16. A map showing the most important criteria when analyzing the existing NPP in Forsmark.
Figure 17. A map showing the most important criteria when analyzing the existing NPP in Oskarshamn.
5.2 Potential Sites
Figure 19 - 21 are the results of the suitability analysis in the site selection process. The scoring goes from 1-50 where a low score is more suitable than a high score. The color scheme is set to go from 15 - 35 because a majority of the scores are between those two values. The existing NPPs and potential sites are marked with purple.
6. Discussion and Conclusion
Of the already existing NPPs the results indicates that Forsmark is the most suitable location for an NPP. Looking at the weighting in figures 19, 20 and 21 Forsmark scored best for all the weighting. On a local level Oskarshamn scored decent for weighting 1 and 2. Ringhals scored decent for weighting 1 but scored bad for weighing 2 and 3. None of the sites are considered a candidate area in the site survey process. Ringhals was not considered a candidate area because there is one or more towns located within the allowed buffer zone today, which can be seen in Figure 18. The locations of both Oskarshamn and Forsmark would have been regarded as candidate areas if the criterion of proximity to transportation was loosened to 30 km instead of 15 km in the calculations which can be seen in Figure 16 and 17. The buffer zone to transportation is set to 15 km in order to eliminate redundant candidate areas because when set to 30 km, there were too many candidate areas to choose from. An alternative solution could have been to set the transportation buffer to 30 km and for example use a greater buffer zone around towns.
One objective of this thesis was to find a potential site for constructing a new NPP using a set of criteria. From an initial site survey process, three candidate areas were investigated further: Mariefred, Sjötorp and Vanneberga, named after nearby populated areas. Between these areas Mariefred is not considered suitable, Sjötorp the most suitable and Vanneberga suitable but possibly too small in size. According to energy.gov a typical nuclear facility in the US takes up an operational area of about 1 square mile.
When comparing potential sites from the results it is obvious that the weighting can have a large impact on the final results. For example comparing the Vanneberga area in Figure 19 where the area is mediocre to Figure 20 where the area is well above average. As mentioned earlier this criterion limits the candidate areas a lot which is good with regards to the time limit for this project. However, for a more thorough study, it could be beneficial to explore loosening this criterion. Neither Forsmark nor Oskarshamn are adjacent to any railroad and therefore the whole railroad raster layer is set to the highest score, 50, in the site selection process. Depending on the weighting for railroad, this could have a higher impact on the suitability of the area than anticipated.
With the criteria used in mind, in practice there are unlimited factors to take into consideration when choosing the potential sites so it would be impossible to completely avoid the subjective aspects of this project. One problem that occured was the lack of information about natural hazards. This criteria would have a large impact on the result because of its great importance to the safety according to many sources. Information that was searched for but not found were for example information about the subsidence risk and seismological activity related to geographical areas.
There are multiple aspects that make verification of the results for this project hard. One aspect is that the data used for this project was collected recently and many data values would be different if the data was collected in the 1970s. This means that one candidate area found today with recent data might not have been a candidate area when an NPP was established and vice versa. By using this logic it is impossible to tell if a candidate area found today would have been better suited for an NPP in 1970. On the other hand safety measures surrounding NPPs have not decreased since the 70s which means that a site chosen for an NPP might still very well classify as a candidate area since there would be no significant buildings in the vicinity etc.
Because of the time limitation of this project not all potential candidate areas could be examined further. The areas chosen were selected because they were all situated in the south part of Sweden, not close to any existing NPPs which means demand would be higher and they all seemed like interesting areas from a geographical perspective.
If we had more general guidelines with actual numbers on what is allowed/not allowed instead of a lot of the reviews being done on a case by case basis this study would have been less subjective.
If we had a super computer we would have been able to convert all relevant data covering Sweden to raster at the same time and use multi criteria analysis for the whole country at the same time.
If there were more time for the project, more time could have been used for trying different weightings and maybe even using different weighting methods. More time would also mean being able to
investigate more potential candidate areas.
Even though help was received from multiple experts in the field, having deeper knowledge about the subject and more weighting insight would probably lead to more accurate results. It is hard to know how accurate the model is but it could definitely be used and modified for further studies, either with the same goal in mind or with a different goal and modified criteria, such as siting small modular reactors (SMRs).
7. Further studies
In a similar way to how weighting and criteria can be changed, as seen under discussion, the model used for this project could be changed and fit for another purpose. The research objectives for this project was to evaluate existing NPP sites and also to find good candidates for hypothetical sites. However, a lot of research is being done in the nuclear sector and if there were to be a new NPP constructed in Sweden today or in the near future, the technical specifications of that NPP might differ from the ones active today which were built decades ago.
7.1 Fourth generation nuclear power
2016). It is hard to predict how a breakthrough in this area would affect legislation and siting criteria for an NPP, maybe a more efficient system would require more strict transportation criteria for example. For that reason it makes sense to rather consider the data and literature available for third generation nuclear for the purpose of this project. However this is a topic that could be studied further in the NPP siting process in the future.
7.2 Small modular reactors
Another subject at the forefront of nuclear research are small modular reactors (SMR). Small modular reactors are much smaller than the reactors active in Sweden today and would potentially use new technology that would make them more cost efficient and widen their use, for example using SMRs in the industrial sector (Värri, 2019). SMRs would have a smaller emergency evacuation area which means that the distance to population criteria could potentially be decreased. Since SMRs are not used commercially in Sweden they have not been taken into account for this project but could potentially be an object for further studies. A study with a proposed quantitative model for siting SMRs by
Devanand 2019 can be found in the sources.
7.3 Cooling towers
The NPPs active in Sweden today are not using cooling towers to remove absorbed heat through evaporation. Instead excess heat is returned to the original water source at a temperature
References
AELB. 2011. Guideline for site selection for nuclear power plant. Malaysia: Ministry of Science, Technology and Innovation.
Available at: http://www.aelb.gov.my/malay/dokumen/panduan/lem-tek/LEM-TEK-63.pdf
Björkman, Fredrik. 2019. M och KD lämnar energiöverenskommelsen. Dagens Industri. December 10th. Available at: https://www.di.se/nyheter/m-och-kd-lamnar-energioverenskommelsen/
(Retrieved: 2020-04-26)
Briney, Amanda. 2014. Overview of Weighted Site Selection and Suitability Analysis. GIS Lounge. April 10th. Available at:
https://www.gislounge.com/overview-weighted-site-selection-suitability-analysis/
(Retrieved: 2020-04-26)
Devanand, Aravind. 2019. Optimal site selection for modular nuclear power plants. Computer &
Chemical Engineering volume 125, p.339-350. June 9th. Available at:
https://www.sciencedirect.com/science/article/abs/pii/S0098135419300997
Energimyndigheten, 2020. 2019 rekordår för svensk elproduktion. Energimyndigheten. February 10th. Available at:
http://www.energimyndigheten.se/nyhetsarkiv/2020/2019-rekordar-for-svensk-elproduktion/
(Retrieved: 2020-02-10)
ESRI. 2020. History of GIS. Available at: https://www.esri.com/en-us/what-is-gis/history-of-gis (Retrieved: 2020-04-29)
Fischerström, Måns. 2018. Kärnkraften får ny sprängkraft i valspurten. Sveriges Natur. September 4th. Available at: http://www.sverigesnatur.org/aktuellt/karnkraften-far-ny-sprangkraft-i-valspurten/
(Retrieved: 2020-04-24)
Fortum. 2013. Hydropower plants in Finland and Sweden. Available at:
https://web.archive.org/web/20150322174048/http://www.fortum.com/en/energy-production/hydropo wer/hydro_riversystems/pages/default.aspx (Retrieved: 2020-04-28) (Archived 2015)
Gåhlin; Emil, Nilsson; Isabelle, Pettersson; Maria, Söderholm; Patrik. 2011. Tillståndsprövning för
kärnkraft i Sverige. Luleå tekniska universitet. Available at:
https://energiforskmedia.blob.core.windows.net/media/21099/tillstandsprovning-for-karnkraft-elforskr apport-11-06.pdf
IAEA. 2015. Site survey and site selection for nuclear installations. Vienna: International Atomic Energy Agency. Available at:
https://www-pub.iaea.org/MTCD/Publications/PDF/Pub1690Web-41934783.pdf
IAEA. 2019. Site evaluation for nuclear installations. Vienna: International Atomic Energy Agency. Available at: https://www-pub.iaea.org/MTCD/Publications/PDF/P1837_web.pdf
Idris, Rosniza. 2012. GIS Multi-Criteria For Power Plan Site Selection. Selangor: Centre of Studies Surveying Science & Geomatics, Universiti Teknologi MARA. Available at:
https://www.researchgate.net/publication/261457191_GIS_multi-criteria_for_power_plant_site_select ion
Jankowski, Piotr. 1995. Integrating geographical information systems and multiple criteria
decision-making methods. International Journal of Geographical Information Systems, 9:3, 251-273. Available at: Link
Lindholm, Kalle. 2019.Vattenkraftproduktion. Energiföretagen. May 23rd. Available at:
https://www.energiforetagen.se/energifakta/elsystemet/produktion/vattenkraft/vattenkraftsproduktion/
(Retrieved: 2020-04-27)
Lindholm, Kalle 2020. Vindkraft. Energiföretagen. April 1st. Available at:
https://www.energiforetagen.se/energifakta/elsystemet/produktion/vindkraft/
(Retrieved: 2020-04-26)
Lindström, Olle. 2019. Rekordfå är emot kärnkraft i Sverige. NyTeknik. November 22nd. Available at: https://www.nyteknik.se/energi/rekordfa-ar-emot-karnkraft-i-sverige-6979356
(Retrieved: 2020-03-25)
Lundberg, Fredrik. 2020. Så spräckte kärnkraften energiöverenskommelsen. Sveriges Natur. January 21st. Available at: http://www.sverigesnatur.org/aktuellt/striden-om-karnkraften/
(Retrieved: 2020-04-24)
Länsstyrelsen Stockholm. 2015. Rekommendationer för lägsta grundläggningsnivå längs
Östersjökusten i Stockholms län - med hänsyn till risken för översvämning. Available at:
https://www.lansstyrelsen.se/download/18.4771ab7716298ed82ba6f17a/1526068446288/Fakta%2020 15-2%20L%C3%A4gsta%20grundl%C3%A4ggningsniv%C3%A5%20ny%20bebyggelse%20vid%20 M%C3%A4laren.pdf
Löwenhielm,Gustaf. 1997. Utsläpp från svenska kärnkraftverk - ett miljöproblem? Analysgruppen
Bakgrund Nr 1. May. Available at:
https://analys.se/wp-content/uploads/2015/05/utslapp-fran-svenska-karnkraftverk-bakgrund1997-1.pdf
Macintosh, Andrew. 2007. Siting Nuclear Power Plants in Australia. Australia Institute. Available at: https://www.tai.org.au/sites/default/files/Nuclear%20siting%2040_8.pdf
Normark, Frida. 2014. Kriserna som fick Sveriges ledning att skaka. SVT. December 2nd. Available at: https://www.svt.se/nyheter/inrikes/kriserna-som-fick-sveriges-ledning-att-skaka (Retrieved: 2020-04-28)
Nylander, Isabella. 2019. Elbehovet kan öka med över 60 procent i Sverige. Dagens Industri. October 11th. Available at: https://www.di.se/nyheter/elbehovet-kan-oka-med-over-60-procent-i-sverige/
(Retrieved: 2020-04-24)
Ryan; Siobhan, Nimick Ellis. 2019. Multi-Criteria Decision Analysis and GIS. ArcGIS Story Maps. August 11th. Available at: https://storymaps.arcgis.com/stories/b60b7399f6944bca86d1be6616c178cf
(Retrieved: 2020-04-15)
Saaty. T.L. 2008. Decision making with the analytic hierarchy process. Int. J. Services Sciences, Vol. 1, No. 1, pp.83-98.
SLU. 2020. Sök digitala kartor och geodata. Available at:
https://www.slu.se/site/bibliotek/anvanda-biblioteket/soka/digitala-kartor/
(Retrieved: 2020-03-25)
SMHI. 2009. Årsmedelvattenföring Storsjön. August 4th. Available at:
https://www.smhi.se/kunskapsbanken/hydrologi/fakta-om-storsjon-1.5084 (Retrieved: 2020-04-21) SMHI. 2012a. Årsmedelvattenföring Mälaren. May 18th. Available at:
https://www.smhi.se/kunskapsbanken/hydrologi/fakta-om-malaren-1.5089 (Retrieved: 2020-04-21) SMHI. 2012b. Årsmedelvattenföring Vänern. November 6th. Available at:
https://www.smhi.se/kunskapsbanken/hydrologi/fakta-om-vanern-1.4732 (Retrieved: 2020-04-21) SMHI. 2016. Årsmedelvattenföring Siljan. February 8th. Available at:
https://www.smhi.se/kunskapsbanken/hydrologi/fakta-om-siljan-1.100225 (Retrieved: 2020-04-21) SMHI. 2020. Mätningar. Available at: http://vattenweb.smhi.se/station/ (Retrieved: 2020-05-01) UCS, Union of Concerned Scientists. 2010. How it Works: Water for Nuclear. Available at:
https://www.ucsusa.org/resources/water-nuclear (Retrieved: 2020-04-01) U.S Department of Energy. The ultimate fast facts guide to nuclear energy. Available at:
https://www.energy.gov/sites/prod/files/2019/01/f58/Ultimate%20Fast%20Facts%20Guide-PRINT.pd f (Retrieved: 2020-05-01)
Värri, Konsta; Seppälä, Petra. 2019. Small modular reactors, report 2019:625. Available at:
Westlén, Daniel. 2016. Fjärde generationens kärnkraft, rapport 2016:317. Available at:
