Photovoltaics, electric vehicles and energy users A case study of the Royal Seaport Visions and energy user expectations

P. Grahn, M. Hellgren, J. Munkhammar

Abstract This report investigates certain aspects of the projected residential area Royal Seaport in Stockholm. The study contains three distinct components: the inhabitants and their activities, photovoltaic arrays and electric vehicles. The goal is to increase the sociotechnical understanding of the new possibilities of emerging technological solutions in their interaction with humans accustomed to contemporary technologies. The study is based on two system layers. The inner layer is defined as the apartment as well as associated parking space and the end-users. The outer layer consists of the entire residential area consisting of Royal Seaport. The main focus in this study is the inner layer. The project utilizes two different approaches; modeling and simulation as well as interviews. The connecting point between the two approaches lies in the inhabitants. The modeling and simulation approach aims to give information regarding the implementation and matching of photovoltaic arrays and the impact of utilizing electric vehicles in the area. The interview method provides an opportunity to gain an understanding of perceptions about the area, scene of the area and the desirability of the opportunities the modeled system offers. This interdisciplinary approach offers a possibility for obtaining understanding regarding human inhabitants in connection with technological solutions. The possibilities for the intersection of technologicaland human challenges as well as opportunities offers a better understanding of the requirements for creating sustainable technological and behavioral changes. It is shown that matching electric vehicle load and photovoltaic power production is poor for a single household, but significantly better for an aggregate of households.

Sammanfattning I denna rapport studeras vissa aspekter av det projekterade bostadsområdet Royal Seaport i Stockholm. Studie innehåller tre olika komponenter: solceller, invånare, och elbilar. Målet är att öka den sociotekniska förståelsen för de möjligheter av nya tekniska lösningar i deras samspel med människors vana vid modern teknik. Systemet definieras i tvålager. Det yttre lagret definieras som det nya planerade bostadsområdet ”Norra Djurgårdsstaden”. Det inre lagret definieras som byggnaden samt tillhörande parkeringsplats och slutanvändare. Rapportens huvudfokus är det inre lagret. I detta projektet används tvåolika metoder, modellering och intervjuer. Knutpunkten mellan de tvåmetoderna ligger i invånarna. Tillvägagångssättet via modellering och simulering syftar till att ge information om genomförande och matchning av solceller samt påverkan av en framtida elbilsanvändning i området. Undersökningen via intervjuer ger en möjlighet att fåen förståelse för uppfattningar om området och önskvärdheten av möjligheter det modellerade systemet erbjuder. Med detta tvärvetenskapliga angreppssätt ges en möjlighet att utvärdera mötet mellan mänskliga invånare och nya tekniska lösningar. Skärningspunkten mellan både tekniska och mänskliga utmaningar erbjuder en bättre förståelse för möjligheter att skapa hållbara tekniska och beteendemässiga förändringar. Det visar sig att matchning för laddningen av elektriska fordon och solceller är svag for enskilda hushåll men signifikant bättre för aggregat.

Part 1. Preface 1. Energy Systems Program 2. Buildings in Energy Systems

i i i

Part 2. Aim 3. Project team 4. Research questions in Royal Seaport 4.1. End-users 4.2. Photovoltaics 4.3. Electric vehicles 5. Challenges 5.1. Listing the challenges 5.2. Main focuses

1 1 1 1 1 2 2 2 4

Part 3. Methodology 6. Models, scenarios and interviews

5 5

Part 4. Background 7. Royal Seaport 7.1. History 7.2. The Royal Seaport project 7.3. Environmental goals 7.4. Transportation 7.5. Energy production 7.6. Purification of soil 8. Distributed generation and the power system 8.1. Basic properties of photovoltaics 8.2. Photovoltaic system output 8.3. Development of photovoltaics in the world 8.4. Load matching photovoltaics 8.5. Net-metering 8.6. Discussion on the future of distributed photovoltaics 8.7. Photovoltaics at the Royal Seaport 9. Electric vehicles 9.1. The vehicle 9.2. Electric vehicles on the market 9.3. Electric vehicles and climate targets 9.4. The electric motor 9.5. The battery 9.6. The charging 9.7. The end-user 9.8. The electric power system

7 7 7 9 9 10 10 10 11 12 13 14 17 17 18 19 19 19 20 21 22 22 23 24 25

Part 5. Interviews 10. Interviews 10.1. Ethical considerations

27 27 28

10.2. Reliability and validity 10.3. Interview guide

28 29

Part 6. Mathematical modeling 11. Modeling energy consumption from activities 11.1. Time use survey 11.2. Time use data method 11.3. Time use data structure 11.4. Time use data analysis 11.5. Time use data used in this study 12. Deterministic mathematical model of household power consumption 12.1. Household power consumption from activities 13. Discrete time Markov chain model approach to synthetic activity generation 13.1. Estimation of transition probabilities 13.2. W&W-model 14. Photovoltaic power production 14.1. Potential photovoltaic setup at Royal Seaport 14.2. SMHI data for photovoltaic power production 15. Electric vehicle power consumption 15.1. SmartLoad model of electric vehicle charging 15.2. SmartLoad equations 15.3. Electric vehicle assumptions 16. Stochastic model 17. Load matching

31 31 31 31 32 32 32

Part 7. Scenarios 18. Electric vehicle scenarios 18.1. Charging power 18.2. Driving frequency 18.3. Battery depth of discharge 18.4. Remaining simulation data 18.5. Electric vehicle scenarios

47 47 47 47 47 48 49

Part 8. Results and discussion 19. Analysis of interviews 19.1. The interviewees 19.2. Contextualization 19.3. The Royal Seaport area 19.4. Enviromental goals 19.5. Inhabitants 19.6. Transportation 20. Results regarding household load neglecting electric vehicle load

51 51 51 52 53 55 58 62

33 34 35 37 37 38 38 39 40 40 40 41 44 45

63

21. Electric vehicle consumption 21.1. Model behavior 21.2. Scenario simulation 21.3. Area load curve simulation 22. Household energy consumption and photovoltaic production 23. Electric vehicle load and photovoltaic production 23.1. Aggregate photovoltaic production and electric vehicle load Part 9. Conclusion and future studies 24. End-user related 24.1. The inhabitants 24.2. End-users and technology 24.3. Solar power 24.4. Electrical Vehicles 24.5. Final thoughts 25. Photovoltaics and load matching 26. Electric vehicles 27. Main conclusion 28. Future research Appendix A. Interview guide References Contents

64 64 70 77 81 85 86 89 89 89 91 91 92 92 93 94 95 96 97 101

Photovoltaics, electric vehicles and energy users

i

Part 1. Preface 1. Energy Systems Program The Energy Systems Program is a national research program and research school financed by the Swedish Energy Agency. Further parties involved are Linköping University, industry, energy companies and municipalities. The research in Energy Systems Program is interdisciplinary with a systems approach. Within the program, technological sciences is combined with social sciences to attain a broad expertise and an ability to manage the challenges concerning energy systems and the surroundings in which the technology is contained in. The main purpose with this approach is thus to look upon energy systems as sociotechnical systems, and analyse the systems with both technological and economic methods and tools as well as with social sciences. The research program consists of a collaboration between four Swedish universities; Uppsala University, Royal School of Technology, Chalmers University and Linköping University. The fundamental objective with the program is to develop diverse knowledge that favors the development of sustainable and efficient energy systems. This knowledge should enable long-term development of sustainable and resourceefficient energy systems to be used to guide changes in harmony with societal goals and democratic influence. The term ‘Energy systems’ in the program is defined as: Energy systems consist of technical artifacts and processes as well as actors, organizations and institutions which are linked together in the conversion, transmission, management and utilization of energy. The view of energy as a sociotechnical system implies that also knowledge, practices and values need to be taken into account to understand the on-going operations and processes of change in such systems [37]. 2. Buildings in Energy Systems The research within the program is organized into three consortia which each are dealing with energy systems with different focus. Each consortium consists of researchers and graduate students from at least two different departments to encourage interdisciplinary exchange. The consortia leader for the consortia of Buildings in Energy Systems, which this project is part of, is Professor Ewa Wäckelgård from Uppsala University. This consortia has a focus that includes a view upon buildings as an interconnected system in which technical components and social

ii

P. Grahn, M. Hellgren, J. Munkhammar

actors interact. The building as a technical subsystem is thus to be seen in relation to how different actors project and manage it, including the end-user. Buildings are part of a larger energy system why the study of energy concerning buildings should occur at a sufficiently high system level to prevent sub-optimization [37].

Photovoltaics, electric vehicles and energy users

1

Part 2. Aim 3. Project team This project was conducted as a collaboration consisting of PhD student Mattias Hellgren at Linköpings University with supervisor Professor Kajsa Ellegård, PhD student Joakim Munkhammar at Uppsala University with supervisor Professor Ewa Wäckelgård and co-supervisor Assistant Professor Joakim Widén, and PhD student Pia Grahn at the Royal Institute of Technology with supervisor Professor Lennart Söder. The project was part of an interdisciplinary course in the research school Program Energy Systems of 2011. 4. Research questions in Royal Seaport Royal Seaport is an urban area under development in Stockholm, Sweden. In this report the Royal Sea Port project and aspects of the environmental profile and related targets are investigated. The district is intended to be an area built with green technologies. The Mayor of Stockholm, Sten Nordin, was in 2010 quoted to said that Stockholm Royal Seaport will be a spearhead for sustainable urban development, where innovative Swedish clean technologies and creative solutions are developed, tested and presented.[66] 4.1. End-users. The people who are going to live in the Royal Seaport area are ultimately the ones who are going to use the technologies implemented and in the end the ones whose patterns of travel and daily activities is going to effect the substantiality of the area. How the area is zoned, planned and built will also directly have effects. The attitudes, perceptions and views of sustainability and the stacks that the various actors hold within the project is to be explored and one component of this report is to investigate how the area is zoned, planned and constructed and the perspectives those involved in these activities perceived and and view the end-users and what considerations they have take into account while carrying out their roles within their organisations. 4.2. Photovoltaics. Coupled with this is the suggestions for the construction is to implement small scale photovoltaic power generation that will be owned and utilized by the inhabitants. The basic idea is that the inhabitants will be able to use the photovoltaic power to sell on the market, for their own electrical consumption as well as a power source for perhaps their own electric car. Developing such systems give

2

P. Grahn, M. Hellgren, J. Munkhammar

rise to technical challenges that will not only require technical solutions. With the technological aspects of systems such as these are also that the system have a user, the inhabitants; users that will need to be ‘configured’.[65] This report investigates the possibilities and challenges for implementing distributed photovoltaics at the Royal Seaport and estimates reasonable typical sizes of photovoltaic arrays that can be used. One particular feature is how the photovoltaic power generation at the end-user can be matched to the power demand locally. One of the interesting investigations in this study regards the load matching between electric vehicle load and photovoltaic electricity production. 4.3. Electric vehicles. The study also investigates what plans and visions there are for electric vehicles related to the environmental targets set up for the Royal Seaport area. This includes how electric vehicle charging is projected and how it will be made optional for the residents moving to the area. With an electric car fleet the amount of flexible loads will become increased which creates a need to study the behavioral of electric vehicle driving pattern and thus the charging frequency to be able to estimate the power system impact. In this project an aim is therefore to find out how electric vehicle use can be simulated based on how electric vehicle drivers will behave and to analyze the impact of such a system in the Royal Seaport area by having an electric car fleet and estimate possible load curves. Both social aspects such as the impact of people’s behavior as well as technical aspects are considered to find out the electric vehicle drivers behavior.

5. Challenges 5.1. Listing the challenges. The main focus in this report will in general be how to implement micro production such as photovoltaics together with electric vehicles to end-users in the Royal Seaport area. In essence the user of a technological systems needs to be adapted to the system in order to use it. A potential clash here is between on one side the intent and the design of the system and the other the users and their expectations, habits and perspective on the system. One central question to explore is how the developers see that the users with their own understandings, expectations and experience of other technological systems will be ’configured’ to work with the system. It is expected that the designers of the system will have expectations or beliefs about the users and how they will interact with the systems. These reasoning furthermore leads to a wide number of reflections and challenges to adopt:

Photovoltaics, electric vehicles and energy users

3

5.1.1. End-user related. • What expectations, in terms of socioeconomic status and background, technical competence and environmental concerns, do the planners and constructors of the technical systems in Royal Seaport have on the inhabitants? • Given that the inhabitants of the area will be presented with visualization of their energy consumption, how do the designers of the systems picture a regular inhabitants use of the system to lower their energy consumption? • What presumptions are made by the designers about what motivates the inhabitants to be more energy efficient? • What other factors besides the end-users do the organisations take into account? 5.1.2. Photovoltaic related. • What is a reasonable photovoltaic setup for each apartment? • What is the photovoltaic power output and variability over time? • What is the level of match between the electricity consumption by the local end-user and the photovoltaic electricity production? • What is the load matching locally between the electric vehicle load and the photovoltaic electricity production? • What is the load matching on an aggregate basis between the electric vehicle load and photovoltaic electricity production? 5.1.3. Electric vehicle related. • What plans or incentives are there regarding an electric vehicle fleet introduction in the Royal Seaport area? • What charging options will become reality and how will the parking spaces related to charging look like? • How can we simulate electric vehicle use based on how electric vehicle drivers will behave? • What load curve would an electric fleet in the area induce? • Is there a possibility to implement load matching towards photovoltaic production by using electric vehicles in the area?

4

P. Grahn, M. Hellgren, J. Munkhammar

5.2. Main focuses. To sum up, the project will try to answer the above mentioned questions which lies within three components. The first is of the inhabitants and how their role is perceived by various actors. How the new, and in certain areas old, inhabitants are expected to fit in the idea underlying the green Royal Seaport. To explore how the environmental and energy technologies to be impleaded in the area will offer a wider picture of the thoughts concerning the area and thus how electrical vehicles and photovoltaics fit into a bigger picture. Contextualization of the two technologies within the Royal Seaport environment is thus one of the aims and how the wider reasonings affect future and presents implementations of the two technologies discussed in this report. The second and third component is thus how the various actors reason regarding the inclusion of electrical vehicles and photovoltaics as well as how these two technologies are perceived by the various actors. The inner layer representing this investigation is illustrated in 11. Nevertheless, the overall focus could be summarized in the three main areas: • How the end-users are perceived and viewed by various stake holders. • The photovoltaic potential in the Royal Seaport area. • An introduction of electric vehicles in the area.

Figure 1. Illustration of the inner layer

Photovoltaics, electric vehicles and energy users

5

Part 3. Methodology 6. Models, scenarios and interviews An area such as Royal Seaport is obviously not built without intention or consideration. Literally thousand of people have been involved in the various stages and processes that have been and will be underway in the construction of the area. This section will take a closer look at some of the people involved in turning the Royal Seaport area from an idea to a reality. This interdisciplinary project applies three different methods; one utilizing mathematical modeling, a second constructing scenarios for simulation and a third utilizing interviews and analysis. The connecting point between the two approaches lies in the inhabitants. The modeling approach aims to give information regarding the implementation and matching of photovoltaic arrays and the feasibility of utilizing electric cars and their behavior. The scenario construction aims to visualize various outcomes of residents different driving habits. The interview approach offers the possibility of gaining an understanding of the perceptions, understandings, motivations and the desirability of the possibilities the modeled system offers. The intersecting connection point offers the possibility to evaluate the meeting of human inhabitants and technological solutions. The intersection of technological and human challenges and possibilities offers a better understanding of requirements for creating sustainable technological and behavioral changes. Starting with section 10 the methodology for the interviews is described. Further the mathematical model, implemented and developed in this work, is defined including the added electric vehicle and photovoltaic model and the load matching model, starting with section 12. The model is based on time use activity surveys, described in section 11.1 and at last the electric vehicle use Scenarios are constructed in Section 18. Technological implementations have dependencies on their environments and how they are perceived to fit into the constructed area. In this part the focus lies on those who participate in planning, zoning and implementing of the layout off, the technology into and the considerations for the new residential, service and commercial areas. What is studied is the overall process and the connected reasonings towards this rather than the specific actions committed in the construction. The reasoning and experiences are more focal than the actuality decisions that has, and in the future will be made. This report will however start by giving a background to the Royal Seaport area and its history. Followed by this, the electric power system,

6

P. Grahn, M. Hellgren, J. Munkhammar

photovoltaics and electric vehicles will be given a background which is needed to form a foundation on which to make interviews, models and scenarios.

Photovoltaics, electric vehicles and energy users

7

Part 4. Background 7. Royal Seaport

Figure 2. Hjorthagen today and tomorrow [53] Royal seaport (‘Norra Djurgårdsstaden’ in Swedish) is an urban district currently under construction in the centre parts Stockholm, the capitol of Sweden. The Royal Seaport is projected to accommodate 10 000 residences and 30 000 office spaces as well as 600 000 sqm of commercial areas.[48] The area is a former brownfield industrial area of 236 hectares. The area is the primary host of Stockholms ferry and cruise operations and much of the traffic is concentrated in the area. The Royal Seaport consists of four areas, Hjorthagen, Värtahamnen, Frihamnen and Loudden (see figure 4). As of 2011 The Hjorthagen has started building its first stage and the commercial city block Riga in southern Värtahamnen. The second stage in Hjorthagen is planned to start in 2012, followed by the Värtahamnen area in 2013, the Frihamnen area in 2017 and finally the Loudden area in 2020. The whole area is projected to be completed by 2025 [47]. 7.1. History. The area where Royal Seaport is located is a part of the government owned ‘Kungliga Djurgården’ (Royal Djurgården) which has been owned by the Swedish regent since the 15th century. In the 17th century The Swedish king Karl XI raised a 20 kilometre long fence around the area for keeping deer and roe deer for hunting. Later the deer was moved to the area of Hjorthagen in 1803 [63] giving the area its name1. At the end of the 19th century the deers where moved further outside of the area. At the end of the 19th ownership of the area that will become Royal Seaport was transferred to the City of Stockholm. The areas has been a focal point for the harbour operations in Stockholm since the late 19th and early 20th century. 1Hjorthagen

can loosely be translated as ‘deer enclosure’

8

P. Grahn, M. Hellgren, J. Munkhammar

Figure 3. Royal Seaport location in centre Stockholm [48] At the end of the 19th century the area developed into on of the Stockholm areas largest industrial and harbour areas. The harbour areas Värtahamnen was built in 1884, Frihamnen in 1919, and construction of Loudden oil harbour started in 1927. The industrial development in the area was for distribution of gas and production of electricity. Followed later in the Loudden area by oil trade and storage. The area has housed town gas producer and distributor Värtagasverket since 1853, which was closed down in January 2011. The close down was due to the use of town gas has been declining in use over the years and the fossil based gas is increasing being replacement by biogas. The work on the area started already in 1972. It was however interrupted by the 1973 oil crisis. The oil crisis made ceasing the operations of the gas power plant problematic since it was the city distributor of city gas. A new attempt was done during the early 80s, this time another crisis foiled the attempt. A major architectural competition were held in 1988 but didn’t fall through due to contracts disputes with the gas production company. The contract was finalised early 2000. In 2001 a program for the city development was made and the work was finalized in 2003.

Photovoltaics, electric vehicles and energy users

9

7.2. The Royal Seaport project. The zoning for Royal Seaport2 was approved by the city council in November 2008. The decision was appealed to the county administrative board who dismissed the appeal. The ruling was appealed to the Ministry of the Environment. In Maj 2011 the appeals were dismissed by the Ministry and the zoning had gained legal force. The zoning was later realized by the Development Committee in Mars 2011. During the summer and fall of 2011 the construction of the area started [49]. The population of the area is to be substantially expanded. The Hjorthagen area was in 2010 home to approximately 2200 inhabitants but will in the Royal Seaport project gain 15 000 new residents.[58]

Figure 4. The areas of Royal Seaport[48]

7.3. Environmental goals. The city of Stockholm have as an expressed goals with Royal Seaport to create an environmentally sustainable city area. The area is to be prepared for climate changes, particulary increases in precipitation. Concerning CO2 the goal is that annual CO2 emissions should be below a level of 1.5 tonnes per inhabitant by 2020 (in 2008 the average emissions per capita in Sweden was 5.3 [57]) and the area is to be free of fossil fuel by 2030. The Royal Seaport project have high goals for energy efficiency and the projected energy consumption in the area is set to 55 kWh sqm/year. As Royal seaport is a harbour area an expressed goal is to make the passenger ferry and cruise ship harbour area among the most environmentally sustainable ports in the world. 2The

name Swedish name, ‘Norra Djurgårdsstaden’, was first used in 2006 to designate a certain area of the development zone but was in 2010 expanded to include the whole area.

10

P. Grahn, M. Hellgren, J. Munkhammar

7.4. Transportation. The Royal Seaport area contains a substantial part of the harbour traffic in the Stockholm area. The area houses the main ports for ferry and cruise travels in the Stockholm area which generate traffic through the area. The harbours in Loudden and Norra Värtahamnen are important for the supply of oilproducts and coal for the Stockholm area. The container harbour in the Frihamnen area is currently one of the more central on the Swedish east coast. The area is projected will retain its ferry and cruise traffic but move freight transport to new areas outside the city. The non-personell harbours are projected have the majority of their operations to be moved to areas outside the main city.[50] A motorway project, Norra Länken (‘Northern Link’), is planned and will connect the area to the greater Stockholm. In addition the area will require new transportation for the planned commercial, habitual and service areas. For public transportation the area is planned to prioritize public transportation and lanes for bicycles and walking [source: interview]. The currently (fall 2011) political debates for the area to have electric trams serving the area as well as debates regarding expansions of the subway. Among other substantiable transports that has been discussed are biogas powered buses as well as expansions of the boat traffic in serving the area. There is currently a pilot project connecting the Hjorthagen area with the centre parts of Stockholm. 7.5. Energy production. The Royal Seaport area houses Värtaverket which is important for the generation of electricity, district heating and chilling for the Stockholm area. There has been multiple calls to close down the power plant over the years due to its envorimental impact. This impact is high due to its reliance on fossil fuels, mainly coal and oil. According to Swedish Society for Nature Conservation (Naturskyddsföreningen) Värtaverket produced 610 thousand tons of fossil CO2 in 2008 making it one of the most CO2 contributing power plants in Sweden. [31] According to the owners, Fortum, there are plans to increase the percentage of bio fuels to lower the amount of coal currently used to 50 percent by 2015 lower the environmental impact. [9] 7.6. Purification of soil. The area is a former brown field area and thus the soil is contaminated. Five hundred million Swedish Krona is planned to be invested to the process of soil purification. This process is one of the larger investments into the area. In the Hjorthagen area the ground surrounding the gas bells, which is at the present a park, is a former quarry. The quarry has over the years been filled with lime mud, limonite and other waste.[49]

Photovoltaics, electric vehicles and energy users

11

The Loudden area has been housing oil production and thus the area is expected to have soil contaminations. Oil contaminations however are not static but move and thus purification can not be started until the area have been cleared of oil operations. At the present the current land users are responsible to bring the ground to less sensitive levels. Since housing is planned for the area the ground needs to be purified further to gain levels for housing [56, p. 30]. 8. Distributed generation and the power system Power systems are the most complex systems ever created and operated by humans [44, p.139]. Thus the addition of new electrical power sources in such complex systems requires considerable amount of theoretical investigation prior implementation. Consumer demand requires stability of the power system; for example that the active and reactive power is kept at a constant frequency and constant voltage. One should keep in mind that the various forms of consumer load on the electric power system are intermittently switched on and off, thus complicating the situation for the producer. This calls for power system control [44, p.139]. Photovoltaics connected to the power system is a special case of the more general concept of distributed generation. Distributed generation in practice is the generation of electric power in the power system from various small energy sources often situated at the end-user in the grid [16, p.25]. Power sources that are distributed are not by definition renewable. Even a connected diesel engine would be considered distributed. Instead distributed generation is perhaps most canonically defined as the opposite of centralized distribution. The foremost current reason for installing distributed photovoltaic is not economic but rather mainly environmental as it is regarded as a renewable energy source [16, p.26]. The power output from large-scale power generation is at 400 kV which is distributed to a major transformer station where it is transformed to 130 kV which in turn is distributed to a regional transformer which transforms it to 40-70 kV [52]. The power is then transmitted to a distribution station which transforms the power to 10 kV and the eventually a substation transforms it to 400 V which is distributed at the household level [52]. Distributed generation may occur at any level of the grid, but we focus on the lower levels. In the lower levels we identify three particular levels as high-, mid- or low voltage. These levels may be characterized as follows [52]. - High volage ∼ 50 kV - Medium voltage ∼ 10 kV

12

P. Grahn, M. Hellgren, J. Munkhammar

- Low voltage ∼ 0.4 kV One may also identify voltage levels on the order of 130 kV and 40-70 kV in the grid, which constitute regional levels [52]. The lack of central top-down control in distributed generation might cause problems if the output from distributed generation becomes significant in the power system. Problems regarding issues such as unintentional islanding, grid stability and power quality might occur [20, p.1]. Distributed generation of certain energy sources may also be very intermittent which creates the need for more complex load matching strategies. Load matching is by definition the matching of demand (load) and supply (production). It is perhaps possible that with demand side management (DSM) the consumer created energy load might change to match the production of energy in the grid, rather than the other way around [61]. Demand side management is a concept that aims to - with the use of incentives - change the consumer side of energy usage. This field of study could perhaps prove efficient for shifting various forms of consumer activities such as washing, drying and dishwashing to times of lower load [16, p.76]. In total there are a number of issues arising when using distributed generation, Widén concludes the following list [60, p.20]: - Altered quality of electric power - Reactive power compensation - Altered reliability - Altered security - Lack of responsibility - Altered local balance between production and load In general the sources for distributed generation may or may not be renewable and they may or may not be intermittent which poses great problems for the design of a power system that can handle distributed generation on a large scale. The problems arising in distributed generation needs to be resolved in order to achieve an effective and reliable power system which is based, at least partially, on distributed generation. 8.1. Basic properties of photovoltaics. Most energy sources we utilize today have their origin in solar radiation. This fact is not related to the current usage of photovoltaics, but rather that fossil fuels are accumulated energy from the sun to some extent. The direct conversion of solar energy into electricity can be done in two different ways. The common consensus when speaking of solar energy is that it regards

13

Photovoltaics, electric vehicles and energy users

direct conversion from solar radiation into heat and/or electricity [44, p.61]. It is in this category photovoltaic technology - or simply photovoltaics - is a special case as it transforms the energy of solar photons into direct current using semiconductor materials [44, p.62]. The smallest cell of a photovoltaic system is called a PV cell or simply a solar cell. Its function is to absorb the incoming photons which then frees electrons from the semiconductor material [44, p.62]. This in turn generates a DC-current. The most common semiconductor materials used in PV cells are the following [44, p.62]: single-cell crystal silicon, amorphous silicon, polycrystalline silicon, cadmium telluride, copper indium diselenide and gallium arsenide. PV cells are connected together and sealed to make a panel. Depending on size and efficiency these panels range from a few watts up to around 100 W. Unfortunately current solar panels only convert a fraction of the sunlight into electricity [44]. Thus active research in photovoltaics is valuable for increased future efficiency of solar panels. Solar panels can collectively be arranged to assemble what is called a photovoltaic array. H

H

Figure 5. A schematic picture of a photovoltaic cell. 8.2. Photovoltaic system output. A photovoltaic panel may be setup in any given orientation, ie with any azimuth and tilt angles. It is also possible to have a panel that follows the sun to some extent. The photovoltaic output is strongly variable even for a single fixed photovoltaic array. The radiation that reaches a PV cell is much higher than the electric output due to a number of losses. First of all the angle of incidence of the different radiation components is related to the amount of energy that is reflected away [15][p.39]. Second of all regarding the amount of radiation that reaches the PV cell, it is only certain wavelengths that contributes to the light-generated current.

14

P. Grahn, M. Hellgren, J. Munkhammar

Figure 6. A typical photovoltaic array (Photo by Joakim Munkhammar 2011) There are also additional losses related to for example recombination in the semiconductor material and high temperature. Ultimately it is the surface area for mounting PV cells that limits the total power production. Since the photovoltaic panel delivers direct current (DC) it is necessary to convert it into alternating current (AC). In order to convert photovoltaic DC (or any DC) current to AC one needs to have a so-called inverter and to connect the alternating current to the three phase grid. A possible setup, which was assumed by for example Widén in [15], is to directly connect the photovoltaic array to the grid with the assumption of constant loss. 39$UUD\

,QYHUWHU

*ULG

Figure 7. A schematic figure of a possible setup for connecting a photovoltaic array to an AC electric grid. (Figure made by Joakim Munkhammar 2011) 8.3. Development of photovoltaics in the world. An effective and reliable power system is a necessity for global prosperity and for providing a decent life for any citizen [22, p.9]. Thus the production of electricity is paramount, and if many options are available, a preference relation among the options is desired. In face of abundance limitations on fossil fuels combined with possible environmental problems related to the combustion of them a considerable amount of attention has been generated for the field of renewable energy sources [39]. The definition

Photovoltaics, electric vehicles and energy users

15

Figure 8. An example photovoltaic output over a day from the minute-based resolution of the meteorological database and simulation software Meteonorm 6.0 for the location Stockholm Sweden, [15, p.44] of renewable energy sources is the following according to encyclopedia Britannica [6]: Renewable energy is usable energy derived from replenishable sources such as the sun (solar energy), wind (wind power), rivers (hydroelectric power), hot springs (geothermal energy), tides (tidal power), and biomass (biofuels) The problem is that today most electricity production in the world comes from non-renewable energy sources [39, 15]. The distribution of energy production among the different sources of energy in the world was in 2008 as follows from figure 9. This prompts investigation into a possible increased level of implementation of renewable energy sources in the power system. One of the strongly expanding and promising renewable energy sources is photovoltaics (PV) [15]. The increased implementation of photovoltaics during the last few decades in certain countries is shown in figure 10. Most of the installed PV power is in the in the International Energy Agency Photovoltaic Power System Programme (IEA-PVPS) countries, which in particular includes the biggest producers Germany and Japan [15, p.25]. A majority of these systems are grid-connected and distributed while being connected at the end-users located at the very end of the distribution grid [15, p.7]. Thus the usage of photovoltaics has further pushed for the so-called distributed generation view of the power system. From that perspective the classic difference between producer and consume is less clear. It has even been suggested that distributed generation constitutes a power systems paradigm shift that prompts rethinking of supply, demand and efficiency [15, p.25-27]. The application of photovoltaics in the power system also has a wide range of inherent problems regarding for example load matching and

16

P. Grahn, M. Hellgren, J. Munkhammar

Figure 9. The energy production in the world 2008 [39, p.16] transformer optimization. Regardless of the current problems of photovoltaics the prospects are good on a global scale. The total solar radiation that hits the earth over a few days is equivalent to the total energy consumption throughout the entire human history [51]. Together with sharp decline in prices for photovoltaic cells and the increasing trend towards generous feed-in tariffs suggests great prospects for an increased implementation of distributed photovoltaics on a global scale.

Figure 10. Total PV peak power installed in the IEAPVPS countries between 1992 and 2007 [15, p.25]

Photovoltaics, electric vehicles and energy users

17

8.4. Load matching photovoltaics. A major challenge in constructing a well functioning power system is to match the load with the production of energy [16, p.1]. In connection with photovoltaics and net zero energy buildings this has been studied by Widén and Wäckelgård [18]. In particular the impacts of various options for obtaining a lower mismatch between production and consumption was investigated by Widén et al. [60]. That study focused on the following three possible options: PV array orientation, demand side management and electrical storage. The PV array orientation is merely the azimuth and tilt angles that a photovoltaic array are aligned [60, p.10]. The optimization of this is in practice the optimization of the solar fraction [15, p.60]. Demand side management aims directly at shifting the demand/load usually by raising load on off-peak periods and lowering it on on-peak periods [60, p.13]. This can technically be achieved in a number of ways by various algorithms. In general demand side management aims at changing the electric power consumption pattern of consumers to better fit the production of electricity by means of for example economic incitements. Such economic incitements might perhaps more effectively be utilized with an hourly price on the electricity consumption [19]. As we have previously stated the consumer side of the problem is not always as flexible as desired and perhaps only a few appliances are reasonably shiftable [60, p.13]. Electrical storages such as for example batteries were concluded by Widén andWäckelgård [60] to be the most effective option for load matching, at least for a larger penetration level. There are a number of possible types of electrical energy storages that can be used. The most obvious example is a battery, but there also exists for example fly-wheel based energy storages [26, p.251] (This reference even includes a great illustration of a fly-wheel connected with photovoltaics). Also electrical vehicles might be used for temporary electrical energy storage for photovoltaics [36]. In general the matching of photovoltaic output on the power system might come from other sources of renewable energy such as wind power, hydro power and perhaps even wave power. In this project the aim is to study the matching locally between the photovoltaic production and load, where the load is from both household activities and electric vehicle load from charging. 8.5. Net-metering. There is a particular problem regarding the connection between the electricity market and distributed generation of photovoltaics, namely how to capitalize on ones own production of electric power in the grid [1, 19]. One solution to this problem is to introduce so-called net-meters for consumers that are both consumers and producers [20, p.477]. A net meter is designed to record the net consumption of electric power for a consumer [19]. If a consumer in

18

P. Grahn, M. Hellgren, J. Munkhammar

addition to consuming electricity also produces it to the grid, via for example photovoltaics, then this amount will be subtracted from the meter status [20]. The use of net meters is an economic incitement for distributed photovoltaics but without subsidies it would still be several times more expensive than the traditional utility electricity [20, p.477]. For this reason several so-called plus-metering systems have been implemented where the photovoltaic electricity producer receives a higher payoff for the produced electricity compared with the general market price. A particular example of a country that has adopted this policy is Germany. In Sweden recently a number of major power companies have feed-in tariffs just below net-metering standard. Although Sweden has no official strategy for financing PV installations most installations made have been subsidized by the ’local investment programme’ LIP [46]. The subsidy is on the order of 3–4 Euro per installed Watt.

8.6. Discussion on the future of distributed photovoltaics. Limitations on non-renewable energy sources combined with an ever growing global demand for electric power suggests that electric power production from renewable energy sources will be more prominent in the future. Photovoltaics, by being the most direct utilizer of the energy from sunlight, is a possible candidate for significant future electricity production. Already today photovoltaics has seen a massive expansion internationally [15, p.7]. Unfortunately a considerable amount of this expansion is most likely due to governmental subsidies of photovoltaic systems [15]. The economic problems of photovoltaics are thus in need of viable and sustainable solutions on the market. Possibly this is done in some form via distributed generation where the end-consumer ultimately ends up not only consuming electric energy but also producing it, thus reversing age-old conceptions about the roles of consumer and producers. Generally distributed generation is in many ways a paradigm shift from a centrally controlled electric energy production to a distributed one. How to setup photovoltaics in the power system via distributed generation is then naturally in need of considerable investigation. It has been concluded that a number of unresolved issues needs to be investigated and taken into consideration before implementing it. For example problems involving load matching and power system control of distributed photovoltaics needs to be considered. Also economic factors such as costs for implementation and possible net metering systems also needs to be taken into consideration. Indeed, as was previously stated, perhaps the primary challenge involved in the large-scale implementation of distributed photovoltaics in the power system is in fact the

Photovoltaics, electric vehicles and energy users

19

cost efficiency of the setup. Today without subsidies distributed photovoltaics is not nearly economically feasible [20, p.477]. Perhaps future photovoltaic systems will be cheaper due to technological advances. Other possibilities include net- or plus-metering with or without subsidies in order to achieve widespread photovoltaics. Other mechanisms for improving photovoltaic input in the power system might be to utilize demand side management in which the behavior of the customer is steered with various incitements such as economic incitements for example. Load matching is only one of the many challenges that needs to be undertaken in the distributed generation revolution ahead. The future Swedish power system is difficult to properly envision, but if we follow the current trend renewable energy sources will become more prominent over time. The trend also suggests that photovoltaics would become increasingly prominent - in particular in the form of distributed photovoltaics. Given a possible future with cheaper or more efficient photovoltaic cells the need for economic incitements for a large scale implementation of photovoltaics would decrease. Given the current conditions and future prospects one of the main challenges today are indeed to discover proper mechanisms which would assist a widespread implementation of photovoltaics in the future. 8.7. Photovoltaics at the Royal Seaport. The Royal seaport intends to have photovoltaic panels on the roofs that are grid-connected. One of the aims of this report is to model different sizes of photovoltaic panels in order to estimate the match between power production and consumption from activities on a stochastic basis. This will mainly be discussed in the model section. 9. Electric vehicles In this section a background is given concerning electric vehicles (EVs), related challenges, environmental targets, vehicle driving behavior and electric power system impact. This background is made in order to attain a foundation in why there are so few electric vehicles driving on the streets of today in 2011 and to perceive an outlook of the opportunities to change this in the future. The EV background sections starts with the vehicle in general and EV’s at the market whereas the main contents are summarized in Table 4. 9.1. The vehicle. When discussing electric cars, this definition covers a wide number of different solutions. The plug-in electric vehicles (PEVs) have an electric motor and a battery to be charged from the

20

P. Grahn, M. Hellgren, J. Munkhammar

Table 1. Electric vehicle background summary Section Summary 9.1 9.2 9.3 9.4 9.5 9.6

9.7 9.8

Vehicle nomenclature variation of ICEV, PHEV, PEV and HEV A couple of EV’s exists on the market and the numbers are low but increasing Opportunity to reduce CO2 emissions and meet climate targets The efficiency of the electric motor is valuable The batteries makes current EVs expensive because of limited lifetime and high costs The charging modes vary with different power levels from available outlets and this together with the driving distance and the battery size determine the time for charge Driver behavior opportunities to impact vehicle fuel consumption Charging impact to the electric power system and current research in the area of the car batteries as flexible loads

electricity grid. The hybrid electric vehicles (HEVs) have both an electric motor and an internal combustion engine (ICE) which charges the battery. The plug-in hybrid electric vehicles (PHEVs) have both a combustion engine and an electric motor run by energy from a battery, which can be charged from the grid. With a PHEV the driver gets to decide what the car should be run by whereas the driver also could reduce the fossil fuel needed by adjusting charging pattern. Both PHEVs and PEVs have the option to stay connected to the electricity grid while the battery is being charged until the vehicle is disconnected. In this report the concept will however be delimited to cover only those vehicles which may charge the battery from the electric power grid, and hence would have an impact to the overall electricity consumption when and where they are connected. This means that for hybrid electric vehicles (HEVs) like for example the Toyota Prius hybrid electric vehicle is excluded, because of the fact that its battery is directly charged from the combustion engine. Electric vehicles with the opportunity to charge the battery from the grid are the plug-in electric vehicles (PEVs) and the plug-in hybrid electric vehicles (PHEVs). 9.2. Electric vehicles on the market. Major automotive manufactures have produced electric vehicles rather recently, example of cars that have been available are GMEV1, Ford Think City, Toyota RAV 4, Nissan Hypermini and Peugeot 106 Electric [14]. Many prototypes and

Photovoltaics, electric vehicles and energy users

21

experimental vehicles are also presently being developed and electric vehicles that are planned are Nissan Leaf, GM Chevy Volt PHEV40, VW Golf PHEV30, Ford PHEV30 and Toyota Prius PHEV [14]. Despite this, electric vehicles in Sweden have only in significantly small scale entered the market. During the last 30 years the amount of personal cars in Sweden have increased with 55 % [38], and today there are around 4.3 million personal vehicles out on the roads [54]. Of this entire personal vehicle fleet, there were only 190 electric vehicles in total traffic [43] which thus represent a small fraction of 0.0044 % of the entire fleet. In total there were around 100 000 [43] environmentally rated vehicles in the end of 2010. During 2010 the number of newly registered personal vehicles were 308 734 [54]. Out of these there were only 11 electric vehicles newly registered during the year, representing 0.036 %, and at the end of 2010 only 18 vehicles was purely electrically driven, PEVs according to [43]. This means that if 50 % of all newly registered personal vehicles were electric vehicle, instead of 0.036 %, the fraction of electric vehicles in 2030 with 19 years to come, with assumably around 300 000 in total each year, corresponds to 2.85 million electric vehicles and 66 % of the entire fleet. Indeed, this is would require a rapid increase, to go from 190 personal electric vehicles at the streets today. 9.3. Electric vehicles and climate targets. Sweden have formulated a climate and energy proposition from 2009 with the aim that Sweden should have a car park independent of fossil fuel by 2030 [21]. Depending of the definition of fossil independency it seems like there are some challenges along the way to realize the goal of a Swedish fossil independent fleet in 2030 even if adding the 100 000 environmentally rated vehicles of 2010. This implies that deployment of new subsidies is needed to approach the target. In the report [32] for the Royal Seaport a fossil free target for the area in 2030 is defined to mean that the use of gasoline, oil, coal, gas and other fossil fuels in the area should be phased out. In the area there will however not be any demand on how people should live their lives. Another target in the Royal Seaport is that the carbon dioxide emissions should be reduced from the average 3.4 ton CO2 /inhabitant and year in 2010 to only 1.5 tonnes CO2 /inhabitant and year in 2020 [7]. According to [55] an average yearly driving range is 15 000 km. If the consumption was 2 kWh/10 km this means a yearly consumption of 3 MWh excluding losses for an average electric vehicle if all kilometers were driven on electricity. If a comparison to a gasoline driven vehicle with the same distance driven and a consumption of 1 liters/10 km this means a yearly consumption of 1500 liter of gasoline. If an assumed

22

P. Grahn, M. Hellgren, J. Munkhammar

2 kg CO2 /liter gasoline is released this adds up to a yearly emission of 3 tonnes CO2 /year and car. This could be compared to the carbon dioxide emissions of the electricity production. According to [5] the average CO2 emissions from electricity production was in 2005 in EU(25) 415 kg CO2 /MWh, in the nordic countries 58 kg CO2 /MWh and in Sweden 10 kg CO2 /MWh. The difference emission rates depend on where the boarders are set due to the power production sources in different areas of the highly integrated power system. The rates could thus be derived from the power production when the nordic countries have a high level of hydropower whereas several other EU countries have a high rate of coal-condensing power plants. If we would assume Swedish production a yearly CO2 emission would with this numbers become 0.03 tonnes CO2 /year and car, and with the EU(25) production 1.245 tonnes CO2 /year and car. In this simplified example, these emissions would be reduced with at least two thirds with the electric vehicle compared to the gasoline vehicle. It is also shown that the CO2 emissions from the car is significant in relation to the 2020 target of each person in the Royal Seaport area of 1.5 tonnes CO2 /inhabitant and year. During a year with extremely high CO2 emissions, the number was 750 kg CO2 /MWh according to [5] and with this base, an electric vehicle would nevertheless emit 2.25 tonnes CO2 /year and car which still does not reach the value for the gasoline vehicle. Yet an ambition from then Swedish ministry is a subvention for highly environmental friendly vehicles ??, to which electric vehicles could count. The subvention is planned to effect starting with 1:st January 2012 and lasting until 2014, allowing vehicle buyers a reduction of 40 000 SEK from the initial cost. However, the subvention is limited to 200 million SEK which in practice is sufficient for 5 000 vehicles, which only is a small fraction in relation to the amount of the entire car park of 4.3 million private vehicles. 9.4. The electric motor. The electric motor has an efficiency of 80 % to 90 % in comparison to the internal combustion engine which has an efficiency of around 30 % [33]. The electric motor is more silent than the combustion engine and this could be an advantage if considering a high level of electric vehicles in an urban area such as the Royal Seaport, when the traffic noise from nearby roads and highways could be reduced. 9.5. The battery. The existing battery technologies for electric vehicles are currently costly giving also the vehicle a relatively high cost

Photovoltaics, electric vehicles and energy users

23

in comparison to internal combustion engine cars. This is both due to a large scale production of internal combustion engine cars which lower their cost, when electric vehicles including their batteries still are produced in small numbers, but also a couple of other few factors concerning the battery. Both lifetime and performance are also reduced with deep discharging cycles and affected by external temperatures [27], why a deep dept of discharge (DOD) should be avoided. In [4] it is discussed that because of these factors it would be advantageous to customize vehicle batteries based on climates and driving patterns. In Sweden the season is varying from summer days through autumn to winter and spring whereas this could be considered during the battery design. Battery lifetime, and consequently the cost, is set by some delimitation. Lithium-ion chemistries are interesting for electric vehicle applications due to their relatively high specific energy, specific power, energy efficiency and cycle lifetime in comparison to other battery technologies. Lithium-ion batteries have a cycle lifetime of around > 1000 cycles, an energy efficiency of > 90 %, specific power 200-350 W/kg and specific energy 90-160 Wh/kg and Lithium-polymer batteries have a cycle lifetime of around 1000 cycles, a specific power of 350 W/kg and specific energy 150-200 Wh/kg [14]. However, if compared to fluent fuel, the energy density is significantly lower for energy stored in batteries in comparison to fluent fuel. For example, the specific energy density for gasoline is 12.5 kWh/kg [14] and this could be compared to the lithiumpolymer battery with the specific energy density of 0.2 kWh/kg. Because of these restrictions batteries become heavy and large in volume to attain the same ranges which cars with internal combustion engines achieves. 9.6. The charging. Charging of an electric vehicle could be made using several modes with varying power. The first option could be with the use of regular one phase outlets which could allow charging at a power of 2.3 kW (230 V, 10 A). A second alternative is three phase charging, a third is fast mode charging at even higher power. The charging could hence be made with conductive technology but also induction solutions are being considered [23] as well as battery swapping or exchange stations [25]. The battery charging time period will depend on the storage capacity, the charging outlet, and the SOC when connecting. For example, if a battery with energy storage of 25 kWh is considered, and the DOD is kept at 60 % an energy amount of 10 kWh could be used for driving excluding losses. To fully charge this type of battery from a state of charge (SOC) of 15 kWh up to a SOC of 25 kWh would with the slow mode of 2.3 kW charging take around 4 hours and 20 minutes.

24

P. Grahn, M. Hellgren, J. Munkhammar

If charging at home the driver could charge during parked hours and the energy cost could be added to the regular electricity bill. If the driver also has the option during working hours, or at shopping malls, the energy cost could be included in a parking fee if parking space companies offer charging which would give them extended revenues. Charging at high power would require safety equipment both for the charging devises and the grid impact which could make fast charging expensive for car owners to utilize [11]. The infrastructure for slow mode charging would be easy to construct for many Swedish houses thanks to the system for car engine heating. In the urban area the engine heating system is not expanded to the same level and apartments might have less parking opportunities where the car could be connected to the grid. However in the Royal Seaport project there are plans for slow mode charging poles in the hole area [41]. Despite this ambitious plans, unfortunately no numbers of poles are mentioned, only that it should be demand-driven and organized within each property developer ??. However, there are 0.5 regular parking spaces projected per apartment [32]. With 10 000 new apartments [66] and a high penetration level this could mean as much as 5000 electric vehicles in the area in the future. After all, the goal to be a fossil free area already in 2030 is defined as ’The use of gasoline, oil, coal, gas and other fossil fuels has been phased out’ [32]. 9.7. The end-user. If owning and driving an electric vehicle the concerns could become how the behavior would have to adapt to a change when it comes to handling of the new technique. With an electric vehicle these changes because of new implementations would for example be the charging, the limited vehicle range before charging is needed, the adaptation in driving behavior to prevent battery degradation and the change in energy payment. The range of electric vehicles is limited by their battery size and restricts the driver flexibility which could mean that more planning is needed for future trips.

Figure 11. Electric vehicle illustration People are used to the range of an ICEV of around 50 liters in the tank which could last for about 500 km, in comparison to a 10 kWh

Photovoltaics, electric vehicles and energy users

25

PEV with a consumption of 2 kWh/10 km and a range of 50 km. However this reasoning considers large safety margins and opportunities for spontaneous long trips, when according to statistics in [8] the people in Sweden do not travel more than 50 km in average every day. Indeed a PHEV could cover the average daily commuting to work and shorter driving with energy from the battery and have the second engine to cover less frequently longer trips. For PEV owners the charging would remove the gas station visits and for PHEV owners decrease the frequency of them due to the driving and charging habits. Furthermore if we assume a gasoline price of 14 SEK/liter and a vehicle with internal combustion engine with a consumption of 1 liter/10 km this becomes a cost of 14 SEK/10 km. This could be compared to an electric vehicle with an assumed energy cost of 2.5 SEK/kWh and a consumption of 2 kWh/10 km adding up to a cost of 5 SEK/10 km for the customer, which is less than half the cost. Nevertheless, this price example is reflecting prices in 2011, whereas the situation in 2030 could become another due to increased energy efficiencies, changes in oil- and electricity demand and supply affecting there respectively prices including taxes and subventions. In Figure 12 an illustration of which time most main trips including trips not made by car starts in Sweden, the statistic comes from the Swedish travel survey made in 2005-2006 [45]. This graph shows that most trips starts in the morning or in the evening which could be derived from commuting traffic to work and back home. It is assumable that this pattern could be generalized to the behavior of cars also if all other main trips are excluded to make it comparable to the activity based data in this study. An interesting option is the concept of car pools. If joining a car pool the resident would be able to use the public transport system during the daily commuting to work when the mobility level is high. The resident would further also have the opportunity to book or use available vehicles during week-ends and evenings. If we consider that kind of behavior, the vehicle could be parked mostly during the day and if the vehicles where electric cars they could charge during the day and thus meet a conceivable photovoltaic power production when the sun is up. This scenario is however not covered in this report but would be interesting for future research. 9.8. The electric power system. With electric vehicles entering the market the overall load curves will be affected in relation to the vehicle charging magnitude over time. With a small number of vehicles the electric power system will not be affected particularly but with larger penetration levels, the peak load could become significantly high due to drivers charging schedule. Uncontrolled charging in this manner and its

26

P. Grahn, M. Hellgren, J. Munkhammar

0.5 0.45

Probability of starting main trip

0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0

0

5

10 15 Time of day [Hours]

20

Figure 12. Probability of which time during the day most trips starts in Sweden [45] impact to the load curves is studied in [10]. Other studies investigates opportunities for controlling the vehicle charging and redistribute loads from the peak load and also to offer vehicle to grid (V2G) services [30]. V2G is based on that electric vehicles should be constructed to allow a reversible power flow and furthermore be able to discharge electric energy from the battery to the grid while connected, which could enable them to act as ancillary service providers to the grid. Further studies investigate how electric vehicles could be used by aggregating them to meet bids at the control power market [12]. Nevertheless in this case study of the Royal Seaport, controlled charging, aggregators and V2G are only briefly mentioned, and the main focus lie on the vision to become a fossil independent area and how electric vehicle would be a part of this target in the projected area. In the simulation only uncontrolled charging is therefore investigated. The potential energy replacement from fluent fuel would however give electricity producers and grid owners’ new revenues, received from the mobility sector.

Photovoltaics, electric vehicles and energy users

27

Part 5. Interviews 10. Interviews The chosen methodology for this part of the project is qualitative interviews. The method was choose due to a generalized result was not desired but rather the experiences and perceptions of various organizations and people involved in the construction of the Royal Seaport area. The approach is more explorative than confirmatory which as well played into the decision of method. Interviews are generally more sensitive and can capture more nuances than explorative survey studies [24, p. 70]. While the area of inquiry needs to be decided prior to data collection in both cases an qualitative approach offers a more open ended approach; an approach that is more adaptable during the course of the study compared to quantitative survey. This adaptability and the ability to be open for new information was instrumental to the choice of methodology. The structure of the interviews were semi-structured. This approach was chosen due to the explorative nature of the inquiry and to give some structure to the inquiry. A few main areas were mapped out (see 10.3 below) in conjunction with the overarching research question. 10.0.1. Interview methodological approach. A approach drawing inspiration from phenomenology has been used. The aim of phenomenological is to grasp the meaning, structure and essence of the lived experiences of a phenomenon of a single or a collective of individuals [34, p. 482ff]. Phenomenology was founded in the early years of the 20th century by Edumnd Husserl in Germany. The tradition has been expanded by a circle of his followers and spread across the globe thereafter. The tradition of phenomenology is not a singular one. Multiple interpretations exist and several traditions have carried the ideas presented by Husserl onwards [24, p. 54ff]. Contemporary phenomenology is diverse but some common areas can be found: i) anti-reductionism which essentially means resistance against generalized explanations. ii) rejection of speculative thinking. The main issue here is considered to be the phenomena themselves. Another common interest is iii) the process of conscious awareness. Finally iv) focus is on descriptive analysis and as such they forgo inferential explanations [34]. It should be noted that a phenomenological approach does not primary study causal relationships. A point made by proponents of phenomenology is that ‘pure’ knowledge regarding a phenomenon can not be reached since all knowledge concerning the phenomenon is related via the experience thereof. The approach is concerned with the experiences of phenomena rather than the state, the subjectively

28

P. Grahn, M. Hellgren, J. Munkhammar

perceived relations rather than the objective one. Simplified the object of study is perceptions of phenomena rather than the object itself. Phenomenology, in Husserl’s conception, is primarily concerned with the systematic reflection on and analysis of the structures of consciousness, and the phenomena that appear in acts of consciousness. The approach of this study will differ from Husserl’s original conceptualizing in that what will lay as the focus is how the interviewees perceive and reason regarding the Royal Seaport project from their viewpoint. The analysis is thus centered on their experiences, their work situation and their organisations stake in the process of constructing the Royal Seaport area. The focus is partly drawn towards their view of the bigger project. The interest in this study is how various actors in the project see the issues at hand from their perspective and how their understanding, constrains, opportunities and considerations shape their work on the Royal Seaport project. 10.0.2. Interviews. The interviewees were contacted prior to interviews and arrangements for the interviews were undertaken. Two of the interviewee were interviewed via telephone whereas the other were interviewed at the place of work. A total of five people were interviewed (see 19.1 for a brief presentation). The difference in structure of the interviews were due to circumstance or request of the interviewee. The interviews lasted for approximately one hour each and the interviews were conducted in Swedish. Interviews were digitally recorded. 10.1. Ethical considerations. The subject itself is not controversial and the interviewee meet the interviewer in their role as representatives of their organisation rather than private citizens. All interviewee were contacted and volunteered. The specific identity of the interviewee will be withheld to guarantee anonymity. The interviewee has not been requested to indulge information that can be potentially harmful or disadvantageous for themselves within or outside their organization. In the cases information of this kind of information was given steps have been taken to protect the interviewee. 10.2. Reliability and validity. Factual statements of the interviewee have treated as statements given during an interview situation rather then explicit statements of fact. The factual statements are considered to have a high degree of accuracy due to that all interviewees have been very clear when the information they provided is uncertain and have given advice or offered to supply the information at a later date. Due to specifics have not been a goal but rather experiences and perspectives no weight have been put at specific numbers or data in the analysis.

Photovoltaics, electric vehicles and energy users

29

The focus has rather been at more general reasoning and descriptions of processes and events. The two different interview approaches, telephone and personal, can plausibly have an effect on the results. However the focus off the interviews have not been focused on the interviewee as personal individual but rather as professional actors. The possible effect has been considered to be smaller since the interviews have been in regards to the actions, reasoning and position the individual interviewee uphold within their organisation. Follow-up question to ensure correct understanding of statements have been used through the interviews and the core areas have been revisited on several occasions during the interviews. The intent for this approach has to been to allow the interviewee to talk about the specific subject from i) various angles, ii) interweave components from other subjects treated during the interview and iii) offer during the analyze multiple discussions concerning the subject. The focus has, as previously noted, not been on statements but rather their experience and insights as participating actors in the Royal Seaport project.

10.3. Interview guide. Prior to the interviews an interview guide was designed. The aim of the guide was to cover several areas: i) some background information about the interviewee, ii) The Royal Seaport area and the interviewee and his/her organisations involvement in the area, iii) experiences from earlier projects and its effect on the current project and lastly iv) how the interviewee perceives the future inhabitants. The background questions were aimed to position the interviewee in the organisation and within the greater Royal Seaport project. The interviewees role within the organisation and the organisations involvement and its stakes in the projected were explored as well as the importance of the project for the organisation at large. A larger area of questions were regarding the interviewee and his/her organisations involvement in the area. Questions here touched upon the zoning, planning and structuring of the area and which components that were of relevance for the interviewee and organisation. Transportation, energy and environmental issues were also handled within this area. Another central area concerned the inhabitants of the area. This area of questions revolved around the organizations perspective on the inhabitants and how they were relating to them. The expectations on the inhabitants on subareas such as what they expect making the area attractive to the future inhabitants as well as whom can be attracted to area. How the envision the area to support an sustainable life style as well as expected traveling habits.

30

P. Grahn, M. Hellgren, J. Munkhammar

Considerations were taken during the interview to whenever questions were suitable for the interviewee. As such certain questions were omitted during some interviews and at focus were more firmly on areas of direct relevance for the interviewees. The order of the interview guide acted more as a guide line and the flow of the conversation were prioritized above strictly following the guide. Several of these areas of question became interlaced during the interviews. A copy of the interview guide (in Swedish) can be found in appendix A. The analysis can be found in section 19.

Photovoltaics, electric vehicles and energy users

31

Part 6. Mathematical modeling 11. Modeling energy consumption from activities This section describes the data which was basis for the time use modeling. The time-use survey was thus not performed in this report, but the results from the survey were used as a basis for modeling synthetic activity data. 11.1. Time use survey. Time-use surveys measure the amount of time people spend doing various activities, such as work, childcare, housework, watching television, volunteering, and socializing. Time use surveys are used worldwide for research by a broad range of disciplines. Fields such as economics, business administration, gerontology, urban planning, political science, occupational therapy, nursing and medicine, recreation and physical and health education, sociology/anthropology, and psychology utilized time use surveys. Time use surveys provides replicable data that are the output of activities, decisions, preferences, attitudes, and environmental factors. It can be utilized to describe, examine and compare such diverse areas as performance of roles, cultures and lifestyles, demands for goods and services, poverty, household and community economies, and more specifically for this text, the energy usage of individuals daily activities. There have been significant advances in the field, during the last decade. New and more sophisticated data collection methods are employed, an increasing diversity of strategies for analysis, more international harmonization, as well as increasing interdisciplinary collaboration. There is an International Association for Time Use Research (IATUR), with an annual scientific conference. 11.2. Time use data method. Time use data are collected by the means of time use diaries. Time use diaries come in several variants but have a few things in common. Firstly the diaries are require the respondents to record their activities in the diaries using their own words. A diary commonly covers a specific period of time such as a day, week or month. The resolution is depending on the resolution of measurement. Exclusive studies of specific activities, such as noting time used to answer e-mails, usually run for longer periods of time than studies that are more inclusive, general aimed at capturing general activities, such as all activities during a day. Common general studies capture activities during a 24 hour period. Mostly two days, a week day and weekend day is recorded. Differences exist between various studies. The harmonised european time use survey (HETUS) uses the four following recording domains [13]:

32

P. Grahn, M. Hellgren, J. Munkhammar

(1) Main activity: “What did you do?” (2) Parallel or secondary activity: “Did you do anything else? If so, what?” (3) Who with: “Were you alone or together with somebody you know, if so, who?”, (4) Location (incl. mode of transport) 11.3. Time use data structure. The collected diaries consists of sequences of activities or events noted by the respondents. Commonly background information on the respondents are collected as well. Usually the purpose of the collected background information is to form population groups for which aggregates can be calculated as well as complementing the time use study. As an example the Swedish time use survey of 2010/11 included a measurement of usage of computers and internet. Commonly control questions to ensure and strengthen data validity is included. Time use data is collected to measure the time used and the order off said data through a predetermined period. The main aim of collecting time use data is to gain a deeper understanding of how the respondents utilize and structure their time. Time use data have advantages over estimates collected via interviews or surveys due to lesser impact of respondent bias and over- or underestimates of actual time used for activities. Time use diaries also have the advantage of offering the possibility to recording the activities as they occur rather than recollecting after the events has transpired. Time use data also commonly record minor activities that are commonly ‘disappear’ in the noise of every day life activities. A disadvantage of time use diaries is that the recording of events interfere with common day to day activities and requires the respondents to be stringent in their notation. This disadvantages interference and effect on the data can however be considered minor in comprisal to the recollection methods employed. 11.4. Time use data analysis. Time use data is generally analyzed on an aggregated level. Common analysis are comparable studies of gender, social factors, income levels and time utilization over a specified period. Examples of other approaches includes visualizations [59] and modelling of residential electricity demand [17]. 11.5. Time use data used in this study. The time use data utilized in this study is collected by Statistics Sweden. Full time use data sets have been collected in Sweden in in the periods 1990/91, 2000/01 and 2010/11. The 2010/11 study was recently concluded and the coding of the data is complete. The data is expected to be made available for

33

Photovoltaics, electric vehicles and energy users

researchers in late 2011 or early 2012. Due to this the data utilized in the study is from from the 2000/01 period. The time use data collected by Statistics Sweden the respondents record their activities through the day in 10 minutes intervals in a time use diary 12. Deterministic mathematical model of household power consumption Our innermost system S investigated in this project is a household connected with a photovoltaic panel and an electric vehicle. In this setup there are three main contributions to the total balance of power for this system. First off we have the photovoltaic production of electric power P1,i where i is the index for time step i. We shall assume that there is a fixed time step length of typically 1 minute. Second we have the household power consumption from activities P2,i . Third we have the power consumption from the electric vehicle when connected P3,i . The discrete time evolution of power production of the entire system S is given by the following equation: Pi = P1,i − P2,i − P3,i .

(1)

Negative production means in effect consumption. For simplicity we assume that i is a natural number (i ∈ N). This makes P a T ×1 matrix where T is the total number of time steps given that i = 1, ..., T . The equation (1) has three terms which have to be quantified in order to estimate the electric power consumption/production of the particular household. ,QVRODWLRQ  7HPSHUDWXUHGDWD

7LPHXVH GDWDIURP LQGLYLGXDOV

39PRGHO

(OHFWULFLW\ SURGXFWLRQ

3

7LPHXVH GDWDIURP LQGLYLGXDOV

6WRFKDVWLF PRGHO

(OHFWULFLW\ FRQVXPSWLRQ

3

6WRFKDVWLF PRGHO

0RGHOIRU (9XVDJH

(OHFWULFLW\ FRQVXPSWLRQ

3

7RWDO HOHFWULFLW\ FRQVXPSWLRQ

3L

Figure 13. A schematic illustration of the model, explaining (1).

L

L

L

34

P. Grahn, M. Hellgren, J. Munkhammar

12.1. Household power consumption from activities. In order to estimate the household power consumption from activities we have to estimate the power consumption associated with each activity. Based on the time-use data in a previous section we may calculate the electric power consumption associated with each activity. We assume that the set of activities Ci ∈ Activities for each time step i is modeled to power consumption over time P2,i ∈ R via the mapping: Ci → P2,i

(2)

Since there is no production of electric power from household activities we may in general assume that P2,i ≥ 0. In order to properly quantify the mapping (2) we need to adopt some simplifying assumptions: 1. Each individual may only occupy one activity at each time. 2. Each individual may only be in any of the predefined activities at any given time. 3. To each activity there is a predefined power consumption associated with it. Let us formally assume that there are N activities defined in the set of activities. Then we setup the power consumption of these activities according to the following equation:

(3)

P2,i =

N ∑

Aji Lj + PSpecial,i + PBase,i .

j=1

In this setup we have divided the household power consumption P2,i from activities into three terms. The first term represents those activities that consume a particular fixed amount of power when active (eg TV) and the second represents those appliances that might require more complex power consumption patterns once activated (eg washer). Let us review these terms. The first term consists of two factors, one covariant vector Lj and a matrix Aji . The vector Lj which is of size N × 1 and contains the instantaneous power consumption from appliances that is running for each activity. Each activity has a corresponding entry in Lj , and the notation Lj is the transposed version of Lj (in order to make the multiplication correct). The second factor is the matrix Aji which is a boolean matrix with entries defined by time i and activity j. The size of A is N × T where N is the size of vector L and T is the total number of time steps. A typical setup on the activity matrix A is the following:

Photovoltaics, electric vehicles and energy users 

1  1  A=  .. .  0

(4)

0 1 .. . 0

··· ··· .. .

35



1  0  ..   .  ··· 1

In this particular example A10 = 0 and A01 = 1 which in practice means that activity 1 is off at time 0 and that activity 0 at time 1 is on. The second term of P2,i is the special activity power vector PSpecial i which in similarity with Pi has size N × 1. It collects all activities with complex activity patterns and has to manually be setup from the particular appliances. The total power production containing all terms becomes:

(5)

Pi = PP V,i −

N ∑

Aji Lj − PSpecial,i − PBase,i − PV ehicle,i

j=1

It should be noted that PSpecial i can approximately be omitted under many circumstances. In many cases certain complex energy consumption appliances might either be omitted or simplified to be meshed into the first term of (3). The third term PBase,i of P2,i is just a term for basic background appliances such as for example standby. 13. Discrete time Markov chain model approach to synthetic activity generation The model used in this paper - a modification of Widén & Wäckelgård’s model - is a Markov chain model, whereas we begin by describing a general discrete time Markov chain model. Generally a discrete-time Markov chain Xt is a discrete stochastic process based on the premise that each time step t = i is occupied by one state Eµ in a number of states defined by E1 , ..., EN and that each state is defined stochastically on the basis of the previous state only - the so-called Markov property. The number of time steps is defined via i ∈ {1, ..., T } and thus µ ∈ {1, ..., N } is the index of state. The probability that the process occupies a particular state at time step t = i is [2][p.228]:

(6)

pµ,i = p(Xi = Eµ )

And naturally since the system is closed the probability for an individual to occupy any state in the predefined set of states at a given time t = i is unity:

36

P. Grahn, M. Hellgren, J. Munkhammar

6WDWH        

7LPH



L

L

L

L

L

L

Figure 14. A schematic illustration of the Markov chain model. See table 2 for legend of the states.

N ∑

(7)

pµ,i = 1.

µ=1

As the process evolves from one time step t = i to the next t = i + 1 the state of the process at t = i + 1 is determined from the state at t = i via the transition probabilities:

(8)

Pµν,i = P (Xi+1 = Eν |Xi = Eµ ).

Here µ ∈ {1, ..., N } and ν ∈ {1, ..., N } since (12) is a square matrix of size N × N . We can see that (12) satisfies the Markov property: the state at t = i+1 is only dependent on the state at t = i. The transition matrix Pµν,i can then be written as [17][p.50]: 

(9)

Pµν,i

P11,i P12,i   P21,i P22,i  = ..  ..  . .  PN 1,i PN 2,i



· · · P1N,i  · · · P2N,i   ..  ..  . .   · · · PN N,i

The distribution of states is the following [17][p.50]:

Photovoltaics, electric vehicles and energy users

(10)

pµ,i =

N ∑

Pµν,i−1 pµ,i = pµ,1

µ=1

i−1 ∏

37

Pµν,τ

τ =1

where pµ,i is the probability of occupancy in time step t = i (6). 13.1. Estimation of transition probabilities. A Markov chain model has to be calibrated for a particular purpose. This "calibration" is determining the transition matrix Pµν,i for each time step [17][p.50]. Furthermore given a set of empirical data it is a straight forward task to estimate the transition probabilities from there. Suppose that we have a series of data sµ,i representing the activity for µ ∈ {1, ..., N } number of individuals and i ∈ {1, ..., T } time steps. Lets assume that from time step i to time step i + 1 the total number of transitions nµν,i from µ to ν are summed up over all individuals. This gives the total number of transitions from state i:

(11)

nµ,i =

N ∑

nµν,i ,

µ=1

from which we get the transition probability estimates: (12)

Pµν,i =

nµν,i . nµ,i

However this expression only holds for when the state i is occupied initially. The probability for transition from i when there is no occupation of i is naturally zero. Thus we get the following setup: {

(13)

Pµν,i =

nµν,i /nµ,i ∀nµ,i > 0 0 ∀nµ,i = 0.

An hourly average is used in the model of (12) in [17][p.50]. 13.2. W&W-model. In this section we shall describe Widén&Wäckelgård’s discrete time Markov chain model (from now on described as the W&Wmodel) for generating synthetic activities over time and how this can be used in the power equations in section . The W&W-model is built on the assumption that the transition probabilities are derived from time use data. This adds one component to the mapping of time-use data to electricity consumption (2). We assume that the set of activities Ci ∈ Activities for each time step i is defined by the discrete time Markov chain process Si which in turn is modeled to household power

38

P. Grahn, M. Hellgren, J. Munkhammar

consumption over time P2,i ∈ R via the mapping (For explanation of P2,1 see section 12.1): Si → Ci → P2,i .

(14)

The transition probabilities pµν,i used in the W&W-model are derived from the equations in section 13.1 based on the data set described in section ??. The set of states is the following in the model used in this particular model [62]: Table 2. Activity states Code Activity 1 2 3 4 5 6 7 8 9

Away Sleeping Cooking Dishwashing Washing TV Computer Audio Other

14. Photovoltaic power production The photovoltaic power production P1,i is generated from observation data over time for some specific region. The photovoltaic power production naturally depends on many parameters such as location, the diurnal and seasonal variability of daylight insolation and panel tilting. We review this below. 14.1. Potential photovoltaic setup at Royal Seaport. In this report we focus on two systems. The inner system being a single household, its members, a photovoltaic panel and an electric vehicle connected to the grid. The outer system is the entire Royal Seaport region. The analysis of photovoltaics in this report will focus on the inner system; the household. The three main power-components of the "local" power system will be load from activities, load from electric vehicle and power production from the local PV-array on the roof, the equations regarding this is discussed in section 12. The setup on the photovoltaic array will be the following. Since this study is based on preliminary estimates on the setup of various components at the Royal seaport we shall assume a reasonable size of the photovoltaic panels

Photovoltaics, electric vehicles and energy users

39

associated with each apartment. We shall assume that each apartment will have a photovoltaic panel on the size of 2m2 , also that it is tilted 45 degrees and facing south. 14.2. SMHI data for photovoltaic power production. In order to estimate the data for the PV-panels we have utilized Swedish irradiation and temperature data 1992-1999 from the Swedish Meteorological and Hydrological Institute (SMHI) [35] combined with a model for PV electricity production developed by Widé n [19]. This output data from the model is estimated from insolation data, panel size, location, tilt, and declination on an hourly basis during 1992 − 1999. The data setup is on the form: [P Pnom Ac = smhi_pv_output(panel_tilt, panel_azimuth, time, station, track] Here panel_tilt and panel_azimuth are tilt and azimuth angles of the PV-panel. time is the vector of time of the data, station is the data on insolation at various locations in Sweden: Table 3. Location Code Location 1 2 3 4 5 6 7 8 9 10 11 12

Lund Växjö Göteborg Visby Norrköping Karlstad Stockholm Borlänge Frösön Umeå Luleå Kiruna

Finally track is a boolean variable regarding whether or not the PVcell follows the sun. In our scenario with the 2m2 photovoltaic array we have assumed that the panel_tilt angle is 30 degrees and that the azimuth angle panel_azimuth is zero. The station is set to station(7) and the track is set to false. The primary data used is that of the first year (1992) in the data-list. Since the stochastic model is dependent on season the data used is matched as to correlate exactly day-by-day over any period used. The model was also used in [19].

40

P. Grahn, M. Hellgren, J. Munkhammar

15. Electric vehicle power consumption 15.1. SmartLoad model of electric vehicle charging. In this project an extension to the W&W model, described in section 13.2, is made which includes the electric power consumption of an electric vehicle.The extended model will be called the SmartLoad model (Stochastic Markov Activity Residential Transport Load model). The model is a bottom up approach, where the residential activities are used to determine use of electric products and therefore the electricity consumption. For the electric vehicles in the model, the consumption of electric energy by a car will take place when the resident is away, thus when A1i = 1, with a probability rate of pncar that the resident is driving, hence takes the car when away. The index i represents each time step for i = 1, ..., T where T is the total number of time steps. The time step length is 1 minute. The state of charge SOCi , i.e. the energy level in the battery, will decline during the away period depending on the distance driven δ, the velocity v and the consumption ζ when driving. The charging of the electric vehicle will occur when the consumer is at home, that is to say when A1i ̸= 1. During the charging at a power of Cp , the SOCi will increase until the battery is fully charged, thus SOCi = SOCmax , or the consumer decides to use the car again. The starting time of a trip and also the returning time after a trip with the vehicle is decided by the time use data, as well as the charging time which takes place when the car is parked at home, connected and not yet fully charged. The load from the vehicle battery charging at time step i is PV ehicle,i and is based on the charging rate of Cp and the connected time period. To avoid decreasing the lifetime of the battery, the state of charge will be limited to a minimum state of charge level which is decided by pSOC min . The consumption ζ while driving will be distributed close to an average consumption and will depend on season; winter, summer, autumn, spring and respectively seasonal coefficients. The velocity v while driving will be assumed to be closely distributed to an average velocity. The parameters are listed in Table 4.

15.2. SmartLoad equations. The power when connecting the battery to the household for charging is Cp with optional charging rate. With this the load PV ehicle,i for each connected vehicle is determined by equation 15:  

(15)

PV ehicle,i =

0  Cp

if A1i = 1, K < pcar if A1i = ̸ 1, SOCi ≤ SOCmax

Photovoltaics, electric vehicles and energy users

41

Table 4. Parameters Parameter

Symbol

Initial state of charge [kWh] State of charge at time i [kWh] Minimum SOC fraction [%/100] Depth of discharge [kWh] Trip distance [km] Charging power [kW] Vehicle load [kW] Probability to take car [%/100] Driving velocity [km/h] Engine consumption [kWh/km] Season coefficient [%/100]

SOCmax SOCi pSOC min DOD δ Cp PV ehicle,i pcar v ζ Scoef f

where K is a stochastic variable which is uniformly distributed between 0 and 1, randomized each loop of the iteration, K ∈ U (0, 1). In general it could be optional to use the battery not only as load, but also as temporary electric power production. However in this project we only assume a load when connected, thus that PV ehicle,i ≥ 0. The battery state of charge is determined by equation 16. The level of energy in the battery SOCi , will thus decrease while out driving and increase while charging each minute according to:

(16) SOCi+1 =

  SOCi −  SOCi +

ν·ζ 60 Cp 60

if A1i = 1, k < pcar , DOD < SOCi if A1i ̸= 1, SOCi ≤ SOCmax

where DOD = SOCmax · pSOC min to reduce the risk of decreased battery lifetime, see more information in section 18.3. 15.3. Electric vehicle assumptions. 15.3.1. Simplifications and delimitations. Some simplifications in the SmartLoad model are made to be able to make feasible iterations for the simulation. It will be assumed that there exists one electric vehicle in the household which only one of the residents will drive. According to the statistics in [45] the number of vehicles per household is most likely to be one compared to households none or more than one. The W&W model allows a number of residents to be chosen to live in a household, but in the SmartLoad model the driver is limited to only one of the persons to reduce the number of iterations and hence the simulation time. It is assumed that the driver only charge the vehicle

42

P. Grahn, M. Hellgren, J. Munkhammar

at home. In future studies, charging opportunities at work or other charging stations could be included. Another assumption that is made in the simulations is that the driver always connects their vehicle for charging when arriving to the parking place at home. This could be modified in future studies. A further opportunity would be the possibility to charge the vehicle when the overall electricity demand is low and hereby also the prices are low which could be beneficial for the vehicle owners. Yet an additional option would be to implement vehicle to grid technology in the area and let vehicle batteries also offer electricity to be injected to the grid by discharging the battery when needed. These two opportunities will nevertheless only be mentioned briefly in this report. Following sections further describes and motivates additional assumptions.

15.3.2. State of charge. The state of charge in the battery in reality could behave similar to one of the graphs in Figure 15, illustrating when the car is charging, out driving or parked. The first one could illustrate a plug-in hybrid electric vehicle that drives to work at 8.00 in the morning whereas the energy level decreases along the route. Afterwards the vehicle is parked without charging between 9.00 and 13.00, and then the vehicle is out driving again until state of charge reaches the DOD. The vehicle could now be driven on a second propellant until connecting for charging at the house 19.00. In the second graph in Figure 15 an electric vehicle is illustrated that drives to work at 8.00, is parked there without charging between 9.00 and 18.00 and then drives back home at 19.00 and is connected for charging. The energy level of the battery could also be assumed to increase during the time away, if charged during regenerative breaking or charged when parked for example at work or at a supermarket, but these cases are delimited from this study. The state of charge in the SmartLoad model is a simplified version of reality and is modeled according to Figure 16. The limitations of the model makes it difficult to illustrate a realistic behavior of the state of charge in a battery over time, but in the scope of this study this scarcity is irrelevant. Instead, the illustrations show that the model is independent of when the car have been driving during the time away. The same amount of energy from the battery is consumed as in Figure 15 and then the battery is charged to the same level again. The difference is that the decreasing of the state of charge could take place anytime during the time away. The SmartLoad model is only dependent of the state of charge when connecting for charging at the household site, and the time parked before leaving again, to determine the charging time of the battery.

StateOfCharge [kWh]

StateOfCharge [kWh]

Photovoltaics, electric vehicles and energy users

43

25

20

15 5

10 15 Time of day [hours]

20

5

10 15 Time of day [hours]

20

25

20

15

Figure 15. State of charge in reality

26

24

StateOfCharge [kWh]

22

20

18

16

14

12

5

10 15 Time of day [hours]

20

Figure 16. State of charge in the SmartLoad model

The state of charge does not always descend to the minimum level, SOCmin after being out driving. The decreasing SOC is instead decided by the period of time away, the velocity and the consumption while driving. This means that the vehicle could return home after a short trip, with high amount of energy left in the battery which then needs to be charged for a relatively short time. In the model the load at the certain site at time i will hence follow the vehicle state of charge

44

P. Grahn, M. Hellgren, J. Munkhammar

behavior shown in Figure 16 which is representing the behavior of the vehicle load at a certain site, which is of interest in this project.

15.3.3. Consumption. The consumption ζ in the SmartLoad model while driving will be assumed to be distributed close to an average ¯ To follow season variation the consumption will be consumption ζ. assumed to depend on winter season, summer season and according to a season coefficient Scoef f with ζ = ζ¯ · Scoef f , listed in 5. This was done to reflect that more energy is needed during cold external condition, such as during winter in comparison to summer, which is shown in [28]. The season coefficient was chosen after inspiration by this system in [40]. The velocity v while driving will be assumed to be an average velocity and will be allowed to variate. Table 5. Season dependent average consumption Season

Season coefficient, Scoef f

Jun-Aug 0.8 Sep-Nov, Mar-May 1.0 Dec-Feb 1.2

16. Stochastic model The model described above suggests how to calculate the energy production/consumption from for example observed data regarding photovoltaic power production, household activities and electric vehicle use. In this section we shall describe how instead a stochastic approach may be used to generate this input data. Note here that the photovoltaic contribution P1,i will not be stochastically modeled, only the household power consumption from activities P2,i and the load from the electric vehicle P3,i . The above models might be described as deterministic which in effect means that the outcome is always the same given the same input. In contrast to deterministic models a stochastic model is not deterministic but based on indeterminacy of outcome [2][p.15]. A stochastic model will only give output in the form of a probability distribution. There exists several types stochastic processes, such as Markov chains and Poisson processes for example [2, 15]. There is a fundamental distinction between using so-called time-continuous simulations - where time proceeds in discrete steps - and discrete-event simulations where there is a discrete sequence of events and time proceeds according to the time between events [15].

Photovoltaics, electric vehicles and energy users

45

In this section we shall primarily discuss how synthetic activity generation from stochastic modeling might be used in the models for estimating power consumption described above for household power consumption in section 12.1 and power consumption associated with an electric vehicle in section 15. 17. Load matching Load matching is the process of estimating the fraction of load that is matched by some power source. In this project we shall in particular use this general concept in order to estimate the degree to which the household load is matched by photovoltaic power production. For simplicity we assume that the electric power consumption of the electric vehicle P3,i is included in the household power consumption from activities P2,i . This gives the following total production of electricity for the household:

(17)

Pi = P1,i − P2,i ,

where P1i is the photovoltaic power production. The task of load matching is to analyze to what degree κi the power consumption P2i is matched by the photovoltaic production P1i at each time step (Naturally when overproduction, or P1,i > P2,i , occurs the ratio κi does not exceed 1). Lets define PLack,i as the amount of power at a particular time that is not matched by P1i . Then we have that PLack,i /P2i is a measure of how unmatched P1,i and P2,i are. According to the normalization criterion we then have the matching as the complement:

(18)

κi ≡ 1 −

PLack,i . P2i

We have the following identity for the unmatched ratio:

(19)

PLack,i =

|Pi | − Pi 2

which can be proven according to the following arguments. If Pi ≥ 0 then PLack,i = 0 and conversely if Pi ≤ 0 then PLack,i = |Pi |. This brings the following expression for the match ratio κi (18):

(20)

κi = 1 −

|Pi | − Pi 2P2,i

46

P. Grahn, M. Hellgren, J. Munkhammar

If this is summed up over time and divided by the number of time steps one gets the matching over that particular time: ∑T

(21)

κ=1−

(|Pi | − Pi ) ∑ 2 Ti P2,i

i

If the matching is to be calculated only between the photovoltaic power production and the household power consumption and excluding the electric vehicle then P3,i is simply omitted from Pi . Conversely if we wish to estimate the match between the photovoltaic power production and the electric vehicle we omit the household power consumption and arrive at:

(22)

κi = 1 −

|Pi | − Pi 2P3,i

This brings the corresponding matching summed up over a given time interval:

(23)

κ=1−

∑T

(|Pi | − Pi ) . ∑ 2 Ti P3,i

i

This is the average match percentage. Again it should be emphasized that the match percentage only measures if the energy consumption has been matched which means that the measure may never exceed 1 even if the photovoltaic production exceeds the consumption.

Photovoltaics, electric vehicles and energy users

47

Part 7. Scenarios 18. Electric vehicle scenarios Following sections describes and motivates additional assumptions made, and input data used, in order to be able to construct Scenarios of peoples driving patterns and corresponding charging behavior. The purpose of constructing scenarios is to visualize possible variations in the end-user behavior, and thus vehicle charging patterns which further is basis for plausible load curves and the electric power system impact in an area. The electric vehicle charging simulation can be found in section 21 whereas one reference scenario is used as basis for matching of photovoltaic power production to be found in Section 23. 18.1. Charging power. The charging power Cp in the scenarios is based on two different charging rate possibilities. These different charging rates are considered to represent a regular charging mode and a three phase charging mode. The regular charging mode is based on an outlet with 230 V and 10 A in comparison to the three phase charging mode which is based on a three phase outlet with 230 V and 16 A. Fast charging representing even higher power could also be included, but here the scenarios are limited to simulate these two options. The charging rates for the scenarios are summarized in Table 6. Table 6. Charging rates Charging outlet

Power, Cp

One phase 230 V, 10 A 2.30 kW Three phase 230 V, 16 A 11.04 kW

18.2. Driving frequency. When the consumer is away and the car also is away the consumption from the battery will take place. There are different probability rates that the consumer actually took the car when away in the model and the results are simulated according to Table 7. Thus when A1i = 1 and the probability rate that the consumer uses the car when away. The frequency probability of 0.1 will allow that when the resident went out of the house there is a chance of 10 % that she or he also took the car. The simulations are done also with rates of 0.2 and 0.5. It is not likely that the consumer will take the car every time when away, this scenario would be with a rate of 1. 18.3. Battery depth of discharge. The simulations are done with three different minimum battery storage levels. To avoid decreasing

48

P. Grahn, M. Hellgren, J. Munkhammar

Table 7. Probability for taking the car when away Car frequency pcar 10 % 20 % 50 %

0.1 0.2 0.5

the lifetime of the battery, the state of charge will be limited to a minimum state of charge level which is decided by the depth of discharge, DOD = pSOC min ·SOCmax . The study in [29] show that lithium-ion battery life is maintained if deep cycles, with a DOD less than 60 % is avoided, and here the storage minimum level of 50 % DOD will be simulated. According to another study, [3], the minimum SOC for a Li-ion battery is not to be less than a DOD of 80 % to preserve a lifetime of approximately 3000 cycles so the model will also be simulated for this. A third case that not takes the lifetime maintenance into account with a DOD of 20 % is also simulated. These simulations with different levels of DOD could also give a brief understanding of how the load could variate with different electric vehicle types, when a varying amount of energy in the battery is allowed to be used before charging. The battery storage minimum fraction limit used in the simulations are listed in Table 8. Table 8. Battery minimum level Minimum SOC pSOC min 20 % 50 % 80 %

0.2 0.5 0.8

18.4. Remaining simulation data. For the simulation the data listed in Table 9 will be used. In the Swedish travel survey from 2005-2006, [45] statistics shows that 30 km was the average distance per day for a driver and the average time of travel with car was 39 minutes each day. This adds up to an average velocity around 46 km/h. Therefore the vehicle velocity v, was simulated in a standard scenario to be 50 km/h and in two additional scenarios the simulation was done with average velocities of 25 km/h and 75 km/h in order to catch the alteration this would indicate. The consumption ζ, is based on the estimated average consumption of a mid-sized electric vehicle of 0.187 kWh/km in [42]. The consumption is also depending on the seasonal coefficient Scoef f defined in Section 15.3.3. An altering consumption depending also on

49

Photovoltaics, electric vehicles and energy users

more condition such as the velocity or different vehicle sizes could be included to the model, but are not simulated here. Table 9. Simulation data Parameter Quantity v ζ SOCmax

25, 50 resp. 75 km/h 0.187 kWh/km 35 kWh

18.5. Electric vehicle scenarios. Out of these assumptions and reasoning, eight scenarios were constructed allowing variations based on mentioned parameters. The electric vehicle scenario of number 2 acts as a reference scenario and forms the basis for comparison to the remaining scenarios with altering parameters. The reference scenario was to be number 2 because of the probable larger occurrence of slow charging poles compared to other charging, the 20 % rate of taking the car was which was validated towards unpublished electric vehicle use data, and the velocity of 50 km/h which were a likely average velocity as well as the minimum state of charge of 50 % to reduce the risk for lifetime reduction. The purpose for constructing the scenarios is to visualize variations in end user behavior, and thus vehicle owner or drivers charging pattern. This further makes it possible to estimate corresponding impact in the electric power system in the area and plausible load curves. The scenarios for the simulation are summarized in Table 10 and the results of the simulation are found in section 21. The reference Scenario 2 will furthermore be used as basis for matching of photovoltaic power production in Section 23. Table 10. Electric vehicle scenario construction and parameters Scenario Charging power Car frequency Velocity Minimum SOC 1 2 3

2.30 kW 2.30 kW 2.30 kW

10 % 20 % 50 %

50 km/h 50 km/h 50 km/h

50 % 50 % 50 %

4

11.04 kW

20 %

50 km/h

50 %

5 6

2.30 kW 2.30 kW

20 % 20 %

25 km/h 75 km/h

50 % 50 %

7 8

2.30 kW 2.30 kW

20 % 20 %

50 km/h 50 km/h

20 % 80 %

Photovoltaics, electric vehicles and energy users

51

Part 8. Results and discussion 19. Analysis of interviews This section contains the results of the interview study conducted. The results are structured in two sections. In the beginning will be some contextual information that have been considered of importance for the further discussion. Part of these have been brought up by the interviewee and part of them have been interjected for a broader contextualization of the Royal Seaport area and the specific parts that is of interest in this study. The second section contains the main body of information given by the interviewee. This section have been structured in four parts; one dealing with the greater Royal Seaport area. This part is to give some wider understanding of the more specific components. The second part of is focused on the buildings; the considerations and reasoning that interviewee have given regarding these. Discussions regarding photovoltaic arrays can be found here. The third part focuses on the transportation in the area inhabitants, how the planning has been made and what the interviewee brought up regarding this. Talks regarding electrical vehicles and car pools can be found in this part. The fourth and final part concerns the interviewees reasoning, talk and discussions regarding the to be inhabitants. A final and broader discussion of the results can be found in 24. 19.1. The interviewees. The interviewee were selected due to their various involvements in the Royal Seaport area. It was desired to have representatives from the political decision and planning process as well as the construction processes. A representative of the development of technology within the Royal Seaport area was also considered of importance due to their direct impact on the inhabitants and their usage of developing green technology. Below is a short description of each interviewee. The names of the interviewee have been replaced with names in alphabetic order to anonymize. The interviewee have been ordered after their workplaces and retains their gender. Anna: Works at Stockholm municipal. Area of expertise is environment. Ben: Employed by Stockholm municipal. Works with city planning. Carl: Works at a real estate development which has interests in the area.

52

P. Grahn, M. Hellgren, J. Munkhammar

Dani: Manages the environmental area at a real estate development. Eric: Employed as a researcher at a research institute conducting research for the Royal Seaport area.

Figure 17. Resident

19.2. Contextualization. The Royal Seaport project contains quite possibly millions of events, decisions and considerations. It also does not exist in a vacuum, separated from the external world. Covering all factors of relevance is not plausible within the context of this papper but there are however a few things of note that can carry importance for the further discussion. Stockholm area and its population. The Stockholm region is Swedens most populated area and is the home to 851,155 in the municipality (2010), 1.37 million in the urban area (2010), and around 2.1 million in the 6.519 km2 metropolitan area (2010).The Stockholm metropolitan area is the home for 22 procent of the Swedish population. [64] Que for living accommodations. According to one of the interviewee (Dani) there is approximately a quarter of a million people in search of living accommodations in the urban area alone. This creates a high pressure on the available and planned housing. Globally there are an trend of increased urbanization and Sweden is no exception to this trend. While housing in the major region increases the construction is outstripped by the demand. Hammarby Sjöstad is a residential area in Stockholm. The area was constructed, as the Royal Seaport area from former industrial areas and also was constructed to have a low environmental impact. The success of the area can be debated but it has had an impact on the

Photovoltaics, electric vehicles and energy users

53

zoning and planning for the greater Stockholm region as well as serves a source of inspiration for the Royal Seaport project. Public transportation in the urban and metropolitan areas of Stockholm has, as in many larger cities worldwide, been greatly debated. There are currently several discussion involving extensions of the subway system, implementation and expansions of electrical trams, implementation of bicycle roads, and construction of a major ring way through the center part (that will connect to the Royal Seaport area). Värtaverket is a combined power and heat (chp) plant residing within the Royal Seaport area. The chp is a vital part of and large contributor to the Stockholm remote heating system. The plant is however, as noted in the background (see 7.5) also a major contributor of CO2 . Parking space is always a premium commodity in major cities. The number of required parking spaces allocated per household has steadily declined in the Stockholm area. The current number is 0.5 parking spaces per household in the Royal Seaport area. Historically the number has been as high as 1.2 in the outer areas. Environmental demands on buildings. The strikt environmental requirements for the Royal Seaport area was not fully in effect for the first stage of construction. The specifics for the buildings in the first stage were finalized first and the more overarching later. This was due to the decision to make the area that would become the Royal Seaport was done during the exploitation phase had started. 19.3. The Royal Seaport area. The Royal Seaport area differs from many other exhibit projects in the regard that the area is an integrated part of the central parts of Stockholm. The area contains harbors and while certain parts are projected to be moved parts will remain as a core part of the city. Det ska vara en integrerad del av Stockholm. Norra Djurgårdstaden ska inte vara avskilt ifrån resten utan det ska vara en fungerande del av Stockholm. Tittar man påandra projekt runt om i världen såär det ofta en isolerad ö eller fyrkant ut i öknen. Som är helt avskilt från resten av världen. [. . . ] Det ska fungera ihop med staden och det ska fungera med hamnen. Stad och hamn i utveckling tillsammans har varit ledorden länge och är det fortfarande. Ben A concept that was expressed during the interviews was that the new area is not to be a specific ‘green’ area but rather an integrated part

54

P. Grahn, M. Hellgren, J. Munkhammar

of the city that just ‘happens’ to be a sustainable area. The areas is also aimed to be an example for future city building projects. In the overarching discussions of the area its proximity to the Royal National City Park (Kungliga nationalstadsparken) area for providing outdoor activities and the nearness to the down town areas that has been in focus. Closeness to vital parts enables the inhabitants to use bicycles, or walk, rather then using the car for transportation to recreational, commercial or business areas. A theme for the area has that it should be ‘easy to do right’ (Lätt att göra rätt). Tanken är att man sak kunna flytta dit och vara helt omedveten och sedan blir det rätt för att det är det enklaste. Anna One intended goal is to make it natural for the inhabitants to act in a sustainable way. How the area is structured and its layout influences behavior appears to be a focal point in the structuring of the area. Creating nearness to areas that are common visited by car and to make bicycles, public transportation and walking more desirable option rather than using a car is a common talked about. The behaviors of the inhabitants and how to encourage, while not enforcing, substantiality is something that has been a reoccurring theme. 19.3.1. Developing Royal Seaport. An earlier project in Stockholm, Hammarby Sjöstad, had an overarching goal for the whole area. One of the wisdoms from that project was that by setting a standard for the whole area the incentives for incorporating new knowledges, ideas and technologies were quite weak. By learning from the stages of production and utilize new knowledge, new technologies and by evaluation see what recommendations and requirement that will apply to the next stage. Tanken är att det ska vara ett övergripande miljöprogram som inte är sådetaljerat och som ska hålla över hela perioden som projektet håller på. Sedan gör man detaljerade krav område för område. [. . . ] Utvecklingen går framåt, och det vi tänker är att man tar ett område i taget och sedan lär man sig av det. Anna The current plan is to work iteratively, stage by stage, area by area and project by project. The environmental plans for the area is deliberately rather loose in order to take advantage of developments. Om man har hela programområdet såjobbar man med en etapp i taget. Och inom den etappen är det en iterativ

Photovoltaics, electric vehicles and energy users

55

process. För där under planneringsprocessen gång för den etappen är det nya rapporter och nya dokument som hela tiden behöver fyllas på. Vilket gör att planen ändras tills den är färdig och vi försöker hela tiden dåintegrera alla förvaltningar, stadsdelsförvaltning, trafikkontor, miljöförvaltning, ja egentligen alla berörda. Såatt man i slutändan har en såbra plan som möjligt. Såäven den processen är iterativ samtidigt som processen för hela området är det. Ben A lesson learned from previous projects has been the lack of evaluation tools. As an example the municipal is working with the Royal Institute of Technology (KTH) to shape tools to evaluate the progress. The current project is utilizing various focus groups where stake holders in the development project can meet, discuss and share wisdoms. An expressed goal of these focus group have been for the experiences to be passed on to the next stages in for example the form of letting constructors know what has been done, what has been tried, what has worked well, what has been problematic and so forth. All as a part of the iterative approach. 19.4. Enviromental goals. The Royal Seaport area has at the current the requirement of 55 kWh per sqm. The law as of this writing requires newly constructed housing to use a maximum of 110 kWh per sqm. Fifty-five kWh makes the requirement in the area to be close to passive houses. The law seems to have a positive effect on the construction industry and increased the demands political organisations feel they can set. Men det är en stor skillnad mot hur det var innan den här BBR3 kom där det var krav på110 kWh. För när den kom var det ett himmlans klagande från många byggherrar. Men när lagen sedan kom var det inga problem och dåprojekterade alla runt 75–80 för att klara 110. Nu säger de att det inte är några problem att klara 55. Det har gått väldigt snabbt. Från att i Hammarbysjöstad sålåg många runt 150-200 och tyckte att det inte gick att komma ner till 90 som man ställde som krav där. Det första kravet var 60, och dåtyckte de att det var larvigt. Och nu säger vi 55 och det klarar de utan problem. Såbara på5–10 år har det blivit jättestor skillnad i attityderna från byggherrarna. Anna 3Boverkets

Byggregler, see [?].

56

P. Grahn, M. Hellgren, J. Munkhammar

In a fairly low period of time, 5–10 years, the perceived required energy usage for buildings have dropped form 150 kWh to 55 kWh. From being impossible to be ‘easy’. How well the construction actually does is something for the future to answer but the shift in perceptions notable. Serval of the interviewee attributed this to two things; the changes in the law and the appearance of example project that has showed the viability of construction. During the last decade new technologies and demonstration projects have showed the viability of technologies such as passive houses. Interviewee pointed towards a shift or an increase within the business itself to explore options to lower energy usage. Several pointers has been made towards demonstration projects and their success. The viability of new technologies and changes in construction routines seems to have had a large impact on the perspective of constructors as well as policy makers in what demands can be fulfilled and what demands that can be set. Not all feel that it is easy however. There seems to be conflicts between parts of the planning process and part of the construction. Some comments have been made towards that the energy usage requirement are constructed in such was that fulfilling them is problematic. Heat pump technology has been desirable to use but has been hampered by a set maximum of electricity use. There has also been some problem in that the constructors have been given restrictions on thickness of external walls, window size, and whenever the roof area can be used for production of electricity via photovoltaic arrays or not. A comment given was that “you don’t build a building and then make it energy efficient” (Dani). 19.4.1. Solar based technologies. At the current photovoltaic arrays are not considered economically viable. Part of the problem is attributed to investment costs but also the current economical incentives to produce electricity in small scale. The cost of implementing the arrays has to be carried by the incomes post-construction. The current economical model for feeding electricity into the grid and the investment costs make them too steep to be considered viable. Solar thermal collectors are seen as more viable but since many buildings use remote heating there is a resistance towards them from the deliverer of remote heat. El-produktion med solceller är inte ekonomiskt lönsam idag. De kostar ganska mycket. Dåär det solfångare som är aktuellt och dåär orienteringen av huset viktig. Dani

Photovoltaics, electric vehicles and energy users

57

An aliment suggested is if these technologies become serialized the cost of implementing them will drop. The cost can be lowered by serialization of the production. Initialt är det en merkostnad. De här systemen måste komma i system och bli serieproducerade. Men de första åren är det en merkostnad pånågra procent. Men om det blir en standard och blir långa serier påde här produkterna kommer det säkert att jämna ut sig. Eric A slight Catch 22 can be seen here, no one want to invest in photovoltaic arrays since they are costly and thus serialization of arrays are harder to achieve. The current economical model in Sweden works counterproductively towards small scale energy production. A factor that can work as an incentive is the projects participation in the Clinton Clime Initiative. Part of the requirement of the initiative is that energy should be produced locally. There has been some mentioning that points towards that a contributing factor for making photovoltaic arrays not cost-efficient enough to implement is the current policies. In particular how small scale electrical generation is handled in the system. Arguments have been made that the current system offers incitaments for local production alone and not producing more energy than is utilized. With a change in policies this could be altered so the economical premisses of the utility of photovoltaic arrays can more geared towards small scale production. There has also been steps taken by energy producers to buy small scale locally produced energy as well as political debates and pressure to remake then current economical model. The German development with the planned closing of nuclear power plants and increase of renewable energies can generate an international market that will increase the possibility of serialization of photovoltaic arrays. An aspect of utilizing photovoltaic arrays is placement. To gain maximum effect from photovoltaic arrays the buildings and the roofs have to be aligned optimally towards the sun. Vi strukturerar inte byggnaderna såatt man ska fåmaximala solvinklar. Det har vi inte gjort eftersom det är såpass lite energi. Utan dåär det kanske ännu viktigare aspekter i det sammanhanget. Utblickar. Alltsåsocial värden eller att vi har en huvudgata som bara kan ligga påett sätt genom området. Den typen av aspekter. Sen handlar det om att taken ska vara förberedda eller platta tak eller rätt vinklar påtaken. Men inte såmycket

58

P. Grahn, M. Hellgren, J. Munkhammar

att byggnaderna vrids till rätt vinkel. Det ger ganska lite påden skalan. Ben In this aspect the social values and overreaching planning has been determinant. Its is however worth noting that considerations for the roofs to be viable for solar power technologies are a part of the planning. There does however also seems to be a view that for the technologies to be considered the conditions should be nearly optimal. The houses will however, hopefully, remain standing for 80–100 years. Given the development pace of solar technologies it is quite plausible that with a short period of time the economical viability of the technology will increase. By building and construction the area for possible future solar power would seem like a good approach. 19.5. Inhabitants. An area that is being constructed for 10 000 new residences obviously need someone to live there. The previous area of Hammarby Sjöstad had expectations of new inhabitants being mainly being older couples, who sold their house. In perspective families with children and and younger couple were a significant group. Stockholm municipal have as a goal for social intermixing. Attempts are being made to mix various social classes to lessen stratification and social exclusion. However large portions of the Royal Seaport is a former brownfield area which requires the land to be purified before construction can start. Det är dels att fåtill en blandning av människor. Vilket är svårt eftersom markrening kostar mycket pengar, det är dyr mark från början, samtidigt som det är ett attraktivt läge i staden. Vilket gör att marknadspriserna blir ganska höga. Vilket gör att det troligtvis är medeloch höginkomstagare som kommer att ha tillgång till området. Ben Coupled with an attractive location, purification of soil and a high demand for living space in the Stockholm region serves to ensure that the price range of the area hardly will be low. Building energy efficient builds tend as well to incur a higher cost adding to the price. The amount spent by an average family on housing in the Stockholm region also seem to have changed. Pååttiotalet, dåtror jag en genomsnittsfamilj lade ungefär 25 procent påboende, idag lägger man nästan 50 procent i Stockholm. Men det är flera faktorer, produktionen har blivit lite dyrare men vi har fått högre levnadsstandard och högre löner. Vi bygger lite exklusivare

Photovoltaics, electric vehicles and energy users

59

lägenheter idag än på80-talet, lite mer påkostade ytskikt, vackrare hus. Mer arkitektoniskt sköna. Carl Factors that should be weighted into this is changes in the housing market for the last two decades as well as changes in policy and regulation. But the two interesting factors in this is i) that living is using up a larger proportion of the income and ii) that that it coincide with a rise in incomes as well as an increase in living standards. It is quite plausible that while income increases is a contributing factor, changes in the perspective of what the home symbolizes and what role it play in the everyday life have changed. It can be argued that the home has increasing become a part of the individuals identity and how this identity is expressed. As such investments into the home is an investment into the self.

19.5.1. Lifestyles. There has been rumors that there would be specific requirement for people moving to the area such as having to maintain certain levels of exercise, required garbage sorting and buying and using eco-friendly products. According to the interviewees these rumors are the result of misinterpretations. The stated goal has been to make it easy to act in a sustainable way. Requiring it has never been a part of the debate. There is however a group that values a sustainable life style. Multiple interviewee stated that the question of substantiality no longer is a hot issue. The discussion seems to have been mostly settled. Tror de flesta är intresserade men det är även en fråga om hur ska man kunna hjälpa till eller bidra. Avfallshantering är dånog en av de enklaste att jobba med. Vi har sorteringshus för de boende. De byggnadsprojekt som vi har där vi tilläggisolerar och vi gör dem energieffektivare än tidigare får en bra respons. Anna How to help seems to be a reoccurring question among residents. While knowledge about ‘minor’ activities such as garbage sorting and using less heated water there seems to be a bit of distress among inhabitants that they are unsure of how to help and what impact their action have. Living green can also be interpreted as having a certain social status. Where living in a sustainable area is beneficial to this status. From this perspective cultivating the air of substantiality as social status marker may be beneficial for marketing, selling and in the end obtaining and paying for sustainable living.

60

P. Grahn, M. Hellgren, J. Munkhammar

The constructors and land lords interviewed expressed that the common questions connected to substantiality is centered around ‘visible’ factor or more drawn towards health based factors. Questions regarding waste, sorting and recycling, construction materials and indoor quality seem to be on top. Distances and availability of public transport have repeatedly been mention as important factor for future inhabitants. The constructors have all pointed towards that having close access to the subway is an important factor in pricing and value of property as well as desirability of the dwelling for the residents. There is however some indicators that the perspective of who is the responsible part of the sustainable development is focused on the municipal and its agencies and not the constructors and landlords. De miljöintresserade. Ibland fungerar jag som någon slags mäklare. Folk ringer och frågar hur man får en lägenhet. Men det har varit ganska många som ringt och frågat just med anledning av skriverier i tidningar eller påTV. Och de vill leva påett hållbart sätt. Det kommer nog att vara en del som aktivt söker till sig området pågrund av det. Ben Considering that all participants together are the base for the sustainability in the area it would seem that the process from zoning to finalized construction is obscure and uncertain for the potential future inhabitants. There does however seem to be implicit demands from potential future residents that the newly constructed dwellings are sustainable or are at minimum not ‘bad’. Undermedvetet hos kunderna ställs kravet. Vi får lite frågor om just detta. Men jag tror att kunderna beaktar det vid köp. Om man inte har det tänket är det svårare att sälja lägenheter. Carl Sustainability seems to be a sales argument for new housing. A theme that is reoccurring is that factors such as construction materials, materials used in the homes and the building itself have a part in the potential price that can be requested for the dwelling. However given the cities long que for accommodations there also seems to be one factor among other more pressing, such as having a home at all, and not a top one either.

Photovoltaics, electric vehicles and energy users

61

19.5.2. Home technologies. Development and usage of technologies in the home to aid in visualizing the energy usage of the residents is a key development. When the buildings themselves lower the energy usage the usage of the inhabitants become more dominating. Det blir väldigt tydligt med mätsystem. När vi bygger husen tillräckligt energisnåla dåblir boven den boende. Det är där energiförbrukningen sitter. Får man det väldigt tydligt redovisat påverkar man nog sig själv att bli bättre påatt minska förbrukningen i sitt hem. Eric Visualization of energy use is commonly seen as an important factor to lower households use of energy. If the end-users can compare their energy utilization with neighbors or with a base-line the use can be evaluated. Currently the main source of information is the billing of heat (where applicable), heated water, and electricity. A source that can be problematic to interpret due to the great variability in the cost. Current web based solution have been seen as problematic as it requires the end-users to actively seek the information. The current leaning is to make the information readily available in a easy accessed and commonly passed by area of the dwelling. A location that has been mention is close to the entrance of the dwelling. This forms of information is not common in contemporary housing and thus seems to cause some trouble. Man går igenom tekniken i samband med inflytt. Oavsett om det är traditionellt eller ny teknik. Dock inget krångligt med den nya tekniken. Inga konstigheter egentligen. Man behöver veta var panelen finns, vilka möjligheter man har själv. Hur man läser av och såvidare. Det gäller att systemen är enkla. Carl Some implementations of systems have in the past been lacking in that the inhabitants have received little or no information about the system. Worries concerning the technology, what the residents are supposed to do and how they can impact their own energy use with the system, have been problematic. The issue seems to be that the insecurity regarding the new technology rather than a distrust of the system themselves. More of a worry of doing wrong or breaking the system than the systems failing. There has as well been talks about visualizing the impacts of travels. If the population of the area uses non-sustainable means of transportation a use of aggregated indicators can visualize the impact of transportation in the area. Erectly how system like this will be implemented

62

P. Grahn, M. Hellgren, J. Munkhammar

is yet to be decided. Ideas of making it visitable at nodes for public transportation is one approach. The base for this talks seems to be that by making inhabitants aware of their travels impact will opt for more sustainable transportation.

19.6. Transportation. The main mode of transportation in the Stockholm region is public transportation. Sixty-five procent of travels are done via the public transit system. For transits to and back work the number is higher, close to 85 procent. According to Ben this is the highest rate in the Nordic countries and he believe in Scandinavia as well. Compared to other cities the usage of bicycles and walking is lower in Stockholm, a number given was about 11 procent. However bicycling is the fastest growing mode of transportation. Since 2008 the number has been tripled and it is still increasing. Vi har försökt att vända påen trafikhierarki. 50–, 60– och 70–talet har det varit bilar, vid sidan av cykel- och gångtrafik kommer ganska lågt ner pålistan. Vi har försökt vända pådet där, det är gång och cykel som ska prioriteras. De ska ha de genaste vägarna. Det ska vara breda cykelstråk och gångbanor genom området för att röra sig påett snabbt och effektivt sätt. Därefter kommer kollektivtrafiken. Där har vi jobbat mycket med att försöka fåtill den här stadsspårvägen genom området. I fjol [2010] beslutades det om förlängning genom Värtahamnen upp till Ropsten. Nu jobbar vi påden sista delen genom Hjorthagen. [. . . ] Därefter bussar, biogasbåtar är något som diskuteras också. Ben The demand for parking space has declined during the last year and parts of this is attributed to a shift in generations. Comments have been made that the younger generations are more concerned with closeness to public transport system. The car is perceived a bothersome mode of transportation due to various fees, finding parking space and cost. Using public transport systems are seen as easier and less problematic. There are indicators that the car is increasingly loosing its center place in the city landcape. Det är många som har sålt bilen och sedan upptäckt att de inte har något behov av den. De kan uträtta sina ärenden utan bil. Eric

Photovoltaics, electric vehicles and energy users

63

Another contributing factor may be the rise of car pool services. Car pools services offers the ability to use a car when required or desired without requiring the end-user to actually owe a car. The flexibility of the car is retained and common travels are via the public transport system or via bicycle or walking.

19.6.1. Car pools. A thread through all interviews was the mentioning of car pools and the near universal praise of them. Decisions have been made for the existence of car pool services in the area. There are areas in the zoning that have been intentionally designed to accommodate car pool services. The important factor noted about car pools services is closeness. Det jobbar vi med, men jag tror man kan göra mycket mer påbilpoolsverksamheten. Det gäller ju med bilpooler att man har de här hämtstationerna åbekvämt avstånd annars fungerar det inte. Men där de finns inom rimligt gångsavstånd är det ett bra alternativ. Carl The types of vehicles such car pool services would use is not set in stone but a universal view is that the vehicles have to be of the green variety. Electrical vehicles have been discussed but there seems to be a diversity between the perspectives of their usability. Objections have been towards the production of electricity for the vehicles as well towards the alternatives offering equal environmental effects. Other perspectives have been that electrical vehicles can be seen as all other green vehicle technologies. The common theme has been that the vehicles should be sustainable primary and which ‘fuel’ they use is a secondary question.

20. Results regarding household load neglecting electric vehicle load This section shows the results from simulations using the SmartLoadmodel for the special case of one household with two inhabitants neglecting the electric vehicle load. Since the special case of the SmartLoadmodel for generating household energy consumption from activities when neglecting the load of the electric vehicle is equivalent to the W&W-model, the output of the simulations is equivalent to that of the W&W model [62]. First off we have the average load for the household over a year. The table 11 is a reference table for the nomenclature used in the table over power consumption in the figure 18.

64

P. Grahn, M. Hellgren, J. Munkhammar

Table 11. Appliances Symbol Appliance A B C D E F G H I Tot

Cold appliances Cooking Washing Dishwashing Television Computer Audio Lighting Additional appliances Total energy consumption

2

Cold Appliances Cooking Washing Dishwashing Television Computer Audio Lighting Additional Appliances

1.8 1.6

Power (kW)

1.4 1.2 1 0.8 0.6 0.4 0.2 0

0

5

10 15 Time (Hours)

20

Figure 18. A one year average consumption for the different categories A B C D E F G H I Tot 1.97 1.30 0.07 0.04 0.82 1.03 0.15 3.22 2.55 11.15

21. Electric vehicle consumption 21.1. Model behavior. Example of one weekday activity in test simulations with the SmartLoad model and assumptions from Scenario 2 are to be seen in this section. The household consumption is represented by the blue line and the vehicle consumption by the black line. In Figure 19 and Figure 21 two different random profiles with the outcome of one charging period is to be seen. In Figure 19 the vehicle is connected for charging at around 8.00 in the morning and disconnected

65

Photovoltaics, electric vehicles and energy users

or fully charged at around 12.00 when the vehicle load is zero again. The vehicle could be out driving all other times. 2.5

Power (kW)

2

1.5

1

0.5

0

0

5

10 15 Time (Hours)

20

25

Figure 19. Test simulations with the SmartLoad model for an apartment with one inhabitant and one electric vehicle over one day, household consumption in blue and EV consumption in black

In Figure 21 the vehicle is connected for charging at around 17.00. The corresponding state of charge for the electric vehicles are to be seen in Figure 20 and Figure 22. These figures does not take into account when the energy was consumed during the time away from the charging location, as described in Section 15.3.2. This is because of that the SmartLoad model depend on the state of charge when connecting for charging, and the time parked before leaving again, to determine the charging time of the battery, but not on the driving respectively pausing pattern when away from the household.

21.1.1. Monthly variation. In Figure 23 the simulation is done for 30 days to illustrate the charging pattern for a month. The corresponding state of charge pattern is shown in Figure 24. What should be highlighted is that according to the graphs the electric vehicle consumption will vary with significant higher maximum load and minimum load

66

P. Grahn, M. Hellgren, J. Munkhammar

36 34 32

Energy (kWh)

30 28 26 24 22 20 18 16

0

500

1000

1500

Time (Minutes)

Figure 20. Test simulations with the SmartLoad model for an apartment with one inhabitant and one electric vehicle over one day, state of charge of the first EV

2.5

Power (kW)

2

1.5

1

0.5

0

0

5

10 15 Time (Hours)

20

25

Figure 21. Test simulations with the SmartLoad model for an apartment with one inhabitant and one electric vehicle over one day, a second case, household consumption in blue and EV consumption in black

than the household load based on the activity patterns. This kind of

67

Photovoltaics, electric vehicles and energy users

36 34 32

Energy (kWh)

30 28 26 24 22 20 18 16

0

500

1000

1500

Time (Minutes)

Figure 22. Test simulations with the SmartLoad model for an apartment with one inhabitant and one electric vehicle over one day, state of charge of the second EV increased fluctuation in the load curve could have an impact to the electric power system and its infrastructure, concerning both components and stability. 2.5

2

Power (kW)

1.5

1

0.5

0

0

100

200

300

400 Time (Hours)

500

600

700

800

Figure 23. Test simulations with the SmartLoad model for an apartment with one inhabitant and one electric vehicle over 30 days, household consumption in blue and EV consumption in black

68

P. Grahn, M. Hellgren, J. Munkhammar

36 34 32

Energy (kWh)

30 28 26 24 22 20 18 16

0

0.5

1

1.5

2 2.5 Time (Minutes)

3

3.5

4

4.5 4

x 10

Figure 24. Test simulations with the SmartLoad model for an apartment with one inhabitant and one electric vehicle over 30 days, state of charge of the EV

21.1.2. Activity comparison. The behavior of the Smartload model could further be illustrated as shown in Figure 25 and Figure 26. Here a comparison is made with two of the scenarios, Scenario 1 and Scenario 2, and raw data. The model is described to not take into account when the energy is consumed during the time away from the home i.e. the charging location. This could be compared to the dotted line in the graphs which shows one specific activity from the raw data, the activity "Took the car", in the activity diaries. That is to say the activity "Took the car" does not include all activities that could imply that the car was driven, because this might also be embedded in other activities and thus be more complicated to extract.

21.1.3. Standard deviation. In order to quantify the variability of the power consumption from the different categories we calculated the standard deviation for their time series. The standard deviation for a variable xi is defined by:

Photovoltaics, electric vehicles and energy users

60

50

Frequency

40

30

20

10

0

0

5

10 15 Time (Hours)

20

Figure 25. Scenario 1 EV average load straight line, compared with activity "Took the car", dotted line

60

50

Frequency

40

30

20

10

0

0

5

10 15 Time (Hours)

20

Figure 26. Scenario 2 EV average load straight line, compared with activity "Took the car", dotted line



(24)

1/2

N 1 ∑ s= (xi − x¯)2  N i=1

where x¯ is the mean value:

69

70

P. Grahn, M. Hellgren, J. Munkhammar

(25)

x¯ =

N 1 ∑ xi N i=1

The standard deviation is shown in table 13 below. The nomenclature of the table 13 is presented in Table 12. The standard deviation for the electric vehicle energy consumption is considerably higher compared to the standard deviation for the energy consumption of the other activities, this is due to the stochastic behavior of vehicle use. Table 12. Appliances Symbol Appliance A B C D E F G H I EV

Cold appliances Cooking Washing Dishwashing Television Computer Audio Lighting Additional appliances Electric vehicle

Table 13. Standard deviation A

B

C

D

E

F

G

H

I

EV

1.96 52.91 5.16 4.32 16.62 3.46 0.44 54.41 0.00 141.68 21.2. Scenario simulation. 21.2.1. Reference scenario 2. The standard setup is shown in Figure 27 and this is Scenario 2 which will be compared to the other scenarios with differing parameters. In this figure the household load is included to illustrate the impact of electric vehicle charging in comparison with the household consumption. In Figure 28 the electric vehicle load for scenario 2 is shown excluding the household load. In the reference scenario including the household load in Figure 27 it could be seen that during time periods in between around 18.00-21.00 the load curve is increased approximately with 0.9 kW each time step, which would represent almost 70 % of the load during the same time. This means that during certain periods the load is increased from a business as usual scenario without electric vehicles with an upper limit, a peak

Photovoltaics, electric vehicles and energy users

71

load of around 0.55 kW up to 1.35 kW which would mean an increase of around 2.5 times the business as usual load peak. 1.4

Cold Appliances Cooking Washing Dishwashing Television Computer Audio Lighting Additional Appliances Electric vehicle

1.2

Power (kW)

1

0.8

0.6

0.4

0.2

0

0

5

10 15 Time (Hours)

20

25

Figure 27. Scenario 2, vehicle load and household load based on monthly average, reference scenario

1 0.9 0.8

Power (kW)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

0

5

10 15 Time (Hours)

20

Figure 28. Scenario 2, vehicle load excluding household load based on yearly average, reference scenario

72

P. Grahn, M. Hellgren, J. Munkhammar

21.2.2. Scenarios 1-8. In the following section the scenarios listed in Table 10 are illustrated. Scenario 1, 2 and 3 in respectively Figures 29, 30 and 31 shows the load curve with differing probability of taking the car, and if these scenarios are compared it could be concluded that with a lower frequency of taking the car when away, and instead traveling in other ways, the load obviously will be less. If 10 % of the time the driver would choose the car when away this leads to a peak load of around 0.25 kW, 20 % leads to 0.41 kW and when the amount whom takes the car when away is 50 % the peak load reaches around 1 kW, based on an yearly average, thus 365 days. 1 0.9 0.8

Power (kW)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

0

5

10 15 Time (Hours)

20

Figure 29. Scenario 1, 10 % chance of taking the car, electric vehicle daily load based on one year average

In Scenario 4 in Figure32, the charging power is assumed to be 11.04 kW. With this power the fluctuations in the power curves are increased, as seen in the figure. This implies that with faster charging, and more frequent fast charging outlets, the power system impact due to variative loads becomes higher. With charging outlets placed in a distributed plan with slow charging outlets, the variation could thus be lower.

Scenario 5 and 6 in Figures 33 and 34 shows the load curve if changing the velocity while out driving to 25 km/h respectively 75 km/h which affects the energy amount consumed during the time period of away.

Photovoltaics, electric vehicles and energy users

73

1 0.9 0.8

Power (kW)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

0

5

10 15 Time (Hours)

20

Figure 30. Scenario 2, 20 % chance of taking the car, electric vehicle daily load based on one year average

1 0.9 0.8

Power (kW)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

0

5

10 15 Time (Hours)

20

Figure 31. Scenario 3, 50 % chance of taking the car, electric vehicle daily load based on one year average If driving at a lower velocity the energy consumed is less as seen for Scenario 5 in comparison to Scenario 6. This is due to that the distance covered during the time away in the simulation becomes longer with higher velocity but the same consumption rate per kilometer.

74

P. Grahn, M. Hellgren, J. Munkhammar

1 0.9 0.8

Power (kW)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

0

5

10 15 Time (Hours)

20

Figure 32. Scenario 4, three phase charge, electric vehicle daily load based on one year average 1 0.9 0.8

Power (kW)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

0

5

10 15 Time (Hours)

20

Figure 33. Scenario 5, velocity of 25 km/h, electric vehicle daily load based on yearly average

In Scenario 7 and 8 in Figures 35 and 36 the household load is also included in the illustration. In the Scenario 7, with the low minimum SOC of 20 %, the vehicle could drive for a longer time before there is a need for charge than in for example Scenario 8 with the minimum SOC of 80 % or in the other Scenarios with the minimum SOC of 50 %. This

Photovoltaics, electric vehicles and energy users

75

1 0.9 0.8

Power (kW)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

0

5

10 15 Time (Hours)

20

Figure 34. Scenario 6, velocity of 75 km/h, electric vehicle daily load based on yearly average is because of the assumption that if a car is out driving the battery will be reduced until it is empty if the vehicle drives for long enough time, and a vehicle with the minimum SOC of 20 % has significantly more energy to use before there is a need for charge than a vehicle with the minimum SOC of 80 %, in this case the difference is 21 kWh which with the consumption of 0.187 kWh/km would last for another 112 km. These figures reflects that the minimum level of the possible state of charge obviously will impact the load curves because of the differing energy amount that is available. This indicate that the power system impact will depend on car manufacturer and battery producer differences, for example if there are standards set for battery size and power outlets, the fluctuations would become lower.

21.2.3. Data results. In the Table 14 the daily distance, daily consumption and number of trips per year are summarized. The consumption varies in a span from 2.3 kWh a day in Scenario 8 up to 11.3 kWh a day in Scenario 3. This is of course related to the allowed minimum state of charge as well as the probability rate of taking the car which determine the frequency the trips. Note that the number of times the vehicle has been out driving in the scenarios 2, 4-6 where the probability that the car was taken is 20 % still differs and this is due to the stochastic behavior of the model in each simulation. The daily consumption in

76

P. Grahn, M. Hellgren, J. Munkhammar

2 1.8 1.6

Power (kW)

1.4 1.2 1 0.8 0.6 0.4 0.2 0

0

5

10 15 Time (Hours)

20

Figure 35. Scenario 7, minimum SOC 20 %, electric vehicle daily load including household load based on yearly average 1 0.9 0.8

Power (kW)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

0

5

10 15 Time (Hours)

20

Figure 36. Scenario 8, minimum SOC 80 %, electric vehicle daily load including household load based on yearly average Scenario 2, 4, 5 and 6 would indeed become the same if the numbers of iterations were increased even further. The number of trips taken during one year, in column Away/year in Table 14, reflect how often the car is away from the house, but could not

Photovoltaics, electric vehicles and energy users

77

be based for calculation of how many trips that are made with the car. In general many commuting trips could be assumed to take place so each time the vehicle is away it could be assumed that it makes at least one trip away from the household, and another back home. The daily distance drive could be validated when compared to the average daily distance from the Swedish Travel Survey [45] where in larger cities, as Stockholm, the distance for main trips by personal vehicle drivers were 27 km. This could be compared with the average daily distance in the scenarios of 14.3 km up to 62.0 km. Table 14. Scenario results Scenario Distance/day Away/year Consumption/day 1 2 3 4 5 6 7 8

14.3 26.2 62.0 28.4 30.3 23.1 42.3 12.4

km km km km km km km km

66 135 344 134 137 134 137 134

2.6 kWh 4.9 kWh 11.3 kWh 5.3 kWh 5.5 kWh 4.3 kWh 7.7 kWh 2.3 kWh

The scenarios are assumed to take place in a city environment where the possibility to choose another way, metro, buses, why the probability that the car is used should be less when away in comparison to a case study in the countryside, where not many other alternatives exists and the probability to take the car when away is higher. Other geographical areas with differing demographics of inhabitants could also be simulated with the SmartLoad model, but this study is delimited to the urban Royal Seaport. 21.3. Area load curve simulation. In this section a different number of apartments with reference Scenario 2 electric vehicles are simulated to illustrate the magnitude fluctuations based on the electric vehicle penetration in areas of different sizes. At first an example of the consumption for one apartment and an EV is simulated, see Figure 37. This load curve is then added up to larger numbers of apartments and additional electric vehicles to give a brief understanding of the impact the number of electric vehicles and households in different areas could induce. The following load curves could also be simulated by aggregating apartments in the model which would result in smoother load curves, but this case study is a simplification which with this approach reduces the number of iterations significantly.

78

P. Grahn, M. Hellgren, J. Munkhammar

1.8

Cold Appliances Cooking Washing Dishwashing Television Computer Audio Lighting Additional Appliances EV

1.6

1.4

Power (kW)

1.2

1

0.8

0.6

0.4

0.2

0

0

5

10 15 Time (Hours)

20

25

Figure 37. ousehold load curve example for 1 apartments with and one Scenario 2 electric vehicle load In Figure 38 the curve shows the apartment simulated without electric vehicles and this is further summarized to show 10 000 apartments load which is the number of apartments that is projected for the Stockholm Royal Seaport area, see Figure 39. It can be seen that the peak load for this example reaches 6 500 kW for 10 000 apartments without electric vehicles. 0.7 0.65 0.6

Power (kW)

0.55 0.5 0.45 0.4 0.35 0.3 0.25 0.2

0

5

10 15 Time (Hours)

20

25

Figure 38. Household load curve example for 1 apartments without EVs

79

Photovoltaics, electric vehicles and energy users

7000 6500 6000

Power (kW)

5500 5000 4500 4000 3500 3000 2500 2000

0

5

10 15 Time (Hours)

20

25

Figure 39. Household load curve example for 10000 apartments without EVs

In Figure 40 the graph show how the load could look like with 1 000 apartments and 100 electric vehicles added to the area whereas Figure 41 illustrates 1 000 apartments with 200 electric vehicles. The difference between the peak load for the examples 40 and 41 turned out to be around 100 kW as seen in the graphs, based on the difference in impact of 100 EVs. Figure 42 shows the load curve for 100 apartments and 10 electric vehicles.

In Figure 43 it can be seen that with a rather low number of vehicles of 200 in an area with 10 000 apartments, this will not have a large impact of the load magnitude. If we instead have a higher penetration level of 1 000 EVs, illustrated in Figure 44, the impact is significant as seen in the graph. If the numbers of EVs is increased furthermore, to an amount that would correspond to an electric vehicle in each parking space in the area, which is 0.5 per apartment, this means 5 000 electric vehicles. This case is illustrated in 45, and shows a non-neglectful impact to the load curve with a peak load increase of 4 500 kW. This should be kept in mind when planning for electric vehicle charging outlets and opportunities for electric vehicle interested residents.

80

P. Grahn, M. Hellgren, J. Munkhammar

800

700

Power (kW)

600

500

400

300

200

0

5

10 15 Time (Hours)

20

25

Figure 40. 1000 apartment household load in blue and 100 Scenario 2 electric vehicle load added to household load in red

800

700

Power (kW)

600

500

400

300

200

0

5

10

15

20

25

Time (Hours)

Figure 41. 1000 apartment household load in blue and 200 Scenario 2 electric vehicle load added to household load in red

81

Photovoltaics, electric vehicles and energy users

80

70

Power (kW)

60

50

40

30

20

0

5

10 15 Time (Hours)

20

25

Figure 42. 100 apartments household load in blue and 10 Scenario 2 electric vehicle load added to household load in red 7000 6500 6000

Power (kW)

5500 5000 4500 4000 3500 3000 2500 2000

0

5

10 15 Time (Hours)

20

25

Figure 43. [10000 apartment household load in blue and 200 Scenario 2 electric vehicle load added to household load in red 22. Household energy consumption and photovoltaic production In this section the results from the simulations of PV production are compared with the household load including the electric vehicle load

82

P. Grahn, M. Hellgren, J. Munkhammar

8000

7000

Power (kW)

6000

5000

4000

3000

2000

0

5

10 15 Time (Hours)

20

25

Figure 44. 10000 apartment household load in blue and 1000 Scenario 2 electric vehicle load added to household load in red

12000 11000 10000

Power (kW)

9000 8000 7000 6000 5000 4000 3000 2000

0

5

10 15 Time (Hours)

20

25

Figure 45. 10000 apartment household load in blue and 5000 Scenario 2 electric vehicle load added to household load in red generated by the SmartLoad model. The PV production estimate was developed by Widé n and based on the SMHI data for hourly insolation in Sweden. The SmartLoad model stochastically generates synthetic activity patterns for two inhabitants where one inhabitant is using an

Photovoltaics, electric vehicles and energy users

83

electric vehicle. As examples of output from the models we have plotted the simulated load for one day each season, and the corresponding PV production that day. 2 1.8 1.6

Power (kW)

1.4 1.2 1 0.8 0.6 0.4 0.2 0

0

5

10 15 Time (Hours)

20

Match 2.5m2 Match 10m2 0.14 %

0.58 %

Figure 46. Plot for one day in January for 10m2 PV area. The solid line represents household load with standard setup and the dashed line the PV production These plots are selected samples from a yearly data set with standard setup. As an alternative to 10m2 PV area we have also considered 2.5m2 PV area. The year average match percentage for the two different scenarios is shown in table 15 below. For more information regarding match percentage see section 17. Match 2.5m2 Match 10m2 7.47 %

20.87 %

Table 15. Match Clearly we see that the increase in match percentage from 2.5m2 to 10m2 PV area is considerable. However, without energy storage of some kind it is impossible to get a perfect match of 100 % with only photovoltaic power when matched with individual household energy consumption because a large percentage of the load is concentrated on the hours when there is no PV production. If the PV system in the local household overproduce during some time the overproduction is either wasted or distributed to the grid unless the energy is stored. In

84

P. Grahn, M. Hellgren, J. Munkhammar 2 1.8 1.6

Power (kW)

1.4 1.2 1 0.8 0.6 0.4 0.2 0

0

5

10 15 Time (Hours)

20

Match 2.5m2 Match 10m2 3.44 %

13.43 %

Figure 47. Plot for one day in March for 10m2 PV area. The solid line represents household load with standard setup and the dashed line the PV production 2 1.8 1.6

Power (kW)

1.4 1.2 1 0.8 0.6 0.4 0.2 0

0

5

10 15 Time (Hours)

20

Match 2.5m2 Match 10m2 9.77 %

33.64 %

Figure 48. Plot for one day in July for 10m2 PV area. The solid line represents household load with standard setup and the dashed line the PV production case the payoff for distributing it to the grid is dependent on the policy of the energy companies in terms of feed-in tariffs.

Photovoltaics, electric vehicles and energy users

85

2 1.8 1.6

Power (kW)

1.4 1.2 1 0.8 0.6 0.4 0.2 0

0

5

10 15 Time (Hours)

20

Match 2.5m2 Match 10m2 2.86 %

11.44 %

Figure 49. Plot for one day in October for 10m2 PV area. The solid line represents household load with standard setup and the dashed line the PV production 23. Electric vehicle load and photovoltaic production This section shows the results regarding the possibility to match the electric vehicle load with the photovoltaic production for the inner layer consisting of the household and one electric vehicle operator. The match coefficient for matching 10m2 PV with only EV is 3.48 %. The average power production was at 3.35 kWh per day and the average power consumption from the electric vehicle was 4.18 kWh per day, see Figure ??. This simulation shows that the level of match over a year is roughly 3.5 % between electric vehicle load generated from the stochastic model and the PV production. Considering the year average plot in figure ?? the match percentage appears higher. The main reason for the poor match is the extremely high variation on both the photovoltaic power production and the electric vehicle power consumption. The extreme intermittence combined with the varying power-demand due to the EV charging becomes a low matching fraction. Charging of the electric vehicle is rather energy intense, in the standard setup it is on the order of 2.3 kW. Demand side management strategies aiming for a better match between EV load and PV production should perhaps focus on reducing the power for charging and relocate the charging as close to mid day as possible. One should remember that this regards the innermost system layer consisting of one household. As will be

86

P. Grahn, M. Hellgren, J. Munkhammar 0.5 0.45 0.4

Power (kW)

0.35 0.3 0.25 0.2 0.15 0.1 0.05 0

0

5

10 15 Time (Hours)

20

Figure 50. The solid line is the year average standard setup electric vehicle load generated by the SmartLoad model and the dashed line is the average PV production shown in the next section the situation is dramatically different for a system containing a large amount of PV production and EV users. Another issue is that the average energy produced by the 10m2 PV system was similar to that of the average electric vehicle load - the load being only about 24.8 % larger - one might assume that a merely a few square meters larger PV array could make the EV-PV system ’net-zero energy’. By that we mean that the total amount of energy produced by the photovoltaic array would be equivalent to the EV load over a year - neglecting the match.

23.1. Aggregate photovoltaic production and electric vehicle load. The SmartLoad model was constructed for simulating a single household and thus does not by design support simulations of aggregate behavior. But as a reasonable approximation of aggregate behavior we have run the SmartLoad model over N number of days without seasonal variation on the vehicle use and electricity consumption. When added up the load produced during these N days were assumed to be equivalent to the load from N models of similar households over one day. The behavior of the Scenario 2 electric vehicle was used, see Section 18.5, but seasonal variation for the electric vehicle was removed and standard setup typically for the summer was adopted. The PV production was averaged over a year. The match was then calculated for the total load of the household including the electric vehicle versus an average photovoltaic production curve. The PV production is from a 10m2 PV array. The average power production was at 3.35 kWh per

Photovoltaics, electric vehicles and energy users

87

household and the average power consumption from the electric vehicle was 5.13 kWh per household, see Figure ??. 0.5 0.45 0.4

Power (kW)

0.35 0.3 0.25 0.2 0.15 0.1 0.05 0

0

5

10 15 Time (Hours)

20

Figure 51. The solid line represents the load of the average household, averaged over 360 households. The dashed line represents the average PV production over one day In this aggregate scenario the match percentage reached around 24.2 %, in comparison with the match percentage for a single household over a year which was about 3.5 % - a near 7 fold improvement. In general this result shows that on a higher system level the power system might be more balanced and local matching might only be of interest to the economy of the local household.

Photovoltaics, electric vehicles and energy users

89

Part 9. Conclusion and future studies The Royal Seaport area will house thousands of residents, all of which will have individual activity- and energy use patterns over time. This will have an impact on the power system locally, especially since local PV power production might be present as well. The estimation of human energy consumption from activities is essential for the design and operation of the power system. In particular when faced with large amounts of intermittent power sources - such as photovoltaics in the power system. Since quantifying predictions regarding human activity- and energy consumption patterns is a complex problem the use of predictive tools based on limiting assumptions is a necessity. For the model of generating synthetic activities - and in turn - synthetic energy consumption patterns Widén& Wäckelgård assembled a stochastic model for calculating household energy consumption based on the assumption that human behavior essentially could be approximated by a Markov Chain process [62]. The W&W model did not take into account the potential energy consumption from electric vehicle use. In this project the SmartLoad model was developed as an extension to the W&W model for incorporating electric vehicle load from synthetic activity generation in the W&W Markov Chain model.

24. End-user related In the interviews a few things have come to light. Part lies in line with the research question stated earlier (see 4). Some aspects also arouse concerting photovoltaic arrays and some minor ones regarding electrical vehicles which are discussed below as well. 24.1. The inhabitants. Expectation on the inhabitants from the various interviewee are rather consistent. The inhabitants are generally seen as coming from the upper economic echelons. Partly due to the cost of newly produced dwellings but also due to the process of soil purification and the sustainability construction increase the cost for the projects. A cost that the constructors will need to redeem via increased costs for the residents. A common perspective is also that the future residents will have view of sustainability more implicitly than explicitly. The fact that the area is planned and designed with sustainability is more a ‘bonus’ or positive aspect rather than a requirement. While there will certainly be groups that gravitate themselves towards the area due to the green profile the main groups seem to be seen as more focused on finding an attractive home.

90

P. Grahn, M. Hellgren, J. Munkhammar

Given this there is however still a common perception that residents are concerned with the environment, considering this questions as important and have the desire to live sustainable as long as this does not have a negative impact on their standard of living. The approach from the actors seems to take this in mind. By designing living environments in such way that acting in sustainable way is easier and rewarded compared to non-sustainable life styles the inhabitants can be lead towards, but not forced, a green life style. The tolerance for spending a larger proportion of the income on living can be argued to have a positive effect on the construction of sustainable living. The construction of sustainable building is more expensive than comparable non-sustainable buildings during the construction phase. There seems to be an openness towards actions taken to lower the energy use of building from residents and actions taken to lower the impact on the environment rarely meet opposition. Rather there seems to be a high degree of acceptance. Another reoccurring theme is that the inhabitants want to improve their life style towards a more sustainable one but lack the tools and knowledge of how to do this. Simpler thing in every day life are usually covered but what impact and effect their own housing and life style have is lacking. Several parts of the rising home technology development is aimed towards this. By presenting the users with information that is accessible and clear they can easier arrange their life style towards a more sustainable life style. A conclusion that can be argued is that the residents are willing but not always able to live sustainable. Demands outside the green questions require considerations as well. Living a full life is of social importance and while a green life style can offer some social status the low impact life styles are usually considered more extreme and requires cut-backs in living standards than most would be willing to sacrifice. It can be argued that society is ultimately about belonging and when the ‘base’ form of social life require living in way that impacts our environment the environmental concerns will still remain, be acted upon but more easily brushed aside. It can be argued that people wish to eat the cake and still have it. Retaining a rich and varied life while still having a low environmental impact. The developments of present seem to point towards that end-users of the technological system will be armed with another component that allows them weight their choices and more clearly see the extent of their actions. The technologies, end-users comfort and increased experience with systems can potentially have beneficial effects for leading them towards a more sustainable life style.

Photovoltaics, electric vehicles and energy users

91

24.2. End-users and technology. The technological competence of the inhabitants is a question that was raised. The answers given by the interviewees points towards that the end-users are wary of the new technologies. Wariness that is more aimed towards the use of the system rather than the systems themselves. Some technologies require minor or at times larger changes in living activities, such as keeping widows open negatively impact the ventilation system, which the users may not always be aware of. The key factor here is information and demystification of the technologies. The new areas incorporate new technologies that require slight alteration and more awareness of the system than traditional housing. Plans exist to arrange meetings where the residents can meet representatives of the land lord, constructor and the municipal to be given information but also ask questions. Meetings with the residents have been used in the past to great effect. The meetings have been appreciated and have had good feed-back. The aim for these systems of information appears to be aimed at making it easier for the end-users to gain more accurate information about how their activities effect their energy use. The goal seems to be to make the information readily available and incorporated into the common, daily, activities thus making it natural to take energy use into consideration. By making the new technologies a common and daily part of every day life the idea seem to be that the effects will be more to adjustments by the end-users themselves rather than a form of enforcement.

24.3. Solar power. Of the solar alternatives solar thermic energy was the only considered plausible in contemporary times. Photovoltaic arrays were considered to expensive to implement in the current construction. There has however been hints that some buildings are constructed with the capacitry to add some solar based power in the future. A hint that can be detected is that there is also a lack of larger demonstration projects that displays the effects and potentials of photovoltaic arrays in a significant way. Building houses are expensive projects and investing in a ‘untested’ technology to power parts of your structure may not be in the interest of the owner nor the constructor. A discussion can also be done considering the maturity of photovoltaic arrays. The arrays are seeing a strong development in recent years. International interest, research and a focus on renewable energies have all contributed to the development. Economic models in countries such as Germany has also been helpful. This does however mean that the

92

P. Grahn, M. Hellgren, J. Munkhammar

technology is making headway and arrays developed last year is potentially less efficient then those of the current year. An argument can be made that the current arrays will in a time frame for house construction investments develop too fast to be worth investing in as more potent versions will soon be available. While the argument can be parallelled to that of computers, which also have a high speed of development, there is a key difference, computers are hard to replace by other tools whereas the electricity produced by photovoltaic arrays is no different than that which comes from fossil fuels, nuclear, wind or water. For the Royal Seaport area a potential conflict in the area is due to the existence of Värtaverket. The chp plant produces heat via the remote heating system and electricity. Locally produced heat via solar thermal energy would directly lower the energy required for the remote heating system. A peculiar problem is that the efficiency of solar energy is highest during the summer months, a period when remote heating is less required. Värtaverket uses garbage incineration for part of the heat produced. Storage of garbage for an extended period of time may be problematic, particulary considering the location of the plant. Either local storage is required, which all the issues of storing garbage for combustion in a parts of a city area entails, or it will require higher degree of transportation during the colder periods. While transportation are required none the less, storage of garbage in an off site location may offer some challenges from a environmental perspective. 24.4. Electrical Vehicles. Electrical vehicles were more brought up in parenthesis the explicitly. Part of this can be attributed to that car travel is increasingly seen as an undesirable form of transportation and increase or status quo in car travel is perceived negatively. Public transportation are more in line with the design and planning of the areas. Electrical vehicles however have two places. One is in the form of electrical tram and a potential for using electrical vehicles in car pools. The technical parts of this study indicates that photovoltaic arrays in combination with car pools with electrical car have a good match. Car pools users also tend to use the vehicle outside the optimal loading time for this technological connection. 24.5. Final thoughts. The catch-phrase for the municipal when it comes to sustainability questions, ‘lätt att göra rätt’ (easy to do right) seems like a successful approach so far. By shaping the area so that sustainable life styles are easier then less such the residents are encourage to act in a sustainable way. The big question for the future seems

Photovoltaics, electric vehicles and energy users

93

to be what effect the information system will have on the actions of the end-users. 25. Photovoltaics and load matching One of the main goals with the SmartLoad model in this project was to be able to match household energy consumption, including electric vehicle load with the local PV production. This was mainly done locally for the innermost system layer considered in this project: a single household with two residents. The outer system limit consisting of the entire Royal Seaport was also simulated to an extent as an aggregate of several households. For the local matching two scenarios were considered for the PV-system: 2.5m2 and 10m2 . The match between household load including electric vehicle load and PV production was on a yearly average about 7 percent for the 2.5m2 option and about 21 percent for the 10m2 option. The match between the electric vehicle and PV production over a year for a single household was very poor at 3.5 percent for the 10m2 photovoltaic array option. The main reasons for the poor match between the electric vehicle load and PV production were the following: - The electric vehicle load appears to be strongest during evenings, nights and mornings. At these hours PV production is poor. - The electric vehicle load was strongly intermittent and when there was load it was very large. This caused problems in matching since PV arrays only produce power during day time with limited peak power. The poor match between electric vehicle load and PV production suggests a large potential for improvement. For example economic incentives for charging the electric vehicle during daytime. Another possibility for increased match between photovoltaic production of electricity and electric vehicle load is to charge the electric vehicle during work hours. Another option is to consider the aggregate demand from a large amount of households alongside a large amount of photovoltaic power production. This was briefly considered in this project via the simulation of 360 households. The setup was 360 households, each with an electric vehicle and each with a 10m2 PV panel. The match of electric vehicle load with PV production in this ’outer layer’ was considerably higher in the aggregate scenario, with about 24 % match between electric vehicle load and PV production. The main difference - compared with the inner layer scenario which had 3.5 % - is that the aggregate electric vehicle load is more smooth typically, which limits the intermittency and average peak load. The conclusions can be summarized as follows:

94

P. Grahn, M. Hellgren, J. Munkhammar

• Since there is no clear directive for the setup of photovoltaics in the Royal Seaport we considered two scenarios. 2.5m2 and 10m2 photovoltaic arrays for a single apartment. • The level of match between photovoltaic power production and electricity consumption from activities (including electric vehicle charging) was about 7 percent for the 2.5m2 scenario and about 21 percent for the 10m2 scenario on average over a year. • The level of match between the photovoltaic power production and the electric vehicle consumption was very poor at about 3 percent for a one year average with the 10m2 photovoltaic array. • The level of match for one day aggregate production of 360 households were about 24 percent. Indicating that at an aggregate system level the match was considerably higher. 26. Electric vehicles In this report we have simulated a future electric vehicle use based on how electric vehicle drivers would behave according to activity data of how people behave today, with the additional condition that when residents leave there home there will exist a probability that the electric car is being used. The simulation of the electric vehicle charging behavior in the scenarios, allowing variation in behavioral parameters, made it possible to visualize the variations in the end-user behavior, and thus vehicle charging pattern. This further made it possible to estimate the corresponding impact to the electric power system in different area sizes and plausible load curves. From the average energy consumption of a period of a year, when split up into household load from activities and electric vehicle load load from charging, it was clear that the electric vehicle load was non-negligible, and that with higher charging power the fluctuations in the load curve was increased. In fact in the standard setup, in reference Scenario 2, it could be compared to the same energy amount consumed in total for an apartment with two residents and corresponding household load due to the activity patterns. Consequently, with an electric vehicle in the household, the EV becomes the single most energy consuming activity in the household. An electric vehicle fleet in the Royal Seaport area would induce a load curve with higher peak load and overall load curve depending of penetration level. With 200 electric vehicles and 10000 apartments there are no significant impact but with over 1000 electric vehicles in the same area have a distinct affect to the load. Load curves induced by electric vehicles are not that easy to match with solar power. With electric car pools, or work parking spaces, with parked cars that

Photovoltaics, electric vehicles and energy users

95

could charge during daytime, the matching rate could however become higher. The visions to become a fossil fuel independent Royal Seaport area in 2030, with the use of gasoline, oil, coal, gas and other fossil fuels phase out was also considered. A related target in the Royal Seaport is that the carbon dioxide emissions should be reduced from the average 3.4 ton CO2 /inhabitant and year in 2010 to 1.5 tonnes CO2 /inhabitant and year in 2020. How the process of reaching these targets and the consequences of doing so is still unknown and with no actual incentives planned for an introduction of an electric vehicle fleet in the Royal Seaport area, the opportunity for electric vehicles to become a large part of this shift is yet to be seen. The closest ambition are plans for slow charging poles in the entire area, but unfortunately no numbers of poles are mentioned, only that it should be demand-driven. This fact could become a barrier for the willingness of residents to own an electric vehicle if there lies an uncertainty with charging possibilities. Another ambition is the subvention which will allow a limited number of inhabitants in Sweden from January 2012 to 2014 to buy in total 5000 environmental friendly vehicles with a cost reduction of 40000 SEK, which is set to improve peoples ability to purchase an electric vehicle despite a rather high initial cost. The analysis in the interview indicates that the opportunity for photovoltaic installations and electric vehicle introduction in the Royal Seaport area will be the end users responsibility, which in reality imply that there are no actual better practical conditions supporting an introduction of electric vehicles or installation of photovoltaics in the Royal Seaport area in comparison to any other area.

27. Main conclusion The actors creating the area, the city of Stockholm, developers, electricity companies and other actors involved will try to interact with each other to make it possible to live in an environmental friendly area with additional photovoltaic and electric vehicles. However, if they succeed or not with this mission, the main conclusion from the interdisciplinary reasonings based on both interview and simulations in this report, is that the responsibility to create an environmental friendly area in the Royal Seaport will lie upon the shoulders of the future individual inhabitants moving into the area. The use of the projected environmental friendly new technology will thus depend on sustainable awareness and attitudes of the inhabitants.

96

P. Grahn, M. Hellgren, J. Munkhammar

28. Future research The SmartLoad model was designed - on the basis of the W& Wmodel - as a stochastic model for generating synthetic load based on data regarding what activities people perform. One feature that the model does not handle is the potential flexibility of the end user. Thus a model of energy consumption from activities which is dependent on demand side management strategies - such as economic incitements for example - would be interesting to construct. With such an extension the model could then also be used to estimate the flexibility of electric vehicle use - and consequent load thereof. In particular this could be interesting in connection with the PV production - both locally and nation wide. In future studies it would be interesting to run the model with SCB activity data from the newly conducted survey, which was completed in late 2011. Future research not covered by this project which would be interesting to study are thus summarized here: • Run the model with SCB activity data from the new survey, which was completed in autumn 2011, to achieve fresh results of people electricity consumption which is based on more realistic data due to behavioral changes over the years. • Investigate if is there a possibility to implement load matching by using vehicle to grid technology by discharging batteries when consumption is high, and implement this technology in a Royal Seaport case study. • Study which opportunities, aspects and constraints there are for V2G technology in the Royal Seaport area. • Study if it will become possible to charge the vehicle when the demand is low and hereby also the prices are low. • Study the end-users perspective on the information technology. • Closer study the end-users views and perspectives on their travel habits and perceived travel needs.

Photovoltaics, electric vehicles and energy users

97

Appendix A. Interview guide In the guide below the abbreviation nds is used for N orra Djurgårdsstaden. Inledande frågor Vem är du? • Namn • Organisation – ”Vilken position har du i organisationen?” – ”Vilka uppgifter har du?” – ”Hur länge har du arbetat med dessa?” • ”Vad har du arbetat med tidigare?” Vilken är din roll i din organisation? • I samband med nds • Utanför nds Betydelsen för den egna organisationen • ”Vad ser ni (som organisation) som ert huvudsakliga intresse i nds?” • ”Vilka är din organisations förhoppningar pånds?” Området Byggnationen/stadsplaneringen • ”Vilka tankar finns det övergripande för området?” Verksamheter i området • ”Finns en övergripande plan?” – Butiker – Kontor – Produktion – Andra utrymmen Transporter i området • ”Hur ser ni påtransporter i nds”? – Kollektivtrafik – Cykel/gång – Biltrafik ∗ Bilpooler

98

P. Grahn, M. Hellgren, J. Munkhammar

· ”Något som diskuterats?” · ”Finns det utrymme för bilpooler?” – ”Inverkar detta påert arbete?” Energi • Produktion inom området? – ”Finns det planer påelproduktion inom området?” ∗ Solceller? • "Vilka tankar finns det om elproduktion utanför området?” – Grön el? • "Vilka möjligheter finns det för eldrivna bilar?” – ”Vem ser du har ansvar för detta?” Miljötekniker • ”Vilka miljötekniker känner du till som planeras för nds?” • ”Hur tror du de kommer att fungera?” • ”Vilka har ansvaret för miljön?” Miljö • ”Vad tänker du pånär du hör miljö?” • ”Vilka är de stora miljöåtgärderna?” • Utrymmen för miljön Andra projekt som kan utnyttjas i nds • ”Har ni (som orgnaisation) dragit lärdomar från andra projekt?” – (Som kan användas i det här) – Vilka då – Påvilket sätt • ”Har tidigare projekt fått inverkan påert arbete i nds?” Boende Vilka är de? • ”Vilka förväntningar har din organisation påde boende i nds?” • ”Det har pratats om speciella krav påde som flyttar till nds.” – ”Stämmer detta?”

Photovoltaics, electric vehicles and energy users

99

– ”Vilka är de?” • ”Vilka tror ni kommer att flytta till nds?” – Ekonomi – Miljöintresse – Närhet till centrum • ”Vad tror du gör nds intressant för dessa?” • ”Finns det någon diskussion om ’sammansättningen’ av boende?” Hur agerar de? • ”Hur passar de in i området?” – Bekvämlighet – Ekonomi – Miljömedvetenhet – Beteenden – Vanor ’Grönt liv’ • Hur är området anpassat till de boende – ”I vilka avseenden är området anpassat till att de kan leva utan att belasta miljön?” • ”Vilka resonemang finns det inom och mellan organisationer, företag och andra intressenter?” Resor • ”Vilka dagliga resor tror du att de gör?” – Sträcka – Färdmedel • ”Hur hoppas ni att de kommer att resa?” • ”Hur tror ni att de kommer att resa?” – Bil ∗ Elbil ∗ Hybrid ∗ Bilpooler – Kollektivtrafik

100

P. Grahn, M. Hellgren, J. Munkhammar

– Gå – Andra transportsätt • ”Service inom området?” Avslutande reflektioner • ”Finns det något som vi inte tagit upp som du tycker är viktigt?” • ”Tack för du deltog!”

Photovoltaics, electric vehicles and energy users

101

References [1] B.Stridh B.Karlsson A.Molin, J.Widén. Konsekvenser av avräkningsperiodens längd vid nettodebitering av solel. Technical report, Elforsk rapport 10:93, 2010. [2] G. Blom. Sannolikhetsteori med tillämpningar. Studentlitteratur, 1984. [3] S. D. Breucker, P. Jacqmaer, K. D. Brabandere, J. Driesen, and R. Belmans. Grid power quality improvements using grid-coupled hybrid electric vehicles. In The 3rd IET International Conference on Power Electronics Machines and Drives, 2006. [4] Andreas Dinger, Ripley Martin, Xavier Mosquet, Maximilian Rabl, Dimitrios Rizoulis, Massimo Russo, and Georg Sticher. Batteries for electric cars - challenges, opportunities and the outlook to 2020. Technical report, The Boston Consulting Group, 2010. [5] ELFORSK. Miljövärdering av el med fokus på utsläpp av koldioxid. In www.elforsk.se, 2011-09-01. [6] Encyclopedia Britannica. Renewable energy. 2011. [7] Energiledargruppen. Nätverk sverige. Möte 2011-01-25. [8] Marika Engström. Transportsektorn idag - om resande och transportmönster. Technical report, Statens institut för kommunikationsanalys SIKA Rapport 1998:3, 1998. [9] Fortum. Kol ersätts med biobränsle vid värtaverket – fortums plan till år 2020. In http://media.fortum.se/2009/11/25/kol-ersatts-med-biobransle-vidvartaverket-fortums-plan-till-ar-2020, 2009. [10] Frederik Geth, Koen Willekens, Kristien Clement, and Johan Driesen. Impactanalysis of the charging of plug-in hybrid vehicles on the production park in belgium. In MELECON 15th IEEE Mediterranean Electrotechnical Conference, 2010. [11] Pia Grahn and Lennart Söder. The customer perspective of the electric vehicles role on the electricity market. In 8th International Conference on the European Energy Market, 2011. [12] Sekyung Han, Soo Hee Han, and Kaoru Sezaki. Development of an optimal vehicle-to-grid aggregator for frequency regulation. IEEE Transactions on Smart Grid, 1:65–72, 2010. [13] HETUS. Hetus introduction. In www.h2.scb.se/tus/tus/introduction2.htmlGuide2, 2010. [14] Iqbal Husain. Electric and hybrid vehicles: Design fundamentals, second edition. CRC press, 2011. [15] J. Widén. Stochastic modeling and simulations. Working Paper no. 45 "Interdisciplinary Energy System Methodology: A compilation of research methods used in the Energy Systems Programme", Linköping University (2011). [16] J. Widén. Distributed Photovoltaics in the Swedish Energy System. PhD thesis, Lic. Thesis Uppsala University, 2009. [17] J. Widén. System Studies and Simulations of Distributed Photovoltaics in Sweden. PhD thesis, Uppsala University, 2010. [18] J. Widén and E. Wäckelgård. Net zero energy solar buildings at high latitudes: The mismatch issue. In EASST Conference, 2010. [19] J.Widén. Correlations between large-scale solar and wind power in a future scenario for sweden. IEEE Transactions on Sustainable Energy, Vol. 2, No. 2:177–184, 2011. [20] J.Widén et al. Impacts of time averaging on statistical analysis of photovoltaic generation, domestic electricity demand and distribution grid voltages. Manuscript (2009).

102

P. Grahn, M. Hellgren, J. Munkhammar

[21] Håkan Sköldberg, Ebba Löfblad, David Holmström, and Bo Rydén. Ett fossiloberoende transportsystem år 2030 ett visionsprojekt för svensk energi och elforsk. Technical report, Elforsk rapport 10:55, 2010. [22] K.Ellegård and E.Wäckelgård. Energianvändning i bebyggelse - boendes och aktörers val av teknik. In Universitetstryckeriet Uppsala Universitet, 2007. [23] Nasser H. Kutkut, Deepak M. Divan, Donald W. Novotny, and Raymond H. Marion. Design considerations and topology selection for a 120-kw igbt converter for ev fast charging. IEEE Transactions on Power Electronics, 13:169 – 178, 1998. [24] Steinar Kvale. Den kvalitativa forskningsintervjun. In Studentlitteratur, 1997. [25] Liu Yongxiang, Hui Fuhui, Xu Ruilin,Chen Tao, Xu Xin and Li Jie. Investigation on the construction mode of the charging station and battery-exchange station. In Asia-Pacific Power and Energy Engineering Conference, 2011. [26] M.A.Green. Solar cells. In Prentice-Hall, inc, 1982. [27] Vincenzo Marano, Simona Onori, Yann Guezennec, Giorgio Rizzoni, and Nullo Madella. Lithium-ion batteries life estimation for plug-in hybrid electric vehicles. In IEEE Vehicle Power and Propulsion Conference, 2009. [28] John M. Miller. Energy storage system technology challenges facing strong hybrid, plug-in and battery electric vehicles. In IEEE Vehicle Power and Propulsion Conference, 2009. [29] A. Millner. Modeling lithium ion battery degradation in electric vehicles. In IEEE Conference on Innovative Technologies for an Efficient and Reliable Electricity Supply, 2010. [30] P Mitra, G K Venayagamoorthy, and K Corzine. Real-time study of a current controlled plug-in vehicle for vehicle-to-grid transaction. In International Power Electronics Conference (IPEC), 2010. [31] Naturskyddsföreningen. Fj ärrv ärmens svarta lista presenteras. In http://mdgs.un.org/unsd/mdg/SeriesDetail.aspx?srid=751, 2008. [32] Norra Djurgårdsstaden. Övergripande program för miljö och hållbar stadsutveckling i norra djurgårdsstaden. Technical report, Royal Seaport, 2011. [33] Örjan Larsson. Ladda för nya marknader, elbilens konsekvenser för elnät, elproduktionen och servicestrukturer. Technical report, VINNOVA Analys, 2010. [34] Micheal Quinn Patton. Qualitative research & evaluation methods. In Sage Publications, 1997. [35] T. Persson. Measurements of solar radiation in sweden 1983-1998. In SMHI, 2000. [36] PG&E. Pacific gas and electric company energizes silicon valley with vehicleto-grid technology. In PG&E News Department (415), 2007. [37] Program Energisystem. 2011-08-25. www.liu.se/energi/welcome?l=sv. [38] Saman Rashid. Fordon tema yrkestrafik. Technical report, SIKA, 2008. [39] REN21. Renewables. In Global Status Report, 2010. [40] Mélaine Rousselle. Impact of electric vehicle on the electric system. Master’s thesis, RTE and KTH, 2009. [41] Royal Seaport. stockholm.se/fristaende-webbplatser/fackforvaltningssajter/exploateringskontoret/ovrigabyggprojekt-i-innerstaden/hjorthagen-vartahamnen-frihamnen-loudden/enmiljostadsdel/. 2011-09-01. [42] Claes Sandels, Ulrik Franke, Niklas Ingvar, Lars Nordström, and Roberth Hamren. Vehicle to grid - monte carlo simulations for optimal aggregator strategies. In International Conference on Power System Technology, 2010. [43] Brita Saxton. Uppföljning av de transportpolitiska målen. Technical report, Trafikanalys, Rapport 2011:1. [44] P. Schavemaker. Electrical power system essentials. In Wiley, 2008.

Photovoltaics, electric vehicles and energy users

103

[45] SIKA. Res den nationella resvaneundersökningen. Technical report, Tabellbilaga, 2005–2006. [46] Solelprogrammet. Ekonomiska frågor. In www.solelprogrammet.se, 20110620. [47] Stockholm Stad. Vision 2030. In Norra Djurgårdsstaden, 2009. [48] Stockholm stad. Norra djurgårdsstaden. In www.stockholmroyalseaport.com/overview-2/, 2011. [49] Stockholm stad, Exploateringskontoret. Nyhetsbrev norra djurgårdsstaden 2010:1. In www.stockholm.se/Fristaendewebbplatser/Fackforvaltningssajter/Exploateringskontoret/Ovriga-byggprojekti-innerstaden/Hjorthagen-Vartahamnen-Frihamnen-Loudden/NorraDjurgardsstadens-nyhetsbrev/Nyhetsbrev-1—2010, 2010. [50] Stockholms Hamnar. Våra hamnar. In www.stockholmshamnar.se/sv/Varahamnar/Stockholm1, 2011. [51] B. Stridh. Hur kan solceller bidra till husets egenanvändning av el? Royal Seaport kickoff presentation (2010). [52] Svensk Energi. Svenska elnätet. In www.svenskenergi.se/sv/Om-el/Elnatet/", 2011-02-25. [53] Sweco. Hjorthagen in royal seaport, 2011-09-14. www.stockholm.se/Fristaendewebbplatser/Fackforvaltningssajter/Exploateringskontoret/Ovrigabyggprojekt-i-innerstaden/Hjorthagen-Vartahamnen-Frihamnen-Loudden/Inenglish1/Norra-Djurgardsstaden/. [54] Trafikanalys. Antal personbilar i trafik ökar mest i Örebro län. Pressmeddelande (2011-02-22). [55] Trafikverket. www.trafa.se. 2011-09-01. [56] Transek. Stockholmsregionens framtida oljef örs örjning - etapp iii - slutrapport. In http://media.fortum.se/2009/11/25/kol-ersatts-med-biobranslevid-vartaverket-fortums-plan-till-ar-2020/, 2006. [57] United Nations Statitics Division. Millennium development goal indicators. In United Nations, http://mdgs.un.org/unsd/mdg/SeriesDetail.aspx?srid=751, 2011. [58] USKAB. Områdesvis statistik, Östermalm. In www.usk.stockholm.se/internet/omrfakta/omradesvis.asp?omrade=10, 2011. [59] Katerina Vrotsou. Everyday mining - exploring sequences in event-based data. In Linköping University, 2010. [60] J. Widén. Distribuerad och intermittent elproduktion i kraftsystemet. Course lecture for "Elektriska nät som system" at Uppsala University (2010). [61] J. Widén. Options for improving the load matching capability of distributed photovoltaics: Methodology and applications to high-laditude data. Solar Energy, Vol.83, Issue 11:1953–1966, 2009. [62] J. Widén and E. Wäckelgård. A high-resolution stochastic model of domestic activity patterns and electricity demand. Applied Energy, Volume 87, Issue 6:1880–1892, 2010. [63] Wikipedia. Hjorthagen. In http://sv.wikipedia.org/wiki/Hjorthagen, 2011. [64] Wikipedia. Stockholm. In http://sv.wikipedia.org/wiki/Stockholm, 2011. [65] Steve Woolgar. Configuring the user: the case of usability trials. Routledge, 1991. [66] www.stockholmroyalseaport.com. Norra djurgårdsstaden. 2011-09-01.