Typology of centralised and decentralised visions for electricity infrastructure

Utilities Policy xxx (2016) 1e8

Contents lists available at ScienceDirect

Utilities Policy journal homepage: www.elsevier.com/locate/jup

Typology of centralised and decentralised visions for electricity infrastructure Simon Funcke a, b, *, Dierk Bauknecht a a b

€ Oko-Institut e.V., Institute for Applied Ecology, Merzhauser Str. 173, 79100 Freiburg, Germany Centre for Renewable Energy (ZEE), University of Freiburg, Tennenbacher Str. 4, 79106 Freiburg, Germany

a r t i c l e i n f o

a b s t r a c t

Article history: Received 26 June 2014 Received in revised form 22 March 2016 Accepted 22 March 2016 Available online xxx

Scientific and public controversies about the design of future electricity systems can be observed, including differences around centralised and decentralised approaches. Taking the German case as an example, we develop a typology of (de)centralisation that distinguishes between (1) infrastructure location (connectivity and proximity), and (2) infrastructure operation (flexibility and controllability). This typology is applied to two competing visions for the future of electricity infrastructure. A differentiated view of the various dimensions can contribute to the current debate, clarify visions for development paths, and inform infrastructure governance. © 2016 Elsevier Ltd. All rights reserved.

Keywords: Decentralisation Centralisation Electricity infrastructure Visions

1. Introduction Scientific and public controversies concerning centralisation or decentralisation can be observed in various utility sectors (Konrad et al., 2008; Truffer et al., 2008). Techno-economic advantages and disadvantages of centralisation and decentralisation (from here on: (de)centralisation) and the governance processes shaping the potential system transformation are at the centre of these debates. This contribution focuses on the techno-economic dimensions of the electricity infrastructure. On the pathway to a cleaner and more sustainable electricity system, increasing amounts of renewable energy sources for electricity (RES-E) generation have been introduced in many countries in recent years. Rising capacities of RES-E power plants with variable output, such as wind turbines or photovoltaic systems, affect and potentially transform the entire electricity system, as can be seen in countries such as Germany and Denmark (Lund et al., 2012). We take Germany as an example, where RES-E generation has strongly increased in recent years and in 2014 reached a share of almost 28 percent in the electricity mix (BMWi, 2015). Concerning the future development of the electricity system, a consensus seems to exist among most actors in Germany

* Corresponding author. Centre for Renewable Energy (ZEE), University of Freiburg, Tennenbacher Str. 4, 79106 Freiburg, Germany. E-mail addresses: [email protected] (S. Funcke), d.bauknecht@ oeko.de (D. Bauknecht).

that RES-E technologies will become the primary source for generation. The strongest disagreements can be found with regard to the time needed for this shift and the actual design of the RES-E system, which includes the question of whether the infrastructure should be centralised or decentralised. Transition scholars define the electricity system as a sociotechnical system that not only consists of the physical infrastructure but that is also strongly influenced by social structures and coevolves with relevant actors and institutions (e.g. Geels, 2002; Goldthau, 2014; Loorbach et al., 2010; Smith et al., 2005). While we acknowledge this perspective, we focus here on the key technoeconomic dimensions of the electricity infrastructure. We argue that a simple dichotomy between decentralised and centralised infrastructure cannot capture the full range of concepts that pertain to electricity infrastructure development. Electricity infrastructure could be organised in a completely centralised or decentralised manner. More likely, however, is a simultaneous combination of centralised and decentralised designs. Furthermore, a single solution can incorporate centralised as well as decentralised characteristics; for example, centralised and decentralised power generation technologies can co-exist within an electric power system. We analyse the following infrastructure dimensions where centralisation or decentralisation can take place: (1) infrastructure location (connectivity and proximity of generation facilities), and (2) infrastructure operation (flexibility and controllability, including reliance on market mechanisms). The social and political

http://dx.doi.org/10.1016/j.jup.2016.03.005 0957-1787/© 2016 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Funcke, S., Bauknecht, D., Typology of centralised and decentralised visions for electricity infrastructure, Utilities Policy (2016), http://dx.doi.org/10.1016/j.jup.2016.03.005

2

S. Funcke, D. Bauknecht / Utilities Policy xxx (2016) 1e8

dimensions of electricity systems that are closely related to the question of (de)centralisation, namely issues of system governance and democratic control, are touched upon briefly in the second section but are not the focus of this contribution. We aim to develop a typology that clearly delineates key techno-economic dimensions of electricity infrastructure and describes potential (de)centralised infrastructure designs for the different dimensions. To demonstrate the typology, we apply it to two related smart-grid1 demonstration projects and their respective techno-economic vision as well as a vision of a “SuperSmartGrid” for a future electricity infrastructure. While German actors within the electricity system usually refer to the term ‘decentralised energy system’ when discussing transformation, the terms ‘distributed’, ‘on-site’, ‘embedded’ or ‘dispersed’ can also be found in the international literature (Ackermann et al., 2001; Pepermans et al., 2005; Alanne and Saari, 2006). In this contribution, we use the term ‘decentralisation’. The article is structured as follows. The following section presents incumbent as well as new actors in the electricity system and their positions on, and visions for, the layout of electricity infrastructure, which allows an understanding of the current context of the (de)centralisation discussion. In section 3, we examine various dimensions of the infrastructure, comparing decentralised with centralised design, highlighting interconnections between the dimensions and providing the structure of the typology. We use the developments in Germany as an example and consider how international developments influence the national level. On this basis, in section 4, we apply the (de)centralisation typology by considering two competing visions for the future of the electricity infrastructure. We finish with some concluding remarks on (de)centralisation and an outlook. 2. Competing visions of the future German electricity system Incumbents on the one hand, and new actors and coalitions2 on the other, pursue different agendas and have diverging if not competing visions for the future electricity system. The question of the degree of (de)centralisation arose in the wake of energy price increases and environmental concerns in the late 1970s and early 1980s (e.g. Lovins (1977) in the USA and Krause et al. (1980) in Germany). As explained in the literature on system transformation and its governance, such visions play an important role in the transformation process. The relevant literature provides insight into the workings of large socio-technical systems and possible transition pathways (Geels and Schot, 2007, 2010). Distinctive visions have already influenced the development of the electricity system for many decades, as Smith et al. (2005) describe for the introduction of nuclear power. A vision of the future can help to mobilise and coordinate actors and resources in the transformation process and provide a stable framework for target setting, but it also functions to support or criticise the status quo (Rotmans et al., 2001; Smith et al., 2005; Berkhout, 2006; Sp€ ath and Rohracher, 2010). The relevance of RES-E to visions for the German electricity system has dramatically increased since the 1970s, and especially since the current feed-in scheme was put in place in 2000. Moreover, there seems to be a consensus among most actors that RES-E will be the dominant source for electricity generation in the future. However, within this consensus there remain proponents of both the centralised and the decentralised vision and

1 By smart grid, we mean the introduction of information and communication technology (ICT). 2 For more details on actors and actor coalitions in the field of renewable energies in Germany, see Hirschl (2008) and Dagger (2009).

thus alternative transition pathways (cf. Verbong and Geels, 2012). The main arguments associated with these visions are presented here. The vision of a mostly centralised system is based on large-scale power plants and balancing measures. Traditionally, nuclear and fossil fuels were essential to this system, but with the political decision to phase out nuclear energy in Germany by 2022 and the goal to decrease greenhouse gas emissions, larger shares of RES-E are expected for the future. Among this group of visionaries are scientists and consultancies that argue for a trans-European reinforcement of transmission grids (e.g. Helm, 2014; Czisch, 2011; PwC et al., 2010; PwC et al., 2011) as well as utilities that invest in largescale power plants. They often ascribe the growing importance to renewables to sites in Europe and North Africa with the highest load factors and the lowest costs for large-scale RES-E deployment, including offshore wind parks and solar systems. Representative projects include Desertec (e.g. Pudlik et al., 2012) as an example of electricity generation, or the interconnector between England and the Netherlands, BritNed,3 as an example of an advancing Europewide transmission grid integration (an example of a centralised vision is provided in section 4.2). Over the past decades, starting with the environmental movement, other actors, supporting a decentralised vision for the electricity system, have entered the energy market (Mautz, 2007; Wissner, 2011). The introduction of feed-in tariffs (FiT) spurred this development as it allowed new actors with less financial resources to invest in generation. This is due to the design of FiTs, which guarantees a calculable remuneration for the electricity produced by smaller-scale and less capital-intensive RES-E power plants. Among these visionaries are private citizens,4 politicians (e.g. Scheer, 2010), NGOs (Paulitzk, 2006), RES-E interest groups (BEE, 2011; Eurosolar, 2012), local initiatives or energy cooperatives (Hauber and Ruppert-Winkel, 2012) as well as companies from the ICT-sector, manufacturers of RES-E equipment, new electricity suppliers that were founded after electricity market liberalisation, farmers or project planners (Mautz, 2007; Wissner, 2011; Erlinghagen and Markard, 2012). While environmental concerns play an important role in the argumentation of these actors, they also stress the importance of regional energy structures as well as the relevance of renewables in self-sufficiency scenarios. They support their position with arguments linked to economic, social and political concerns (such as additional regional added value, a wider distribution of profits, reduction of market power, and a stronger democratic control of the electricity system; an example of a decentralised vision is provided in section 4.1). Many actors consider smart-grids as an essential technological aspect of this vision, flowing from early conceptions of decentralised electricity systems (Lovins, 1977; Krause et al., 1980) with a new emphasis on ICT-technologies and support from ICT interests. The discourse concerning (de)centralisation of the electricity system is mainly driven by the actors and divergent positions described above. Our typology can inform the scientific and public debate because too often only certain dimensions or technologies of the system (such as generation or grids) are considered, while others (such as proximity or storage and demand-side management) are neglected.

3

For more information see: www.britned.com. Private citizens potentially influence the electricity sector as investors in RES-E equipment, as consumers and as voters of political parties that represent their interests. 4

Please cite this article in press as: Funcke, S., Bauknecht, D., Typology of centralised and decentralised visions for electricity infrastructure, Utilities Policy (2016), http://dx.doi.org/10.1016/j.jup.2016.03.005

S. Funcke, D. Bauknecht / Utilities Policy xxx (2016) 1e8

3. Dimensions of (de)centralised infrastructure In this section, we consider whether infrastructure is centralised or decentralised according to four dimensions. Table 1 integrates these dimensions and defines them for both centralised and decentralised designs. The transformation of Germany's electricity system in recent years can be explained with an understanding of change across these dimensions. 3.1. Location of power generation technologies Here we look at the location of power generation technologies in terms of two spatial dimensions, connectivity and proximity. 3.1.1. (De)centralised connectivity When describing the degree of (de)centralisation in electricity infrastructure, a key issue defined in spatial terms is the connection of generation facilities to a power grid. Centralised connectivity can therefore be defined as generation facilities connected to the transmission grid (Type 1C in Table 1). This conception does not distinguish between power sources; centralised generation technologies, although generally considered large in scale, can draw on nuclear and fossil fuels as well as renewable sources. The most concise definition concerning decentralised electricity generation, which we apply here, was put forward by Ackermann et al. (2001). They state that “[d]istributed generation is an electric power source connected directly to the distribution network or on the customer side of the meter” (Type 1D in Table 1). Ackermann et al. found that there is agreement among analysts in this field that the definition only encompasses active power. As a means to distinguish between the transmission and distribution networks, they refer to the legal definition applied in each country, as voltage levels are not the same in all countries. The capacity of the generation unit is not definitive, but in practice is usually limited by technical constraints of the distribution grid to a maximum of 100e150 MW. The conversion technology is also not decisive, and renewable as well as fossil-fuel generation technologies can belong to decentralised generation. Generation units and their typical capacities and characteristics can be found in Ackermann et al. (2001: 198), Pepermans et al. (2005: 792) and Viral and Khatod (2012: 5151). For a discussion of the regulatory implications of distributed generation, see Bauknecht (2012). Ackermann et al. further remark that the power delivery area and environmental impacts are also irrelevant for the definition. Therefore, the generated electricity could, depending on production and consumption patterns, be distributed to other areas via the transmission grid. From an environmental impact perspective, generation plants connected on a decentralised level (i.e. the distribution network) are not automatically more sustainable than centralised ones (Alanne and Saari, 2006; Karger and Hennings,

3

2009); more relevant is the question of how efficient the plants are and the energy sources on which they draw. 3.1.2. (De)centralised proximity The energy transition has a strong spatial element that sometimes is overlooked in scientific and public debates (Bridge et al., 2013). Concerning (de)centralisation of electricity infrastructure, the proximity of power plants to load centres is decisive, influencing the level of transmission capacity that is needed as well as the losses that occur in the network. In Germany, the liberalisation of the electricity sector, the decision to phase out nuclear power, and the commitment to increase the share of renewable energy sources are affecting the spatial profile of electricity generation. Traditionally, large-scale fossil and nuclear power plants were, at least in Germany, built relatively close to consumption centres (Type 2D in Table 1). Renewable energy power plants on the other hand, have predominantly been constructed at sites with favourable natural potential, which are not necessarily close to demand centres (Type 2C in Table 1). Hence, the influence of RES-E is as follows. In Germany, more small-scale and spatially distributed power plants have been built in the last 20 years, often drawing on renewable sources such as wind or solar. This has led to a concentration of onshore wind power in the windier north of the country, close to the sea. Even though southern states like Bavaria and Baden-Wuerttemberg have set goals for increasing the share of wind energy in the future, additional offshore wind projects will most likely strengthen the focus on wind energy in the north. On the other hand, photovoltaic power plants are concentrated in the southern federal states (AEE, 2013). Therefore, the increase in RES-E generation capacity has mostly been disconnected from consumption patterns because the German legislation concerning renewables, as well as unbundling and the grid pricing structure, mostly does not include incentives relating to the geographical location of new power facilities (Groschke et al., 2009; Steger et al., 2008). Increasing amounts of variable RES-E that are geographically concentrated lead to the need for additional grid expansion or storage capacity. As a consequence, the transmission grid capacities in the north-tosouth direction occasionally reach their limits (BMWi, 2011). 3.1.3. Integration of connectivity and proximity The integration of the two electricity infrastructure dimensions (connectivity and proximity) presented in the sections above can lead to various infrastructure designs, depending on whether they are construed in a decentralised or centralised manner. In total, four arrangements are conceivable. The “incumbent infrastructure” can be described as a combination of Type 1C and Type 2D (cf. Table 1) with large-scale generation technologies connected to the transmission grid that are located relatively close to consumption centres. With increasing shares of RES-E in the electricity mix, a shift

Table 1 Typology of centralised and decentralised infrastructure dimensions. Infrastructure dimension

Decentralised

Centralised

Location of power generation technologies

1. Connectivity: “(De)centralised generation” 2. Proximity: “Geographical distribution”

Technologies are connected to the distribution grid [Type 1D] Technologies are located near load (demand) [Type 2D]

Technologies are connected to the transmission grid [Type 1C] Technologies are located near resources (supply) [Type 2C]

Operational methods for system balancing

3. Flexibility

Infrastructure is balanced through distributed resources and demand-side management [Type 3D] Infrastructure is controlled by DSOs and/or prosumers and coordinated through regional markets [Type 4D]

Infrastructure is balanced through large-scale power generation, storage and transmission grid [Type 3C] Infrastructure is controlled by TSOs and coordinated through national or international markets [Type 4C]

4. Controllability

Please cite this article in press as: Funcke, S., Bauknecht, D., Typology of centralised and decentralised visions for electricity infrastructure, Utilities Policy (2016), http://dx.doi.org/10.1016/j.jup.2016.03.005

4

S. Funcke, D. Bauknecht / Utilities Policy xxx (2016) 1e8

can be expected. The second option could be titled “utility-scale RES-E”, with a Type 1C and 2C combination. The focal points of this design are large-scale RES-E plants that are connected to the transmission grid, rely on abundant renewable resources and are therefore located at a distance from consumption centres. This model could be implemented within single countries or across continents; a Europe-wide development would entail a significant expansion of interconnector capacities. Within a “dispersed infrastructure”, the spatially dispersed generation technologies would feed into the distribution grid according to regional demand (Type 1D and 2D). Finally, “distribution surplus” results from generation technologies that are connected to the distribution grid (Type 1D) but are not geared towards the distribution of demand (Type 2C). This situation can already be observed in some of the distribution grids in Germany, where at certain times more electricity is generated than needed and surpluses must be exported via the transmission grid. These four identified options can also exist in hybrid forms. 3.2. Operational methods for infrastructure balancing While the two dimensions of (de)centralisation presented in the previous section mainly refer to the location of generation technologies, the following two dimensions are based on the operation of the infrastructure with respect to balancing generation and consumption at all times. Balancing requires flexibility, which can be achieved through load and generation management strategies, storage systems, and grid expansion. Balancing generation and demand also requires a strong element of (real-time) controllability, which can be coordinated through markets. Flexibility and controllability can both be organised in a centralised or decentralised way. 3.2.1. (De)centralised flexibility In utility systems, electricity generation and consumption must be balanced at all times. In traditional fossil-based and nuclearbased electricity infrastructure, this is achieved by operational measures related mainly to large-scale generation and storage, as well as some load management. The following options can be considered as centralised approaches to flexibility (Type 3C in Table 1); they have in common that they are mostly connected to higher voltage levels of the network, i.e. the transmission grid, and potentially encompass large geographical areas for balancing. Conventional power plants can provide flexibility by reducing or increasing their power output. The conversion technology determines the response time, which is why peak-load plants (such as gas-fired turbines) that react quicker than nuclear power plants can be important for balancing supply and demand. Load management solutions are typically provided by large consumers who can increase or decrease their consumption based on requirements or prices. Large-scale electricity storage can also play an important role in traditional electricity infrastructures. Pumped hydropower storage is by far the most important storage technology, but practical capacity in most countries is limited by geological constraints. Another option for adding centralised flexibility is a Europe-wide grid expansion, so that renewable generation at different locations can be complementary (cf. e.g. Czisch, 2011; Blarke and Jenkins, 2013). Flexibility in production and consumption becomes even more important in electricity infrastructure that increasingly draws on decentralised generation (cf. section 3.1.1), especially if it depends on high shares of variable renewable generation. The decentralised flexibility options presented in the following are smaller in scale, can be used for balancing in small geographical areas, are generally connected to lower voltage levels of the network, i.e. the

distribution network, and mainly depend on ICT as is used in smart grids (Type 3D in Table 1). The smart-grid approach makes it possible to exploit flexibility provided by power plants connected to the distribution grid (such as combined heat and power (CHP) plants), electricity consumers via demand-side management (DSM), or small-scale storage (Wissner, 2011). DSM and storage capacity on the consumer or prosumer5 side could be incentivised through electricity prices that fluctuate with variable RES-E generation (Schleicher-Tappeser, 2012). It is argued by some (e.g. Chicco and Mancarella, 2009; Lund et al., 2012; Richardson, 2013) that the heating, cooling, and mobility sectors should also be considered as they offer additional balancing options with vehicleto-grid technologies, heat pumps, and combined heating and cooling power plants. 3.2.2. (De)centralised controllability A key operational requirement for electricity infrastructure is a mechanism to ensure that generation and demand are always in line and to control and coordinate flexibility options. The incumbent electricity infrastructure is controlled in a centralised manner on the level of transmission grids (Type 4C in Table 1). Coordination of controllability is provided by markets, such as day-ahead, intraday, and balancing markets. In that sense, even liberalised systems today are characterised by trading within larger areas beyond regional monopolies but with centralised control. For the European Union, the completion of the internal energy market is still a key objective. Connecting markets that were previously separated means that infrastructure control becomes even more centralised. With increasing shares of decentralised connectivity as well as flexibility (cf. sections 3.1.1 and 3.2.1), the question arises of whether control responsibility should also be decentralised to some extent. A fully decentralised control would include balancing production and consumption on the distribution level with coordination on local markets (Type 4D in Table 1). From a technological point of view, a more active control of voltage levels by the distribution system operators (DSOs) could be a first step towards this scenario. Today, DSOs often accommodate increasing numbers of generation units (decentralisation, as discussed in section 3.1.1) with costly grid reinforcement. Proponents of regional smart-grid visions argue that technological solutions, such as controlling reactive power and implementing controllable transformers, can help to delay or avoid costly network expansion (e.g. Esslinger and Witzmann, 2011). From an actor-perspective, final consumers and prosumers, next to DSOs, could become more important. They could exert control to a certain degree through energy management programmes that control demand (e.g. household appliances like freezers can be switched off for some minutes at times of peak demand). Prosumers could be incentivised towards self-supply by higher electricity prices and lower costs for on-site RES-E production equipment. An example is photovoltaic systems in combination with electricity storage or heat pumps that, in some countries, have the potential to decrease the reliance on more expensive electricity from the grid (Schleicher-Tappeser, 2012). Small-scale and medium-scale consumers as well as generators and DSOs could be integrated in local markets that coordinate generation and load profiles through local price signals. The recent build-up of significant shares of modern RES-E capacity in some countries has mainly been achieved with support schemes such as feed-in tariffs that were enacted on the national level. According to visions of decentralised control, local or regional energy markets

5 Prosumers in this context are considered to be consumers that also own generation capacity at the site of consumption.

Please cite this article in press as: Funcke, S., Bauknecht, D., Typology of centralised and decentralised visions for electricity infrastructure, Utilities Policy (2016), http://dx.doi.org/10.1016/j.jup.2016.03.005

S. Funcke, D. Bauknecht / Utilities Policy xxx (2016) 1e8

that to date only exist in model communities could in the future help to integrate and commercialise RES-E from variable sources. A discussion and evaluation of possible approaches that take regional infrastructure control options into account can be found in Trepper et al. (2013). DSM, decentralised storage, and flexible endconsumer tariffs are a few of the mechanisms that could spur these developments towards a more decentralised infrastructure, where control mechanisms are organised on local markets. Importantly, decentralisation of resources in terms of generation and flexibility (see sections 3.1.1 and 3.2.1) does not necessarily imply that control is organised in a decentralised way. Rather, decentralisation along these two dimensions is often combined with a centrally controlled infrastructure. Examples of centralised control of decentralised resources can be found in various pilot projects, such as e.g. virtual power plants (VPP) which allow modelling of the combination of decentralised connectivity and flexibility that can technically be controlled either in a centralised or decentralised manner. Commercialisation can also be realised on decentralised or centralised markets (Pudjianto et al., 2007). 3.2.3. Integration of flexibility and controllability Integrating the two dimensions presented above, flexibility and controllability, three potential electricity infrastructure layouts can be identified. Flexibility within an “incumbent infrastructure” is provided centrally with large-scale balancing measures, such as pumped hydro power plants (Type 3C in Table 1). Power-plant dispatch is controlled and coordinated mainly through transmission grid operators and the centralised electricity exchange (Type 4C). The number of decentralised power plants can be increased within this organisational setting without changes in flexibility and controllability. This can already be observed in several regions in Germany that pursue the political goal of becoming ‘energy autarkic’ (autonomous). However, these regions usually produce as much electricity as is needed during one year, rather than trying to achieve full self-sufficiency in supply at all times. Most of these regions do not decentralise flexibility and controllability and therefore are still within the “incumbent infrastructure” concerning these dimensions. The second potential infrastructure layout could be labelled “centralised controllability of decentralised flexibility options” (combination of Type 4C and Type 3D). An example of this is a flexible decentralised electricity storage unit that is controlled centrally in a VPP which participates in a central exchange. The third option draws on decentralised flexibility (Type 3D) and decentralised controllability (Type 4D). This option could be titled “autarky” as it, theoretically, allows for a full self-supply at all times in areas that are equipped with sufficient amounts of decentralised generation. However, full autarky comes with comparably high costs concerning flexibility measures and electricity generation. It is therefore more likely that regions that implement decentralised flexibility and controllability are still connected to centralised back-up options (e.g. the transmission grid) to buy or sell electricity in times of exceptionally high consumption or generation. 4. Analysis of two visions Several different visions for the future of the German electricity system have been proposed. Some of them focus on the national level and others on the international level; some consider electricity only, while others take the heating and cooling sector as well as the transportation sector into account.6 From the wider

6 An overview of visions concerning Germany and Europe can, for example, be found in German Advisory Council on the Environment (2011).

5

discourse about future electricity systems, we selected two cases that represent opposite views of possible developments. The first vision pursues a mainly decentralised and the second one a mainly centralised approach. This allows an analysis that highlights a range of possible developments based on the infrastructure dimensions described in Section 3. Various approaches to decentralisation with smart-grid solutions were developed and tested within the E-Energy programme.7 Two of them, RegMod Harz (renewable energy model region Harz; Fraunhofer IWES, 2012) and eTelligence (Agsten et al., 2013), were selected for this analysis. A centralised approach can be found in a vision that was advanced by several research institutes in the form of a roadmap to 2050 based on a SuperSmartGrid (SSG) and a unified electricity market for Europe and North Africa (PwC et al., 2010; PwC et al., 2011). A summary of the main results can be found in Table 2.

4.1. Decentralised vision: regional solutions The starting point for the E-Energy projects was the current framework of the electricity system and the question of how to adapt and extend it for the uptake of increasing amounts of distributed and renewable generation. The RegMod Harz project took a broad approach that, alongside the technological research, covered development and demonstration of the infrastructure dimensions presented in Section 3. For this project, the generation units were pooled in a VPP, which means that different RES-E generation technologies were combined and controlled together to provide the amount of electricity needed. Included were technologies that are connected to the distribution grid (Type 1D in Table 1) as well as technologies that are connected to the transmission grid (Type 1C). All generation units were situated relatively close to the points of demand in one geographical administrative district (Type 2D). This is not a requirement for VPPs as they could also consist of spatially dispersed generation units that are controlled centrally. Even though the technological developments within E-Energy could be seen as a pathway towards electricity autarkic regions, this was not the aim. On the contrary, areas with high potential for RES-E are expected to provide a surplus to supply other areas with less generation potential but more demand (such as rural areas providing surplus electricity to nearby cities). In this context, connection to a reliable transmission grid is still necessary to reduce the demand for expensive storage systems for balancing over time. A simulation within the scope of the project showed that 60% less storage capacity is needed if a connection to the German transmission grid is still available. Growing amounts of variable generation are increasing the demand for flexibility within the German infrastructure. A number of possibilities exist to balance generation and consumption and often a combination of options could be used. Reliable weather forecasts for photovoltaic and wind-power systems as well as load forecasts are important for controlling VPPs and managing demand. Flexible generation and storage systems are needed for covering the residual load or absorbing peak generation at times of low demand. A dispatch simulation in the RegMod Harz project showed that it may be possible to optimise self-supply in a certain area and thereby reduce imports and exports by about 20%. Fraunhofer IWES (2012) argues against complete autarky, although the tested flexibility measures could be implemented towards this end. However, they emphasise that higher rates of regional production and

7

For more information, see www.e-energy.de.

Please cite this article in press as: Funcke, S., Bauknecht, D., Typology of centralised and decentralised visions for electricity infrastructure, Utilities Policy (2016), http://dx.doi.org/10.1016/j.jup.2016.03.005

6

S. Funcke, D. Bauknecht / Utilities Policy xxx (2016) 1e8

Table 2 Application of the typology on two visions for potential future infrastructures. Infrastructure dimension

Decentralised vision: Regional solutions

Centralised vision: Connecting Europe and North Africa

Connectivity (cf. section 3.1.1)

Type 1D and 1C in Table 1: - Renewable generation technologies pooled in a VPP; connected to the distribution or transmission grid

Type 1C (and 1D): - Renewable generation technologies mainly connected to the transmission grid (Type 1C) - Possibly additional generation technologies connected to the distribution grid (Type 1D)

Proximity (cf. section 3.1.2)

Type 2D: - Generation technologies are located close to demand in one district (the applied mechanisms would also allow the control of spatially dispersed generation units) - Goal: regions with high potential co-supply areas with less potential - Regional autarky possible, but not aimed at

Type 2C: - Renewable power plants concentrated at sites with best potential in Europe and North Africa

Flexibility (cf. section 3.2.1)

Type 3D: - Balancing of generation and demand in one district through VPP - Decentralised storage systems - Potentially: coupling of electricity with heat and gas supply

Type 3C: - Large-scale dispatchable generation: CSP and hydro - Europe-wide balancing with a HVDC overlay-grid

Controllability (cf. section 3.2.2)

Type 4D (and 4C): - Regional market mechanism to control generation and demand - But: low liquidity on regional market; therefore a link to the national power exchange was introduced

Type 4C: - Expansion of existing centralised control mechanisms - Increase of international trade and grid connections - Goal: single market for Europe and North Africa

consumption of electricity could ease the burden on the transmission grid. The coupling of electricity with heat and gas supply could offer additional potential to increase infrastructure efficiency. Electricity generation peaks of a few hours per year (e.g. by wind turbines during a storm) could for instance be taken up by heat sinks or used for gas synthesis, and thereby help avoid grid expansion or the need to shut down renewable generation units at these times. On the demand side, smart meters and flexible tariffs that become available for consumers in real-time on an online market platform could incentivise load shifting (Type 3D in Table 1). In the eTelligence project, a regional market was developed with the aim of integrating decentralised flexibility options, such as load management and power-driven CHP plants (Agsten et al., 2013). This market could in principle be used as a decentralised control mechanism to coordinate the local balancing of demand and generation (Type 4D in Table 1). The local CHP plants, for example, could to the extent possible shift their power generation to periods when local RES-E generation is low. However, for the project, the regional market was linked to the German EEX power exchange through a market maker (i.e., an electricity trader that offers the prices available on the national market on the regional market in order to ensure liquidity on the regional market). This became necessary, as the relatively small number of market participants in the pilot project could not by themselves provide enough liquidity on the regional market. As a result, the deployment of local flexibility options is no longer geared towards the local demand for flexibility, but rather towards price signals on the central market. This leads to the question of whether the project developed a prototype for a regional market with decentralised trading of electricity, or whether the market in the project mainly tested the processes and products needed for the integration of decentralised flexibility in a centrally controlled infrastructure. The E-Energy projects RegModHarz and eTelligence offer technological and organisational pathways towards a (more) decentralised infrastructure in all discussed dimensions. In general, this could lead towards a more flexible infrastructure in which consumption patterns are more strongly aligned to availability of

electricity. However, the authors acknowledge economic restrictions that could hamper development towards a fully decentralised infrastructure, the influence of the existing infrastructure, and the limitations of their research projects. These limitations can be observed in the dimensions connectivity and controllability that include aspects of a more centralised infrastructure approach and were not entirely geared towards decentralisation in the course of the projects. 4.2. Centralised vision: connecting Europe and North Africa Four research institutes that are positioned in the business sector or science community developed a political vision of how to supply Europe and North Africa with 100% renewable electricity by 2050 (PwC et al., 2010; PwC et al., 2011). Even though the authors acknowledge the research and development efforts in the area of carbon capture and storage (CCS), as well as the relevance and plans for nuclear energy in some European countries, the vision is geared solely towards renewables. They contend that electricity demand in Europe and North Africa will increase through 2050 due to economic development and the roll-out of more electricity-consuming devices, including electric vehicles. According to the 2050 vision, rather large-scale generation technologies such as wind and concentrated solar power plants (CSP) that are connected to the transmission grid will dominate the electricity mix (Type 1C in Table 1). However, the build-up of a SSG will also allow the integration of generation technologies connected to the distribution grid, which would help meet off-grid and small-scale local demand. Nonetheless, in contrast to other studies (e.g. Battaglini et al., 2009), the authors only mention this Type 1Doption briefly and do not explain in detail its potential share or its relevance for the functioning of the infrastructure. By 2050, North Africa is expected to export a substantial amount of electricity to Europe based on a surplus of 60% electricity generation anticipated for the region. PwC et al. (2010) state that the erection of the SSG will trigger RES-E generation at geographical locations with the best potential.

Please cite this article in press as: Funcke, S., Bauknecht, D., Typology of centralised and decentralised visions for electricity infrastructure, Utilities Policy (2016), http://dx.doi.org/10.1016/j.jup.2016.03.005

S. Funcke, D. Bauknecht / Utilities Policy xxx (2016) 1e8

This would mean wind energy from the North Sea region, solar energy from the south of Europe and North Africa, wind and biomass from the Baltic Sea region and Eastern Europe, and hydropower as well as storage capacity from the mountainous areas of Scandinavia and the Alps. Imports from North Africa would still be needed, as the authors find that there is insufficient land available in Europe for renewable resource development (Type 2C). For this vision, flexibility in the infrastructure could mainly be provided by dispatchable generation units, such as CSP and hydropower facilities, and the high-voltage direct current (HVDC) part of the SSG. A HVDC overlay-grid would allow transmission from North Africa and other central generation sites to areas with high demand and the means to balance short-term surplus generation in one region with short-term low supply in another region (Type 3C). Managing load in the smart-grid part of the SSG with flexible tariffs for end-consumers is mentioned as an option to increase decentralised flexibility (Type 3D), but this option is given relatively little attention in the roadmap. A single market including Europe as well as North Africa by 2050 is one of the main goals in the roadmap, which means that the current centralised control mechanism would be further developed and expanded (Type 4C). This is supposed to be achieved by increasing international electricity trade and grid connections and by unifying the European markets, subsequently extending them to North Africa. The market would need to be adapted to include incentives for rewarding dispatchability that CSP, hydropower, or biomass resources could offer at peak-load times. This vision for Europe and North Africa through 2050 presents an insight into a development path towards a mainly centralised electricity infrastructure based on power plants that draw on renewable resources. Some aspects of this vision point towards integrating elements of decentralised infrastructure, (e.g. DSM or decentralised generation), but these are only touched upon briefly. 5. Conclusions and outlook The early transformation of entire electricity systems in countries like Germany is manifested in increasing RES-E generation capacities. One of the main issues arising from diverging visions for the (future) system is the degree of centralisation or decentralisation of the infrastructure. At the same time, a clear definition of (de) centralisation does not exist and relevant dimensions often remain vague. In our contribution, we identified four infrastructure dimensions from which we develop a typology concerning their potential (de)centralisation. In terms of infrastructure location, new generation capacity in Germany, if renewable, is often connected to the distribution grid and therefore decentralised. At the same time, the geographical distribution of these new power plants is mainly centralised because they are proximate to natural resource potential (wind, sun and water resources) rather than to demand patterns. In terms of infrastructure operation, increased reliability on variable resources entails changes in flexibility and possibly in control measures. These might include decentralised smart-grid technologies at the distribution level or centralised options, such as European-wide grid expansion. In addition, a combination of centralised and decentralised approaches is possible. The scenarios described in Section 4.1 and 4.2 illustrate the role of centralisation and decentralisation in visions for the future of electricity infrastructure. The examples show how the typology can be applied to the analysis of design alternatives. In Germany, as in many other countries, the co-development of pathways that include various decentralised and centralised elements can currently be observed and both pathways have strong political support. From a policy viewpoint, the question remains as

7

to how the co-development of various pathways can be governed. Our typology can facilitate the analysis of centralised and decentralised infrastructure alternatives, and the combination of these. This in turn can help to elevate the debate among key stakeholders. The literature recognizes that formulating a vision can be important to the process of system transformation (cf. Section 2). A clearer understanding of infrastructure dimensions can help identify and clarify possible development paths. With regard to policy, wellinformed visions have the potential to influence decision-making, financial support, and grid regulation at the national and, in the case of Europe, the supranational EU level. Acknowledgements The authors would like to thank Chantal Ruppert-Winkel, Janice Beecher and two anonymous reviewers for valuable comments on an earlier version of this manuscript. The work presented in this article has been partly supported by the German Federal Ministry of Economics and Technology on the basis of a decision by the German Bundestag as part of the project “Dezentral und zentral gesteuertes Energiemanagement auf Verteilnetzebene zur Systemintegration erneuerbarer Energien” (DFlex), project number 0325578A, 2013e2015. References €der, L., 2001. Distributed generation: a definition. Ackermann, T., Andersson, G., So Electr. Power Syst. Res. 57, 195e204. €nder mit neuer Energie. Jahresreport Fo € deral-Erneuerbar 2013. AEE, 2013. Bundesla Zahlen, Daten, Fakten, Berlin. Agsten, M., Bauknecht, D., Becker, A., Brinker, W., Conrads, R., Diebels, V., Erge, T., Feuerhahn, S., Heinemann, C., Hermsmeier, J., Hollinger, R., Klose, T., Koch, M., Mayer, C., Pistoor, G., Rosinger, C., Rüttinger, H., Schmedes, T., Stadler, M., 2013. eTelligence. New Energy Sources Require a New Approach, Oldenburg. Alanne, K., Saari, A., 2006. Distributed energy generation and sustainable development. Renew. Sustain. Energy Rev. 10, 539e558. Battaglini, A., Lilliestam, J., Haas, A., Patt, A., 2009. Development of SuperSmart Grids for a more efficient utilisation of electricity from renewable sources. J. Clean. Prod. 17, 911e918. Bauknecht, D., 2012. Transforming the Grid. Electricity System Governance and Network Integration of Distributed Generation, first ed. Nomos, Baden-Baden. BEE, 2011. Standortdebatte (“best Sites”). BEE-Position, Berlin. Berkhout, F., 2006. Normative expectations in systems innovation. Technol. Anal. Strat. Manag. 18, 299e311. Blarke, M.B., Jenkins, B.M., 2013. SuperGrid or SmartGrid: competing strategies for large-scale integration of intermittent renewables? Energy Policy 58, 381e390. BMWi, 2011. Monitoring-Bericht des Ministeriums für Wirtschaft und Technologie nach x 51 EnWG zur Versorgungssicherheit im Bereich der leitungsgebundenen €t, Berlin. Versorgung mit Elektrizita BMWi, 2015. Erneuerbare Energien im Jahr 2014. Erste Daten zur Entwicklung der erneuerbaren Energien in Deutschland auf Grundlage der Angaben der Arbeitsgruppe Erneuerbare Energien-Statistik, Berlin. Bridge, G., Bouzarovski, S., Bradshaw, M., Eyre, N., 2013. Geographies of energy transition: space, place and the low-carbon economy. Energy Policy 53, 331e340. Chicco, G., Mancarella, P., 2009. Distributed multi-generation: a comprehensive view. Renew. Sustain. Energy Rev. 13, 535e551. Czisch, G., 2011. Scenarios for a Future Electricity Supply. Cost-optimised Variations on Supplying Europe and its Neighbours with Electricity from Renewable Energies. Institution of Engineering and Technology, Stevenage. Dagger, S., 2009. Energiepolitik & Lobbying. Die Novelle des Erneuerbare-EnergienGesetzes (EEG) 2009. Ibidem-Verlag, Stuttgart. Erlinghagen, S., Markard, J., 2012. Smart grids and the transformation of the electricity sector: ICT firms as potential catalysts for sectoral change. Energy Policy 51, 895e906. Esslinger, P., Witzmann, R., 2011. Improving grid transmission capacity and voltage quality in low-voltage grids with a high proportion of distributed power plants. Energy Proc. 12, 294e302. Eurosolar, 2012. Die Beschleunigung der dezentralen Energiewende spart Kosten und ist der wirtschaftlich vernünftigste Weg. Eckpunkte für die Fortentwicklung des EEG, Bonn. Fraunhofer IWES, 2012. Landkreis als Vorreiter. Regenerative Modellregion Harz. Abschlussbericht, Kassel. Geels, F.W., 2002. Technological transitions as evolutionary reconfiguration processes: a multi-level perspective and a case-study. Res. Policy 31, 1257e1274. Geels, F.W., Schot, J., 2007. Typology of sociotechnical transition pathways. Res. Policy 36, 399e417.

Please cite this article in press as: Funcke, S., Bauknecht, D., Typology of centralised and decentralised visions for electricity infrastructure, Utilities Policy (2016), http://dx.doi.org/10.1016/j.jup.2016.03.005

8

S. Funcke, D. Bauknecht / Utilities Policy xxx (2016) 1e8

Geels, F.W., Schot, J.W., 2010. Part 1: the dynamics of socio-technical transitions: a sociotechnical perspective. In: Grin, J., Rotmans, J., Schot, J.W. (Eds.), Transitions to Sustainable Development. New Directions in the Study of Long Term Transformative Change. Routledge, New York, pp. 11e104. German Advisory Council on the Environment, 2011. Pathways towards a 100 % Renewable Electricity System (Special Report, Berlin). Goldthau, A., 2014. Rethinking the governance of energy infrastructure: scale, decentralization and polycentrism. Energy Res. Soc. Sci. 1, 134e140. €st, D., Fichtner, W., 2009. Neue Anforderungen an optiGroschke, M., Eßer, A., Mo mierende Energiesystemmodelle für die Kraftwerkseinsatz- und Zubauplanung bei begrenzten Netzkapazit€ aten. ZfE Z. für Energiewirtschaft 33, 14e22. Hauber, J., Ruppert-Winkel, C., 2012. Moving towards energy self-sufficiency based on renewables: comparative case studies on the emergence of regional processes of socio-technical change in Germany. Sustainability 4, 491e530. Helm, D., 2014. The European framework for energy and climate policies. Energy Policy 64, 29e35. Hirschl, B., 2008. Erneuerbare Energien-Politik. Eine Multi-Level Policy-Analyse mit Fokus auf den deutschen Strommarkt, first ed. VS Verlag für Sozialwissenschaften; GWV Fachverlage GmbH, Wiesbaden. Karger, C.R., Hennings, W., 2009. Sustainability evaluation of decentralized electricity generation. Renew. Sustain. Energy Rev. 13, 583e593. Konrad, K., Truffer, B., Voß, J.-P., 2008. Multi-regime dynamics in the analysis of sectoral transformation potentials: evidence from German utility sectors. J. Clean. Prod. 16, 1190e1202. Krause, F., Bossel, H., Müller-Reissmann, K.-F., 1980. Energie-Wende. Wachstum und € €l und Uran (Ein Alternativ-Bericht des Oko-Instituts/ Wohlstand ohne Erdo Freiburg. S. Fischer, Frankfurt am Main). Loorbach, D., Frantzeskaki, N., Thissen, W., 2010. Introduction to the special section: infrastructures and transitions. Technol. Forecast. Soc. Change 77, 1195e1202. Lovins, A.B., 1977. Soft Energy Paths. Toward a Durable Peace. Penguin Books, Hamondsworth. Lund, H., Andersen, A.N., Østergaard, P.A., Mathiesen, B.V., Connolly, D., 2012. From electricity smart grids to smart energy systems e a market operation based approach and understanding. Energy 42, 96e102. Mautz, R., 2007. The expansion of renewable energies in germany between niche dynamics and system integration - opportunities and restraints. Sci. Technol. Innov. Stud. 3, 113e131. €ltnisse durch Paulitzk, H., 2006. Revolutionierung der gesellschaftlichen Verha dezentrale Energiegewinnung. Solarzeitalter 18, 9e11. Pepermans, G., Driesen, J., Haeseldonckx, D., Belmans, R., D’haeseleer, W., 2005.

Distributed generation: definition, benefits and issues. Energy Policy 33, 787e798. Pudjianto, D., Ramsay, C., Strbac, G., 2007. Virtual power plant and system integration of distributed energy resources. IET Renew. Power Gener. 1, 10e16. Pudlik, M., Ragwitz, M., Sensfuß, F., Zickfeld, F., Wieland, A., 2012. 2050 Desert Power. Perspectives on a Sustainable Power System for EUMENA, München. PwC, PIK, IIASA, 2011. Moving towards 100% Renewable Electricity in Europe & North Africa by 2050. Evaluating Progress in 2010, London. PwC, PIK, IIASA, ECF, 2010. 100% Renewable Electricity. A Roadmap to 2050 for Europe and North Africa, London. Richardson, D.B., 2013. Electric vehicles and the electric grid: a review of modeling approaches, Impacts, and renewable energy integration. Renew. Sustain. Energy Rev. 19, 247e254. Rotmans, J., Kemp, R., van Asselt, M., 2001. More evolution than revolution: transition management in public policy. Foresight 3, 15e31. €hrdet dezentrale Scheer, H., 2010. Fata Morgana statt Wüstenwunder. Desertec gefa Energie. In: AEE (Ed.), Kraftwerke für Jedermann. Chancen und Herausforderungen einer dezentralen erneuerbaren Energieversorgung, Berlin, pp. 52e55. Schleicher-Tappeser, R., 2012. How renewables will change electricity markets in the next five years. Energy Policy 48, 64e75. Smith, A., Stirling, A., Berkhout, F., 2005. The governance of sustainable sociotechnical transitions. Res. Policy 34, 1491e1510. Sp€ ath, P., Rohracher, H., 2010. ‘Energy regions’: the transformative power of regional discourses on socio-technical futures. Res. Policy 39, 449e458. Steger, U., Büdenbender, U., Feess, E., Nelles, D., 2008. Die Regulierung elektrischer € sungsans€ Netze. Offene Fragen und Lo atze. Springer, Berlin, Heidelberg. Trepper, K., Weber, C., Kallen, D., 2013. Integrationsmanagement für Erneuerbare Energien e dezentrale Koordination im Strommarkt der Zukunft. uwf UmweltWirtschaftsForum 21, 209e218. Truffer, B., Voß, J.-P., Konrad, K., 2008. Mapping expectations for system transformations. Lessons from Sustainability Foresight in German utility sectors. Technol. Forecast. Soc. Change 75, 1360e1372. Verbong, G., Geels, F., 2012. Future electricity systems. Visions, scenarios and transition pathways. In: Verbong, G., Loorbach, D. (Eds.), Governing the Energy Transition. Reality, Illusion or Necessity? Routledge, New York, pp. 203e219. Viral, R., Khatod, D.K., 2012. Optimal planning of distributed generation systems in distribution system: a review. Renew. Sustain. Energy Rev. 16, 5146e5165. Wissner, M., 2011. The smart grid e a saucerful of secrets? Appl. Energy 88, 2509e2518.

Please cite this article in press as: Funcke, S., Bauknecht, D., Typology of centralised and decentralised visions for electricity infrastructure, Utilities Policy (2016), http://dx.doi.org/10.1016/j.jup.2016.03.005