THE IMPACT OF CARBON CAPTURE AND STORAGE ON A DECARBONIZED GERMAN POWER MARKET

Transcription

1 Chair for Management Science and Energy Economics Prof. Dr. Christoph Weber Chair for Management Science and Energy Economics University of Disbrg-Essen EWL Working Paper No. [04/13] THE IMPACT OF CARBON CAPTURE AND STORAGE ON A DECARBONIZED GERMAN POWER MARKET by Stephan Spiecker Volker Eickholt and Christoph Weber

2 The Impact of Carbon Captre and Storage on a Decarbonized German Power Market Stephan Spiecker, Volker Eickholt and Christoph Weber Abstract The Eropean energy policy is sbstantially driven by the target to redce the -emissions significantly and to mitigate climate change. Nevertheless Eropean power generation is still widely based on fossil fels. The carbon captre and storage technology (CCS) cold be part of an approach to achieve ambitios redction targets withot large scale transformations of the existing energy system. In this context the paper investigates in how far the CCS-technology cold play a role in the Eropean and most notably in the German electricity generation sector. To accont for all the interdependencies with the Eropean neighboring contries, the embedding of the German electricity system is modeled sing a stochastic Eropean electricity market model (E2M2s). After modeling the Eropean side constraints, the German electricity system is considered in detail with the stochastic German Electricity market model (GEM2s). The focs is thereby on the location of CCS plant sites, the strctre of the -pipeline network and the regional distribtion of storage sites. Reslts for three different Eropean energy market scenarios are presented p to the year Additionally, the se of CCS with se of onshore and offshore sites is investigated. Index Terms stochastic optimization, carbon captre and storage, power system economics JEL-classification: Q3, Q4, C61 STEPHAN SPIECKER, VOLKER EICKHOLT AND CHRISTOPH WEBER Chair for Management Science and Energy Economics, University of Disbrg-Essen (Camps Essen) Universitätsstr. 11, Essen (0)2 01 / stephan.spiecker@ni-de.de The athors are solely responsible for the contents which do not necessarily represent the opinion of the Chair for Management Science and Energy Economics. 1

3 1 INTRODUCTION In the latest Energy Roadmap (EC 2011) the EU emphasizes the objective to transform the Eropean energy sector into a low carbon economy. Hence, a redction of greenhose gas emissions by 80 to 95% compared to the levels of 1990 is the ambitios goal for This mitigation target determines efforts in all relevant energy sectors. Simltaneosly an affordable, sstainable, competitive and safe energy spply shold be garanteed (EC 2011). It is a matter of fact that large scale emission cts and the transformation of a whole economy into a decarbonized system is a mega-project that reqires years if not several decades and billions of Ero. Despite the ambitios climate targets of the EU it is very likely that fossil fels will have a certain share in the Eropean energy sector for the next decades. The IEA projects in the main scenario of the latest World Energy Otlook (IEA 2012) a 75%-share of fossil fels in 2035 concerning the world primary energy demand. The energy mix in the EU in 2011 is dominated by abot 75% of fossil fels. Whereas the transport sector is predominantly oil-driven, the electricity generating sector is to abot 50% based on coal and gas plants (IEA 2012). The Carbon Captre and Storage-technology (CCS) cold be instrmental to ct global or regional -emissions significantly withot removing fossil fel plants from the energy system overhastily and jeopardizing the secrity of electricity spply. CCS provides a technical soltion to ct -emissions from pnctal large-scale emissions sorces like fossil power plants or large indstrial prodcers. The CCS-technology reqires a transport-system (e.g. a pipeline-net) to transport the separated carbon dioxide to geological storage formations like saline aqifers or exploited gas fields on- or offshore (Metz et al. 2005). CCS cold therefore facilitate the transformation to a low carbon energy system withot massive redeployments in the infrastrctre of the electricity generating sector simltaneosly providing the advantages of a controllable energy generation backing p the ongoing proliferation of renewable energies. Germany, the biggest member state in the EU, realized a renewable share of more than 20% of the electricity prodction in 2011 (AGEB, 2012). Despite the nprecedented advance of the renewable energy sorces in Germany, the fossil fels lignite, hard coal, gas and oil stand p for more than 58% of the gross electricity prodction (AGEB, 2012). According to the new energy concept that has been annonced by the German administration in the middle of 2011, the German nclear phase ot has to be completed till The loss of rondabot 20 GW installed capacity of the carbon free nclear technology compared to 1990 makes the -redction targets of mins 40% till 2020 and of mins 80% till 2050 (compared to 1990 levels) considerably more challenging. Until 2012 already abot 10 GW of nclear have been phased ot and energy-related CO2 emissions have decreased by 22 % compared to However most of the emission redction has been achieved ntil 2005 (-18 %) and between 2011 and 2012 there has been even an increase in emissions by 2.2 % (BMWi 2013). Ths the 2050 target is still ambitios. Notwithstanding the fast diffsion of renewable technologies it is qite likely that fossil fel plants will contine to play a fndamental role in electricity generation in Germany for the next decades. Therefore, CCS might be an interesting soltion for the German electricity market. In 2

4 accordance with the Eropean Union, the German administration considers the CCS technology as a possible bridge into the forthcoming era of renewable energies (BMWI, 2010). The implementation of the CCS technology in conventional power prodcing sectors has been topic of many modeling approaches. Gogh et al. (2006), Fowler (2008) and McFarland and Herzog (2006) presented bottom-p models as well as top-down and integrated assessment model approaches. A variety of linear optimization model approaches can be fond in Bakken and von Streng Velken (2008). An interesting approach of a scenario based policy analysis is presented e.g. in Capros et al. (2008). In the context of a mixed integer and nonlinear programming problem, Chen et al. (2010a) integrate the CCS technology as an investment alternative in the electricity sector. Szolgayova et al. (2008) and Chen et al. (2010b) discss CCS in the frame of the real options method. Differences in the regional observation occr beside different methodology approaches. Labriet et al. (2012) analyze the impact of technology and climate ncertainties on the ftre global energy system which may also inclde CCS power plants. A Eropean perspective can be fond in Lohwasser and Madlener (2012). They show the impact of techno-economic assmptions on the diffsion of the CCS technology in Erope sing a bottom-p electricity sector model. Cremer (2007) formlates a modeling approach with special focs on the German power market. The presented bottom-p and mltiperiodic linear optimization modeling approach contains a copling with a geographical information system (GIS) based model for location decision spport. That approach incldes the opportnity to analyze the inflence and importance of different geographical parameters e.g. the distance between the -sorce and the -storage site. Frther single contry analyses especially for Eropean and Asian contries exist which analyze the ftre potential of the CCS technology (i.a. Lnd and Mathiesen 2012, Gerbelova et al. 2013). A crcial point in all these analyses are the cost assmptions for the CCS technology. This is frther stressed by Voll et al. (2012) and Rbin (2012). Bt social acceptance and policy spport has also to be taken into consideration when it comes to the introdction of new technologies into the market. The social acceptance of CCS and the willingness to pay have been investigated by Kraesel and Möst (2012). They fond ot that CCS technologies fall behind renewable technologies in pblic opinion. This makes policy spport even more important if the CCS technology shold obtain market matrity (cf. Nykvist 2013). A decision framework to answer the qestion whether at all CCS technology shold be spported is presented by Torvanger and Meadowcroft (2011) while Lohwasser and Madlener (2013) analyze the effectiveness of different policy measres in order to promote the CCS technology. Yet so far an approach is missing which combines a consideration of the interdependencies in the Eropean electricity market with a detailed analysis of location and transportation isses inside Germany. Indeed, a carbon captre, transport and storage network model can be fond in Mendelevitsch et al. (2010), bt they set prices exogenosly and operational details are only roghly approximated. 3

5 The fast proliferation of intermittent renewable energies in Germany pls the nclear phase ot have a deep impact on the German electricity exchange with the neighboring contries. Moreover, the local availability of storage sites mst be considered as well as the challenging climate targets that have been annonced by the German administration. These different market developments are considered in three scenarios. In addition to that, the political opposition of some German federal states and the local resistance against storing close to inhabited areas is another isse that has to be taken into consideration. De to this, two variations for each of the three scenarios are introdced: A variation with no restrictions for the transport and storage of (all storage case) and an offshore-variation that implies the strict reqirement to store only in offshore sites in the North Sea (offshore storage case). A model based approach is sed to analyze the inflence of a CCS deployment together with the expanding flctating prodction of renewable energies on ftre electricity markets. Thereby an intertemporal optimization model is chosen to ensre a lifetime assessment of the storage sites and capacities as well as to garantee sfficient storage space for the captred from new CCS power plants. The paper is organized as follows: The second section deals with the applied model framework and enhanced methodology to assre a detailed assessment of the CCS options. Section 3 explains the investigated system and the three main scenarios. Frthermore model reslts are shown and their implications are discssed. A brief conclsion on the obtained reslts is finally presented in section 4. 2 FORMULATION OF THE MODEL FRAMEWORK 2.1 Theoretical backgrond A two-stage approach is chosen to analyze the relevance of CCS for the ftre German power market (cf. Figre 1). Therefore, basically a fndamental stochastic electricity market model called E2M2s (cf. Swider and Weber, 2007) is sed with enhanced specifications. Additionally, a version of the model focsing specifically on Germany is employed (GEM2s). The main differences between the two models are the higher regional resoltion and the more detailed representation of CCS related facilities in the GEM2smodel. 4

6 Figre 1: Model framework Both models are based on the E2M2s. The E2M2s assmes fnctioning competitive markets and derives market reslts throgh optimization. That means that a cost minimal power plant configration is chosen to cover electricity as well as heat demand. This approach is formlated as a linear, stochastic programming model and is implemented in the General Algebraic Modeling System (GAMS), (cf. Swider, Weber 2007 and Spiecker, Weber 2012). The specifics and extensions of the respective models are described in the corse of this chapter. Key variables within the optimization are electricity and heat prodction, generation capacities, transmission and pmping qantities. The model simltaneosly determines the yearly vector of variables that minimizes total cost sbject to load restrictions and detailed technical limitations of power system components. Investments in new generation capacities are endogenosly determined in the model beside power plant operation and transmission line loading. Thereby, yearly fll load hors are a key driver of the investment decisions and selection of power plant technology. This decision is represented by annalized investment costs in the system objective fnction. This leads to investment decisions and pricing patterns in line with the peak-load pricing approach as developed by Boitex (1960) and others. Flctations of solar and wind infeed have an impact on ftre power markets. Ths, a modeling of 8760 hors is preferable in order to consider the whole range of possible infeeds. Bt this increases the comptational brden in a way that the problem is impossible to handle. Therefore, typical time segments are modeled bt additionally a stochastic approach is chosen to consider short-term flctations of renewable infeeds from wind and solar. Instead of a deterministic average with a single operation mode the ftre system mst cope in each time segment with varying ftre system states. In this case, the system states are depending on the expected infeeds of wind and solar. The challenge at hand is the representation of stochastics in an extremely large system. By recombining the nodes at each stage (as shown in Figre 2), we avoid the exponential growth in the nmber of nodes and ths the crse of dimensionality. Clster analysis is applied on historical wind and solar prodction data in order to 5

7 derive the probabilities of the nodes and the corresponding transition probabilities between the nodes. A more detailed description as well as frther remarks on consistency can be fond in Spiecker et al. (2013). Figre 2: Concept of recombining decision tree Typical time segments are sed to redce comptational complexity frther. On the one hand, these time segments shold be short enogh to captre flctations of load and renewable infeeds within a day. This is necessary to describe the power plant operation with start-p costs and operation at part load. On the other hand, a whole year is optimized simltaneosly to model investments and the management of large-scale hydro reservoirs adeqately. Therefore, eight typical days within one year are considered. They stand for a working day and a non working day in every third month. These typical days are again divided into seven time-steps. After a first time step of six hors a five hor time-step follows, scceeded by a one-hor noon peak time segment. The remaining twelve hors after the noon peak hor are divided into for segments of eqal length. Ths there are altogether eight typical days with seven typical hors each. Total costs TC y within a simlated year y to be minimized inclde operating costs OC r,,t,y,n, start-p costs SC r,,t,y,n and fixed costs FC r,,y smmed over regions r, nit types time segments t within a year and stochastic nodes n. The smmands are weighted by the dration d t and freqency f t of the corresponding time segment as well as the time segment related stage probability. s ( t ), n TC y r t n ( d t f t s ( t ), n ( OC r,, t, y, n SC r,, t, y, n ) FC r,, y ) (1) Since costs are compted endogenosly in the model, operating costs for power plants can be decomposed in fel costs for operation at minimm load, fel costs for incremental otpt and other variable costs oth c, op : p FUEL r,, t, y onl r,, t, y onl oth, op OC ( P l L ) l L c P r,, t, y, n m r,, t, y, n r,, t, y, n 0 r,, t, y, n r,, t, y, n p FUEL (2) FUEL In this eqation, p is the fel price, r,, t, y efficiency at the minimm load factor l. m the marginal efficiency for an operating plant and P r, t, y, n 6 0 the, corresponds to the electric otpt and the variable L onl represents the capacity crrently online. A linear approximation is sed to consider part r,, t, y, n

8 m 0 load efficiencies. With it is less costly to increase the otpt of plants already operating instead of increasing the capacity online. Ths there is an incentive to redce as far as possible the capacity online in the power system. Additional eqations describe the operation restrictions and the operational costs of CHP plants (similar to Weber 2004). Depending on the flexibility, strctre and size of the CHP system it facilitates or complicates the integration of renewable energies (cf. Andersen and Lnd 2007) 1. Besides operating costs, start-p costs are considered since they may inflence schedling decisions considerably. When it comes to the consideration of long-term developments, for the optimization two aspects are crcial: The consideration of the generation schedling and the treatment of fixed costs FC r,,y,n inclding investment costs. Thereby the choice among different available investment alternatives with specific inv investment costs c is modelled sing the decision variable for newly bilt capacity L new r, y The German power market is embedded in the Eropean power market. Hence, we focs on the Eropean level and interactions between single contries at the first stage. As a reslt power transmission between Germany and other Eropean contries (RoE) is determined from an optimal prodction allocation throgh optimized dispatch and investment-planning. In addition, emissions and prices are derived from Eropean wide bonds. These constraints have a major impact on the ftre investment in conventional power plant capacities and renewable energies. The reslts of the Eropean model are sed at the second stage as side constraints for the German market development. To avoid intertemporal inconsistencies, emissions are transferred instead of prices. At the second stage, intertemporal optimization is considered, inclding trading between several periods. In contrast to the first stage, a detailed description of CCS is implemented and line investments are determined endogenosly in order to represent the trade-off between power transmission and transport. Therefore a higher geographical resoltion is chosen to analyze the regional distribtion of investments in more detail. The same general inpt data is sed for both models in order to ensre consistent reslts. This is frther ensred by the fixation of CO2 emissions and power exchange on the second stage. Nevertheless small differences occr de to the more detailed consideration of side constraints for the CCS technology in the GEM2s and the intertemporal approach. 2.2 Stage I At the first stage power transmission between Germany and other Eropean contries (RoE) as well as the regional distribtion of emissions are compted. De to the fact that a Eropean emission trading system exists, a common -price is derived sing a dynamic recrsive approach based on abatement costs. With a more challenging bond,. B the related price increases. The amont 1 Note that frther flexibilities for RES integration reslting from integrated strategies covering also the heating, transportation and gas sectors are not considered here (cf. e.g. Lnd et al. 2012). 7

9 of emitted is dependent on prodction, power plant efficiencies, captre rate cr of different power plants and emission rates weighted similarly to the operating costs (cf. eqation 2). B. Part load efficiencies are considered and emissions are ( onl t t s ( t ), n m r,, t, y, n r,, t, y, n 0 r,, t, y, n r t y n onl d f 1 cr ( ( P l L ) l L )) (3) RES are also analyzed at a Eropean level becase their expansion has strong inflence on power generation and transmission. Endogenos investments in RES are especially important nder a coordinated Eropean energy policy with largely market based RES expansion. This possibility is considered by implementing cost potential crves for RES. A detailed description can be fond in (Spiecker and Weber, 2012). 2.3 Stage II At the second stage we focs on the German power market with a detailed representation of the CCS technology and investments in transmission infrastrctre. In order to consider the intertemporal aspects of storage an intertemporal approach is chosen, which incldes the whole lifetime of a CCS power plant in a single optimization. In contrast to the first stage, where a dynamic recrsive optimization is sed, a period of 40 years from 2010 to 2050 is integrally optimized at the second stage. With this approach it can be assred that CCS power plants bilt dring the analysis period have sfficient storage capacity over their whole lifetime. Therefore the objective fnction is extended to inclde CCS related costs, in particlar fixed costs FC CCS for CO transmission capacity 2 C r r ',. Investment costs and other fixed costs as well as variable y costs CCS OC of transport are modeled in detail. Moreover, costs for storage determined depending on the stored volme operation of storage oth, fix c interest rate i and lifetime FC CCS r, t, y n STC, Stor and related costs for investment r, t, y, n, are. Investment costs are weighted with the annity factor a depending on life b. CCS life inv oth, fix a ( i, b ) c C c C (4) r, r ',, y r r ', y r r ', y CCS oth, op CCS c E (5) r, r ',, t, y, n r r ', t, y n OC, c inv and STC CCS life inv oth, fix a ( i, b ) c Stor c Stor (6) r,, t, y, n r, t, y, n r, t, y, n In addition capacity restrictions for transmission E C r r ', y have to be considered. C (7) r r ', t, y, n r r ', y These capacities are determined endogenosly by capacities in the preceding period C and r r ', y 1 crrent investments C new : r r ', y C new C C r r ', y r r ', y 1 r r ', y (8) 8

10 Concerning storage operation, the total amont of stored H has to be lower than total storage r, t, y, n capacity, Stor L r., Stor H L (9) s ( t ), n r, t, y, n r t y n Again, injection decisions in the previos period are considered and are weighted according to the time resoltion and the transition probability τ: t 1 n 1 t, t 1, n, n 1 H d f Stor H (10) s ( t 1), n 1 r, t 1, y, n 1 t t r, t, y, n r, t, y, n The amont of captred V CCS r t, y, n, is determined as the complement to the amont of emitted B. CCS onl onl V cr ( ( P l L ) l L ) r, t, y, n m r,, t, y, n r,, t, y, n 0 r,, t, y, n (11) To ensre the overall balance, the captred set eqal to the stored volme V CCS E CCS r t, y, n r r ', t, y, n Stor in each region r, time segment t and stochastic node n. r, t, y, n, and the related exchange are V CCS r, t, y, n r ' CCS CCS E E Stor (12) r ' r, t, y, n r r ', t, y, n r, t, y, n Investments in the power grid are modeled analogosly. Investment costs are following the description in formla (4) where the annity of the new capacity is considered in the objective fnction. Operative costs of power transmission are considered in a different way. Losses of power transmission are sed instead of having a cost nit rate. These losses are depending on the distance between two nodes. 3 SYSTEM AND SCENARIOS ANALYZED 3.1 Geographical scope and data A different regional scope and resoltion is chosen at the two stages. At the first stage, the entire Eropean electricity system is modeled. Every contry is considered as one single node in the model. Only Germany is split in seven regions to consider bottlenecks within the contry. These bottlenecks especially occr with higher infeeds of renewable energies in the north. This energy has to be transmitted to the western and sothern demand centers. Denmark is represented by two nodes in line with its zonal separation. The second stage focses on Germany. Therefore a higher resoltion for Germany is chosen. Here, 18 regions are considered (cf. Figre 3). 9

11 Figre 3: Regionalization of Germany 3.2 Modeled scenarios We se different scenarios to model ncertainty that goes beyond flctations of stochastic renewable energy infeed and power plant otages. These are especially long-term ncertainties which impact the economic framework for new power plant capacities. The scenario CLIMATE describes a development which is in line with today s policy objectives. Electricity demand will slightly increase while emissions are constrained. Until 2050 a redction of 80% compared to 1990 shold be reached. Crrent nclear policy in different member states also determines whether new nclear power plants can be bilt. The same holds for the scenario CONVENTIONAL WISDOM. Here, the focs lies more on secrity of spply. Electricity demand is higher and absolte redctions are lower compared to the scenario CLIMATE. Nevertheless, redctions might be more ambitios de to higher demand growth. On the other side, the scenario CLIMATE+ has a stronger emphasis on sstainability. Energy demand is only slightly increasing ntil 2020 and nearly kept constant after 2020 while the redction target is 95%. Frther on, nclear power plants are no option for low carbon electricity prodction in most Eropean contries. This makes it more difficlt to reach the ambitios political targets. From today s point of view the scenario CLIMATE seems to be the most likely one as it considers economical isses as well as environmental objectives. The CLIMATE+ scenario has a stronger focs on environment and sstainability while the CONVENTIONAL WISDOM scenario focses on economical targets. A main driver for the ftre development is the overcoming of the economic crisis. In this context, less ambitios environmental targets are prsed. The development of fel prices is the same in all three scenarios. Here, we have chosen a data set in line with World Energy Otlook scenarios (IEA, 2012). The same inpt data is also chosen for the expansion of renewable energies. Until 2020, new capacities are installed according to the contry specific national renewable action plans (EC 2010) with partly redced investments in the CONVENTIONAL WISDOM 10

12 scenario. After the year 2020, investments in renewable energies are determined within the model nder market conditions. Henceforward, renewable energies compete with other low carbon technologies like CCS or even efficient conventional power plants on a level playing field, i.e. no qota or spport schemes are assmed to be in place. An important driver for these investment decisions is the price obtained. 3.3 CCS technologies Table 1: Scenario overview Conventional Wisdom Climate Climate + Demand high mid low CO2-redction compared to 1990 Acceptance of nclear power 50% 80% 95% mid mid low The three main -captre technologies, the pre-combstion (pre_c), post-combstion (post_c) and oxy-fel combstion (oxy_c) technology are implemented in the models. From the crrent point of view, there is no clear technological or economic advantage for any one of these three opportnities. Hence, all technologies are considered to have a realistic chance to be similarly competitive in the long term view (ZEP, 2011a), (IEA, 2010b). Table 2 and fel captre technology investment costs max. efficiency annal fix costs other variable costs captre rate /KW % /KW /MWh % Coal oxy_c ,3 93 Coal post-c ,5 93 Coal pre-c ,5 93 Gas oxy_c ,5 89 Gas post-c ,8 89 Gas pre-c ,2 89 Lignite oxy_c ,2 93 Lignite post-c ,4 93 Lignite pre-c ,3 93 give a brief overview of the most important technology parameters sed in this article, inclding also ftre improvements. Table 2: Technology data for CCS with constrction years fel captre technology investment costs max. efficiency annal fix costs other variable costs captre rate 11

13 /KW % /KW /MWh % Coal oxy_c Coal post-c Coal pre-c Gas oxy_c Gas post-c Gas pre-c Lignite oxy_c Lignite post-c Lignite pre-c Table 3: Technology data for CCS with constrction years fel captre technology investment costs max. efficiency annal fix costs other variable costs captre rate /KW % /KW /MWh % Coal oxy_c ,3 93 Coal post-c ,5 93 Coal pre-c ,5 93 Gas oxy_c ,5 89 Gas post-c ,8 89 Gas pre-c ,2 89 Lignite oxy_c ,2 93 Lignite post-c ,4 93 Lignite pre-c ,3 93 The key figres for investment costs, efficiency and the captre rate are based on the figres of varios stdies (e.g. ZEP 2011a, Damen et al. 2006, Abadie and Chamorro 2008, Otto and Reilly 2006, Cozijnsen 2005, Sasaki 2004, BMWI 2003 as well as Christensen and Holloway 2004) in combination with own assmptions (cf. TABLE 2). Beyond the year 2030, slight technological progress on varios parameters is assmed (cf. TABLE 3). The investment costs per kw will be redced by 5% and the captre rate will be increased by three percentage points till Improvements of the maximm efficiency occr in a range of 4% to 6% and depend on the technology. There is also a technology-dependent decrease of the annal fix cost between 5% and 15% and for other variable cost between 11% and 30%. The costs for the pipeline investments and the power grid are given in Investment costs 12 Operational costs pipeline 620 /km/t/a /km/t power grid 500 /km/mw calclated as losses of 1 %/100 km

14 . The assmptions for the pipeline-costs are based on own calclations according mainly to the latest ZEP-report (ZEP 2011b) and frthermore on Metz, (2005), Harrison (2010), McCoy and Rbin (2005) and May et al. (2005). Table 4: Investment costs for pipeline and power grid Investment costs Operational costs pipeline 620 /km/t/a /km/t power grid 500 /km/mw calclated as losses of 1 %/100 km As discssed before, two variants for every main scenario are considered. In the offshore-case dmping in onshore-sites is prohibited de to local opposition or political resentments and only offshore storage sites are available. In the all storage case no political restrictions are considered. For the onshore storage two options are available: deep saline aqifers and exploited gas fields. These options reslt in a cmlative storage capacity onshore of rondabot 8.3 Gigatons (Gt). Approximately 80 storage sites can be fond in Germany (cf. Hipp 2011, WI et al. 2007). The mean storage volme per site is rondabot 100 million tons of. The locations of the sites have been mapped to the regional strctre shown in Figre 3. The offshore capacity adds p to 2.9 Gt and is ths mch more restricted and comes along with higher investment and variable costs. All in all the capacity assmptions stand for a rather conservative assessment and present a lower bond of varios estimations (Hipp 2011, Christensen and Holloway 2004, May et al. 2005, WI et al. 2007). Table 5 shows the assmptions for the storage costs for the on- and offshore variants. These costs imply the investment costs for the storage sites. The assmptions are based on own estimations and ZEP (2011c). Table 5: -storage costs for the on-/offshore variants scenario storage type costs ( /t) exploited gas field 3 onshore saline aqifere 5 exploited gas field 6 offshore saline aqifere 14 The aforementioned scenarios are investigated to identify the impact of CCS on the German electricity system. First, prodction and emissions at a Eropean scale are presented. Next the implications on CCS power plant capacity and prodction are examined. Then the investments as well as the tilization of transmission infrastrctre are analyzed. Finally, storage operation is described in detail. 13

15 Prodction and power plant capacities in Erope Figre 4 shows the development of Eropean electricity prodction. Thereby, total prodction is determined by total demand corresponding to scenario assmptions. Bt differences occr in the strctre of the power plant capacities. Policy targets are a main reason for these differences. Bt also differences according to power plant availability and cost strctre inflence the investment decisions. Figre 4: Electricity prodction development in Erope The highest relative share of renewable energy prodction occrs in the scenario CLIMATE+. Bt also in the other scenarios, prodction of renewable energies increases to a large extent. The absolte prodction from renewables is qite similar in all scenarios. Thereby, especially with market driven investment in renewable energies, available resorces are a key driver for investments in renewable energies. Solar technologies are conseqently mainly installed in the sothern parts of Erope, while wind farms are bilt at places where wind conditions are advantageos. Depending on fel prices and frther investment restrictions, the residal load is covered by conventional power plants. 14

16 Prodction developments go along with corresponding capacity developments (Figre 5). In general flctating renewable energies have rather low fll-load hors compared to conventional power plants and therefore the share of those technologies on prodction and capacities differ. Until 2020 capacities of renewable energies are set exogenosly in line with the objectives of the different contries according to their national renewable energy action plans (EC 2010). Only in the CONVENTIONAL WISDOM scenario, these objectives are not flly reached. After that new investments in renewable energies have to compete on a market basis. In the case of low CO2 prices they compete with conventional power plant technologies and in the case of high CO2 prices they compete with CCS power plants. In addition they are restricted by their availability to provide firm capacity. That means that for later years a combination of renewable energies and peak-load capacities compete with traditional base and midmerit order capacities. Bt also fll-load hors of conventional power plants are redced with increasing flctating feed-in of renewable energies. Especially natral gas fired power plants are sed as back-p capacities for flctating renewable energies and have less fll load hors. Figre 5: Eropean capacity development for different scenarios 3.4 German power system development Regional prodction patterns may differ from the total Eropean patterns (Figre 6) and differences are cleared by electricity transmission. De to a strong extension of renewable energies Germany for example increases its exports in the first years. These exports partly sbstitte domestic prodction in other contries ntil After that this effect is more than compensated by the nclear phase-ot in Germany (Figre 7) and frther investments in renewable energies otside Germany. In the case of an efficient placement of renewable energies in a market based system, investments in RES often occr otside Germany where environmental conditions are more advantageos. From a system point of view, it is optimal to follow these rles and increase imports to Germany. As a reslt, the trade balance becomes negative. Already in 2030, net imports attain 5 % of total demand in the scenario CLIMATE+. 15

17 A frther reason for this development are tight bonds which pt extreme pressre on fossil fel dominated generation systems like the German one. Conversely, with less tight redction targets imports to Germany are lower and the share of fossil fel fired power plants is higher. Ths, not only the total emissions have to be considered bt also the ratio of demand and emissions. In smmary, Germany becomes dependent on electricity imports which lead to a net import in a range of 22 TWh to 35 TWh in 2050 (Figre 8). Figre 6: Electricity prodction development in Germany Figre 7: German capacity development for different scenarios Besides data abot installed capacities of renewable energies, emissions as well as exports and imports are determined at the first stage. emissions for single contries are compted from power plant operation. The transfer of emissions bonds instead of prices allows achieving overall greenhose gas abatement targets. Moreover the emission bdget can be allocated optimally over the 16

18 years. In contrast to the first stage with a recrsive optimization approach, a shift of emissions between different periods is possible.. Total redction is driven by the exogenos Eropean redction objective. The distribtion to different regions is done in the economically most efficient way. In the first decade, redctions are cost-efficiently done by modernization of old and inefficient power plants. Later, redctions are reached by investment in additional renewable energies and CCS power plants which become less expensive with frther progress in research and development. Figre 8 shows the emissions for Germany in different years. Redction efforts are weakened in the second half of the observation period and by 2050 emissions are almost zero at least in the CLIMATE scenarios. One reason are more efficient CCS power plants which allow a cost efficient redction of emissions. Figre 8: Germany s emissions and trade balance 3.5 CCS investment and -prices Investments in CCS power plants occr in all scenarios. One might expect that the highest investments in CCS power plants occr in scenarios with most ambitios emission targets. Bt analyzing investments in CCS power plants in Germany, they are most sbstantial in the scenario CONVENTIONAL WISDOM (Figre 9). This shows the impact of demand development and existing power plant capacities on the investment decision. The impact of demand developments can even be observed today. Crrent low CO2 prices are cased by targets which were set in expectation of a stronger economic growth instead of a decline in demand. In the CONVENTIONAL WISDOM scenario, the emission target is comparatively low bt demand is rather high. In addition, emission redction in the CONVENTIONAL WISDOM scenario otside Germany is disproportionately low compared to the other scenarios. A frther key aspect is that the potentials of renewables available at low cost are limited. This is notably tre for wind onshore. Conseqently capacities of renewable energies are qite similar in all scenarios. Correspondingly the residal load is higher in the CONVENTIONAL WISDOM scenario. So the necessary CO2 abatement in the fossil sector is higher even thogh more emission allowances are available. The CO2 price reflects this reqired abatement in the fossil power sector. In this case the effect on the demand side dominates. As a reslt CO2 prices are the highest in the CONVENTIONAL WISDOM scenario. These high prices 17

19 together with high fll load hors are sfficient to garantee an economically viable operation of CCS plants and allow them to compete with renewable energies as well as other conventional power plants. In the climate scenarios, prices are lower and the operating hors of conventional power plants decrease de to a lower total electricity demand. As a reslt, operation of CCS-power plants is economically hardly viable. For that reason, the following analysis focses on the scenario CONVENTIONAL WISDOM. Within this scenario two cases are considered. In the first case storage of is possible withot any regional restriction (all storage case), in the second case storage of is restricted to offshore storage (offshore case). In both cases total storage capacities are considered. For the all storage case and the offshore case similar investment strctres can be observed. Especially in the years 2030 and 2040 investments are dominated by natral gas fired power plants. Becase of their low investment costs, their flexibility and their low emissions they are especially sed as back p technology for flctating renewable energies. As a reslt of the efficiency gains for all power plant technologies dring the whole period and interest rate effects, cost intensive redction efforts are postponed to the end of the optimization period. Deployment of CCS power plants mostly differs between climate scenarios and the scenario CONVENTIONAL WISDOM. Here, investments are ten times higher. In addition, sbstantial investments in this technology already occr in 2040 while investments are qite low in the climate scenarios (Figre 9). For the case CONVENTIONAL WISDOM all storage CCS power plants become economically feasible from 2040 onwards with 2700 MW of lignite and 3000 MW of coal power plant with captre being newly installed in This lignite capacity increases to 12 GW in 2050, coal fired capacities are even enlarged to 12.3 GW. Gas combined cycle (CC) CCS is not bilt. In the offshore case investments mainly occr in the year Here, coal CCS has capacities of 380 MW while lignite fired CCS power plants reach a capacity of 800 MW in These capacities increase in the year 2050 to 5.1 GW and 7.9 GW, respectively. Figre 9: Power plant investments in different scenarios (case all storage) 18

20 In comparison to the all storage case, investments in the offshore case have a similar strctre bt are significantly lower in all scenarios. The reasons are higher storage costs in the North Sea and additional costs for transport de to longer distances. These higher costs can also be observed in prices which are derived from the shadow price of the intertemporal emission bond. For the offshore case (CONVENTIONAL WISDOM) prices are 35 /t higher compared to the all storage case in the year 2050 (60 /t). prices are lower in the climate scenarios with prices of 28 /t and 36 /t in 2050, respectively. As a reslt investments in CCS power plants are low with 1.4 GW (CLIMATE) and 1.8 GW (CLIMATE+). In the offshore case, these investments are frther redced as a conseqence of higher storage costs. prices are only slightly increased by 3 /t becase of the small share of CCS power plants in total power prodction. 19

21 3.6 storage operation In line with the increased capacities of CCS technologies, the stored amont of increases over the years (Figre 10). Thereby it is ensred that CCS power plants have sfficient storage capacity over their lifetime even beyond the optimization period. Storage operation is mainly driven by assmptions on availability and costs. De to the cost minimizing approach less expensive storage facilities are sed first. Costs consist of storage and transport costs. Therefore onshore storage is sed in the all storage case being cost efficient compared to offshore storage. As to sitable storage sites in Germany, there are two special characteristics that have great inflence on the decision where to store and which transport rotes are preferable. First, the storage sites with the greatest potential are all in the Northern provinces like Lower Saxony and Schleswig-Holstein. Second, Germany has abndant lignite reservoirs in the middle belt of the contry from the Rhineland area in the west to the Lasitz area in the east. Figre 10: -storage operation - scenario CLIMATE (l), CLIMATE+ (m) and CONVENTIONAL WISDOM (r) According to the model reslts the CCS-technology is mostly sed for lignite power plants. De to the fact that lignite cannot be economically transported over long distances, the lignite plants are all linked to the lignite reservoirs. Ths, the CO2 has to be shifted from the lignite regions in the middle of the contry to storage sites either in the north or to the offshore-sites in the North Sea (Figre 11). Compared to a total storage potential of abot 12 trillion t, only abot 10 % of the total storage potential in Germany is sed ntil 2050 in the all storage CONVENTIONAL WISDOM case. Having sfficient onshore storage potential, a sage of offshore facilities is not necessary. In the climate scenario tilization of storage capacities is even lower. 20

22 For the offshore case a storage bottleneck cannot be observed, either. The amont of stored is even lower than in the all storage cases. Notably higher investments in -pipelines are reqired becase of longer transport distances what makes the CCS technology less profitable. 3.7 transport is often not stored in the region where it is captred bt is transported to other regions for storage prposes. Thereby it is obvios that pipelines are in competition with grid extensions. De to losses in power transport, power transmission costs mainly depend on the crrent electricity price. As mentioned in the section before, the reslts show an extensive se of the CCS technology in lignite plants. Figre 11 emphasizes the prominent transport rotes from the lignite reservoirs to the large storage sites in the north. In case of CCS applications to hard coal plants (most notably in Sothern Germany), the nearby local potential for storage trns ot to be sfficient and long range transport is not necessary. Figre 11: transmission capacity in 2050: CONVENTIONAL WISDOM all storage (l), offshore (m), CLIMATE+ all storage (r) A frther point is the need for firm capacity in the different regions which can ensre secrity of spply and system stability in a system with large amonts of renewable energy infeed. This means that back p capacities have to be considered in addition to a pre trade-offs between transport and power transmission. The decision calcls has then to weigh CCS power plants at the storage sites pls power transmission and back-p capacities in the demand regions against CCS power plants in the demand regions and CO2 transport to the storage sites. Therefore is captred in areas with high electricity consmption and transported to regions with sfficient storage capacity. Comparing the all storage and the offshore case, in the latter case the pipeline system has to be developed frther in the north (Figre 11). This leads to higher costs for redction and finally prices increase. As a reslt also operation of conventional power plants becomes more expensive. 21

23 4 CONCLUSION This article presented a model framework for the evalation of CCS with an application to the German power market inclding district heating. Therefore a bndle of scenarios was developed with different emphasis on climate change mitigation. The comptation reslts show that investments in the CCS technology are strongly dependent on the scenario selected, notably with respect to the development of emissions and prices respectively as well as residal demand and fll-load-hors. Investment costs have another considerable impact on investment decisions. The interdependencies between the bond, the price and total demand have been identified as the main inflencing factors. Only in cases with high redction targets compared to electricity demand, investments in CCS technology become economically feasible. Following the model reslts, we come to the conclsion that a comprehensive investment in CCS power plants is not likely in the next decades. This corresponds to other stdies which do not see a bsiness case for CCS power plants nder the crrent framework conditions in Germany (cf. Viebahn et al. 2012). A crcial point is the extension of renewable energies. De to sbsidies they might reach a market share where an investment in alternative technologies with low CO2 emissions does not make sense from an economic point of view. However, this is not the main driver for non adaptation of CCS in or scenarios. Accordingly, the restricted storage potential for is sfficient in these scenarios. Pblic acceptance is important besides technical and economical restrictions. Therefore we have investigated a case where storage of is restricted to offshore areas. We fond ot that costs for transport increase along with abatement costs. As a reslt, the sage of CCS power plants is redced. Again, available offshore storage capacities are no limitation to the diffsion of the CCS technology. APPENDIX Nomenclatre Variables E transmission flow OC operating costs FC fixed costs P power prodction H storage level SC start-p costs L capacity TC total costs STC storage cost V captred emissions Indices 0 minimal op Operating CCS CCS oth Other CO2 r Region cyc cycling pm Pmping 22

24 fix fix res power reserve FUEL fel s stochastic stage inv investment stor -storage life lifetime st start-p n node t time segment new new t final time step onl online nit type m maximm y year Parameters a annity factor l minimm load factor C transmission b life time capacity c specific costs W water inflow cr captre rate η efficiency d dration fraction of prodction D energy demand occrrence probability f freqency availability i interest rate transition probability p price emission factor B bond REFERENCES Abadie, L.M., Chamorro, J.M., Eropean prices and carbon captre investments. Energy Economics. 30(6), AGEB (Arbeitsgemeinschaft Energiebilanzen e.v.), Energieverbrach in Detschland im Jahr 2011, Berlin. Andersen, A.N. and Lnd, H., New CHP Partnerships Offering Balancing of Flctating Renewable Electricity Prodctions. Jornal of Cleaner Prodction. 15(3), Bakken, B., von Streng Velken, I., Linear Models for Optimization of Infrastrctre for Captre and Storage. IEEE Transactions on Energy Conversion. 23(3), BMU, Leitszenario Langfristszenarien nd Strategien für den Asba erneerbarer Energien in Detschland nter Berücksichtigng der eropäischen nd globalen Entwicklng. [Online]. Available: BMWI, Forschngs-nd Entwicklngskonzept für emissionsarme fossil befeerte Kraftwerke. BMWI, Energy Concept for an Environmentally Sond, Reliable and Affordable Energy Spply. 23

25 BMWI, 2013, Energiebedingte CO2-Emissionen nach Energieträgern /Energie/Energiedaten-nd-analysen/Energiedaten/energie-mwelt.html Boitex, M., Peak-load pricing. Jornal of Bsiness. 33(2), Capros, P., Mantzos, L., Papandreo, V., Tasios, N., Klaassen, G., Energy systems analysis of CCS development in Erope. 5th Int. Conf. EEM Chao, H., Peak load pricing and capacity planning with demand and spply ncertainty. The Bell Jornal of Economics. 14(1), Chen, Q., Kang, C., Xia, Q., Zhong, J., 2010a. Power Generation Expansion Planning Model Towards Low- Carbon Economy and Its Application in China. IEEE Transactions on Power Systems. 25(2), Chen, Q., Kang, C., Xia, Q., Zhong, J., 2010b. Real option analysis on carbon captre power plants nder flexible operation mechanism. IEEE Power and Energy Society General Meeting, pp Christensen, N. P., Holloway, S., Geological storage of from combstion of fossil fel (GESTCO). Eropean Union Fifth Framework Program for Research and Development. Project No. ENK6-CT , Smmary Report, vol. 2. Cozijnsen, J., Towards the se of Captre and Storage in the EU Emissions Trading System. [Online]: Cremer, C., Integrale Analyse der Infrastrktr zr -Abtrennng nd Speicherng mit EMET- Captre, in Form für Energiemodelle nd Energiewirtschaftliche Systemanalysen in Detschland (Ed.), Energiemodelle z Innovation nd moderner Energietechnik: Analyse exogenen nd endogenen technischen Fortschritts in der Energiewirtschaft, Münster, pp Damen, K., Troost, M. v., Faaij, A.P.C., Trkenbrg, W.C., A comparison of electricity and hydrogen prodction systems with captre and storage. Part A: Review and selection of promising conversion and captre technologies. Progress in Energy and Combstion Science. 32(2), Eropean Commission (EC), National Renewable Action Plans. [Online] Eropean Commission (EC), Energy Roadmap COM (2011) 885/2, Brssels. Fowler, M., The Role of Carbon Captre and Storage technology in attaining global climate stability targets: A literatre srvey. Clean Air Task Force, Boston. Gogh, C., Shackley, S., Holloway, S., Bentham, M., Blatov, I., McLachlan, C., Klemes, J., Prdy, R.W., Cockerill, T., An integrated assessment of Carbon dioxide captre and storage in the UK. Proc. GHGT-8 Conf., Trondheim, Norway. Gerbelová, H., Versteeg, P., Ioakimidis, C.S., Ferrao, P., The effect of retrofitting Portgese fossil fel power plants with CCS. Applied Energy. 101, Green-X, Deriving optimal promotion strategies for increasing the share of RES-E in a dynamic Eropean electricity market. [Online]: Harrison, L., New Technology Magazine - Half-Billion-Dollar Pipeline Proposed For 2011 Start.. [Online]: [Accessed: 29-Nov-2010]. Hipp, F., Abschätzng der Kapazitäten zr daerhaften geologischen Speicherng von Kohlendioxid 24