Flow based Market Coupling

Transcription

1 2007 Engineering and Policy Analysis Manoj Sharma Flow based Market Coupling W h a t we kn o w, W h a t we d o n t kn o w a nd W h a t we ne ed t o k n o w Faculty of Technology Policy and Management +31 (0) [email protected] To develop a decision support model for evaluating the implications of and impediments to implementing Flow based Market Coupling for alleviating the issue of interconnector congestion in North West European region.

2

3 Flow based Market Coupling What we know, What we don t know and What we need to know Name Manoj Sharma Student number address [email protected] Date August 2007 University Delft University of Technology Faculty Technology, Policy and Management Programme Engineering and Policy Analysis Division Energy and Industry Economics of Infrastructures Intern Ministry of Economic Affairs, The Netherlands (EZ) Graduation Committee Prof. dr. ir. M. P. C. Weijnen Professor TU Delft dr. ir. R. A. Hakvoort First Supervisors TU Delft mw. ir. H.M. de Jong ir. M. Jonker Second Supervisors TU Delft Prof. ir. W.L. Kling drs. André Jurjus External Supervisors EZ drs. Lineke den Ouden mr. drs. Jon Eikelenstam drs. Kick Bruin

4

5 Dedicated to my parents

6

7 ACKNOWLEDGEMENTS I am deeply indebted to my supervisor Dr. Rudi Hakvoort, from the Technical University Delft whose support, stimulating suggestions, encouragement and confidence in my abilities enabled were a source of constant motivation. I cannot possibly imagine completing this work without his guidance and pragmatism. I also thank him for providing me the opportunity of working on a real policy issue, in whatever small measure possible. I also sincerely thank Hanneke de Jong for her daily presence and continuous mentorship through the period of my thesis. Her continuous feedback and comments have played the most important role in shaping up this work. I would like to thank wholeheartedly my supervisors, Professor of the Energy & Industry section, Margot Weijnen for her scientific advice and astute feedback on the thesis and Prof. Wil Kling for the his support on understanding the working of complex electricity systems and components and sharing his knowledge of electricity markets. I would like to thank Martijn Jonker for being on my graduation committee. I would especially like to thank people from the Ministry of Economic Affairs, Andre Jurjus, Lineke den Ouden, Jon Eikelenstam and Kick Bruin for their valuable feedback and insightful discussions. Their presence gave a direction to an otherwise very broad project. Their questions and deadlines were essential to keep the project on tractable both in terms of focus and duration. I would also thank my fellow students especially, Siamak Vaezpour and Paul Dailey, for sharing thoughts and offering an avenue for venting difficulties during much needed and unforgettable coffee brakes. I want to thank them for all their help, support, interest and valuable hints. I feel profoundly indebted to TU Delft and Dutch Ministry of Education for offering me financial assistance during my graduate studies without which I could not have possible undertaken studies in the Netherlands. Finally and especially, I would like to give my special thanks to my family members back home whose patient love and encouragement enabled me to complete this work. August 2007 Manoj

8

9 1 ABSTRACT Since late 90 s the European electricity market has been opened aiming towards development of a single organized competitive internal electricity market. However given the limited capacities on the tie lines between countries achieving a single market is not a possibility yet. Congestion on these tie lines separates geographic markets into submarkets. Indeed cross border congestions run counter to the idea of the creation of a competitive market. Because one cannot build infrastructures of infinite capacities, congestion is unavoidable but should not be excessive and it is desired to economically allocate the limited transmission resource. In that context the optimal utilization of the available capacities is a major issue. To be economically efficient the price of transmission capacity should be based on its marginal utility. However, the path that electricity will take to reach from point A to point B connected by a meshed network is hard to predict on the day ahead basis. Flow based market coupling (FMC) is a suggested mechanism for cross border congestion management for the Northwest continental European region which takes into account both the market and technical aspects of system to calculate dispatch. The Dutch government has expressed interest in possibility of further regional market integration. Though FMC holds a promise of being an upgrade over the current mechanisms, it still is a complex mechanism and is often not completely understood. It is thus necessary to further research FMC to be able to identify and delineate possible technical and economical implications from its implementation and search prognostic solutions prior to its introduction. This master thesis aims first at giving an understanding of the flow based method. Second objective of the work was to try validating the claims of efficiency and welfare gains offered by flow based method. This entailed designing a decision support model based on the technical economic realities which contribute to making some of the major issues transparent to the decision makers in current case Dutch Minister of Economic Affairs. The decision support model was based on realistic data from the four countries involved with FMC and delineates the effects on both regional and national level also offering an understanding into the distribution of benefits and costs amongst the involved actors. These results would help in further progress towards acceptance and implementation of FMC. In the last part of the work institutional aspects involved with adoption of flow based method were studied be highlighted along with listing of important actors, since though FMC might theoretically be efficient to reap these benefits in reality it will require a total coordination of all the actors concerned, and commitment from the involved governments. This report aims to be useful for people working on a project related to the cross border electricity exchangesespecially FMC. The report is written in a manner that it would be understandable to a person with no prior specialized knowledge of electricity networks or economics, though a basic understanding of concepts from electricity markets would be of assistance. Report will be useful for anybody who is interested in cross border electricity exchanges even if it is not his domain of study, with special focus on the North West European markets. Keywords: power system, electricity market, cross border exchanges, flow based market coupling. Chapter: Abstract i

10 ii Flow based Market Coupling

11 2 GLOSSARY 2.1 COUNTRY CODES NL BE FR DE Netherlands Belgium France Germany 2.2 OTHER ABBREVIATIONS Abbreviation Signification HV ISO TSO ETSO UCTE NTC ATC PTDF MCP NWER OECD FC VC TC AD High Voltage Independent System Operator Transmission System Operator Association of European Transmission System Operators Union for the Coordination of Transmission of Electricity Net Transfer Capacity Available Transfer Capacity Power Transfer Distribution Factor Marginal Clearing Price North West European Region Organisation for Economic Co operation and Development Fixed cost Variable Cost Total Cost Average demand Q* Equilibrium demand/supply quantity P* Equilibrium price LMP EC Locational Marginal Pricing European Commission CM FMC Congestion Management Flow based Market Coupling Chapter: Glossary iii

12 iv Flow based Market Coupling TLC EDF RTE MC Trilateral Market Coupling Electricité de France Réseau de Transport Electrique Market Coupling 2.3 DEFINITIONS Power system: Set of means of electricity production, transport and distribution and loads (national electricity network) Marginal clearing price: Price of the last selected bid in the merit order. Spot price: Price set by the intersection of the selling and the buying bids on the spot market. Explicit auction: Auction where the product sold is a right to program an exchange on a given contractual path. Implicit auction: Auction for which the allocations capacities are combined with the functioning of an organized electricity market. Market coupling: Consists in a coupling of N markets, according to a decentralized approach, resulting in a unique virtual market as long as the interconnections capacities are not saturated. Market splitting: One market which is split in N independent virtual markets when the capacities are saturated. Market value: The linear combination between the bid price and the allocated capacity NTC: The net transfer capacity represents the best estimate limit for physical electricity transfer between two areas ATC: The available capacity which can be defined for different periods of time (year, month, day ) Critical branch: A line of the network which could probably limit the commercial exchanges between the areas because it often reaches its physical limits. Participating nodes: These nodes represent the junction point of the units to the network where the production varies significantly when the total production of the zone varies. PTDF: Based on a state of the initial network, this is the influence on the flow on a critical branch of every additional MW injected at a participating node. Netting: Take into account that the exchanges in an opposite direction counterbalance each other.. F max : The maximum amount of electricity that can be exported or imported by a country depends on the capacity of the interconnector, this is called the Flow gate capacity (F max ), calculated according to flow based method

13 . F ref : A part of the maximum allowed flow on the critical branches is already used (i.e. prior to the allocation) by so called already occupied flows, this is termed as F ref, this is calculated considering flow based method Approximation of the direct current load flow solution: approximation where the following hypotheses are taken into account: V i =V N, where V N is the corresponding nominal voltage R=0 No losses The angles θ between nodes are small and thus sin(θ) ~ θ Chapter: Glossary v

14 vi Flow based Market Coupling

15 3 STRUCTURE OF REPORT Section 1 Electricity Markets This section gives a brief introduction to the specifics of the electricity markets. Details the history of congestion management in electricity followed by an introduction to the idea of flow based market coupling for managing congestion. Section 2 Simulation Model A step by step description of how the model was created is presented in this section. Details and relevant data is presented in the appendices. Appendix All graphs and numbers for the results are attached with the appendix. Other important documents can also be found in the Appendix. Report refers to appendix. Section 3 Results Here the results from the model are presented and conclusion are drawn Section 4 Market Design Issues The final section of the report sums up the research by identifying and delineating the risks and issues associated with implementation of a new market design in a multi actor setting. The main focus would hinge around the institutional and informational aspects Conclusions and Reflection The conclusion, reflections and scope for future work is outlined in this section. Chapter: Structure of Report vii

16 viii Flow based Market Coupling

17 CONTENTS 1 Abstract... i 2 Glossary... iii 2.1 Country Codes... iii 2.2 Other Abbreviations... iii 2.3 Definitions... iv 3 Structure of Report... vii Figures... xiv Tables... xvi 1 European Electricity markets Background Electricity The special Case Present European electricity system March towards Internal Energy markets Conclusion Congestion Management Congestion Management Methods Trilateral Market Coupling Market splitting Locational marginal prices (LMP) Conclusion Flow based market coupling Operation of market under FMC Flow Based Modeling Market Coupling Market Process of FMC difference between FMC and Market Coupling Conclusion Problem Formulation Problem Statement Scope Definition Geographical Scope Temporal Scope Perspective/Problem owner Sub Questions System behavior Security of supply Chapter: Structure of Report ix

18 x Flow based Market Coupling Welfare effects Implementation issues Research design Structure of Simulation Model Transmission network module Power Transfer Distribution Factor Use of PTDF Calculation of PTDF in PowerWorld How to get PTDF from PowerWorld Electricity demand and supply module Supply Curve Cost of electricity generation Installed generation capacity Calculation of Supply Curve Slope Demand Curve Tie line capacity Electricity day ahead market module Equilibrium without Market Coupling After Market Coupling Model Outputs Outputs for the Non Coupled markets Outputs for the coupled markets Overall Results Comparative Results Capacity Utilization Results Graphs Validation and Verification of model Validation of transmission model Quality of data used for supply curve Quality of data used for average demand Price of electricity without imports and exports Validation of PTDF values from Power World Conclusions Results Base Case Price of Electricity Electricity Demand Welfare Data... 67

19 7.1.4 Interconnector Capacity Utilization Conclusions from base case Current Situation vs FMC Comparison of FMC with Market Coupling Scenarios Increase generation capacity in Netherlands Increased tie line capacity Discussion on sensitivity of PTDFs Effect of Carbon Market Decoupling of German system into North and South Block Actor Analysis General Overview Structure of the electricity markets List of actors Interests Generators Consumers Regulator Government Traders TSO Power Exchanges Position of Actors Power Exchanges Generators TSOs Strategic Generators Questions raised by the actors Regulator TSO Power Exchange Results from FMC Conclusions Market design issues Legal and Organizational Issues Technical Issues Market related issues Transparency towards the market Chapter: Structure of Report xi

20 xii Flow based Market Coupling Economic signals to market participants and sharing of congestion income Liabilities of TSOs and position of individual regulatory authorities Technical implementation issues Regional Base Case Phase shifting transformers Network representation Generation representation Definition of flow gate capacity (F max ) Type of bids Possible solutions Conclusions System behavior Security of supply Welfare effects Implementation issues limitations and Research recommendations Limitations of the study Research recommendations A Appendix A.1 Cost of electricity generation model... 1 A.2 Data sources and assumptions for Cost of electricity generation model... 5 A.3 National Supply Curves... 8 A.4 Average Historical Market Price of Electricity A.5 Demand Worksheet from Market Module A.6 Tie line Capacities A.7 Power Transfer Distribution Factors for North West European Region A.8 Electricity day ahead market module A.8.1 Constraints A.8.2 Objective Function A.9 Electricity day ahead market Results A.10 Capacity Utilization Results A.11 Comparison of Market Coupling with Flow based Market Coupling A.11.1 Base Case Flow based market coupling A.11.2 Base Case Flow based market coupling with NL reserving 1500MW on DE NL border A.11.3 Market Coupling A.12 Result of FMC after installation of 5000MW of coal based power plants in NL A.13 Result of FMC after installation of 10,000 MW of coal based power plants in NL A.14 Result of FMC after installation of 2000 MW transmission capacity between DE NL A.15 Increase tie line capacity between Fr BE +2000MW... 30

21 A.16 Result of FMC after installation of 5000 MW transmission capacity between DE NL A.17 Result of FMC after installation of 10,000 MW transmission capacity between DE NL A.18 Result of FMC after installation of 2000 MW transmission capacity between DE NL with PTDF from 5000 MW case A.19 Effect of Carbon Market A.18.1 FMC results with emission price of 15 /ton of CO A.18.2 FMC results with emission price of 20 /ton of CO A.18.3 FMC results with emission price of 30 /ton of CO A.18.4 FMC results with emission price of 50 /ton of CO A.20 List of Actors A.19.1 Netherlands A.19.2 Belgium A.19.3 France A.19.4 Germany Bibliography Websites Chapter: Structure of Report xiii

22 xiv Flow based Market Coupling FIGURES Figure 1 Present state of integration of electricity markets in Europe... 3 Figure 2 Percentage of flows in FR BE NL DE CH A I zone for a commercial transaction of 1 MWh from France to Germany (IEA Workshop Presentation 2004)... 6 Figure 3 Planned increase of interconnections in the EU (% of installed generation capacity) Figure 4 Current congestion management methods in NWE region Figure 5 Prices after implementation of TLC showing convergence ( 13 Figure 6 Physical energy flows in 2006 (all values in GWh) from 13 Figure 7. Outline of FMC a combination of flow based modelling that takes care of the physical flow path and market coupling to maximize the competition and efficiency of electricity markets (ETSO and EuroPEX JWG 2004) Figure 8. Market Coupling between two markets leading to increase in the net welfare (based on H. de Jong and R. Hakvoort, 2005) Figure 9. The process of pricing under FMC, this will be modelled for a simplified market case to create a software model that helps understand the market operation better (ETSO and EuroPEX JWG 2004) Figure 10 Transfer capacity definitions Figure 11 Interdependencies of NTC between two araes ( net.org) Figure 12 The system under consideration Figure 13 Research design Figure 14 Design of the simulation environment Figure 15 Interconnected Network of UCTE ( 32 Figure 16 Physicl flows because of a transaction of 100 Units of trade between France and Germany ( net.org) Figure 17 Generic Reference case for calculation of PTDF Figure 18 Schematic of the Transmission Network Module Figure 19 THe utilization of interconnector between the countries in the nort west european region from the modle run in PowerWorld Figure 20 UCTE network with enhanced NL BE border Figure 21 Schematic of the Electricity Demand and Supply Module Figure 22 The equation of the demand function for the North West European Region Figure 23 Tie Line capacity for the north west european electricity network (based on data from website of Tennet, RTE, RWE, Elia and TSO auction website) Figure 24 Schematic of the Market Moduel Figure 25 Principle of MCP (from 47 Figure 26 The definition of consumer surplus and producer surplus Figure 27 Simple Stylized electricity market Figure 28 Effect of import and export on prices and quantity in the country Figure 29 Welfare Calculations after inter country trade Figure 30 Optimization using MS Excel Solver Figure 31 Optimization parameters Figure 32 Effect of an inelastic demand for deciding the objective function Figure 33 Merit order curves for the North West European Region, model data and comparision with actual data from the past Figure 34 Load duration curve for Netherlands Figure 35 Load Duration Curve Belgium Figure 36 Load duration curve for France Figure 37 Load Duration curve for Germany... 59

23 Figure 38 Comparision of wholesale price of electricity based on the model and actual past data from the power exchanges for year Figure 39 MCP, equilibrium demand and surplus calculation without market coupling Figure 40 Traded volumes for FMC and Trilateral Market Coupling comparision Figure 41 results from simulation model for current state of trade in north west European region Figure 42 Flow Direction for comparision of FMC with NTC based Market Coupling Figure 43 New Supply curve with additional 5 GW of hard coal based power plant Figure 44 Price change in NL, Imports and % tie line utilization for scenario with installation of 5000 MW coal based generation in NL Figure 45 New supply curve with additional 10,000 MW of hard coal based power plants Figure 46 Price change in NL, Imports and % tie line utilization for scenario with installation of 10,000 MW Hard coal based generation in NL Figure 47 The dialog in PowerWorld to set up the values of parameters for the transmission line Figure 48 New tie line capacity values Figure 49 Results from FMC after installation of 2000 MW transmission line between DE NL Figure 50 Price change after installation of +2000MW on NL DE Tie Line Figure 51 Price change with installation of +2000MW on BE FR Tie Line Figure 52 Utilization of flow line with MW on BE FR line, COmapre to Figure Figure 53 Results from FMC after installation of 5000 MW transmission line between DE NL Figure 54 Price change and % Utilization from +2000MW on DE NL Border and PTDF's also from +2000MW case Figure 55 Price change and % utilization of interconnector from +2000MW capacity on DE NL border and PTDFs from +5000MW Figure 56 Contribution of Electricity and heat generation towrds total carbon emissions ( 86 Figure 57 Development of Carbon prices in EU, Including the crash in april 2006 (SOurce: Point Carbon) Figure 58 Effect of carbon pricing on the merit order within north west european region Figure 59 Results from FMC after considering carbon pricing Figure 60 The divisionof Germany into north and south block Figure 61 Flows after splitting germany into north and south blocks Figure 62 Annual share of daily wind power in respective daily peak demand on e on grid (Germany) Figure 63 The difference between transit and loop flows (Green arrows mark the transit flows, Grey arrows mark the Loop Flows) Figure 64 The electricity value chain Figure 65 Present position of the power exchanges Figure 66 Main strategic actors for the norty west European region (From vattenfall annual report 2006) Figure 67 The common data for cost of electricity... 1 Figure 68 Fixed cost data for electricity generation... 1 Figure 69 Variable cost data for electricity generation along with emission costs... 2 Figure 70 Formulation for the capital cost calculations... 3 Figure 71 Formulaiton of the Variable and Total cost... 4 Figure 72 Data points for the merit curve (supply curve)... 8 Figure 73 Supply curve for Netherlands... 8 Figure 74 Electricity supply curve for Belgium... 9 Figure 75 Electricity supply curve for France... 9 Figure 76 Electricity supply curve for Germany Figure 77 Equation of the supply curves (linear regression on the merit order curves) Figure 78 Distribution of MCP for Netherlands APX, year Figure 79 Distribution of MCP for Belgium Belpex, year Chapter: Figures xv

24 xvi Flow based Market Coupling Figure 80 Distribution of MCP for France Powernext, year Figure 81 Distribution of MCP for Germany EEX, September 06 Jun Figure 82 Demand Work Sheet Values Figure 83 Demand Worksheet Formulation TABLES Table 1 Senstivity matrix and PTDF based on the UCTE winter peak model from PowerWorld Table 2 Categories of generation technologies Table 3. Total cost of generation for different technologies Table 4 Formulation of Cost sheet Table 5 Installed generation capacities in the north west european region Table 6 Supply curves slope for the four countries Table 7 Average demand of electricity based on the UCTE system adequacy forecast Table 8 Reliable generation capacity based on the UCTE system adequacy forecast Table 9 Output list for model from non coupled markets Table 10 Outputs from the model after FMC Table 11 Comparative results from the model between FMC case and the base case Table 12 Capacity utilization results from the simulation model Table 13 Comparision of average demand of electricity used in the model and historical data Table 14 PTDF value comparison from power world model to actaul published values Table 15 Parameters for the new transmission line between DE NL Table 16 Comparision of the PTDF values before and after installation of 2000 MW transmission line between DE NL Table 17 Comparision of the PTDF values before and after installationof 5000 MW transmission line between DE NL Table 18 Impact of Carbon pricing on the merit order based on fuel and energy technology type (the items marked with olive are the once that moved) Table 19 Average demand in north west european region (UCTE system Adequacy report ) Table 20 PTDF values after splitting Germany into North and south block Table 21 General characteristics of the north west european electricity market Table 22 List of all relevant actors for the north west european region Table 23 Interests of the actors and the quantifying/qualifying parameter Table 24 Share of generaion companies in the total generation within north west european region Table 25 Details about the most strategic players in the North west european market Table 26 Actors perception of FMC Table 27. Total cost of generation for different technologies based on Vattenfall Table 28 Consolidated MCP data for centre west European region for Table 29 TIe line capacity values... 16

25 Section 1: Electricity Markets A general introduction to the electricity markets followed by focus on the European system in general and northwest European region in particular. Congestion management methods are introduced. Flow based market coupling is also explained. The section concludes with problem definition. Chapter: Tables 1

26 2 Flow based Market Coupling

27 1 EUROPEAN ELECTRICITY MARKETS With the opening of the electricity markets in Europe through the Directive (EU96/92) enforced on 19 February 1999, the European market integration could not result into a copper plate. The variety of generation mixes among the fifteen member countries, and the state of interconnection ties between them, has resulted into regional markets interfaced by bottlenecks, rather than into a single market with a unique price. European Authorities have quickly understood this situation and defined a new European Regulation enforced since 1 July 2004, which promotes market based congestion management mechanisms, able to provide efficient use of the interconnection as well as appropriate market signals giving the right incentives for transmission or generation investments. EU Vision 27 Individual electricity markets One Integrated (internal) electricity market Current Progress Several regional markets FIGURE 1 PRESENT STATE OF INTEGRATION OF ELECTRICITY MARKETS IN EUROPE BASED ON (PETER SIMIG, CONGESTION MANAGEMENT PRACTICE AT CROSS BORDER TRADE IN ERRA REGION, ERRA EU INTEGRATION WORKING GROUP, BUDAPEST, HUNGARY, 13 SEPTEMBER 2004) Chapter: European Electricity markets 3

28 4 Flow based Market Coupling 1.1 BACKGROUND Europe faces the problem of congestion 1 on the inter country electricity transmission networks (interconnectors 2 ). Notice that the EU definition on congestion is focused only on interconnector congestion, however it must be realized that congestion can also occur in the internal network. The investment in building new lines is a feasible long term solution; however the capital costs are high and return on investment depends strongly on both the regulatory environment and the future cross country demand supply trends. As a short term operational solution regulators and system operators spend considerable time in trying to optimally allocate capacity at interconnectors. In its preliminary sector enquiry report (ENERGY SECTOR INQUIRY 2006), DG Competition (European Commission) concluded that market integration in the European electricity sector is, among others, hampered by insufficient interconnector capacity. Besides the interconnector congestion issue there are two other important reasons for concern namely; the deficiency of incentives harbored in the current congestion management methods for maintaining a desirable level of investment in additional capacity and a lack of regulations/incentives to enforce/ensure efficient use of existing capacity. The fact that electricity does not follow the contract path, but flows in the network governed by Kirchoffs Laws makes the design of a suitable market mechanism for pricing the transactions difficult. For meshed transmission network, each flow has an effect on power flow conditions over the whole system; the new congestion management guidelines (amending the existing annex to the European Regulation (EC) 1228/2003) entered into force 1 December These new binding guidelines prescribe that TSOs should implement a flow based allocation system as soon as possible. However, currently ministries, regulators and maybe even TSOs have no clear overview of the political and regulatory consequences; e.g. the effect on the security of supply of an individual nation, the effect of transparency, the effect on the national welfare, etc. To be able to answer these questions it is important to understand the behavior of FMC and explore the sensitivities of the system that can be exploited by market participants. The thesis work aims to create a simplified representative model of the north west European electricity market under FMC; this would be used to understand the issues that might arise from implementation of FMC. The results would help the Ministries and regulators in understanding FMC and its behavior and also figuring out what to ask from the involved actors. 1.2 ELECTRICITY THE SPECIAL CASE Several significant features that distinguish the electricity industry from others contribute to the structure of market institutions: 1. Nonstorability 2. Intertemporal and random variability of demand 3. Necessity of balance in an interconnected transmission network 4. Direct connections to customers without any buffers 5. Capital intensity and economies of scale The demand for electricity varies continuously and unpredictably from hour to hour, day to day, and season to season. Practically, however, electricity cannot be stored. 1 Regulation on cross border trade of electricity (Regulation 1228/2003) Article 2 paragraph 2(c), provides definition for congestion involving international transactions: congestion means a situation in which an interconnection linking national transmission networks cannot accommodate all physical flows resulting from international trade requested by market participants, because of a lack of capacity of the interconnections and/or the national transmission systems concerned 2 High voltage links that couple national electricity systems and facilitate international competition

29 In an interconnected network, the demand and supply of electricity must be balanced instantaneously over time at every point of the network to maintain frequency, voltage, and system stability and to avoid power outages. Unlike other types of networks, electricity in an AC electric transmission network flows in directions determined by physical laws rather than by contracts and therefore is more difficult to control. A local variation of demand or supply of electricity affects power flows throughout the interconnected network. An equipment failure in one part of the network can cause the entire system to collapse. Efficiently meeting a new demand may involve coordinated adjustments of generators located far from the source of the demand. Therefore, the transmission system does not provide a simple physical connection between electricity generation and consumption facilities; it also involves active coordination of generating units dispersed throughout the network to meet variations in demand and supply. Due to these technical characteristics, electricity markets are inherently incomplete. The centralized organization afforded under the traditional structure of vertical integration facilitated the essential task of coordinating efficient system operation and balancing the supply and demand of electricity continuously in response to changing system conditions. The challenge now is to design a market organization that can accomplish system coordination in transmission without compromising the opportunities for market competition in energy. 1.3 PRESENT EUROPEAN ELECTRICITY SYSTEM The north west European region is a subset of the larger synchronous UCTE region composed of 18 continental European countries: Portugal, Spain, France (FR), Belgium (BE), Luxemburg, Switzerland, Italy, Netherlands (NL), Germany (DE), Denmark west, Czech Republic, Slovakia, Poland, Austria, Hungary, Slovenia, Croatia and part of Bosnia and Herzegovina. Synchronous interconnection means that individual systems are connected and being run together, at the same frequency, and assist each other if a disturbance occurs in a system. Vice versa this also means that major disturbances might propagate throughout this whole interconnected system and endanger its stability. UCTE is composed of the highly meshed system of 24 European countries consisting of some 200,000 km of 400 and 220 kv lines, hundreds of generating capacities directly linked to the system, and hundreds of substations ( Within a meshed system there will be parallel flows, which are hard to predict in advance. However, they would use up the available transfer capacity on the interconnectors. From an economic efficiency viewpoint both transit and loop flows should be charged for the marginal and opportunity cost of using the available capacity. However the system is governed by physical laws that do not follow economic reasons. Power flows over the multiple paths between supply and use according to least impedance, and not according to the shortest path or the path with the most unused capacity. In the present scenario tie line capacities are based on ATC/NTC values. ATC being the only limit for market operations greatly simplifies cross border trading. Considering that the electric network is a highly complex non linear system, such an abstraction has to be done in a careful manner. For example, one must recognize the difference between the commercial transfer between zones and actual flows in a system as illustrated in Figure 2. For a 1 MWh flow occurring between France and Germany, 0.34 MWh is flowing through Belgium and Netherlands route, while only 0.35 MWh flows directly from France to Germany flows through BE and NL are termed as transit flows. The remaining flows through networks of other neighbours. Up to now, most trade arrangements within Europe have been based on the contract path concept, where electricity is supposed to flow according to the chosen trading arrangements. This concept has been acceptable as long as it has been applied to long term steady transactions between integrated companies. Although it has severe drawbacks as already outlined in the example above, at least it prevents taking maximum benefit from existing transmission facilities in meshed networks. At most it can severely jeopardize power system security (e.g. the chaining contracts without visibility, that have been made possible thanks to the contract path paradigm, have been the origin of several very dangerous operational situations due to Chapter: European Electricity markets 5

30 6 Flow based Market Coupling unexpected flows through Belgium). Moving to the physical flow path concept while managing physical flows through the market itself poses a major challenge. ETSO and EuroPEX joint working group presented the Flow based Market Coupling (FMC) as a solution to congestion management which prices according to the physical path concept (ETSO and EuroPEX JWG 2004). However it is still different from the Location Marginal Pricing (for detail discussion on LMP refer chapter 2) or nodal pricing, as it would aggregate multiple nodes, the once belonging to one control area, into one super node. This super node is expected to cover all nodes that are geographically located in one country. As the EU directive only dictates FMC for interconnector congestion management, hence the prices will only be calculated for each region separately considering the maximum amount of electricity that can flow given the capacity constraints. Next section will discuss FMC in detail. FIGURE 2 PERCENTAGE OF FLOWS IN FR BE NL DE CH A I ZONE FOR A COMMERCIAL TRANSACTION OF 1 MWH FROM FRANCE TO GERMANY (IEA WORKSHOP PRESENTATION 2004) Other important consideration to be taken into account while designing a market based congestion management mechanism for electricity trade are the following: Electricity transmission is a natural monopoly, and hence needs special governing structures while dealing with efficiency issues.(crew and Kleindorfer 1985) Transport pricing issues because of one country one price paradigm do not agree with the economic optimization principles (Crampes and Laffont 2001) Pricing according to the physical flow path and contract path are different (Borenstein 2001; Budhraja 2003) Markets for electricity have undergone recent restructuring and liberalizing. Operations of electricity markets after restructuring is different as compared to vertically integrated companies that proceeded them (Bhattacharya, Bollen et al. 2001) Transmission lines are still regulated (as transmission is natural monopoly it needs to be regulated to prevent abuse of market power) and firm behavior is different under regulation not always maximizing the profit. (Averch and Johnson 1962) In case of Netherlands the system operator

31 TenneT is under a revenue cap with some form of international benchmark. Which is not the most economically optimal solution but limits the gold plating. The structure of the market is such that it will never be perfect competition because of huge capital costs involved. Oligopolies will exist in the generation side which will hinder operation of efficient markets. (Day, Hobbs et al. 2002) TSOs try to reserve some transmission capacity due to security reasons. This further complicates the efficient operation of any market design.(harvey, Hogan et al. 1996) International transmission trade would call in for standardization of the legal processes to enable neutral settling of the trade issues. (Knops, De Vries et al. 2001) The externalities due to loop flows in a transmission network represent a critical issue that must be resolved before competition can be successfully introduced into the electric power industry for longterm economic benefits.(chao, Peck et al. 2000) For the basic understanding of economics concepts book by Samuelson and Nordhaus was used (Samuelson and Nordhaus 2005). Power System Economics was used for more specific understanding into the operations of electricity markets (Kirschen and Strbac 2004). A book by Hunt was consulted for understanding the issues faced by efficient operation of electricity markets (Hunt 2002). To gain broader understanding into risks related to electricity infrastructures within European context book named Critical Infrastructures at Risk was also consulted (Gheorghe, Masera et al. 2006). 1.4 MARCH TOWARDS INTERNAL ENERGY MARKETS In 1987 the Single European Act actively promoted integration of national markets into one single European market and invented the 1992 Internal Market agenda. The Internal Energy Market (IEM) provides new opportunities to energy consumers and to energy undertakings. It has the potential to increase economic and technical efficiency, as well as security of supply, thus improving European welfare and the competitiveness of European industry. To realize Europeanization of energy sector three concepts were focused on liberalization, independent regulation and supra national integration of electricity markets, which are interrelated. The first IEM document (Directive 96/92/EC) defined some common rules to be applied by all Member States in order to open up their energy markets and established the legal basis for the IEM. For instance, the directives defined minimum unbundling requirements applicable to vertically integrated undertakings, minimum eligibility thresholds, a menu of network access regimes, etc. With opening up of market the cases where electricity is significantly cheaper in one member state than in its neighbor, a large demand for crossborder transports could occur. In some of these cases the demand for imports exceeded the available transport capacity on the (cross border) interconnector(s) leading to congestion. Transmission system operators (TSOs) have historically not designed interconnections between their networks with the primary objective of facilitating international power trade. As a consequence, the integration of national markets is impeded by the limited amount of cross border transmission capacity. However there have been plans to increase capacity of the interconnectors, as seen Figure 3, the by 2010 it is planned to raise the capacity of the interconnectors to almost 20% of the installed generation capacity. It becomes imperative to design efficient congestion management and capacity allocation methods for this additional capacity in order to maximize the economic benefit from trade. Chapter: European Electricity markets 7

32 8 Flow based Market Coupling Interconnector capacity as % of total installed generation capacity FIGURE 3 PLANNED INCREASE OF INTERCONNECTIONS IN THE EU (% OF INSTALLED GENERATION CAPACITY). For 2003 a level of approx 10% were foreseen, while the level graph shows that the actual level of interconnection capacity was only 7%. Currently the level was approximated at around 15% with the number growing up to 20% by year This underscored the importance that inter country trade of electricity has on the national markets. This would only become more important in the coming time(source: EU Directorate General for Energy and Transport, 2000) Congestion may hamper the full integration of different national electricity markets into a single market. The existence of negotiated third party access regimes, the limited level of unbundling obligations and the lack of an obligation to establish a national energy regulator were also viewed as obstacles to creation of competitive markets. To address these concerns, further measures were proposed by the Commission leading to the adoption of (Directive 2003/54/EC)( Second Electricity Directive ) and (Regulation (EC) No. 1228/ ) ( Cross Border Electricity Trading Regulation ). The content of the Regulation largely reflects the outcome of the work of the Florence Regulatory Forum (see DG TREN web site transport/). The Forum began its work in 1998, with an aim of identifying obstacles in the way of IEM and explores solutions, and has since issued several studies on the two subjects addressed by the Regulation. In 2004, ETSO and EuroPEX launched a joint proposal on Flow based Market Coupling (FMC) to achieve the goals envisioned by 1228/2003. On 1 December 2006 a new annex to the European Regulation (EC) 1228/2003 entered into force (congestion management guidelines) requiring that al congestion management methods applied must be market based which means that capacity shall be allocated only by means of explicit or implicit auctions. These new binding guidelines prescribe that TSOs should implement a flow based allocation system as soon as possible. 1.5 CONCLUSION Once all the above issues have been studied it would be possible to place FMC into the right perspective while considering the alternatives. This would contribute to one part of the report which is mostly based on literature and expert opinion and would study the long and short term implications of implementing FMC on Dutch interconnectors with Germany and Belgium, and the transactions from France. There is a need to understand what makes a market design successful. It has been proven in literature that theoretically all current congestion management methods today have the capability of being economically efficient (De Vries and Hakvoort 2000b; De Vries 2004). However they differ with respect to distribution of the costs and welfare. This could lead to opportunistic behavior and hence any market design should be closely monitored for any such opportunity. FMC also needs to be studied with respect to the distribution of the benefits and costs. The actors who bear the costs without any benefits will be against any such change and those who enjoy benefits without any cost would strongly support it. A neutral and efficient method would

33 need to distribute the benefits and costs justly. FMC has been said to build upon the present system without major structural changes and it can evolve over time. However it has been shown in past that unsuitable market designs like the one in California could lead to huge damage. Hence FMC should be scrutinized more rigorously especially with respect to the impact it would have on the Dutch electricity system and national energy security. Chapter: European Electricity markets 9

34 10 Flow based Market Coupling 2 CONGESTION MANAGEMENT Congestion occurs when the demand for transmission capacity exceeds the available transmission network capabilities. Every power system will from time to time experience this situation. It is not economical to invest to a level where all constraints are eliminated. The principles for congestion management in a country or a wider area are based on the legislation and regulation in place. An increased focus towards handling of congestions has developed over recent years. In Europe the EU Directive 2003/54/EC and the Regulation 1228/2003/EC draw up the basic principles for congestion management throughout Europe. Guidelines are being developed to detail the principles. Only market based methods for congestion management will be accepted in the future. Congestion separates geographic markets into submarkets. Congestion arises from the saturation of transport infrastructure. It may occur in any spatially distributed sector that requires transport facilities. Because one cannot build infrastructures of infinite capacities, congestion is unavoidable but should not be excessive. The frequency of congestion and the size of the resulting submarkets depend on both the capacity and the management of the infrastructure. This also applies to electricity, with the additional complexity that this product creates. In the next section various congestion management methods would be detailed. This would be followed by description of the current state of congestion management methods employed in north western Europe. 2.1 CONGESTION MANAGEMENT METHODS Congestion management in electricity is far from a new subject in the academic and professional literature. Its analysis began with Hogan s seminal paper (Hogan 1992) on nodal pricing (also referred to as locational marginal pricing (LMP)). The subject has since been extensively debated and there is now considerable theoretical knowledge on it. Market splitting is a simplified form of nodal pricing that was implemented in the Nordic system in 1993 ( New Zealand introduced a first version of nodal pricing as early as The real development began with the implementation of this method in PJM in 1998 and the progressive adoption of it by the voluntary pools of the East Coast of the United States. Congestion management methods can be divided into different sub categories. It can be methods for handling congestions between areas (inter area) or it can handle congestions within an area (intra area). Market based or not market based is another subdivision. Finally it should be divided between methods used to allocate capacity up to the capacity limit (allocation methods) and methods used to alleviate the transmission network down to the capacity limit (alleviation methods). Based on the literature review, the three forms of the congestion management have been adopted in deregulated electricity market (EM) around the world (Christie, Wollenberg et al. February 2000). One form is based on centralized optimization with some form of optimal power flow program or depending upon the control measures executed by the TSO for congestion relief. A second form is based on the use of price signals derived from ex ante market resolution to deter congestion by constraining scheduled generator output prior to real time operation. A third form seeks to control congestion by allowing or disallowing bilateral transmission agreements between a producer and a consumer, based on the effect of the transaction on the transmission system. Within Europe following are the cross border congestion management methods that are currently implemented (ETSO May 2006)

35 Access limitation: Access rationed by one or several independent companies which are not the owner of the network the link is connected to. Priority List: One gets capacity in a priority order until the whole available capacity is allocated. Examples of priority criteria are: chronological order (first come first served), past use of capacity, etc. Pro rata rationing: All requests are partially accepted in the way that each participant is granted a fixed share of his requested capacity amount. Explicit auction (ATC Based): The applicants have to declare the amount of capacity they want to obtain and how much they are willing to pay for this capacity. Bids are ordered and allocated by price. In the SEE region they recently introduces coordinated (regionally optimized) explicit auctions. Implicit auctions (Market splitting; Market Coupling): The electricity markets provide initially a common clearing. If the interconnector is congested, markets split into pre determined price areas cleared individually. Countries apply the methods of accession according to regulation EC 1228/2003. Regarding inter country transmission capacity within EU, regulation 1228/2003 of prescribes that the congestion management methods implemented by its member states must be market based and that allocation of capacity shall be made only by explicit (only capacity) or implicit (both capacity and energy) auctions.(regulation (EC) No. 1228/ ). Priority based rules, such as first come first served, and pro rata rationing are allocation based and non market based methods. This entails that access limitation, priority and pro rata will be soon phased out within EU. In Europe different assignment models are used. For example, the cross border capacity between the Netherlands and Germany is assigned through explicit auctions. Netherlands, Belgium and France day ahead market is assigned through trilateral market coupling; while the long term capacity (month/year) is allocated via explicit auctions. Nordic System uses the model of market splitting. The selection and implementation of a well suited congestion management model is a necessary precondition for developing a successful regional market(etso May 2006). Congestion management, in particular at the European level has become a relevant topic since liberalization of electricity markets is in progress. Boucher and Smeers (2001) analyzed the future organization of cross border trade in the European electricity market concluding that the economic principles as proposed by the European Commission in 2001 are not sufficient. Ehrenmann and Smeers (2004) analyze the regulation of cross border trade of electricity (Regulation 1228/2003) in terms of efficient congestion management. They conclude that market coupling although its implementation is more complex can path the way to a consistent system integrating the energy and transmission markets. Arriaga and Omos (2004) analyzed plausible congestion management schemes for the internal electricity market of the European Union. Taking a joint energy and capacity auction as benchmark they test two alternative approaches, an integrated transmission and energy auction and a coordinated explicit auction of transmission capacity followed by separate energy auctions at the different power exchanges. The authors propose the latter since it is relatively close to the actual market structures. Explicit and implicit auctions are the favored methods by the European Union. They are market based allocation methods. Market splitting and market coupling are special cases of implicit auctions. In an implicit auction energy and capacity are traded at the same time. In an explicit auction only transmission capacity is traded. Locational Marginal Pricing (LMP) is similar to implicit auction since both methods consider both energy and capacity. But LMP is basically the backbone of a market organization with central bid based dispatch. Chapter: Congestion Management 11

36 12 Flow based Market Coupling Examples of alleviation methods are counter trading, re dispatching, coordinated redispatching and transmission loading relief (TLR). Next section would outline the current approach of congestion management that has been implemented between France, Belgium and Netherlands. It is also important to discuss here the main features of two other successful market based methods for congestion management that have been implemented elsewhere. Market splitting (MS) as employed by Nordic countries and LMP as employed by PJM would be discussed subsequently. The chapter would end would an outline of FMC, which would be dealt in detail in the next chapter. 2.2 TRILATERAL MARKET COUPLING On 21 November 2006 the Trilateral Market Coupling ( TLC ) between the Netherlands, Belgium and France was successfully launched. The success of TLC proves the market coupling concept, and paves the way for the creation of a North West European energy market. Trilateral Market Coupling was the first coupling of separate, independent exchanges in Europe, so there were many challenges and new problems to solve. The success of the project was due to increased co operation between the six parties (RTE, Powernext, Elia, Belpex, TenneT and APX) together with the strong support of the Governments, the European Commission, the regulators and the market parties. However it was not a smooth sailing. Market coupling is a way to integrate power markets in different physical areas while requiring minimal changes to the local arrangements. Under TLC the three power exchanges continue to exist as separate legal entities with their own trading platform, contracts and clearing. The markets are nonetheless brought together by using the available transmission capacity to create a single regional market most of the time. The transmission capacity is used in an optimal way, enabling the best bids and offers to be matched from across the region. FIGURE 4 CURRENT CONGESTION MANAGEMENT METHODS IN NWE REGION

37 TLC replaced a two step process: a daily explicit auction of transmission capacity followed by the day ahead energy markets. This sequence has some inherent inefficiency. Market coupling integrates transmission allocation and energy trading, removing many of the inefficiencies at the day ahead stage. Explicit auctions are still used for the monthly and yearly allocation of transmission capacity rights. Trilateral market coupling has already had a clear positive impact. There has been strong price convergence across the three coupled day ahead markets, with a single price the large majority of the time (and separation in to 3 price areas being very occasional). The trend seems to be that convergence is increasing. FIGURE 5 PRICES AFTER IMPLEMENTATION OF TLC SHOWING CONVERGENCE ( As a result of the market coupling mechanism, the use of the available daily capacity (import and export flows between the Netherlands and Belgium) has changed significantly. There has been a marked increase in utilization of cross border capacity, occasionally reaching 100%, which theoretically must be 100% if there is any price difference between the countries. Historically, the Netherlands was a net importer but, while this is usually still the case during peak hours, it is becoming a strong exporter during off peak hours. These effects show that market coupling has contributed significantly to a better integrated and more efficient electricity market. Figure 6 shows the imports and exports between the countries in North West European region for the year Netherlands imports large percentage of electricity from Germany. FIGURE 6 PHYSICAL ENERGY FLOWS IN 2006 (ALL VALUES IN GWH) FROM Chapter: Congestion Management 13

38 14 Flow based Market Coupling Considering the benefits it is planned to extend the system to include Germany. The German power exchange EEX is also considered "a pacemaker" for European power prices. Inclusion of Germany would increase the number of actors involved in the trading and hence make the markets more competitive and hence more efficient. It is also proposed to use Flow based market coupling calculation to calculate the available transfer capacity on the tie lines. This is claimed to be more efficient way of allocating the available capacity as it considers the physical realities of electricity flow based on the Kirchoff s laws rather than the contract path method. Not only would it make more realistic calculations on the flows, it would enable TSO s to reduce the security margins as they would now be able to predict the flows more accurately now. Hence balancing would require reduced amount of reserved margins. The 11th Florence Regulatory Forum held in Rome on the 16th 17th September 2004 welcomed the ETSO EuroPEX joint proposal for addressing cross border congestion. Their jointly developed approach called Flow based Market Coupling or FMC provides a model for cross border congestion management and integration of electricity markets across Europe. It consists of regional price areas with inter regional trading facilitated by market coupling subject to simplified transmission constraints. The model included detailed arrangements for day ahead trading as well as providing for alternative forward market structures. Since then the studies have been ongoing over the possibility of introducing flow based market coupling. It has been now proposed to introduce FMC in the north west European region. On June 6, 2007 ministers of the Pentalateral Energy Forum and the High Level representatives of the Regulatory Authorities, TSOs, PXs and the Market Parties Platform of the Central Western European (CWE) region (Belgium, France, Germany, Luxembourg and the Netherlands), signed a Memorandum of Understanding (MoU) on Market Coupling and Security of Supply in Central Western Europe. The MoU is a step forward towards a more efficiently functioning cross border electricity market in the five countries and towards further European integration. Key elements are the development of a flow based market coupling system for the region and several measures for increased security of electricity supply. However it is still a paper and developments in practice have not been visible yet. The target date for this endeavor is January The basic underlying assumption is that a pricing method that couples national markets and maximizes the available interconnector capacity (using flow based calculations) would be most economically efficient. The flow based market coupling will support the global reliability of the electrical system and increase economic efficiency in the region by introducing a single region wide implicit auctioning system of the cross border interconnection capacities, based on a regional load flow calculation. A working group of TSOs and PXs will analyze the requirements for design and implementation of the flowbased market coupling mechanism. The working group is also in charge of the necessary negotiations, steering and decision making. TSOs and/or PXs will also examine the opportunity of setting up a joint TSO company and/or PX company, which would be in charge of the relevant market issues within the CWE area. However, FMC is still not completely understood and should be analyzed for pitfalls, if any, before the implementation can take place. The initial phase would require the involved actors to agree on the terms and conditions of the FMC. The fact that FMC system is an integrated system governed both the technical realities and market economics makes it tough to get an intuitive understanding. The results are also non linear making predictions impossible without help of a appropriate tool. The complexity inherent to the transmission system, compounded by integration of four independent networks and additional complexity introduced by FMC creates a non transparent system that is complicated for the policy makers to understand, and hence is a stumbling block towards agreements and might lead to rough negotiations. For this it is important to bring out the distributive effects that FMC would entail for the participants. This can only be understood by modeling the real system and studying the impact of introducing a flow based system with market coupling.

39 The next chapter would introduce FMC is more detail. The salient features of Market Splitting and LMP would be discussed next. 2.3 MARKET SPLITTING The market splitting is an implicit auction where the capacity is traded simultaneously with the energy. In cases with congestions, the markets are split into two or more price areas. Each price area is then balanced while fully utilizing the transfer capacities between the areas. As stated earlier, larger areas with uniform prices are important for the competition and the market splitting approach has therefore many advantages from this point of view. The economic outcome from market splitting would be same as market coupling only the process of getting there is different. Market splitting is applied in the Nordic market and some of the designs used there will be described and discussed for illustration purposes. The steps are: The whole market area is divided into smaller bidding areas, mainly defined along the structural bottlenecks. The TSOs calculate transfer capacities and all capacity for each hour is allocated to the power exchange for trade in the day ahead market. Information on capacities are given to the market every morning. Market actors can then submit bids in a bidding area. Bids must be in balance in each area. Market clearing is performed for each hour by the latest 2 am the day ahead of operation. First an uncongested market clearing is performed for the whole market. Flows between areas are checked against available capacity. If there are any violations, the market is split in two areas and separate clearing is done in each area. This iteration process continues until all transfers are within capacity limits. All market actors will be compensated and charged according to the market prices in the area they are located. The market actors obligations within each price area are then the basis for a detailed planning of the generation and loads. These detailed schedules are submitted to the TSO and will be the basis for calculating deviations from the schedules and need for secondary control. The market splitting method requires a centralized market operator that combines the bids in a market clearing procedure. As mentioned the market splitting enhances the competition due to relatively many market actors within the same price areas. The implicit auction principle also guarantees that the capacity is made available to the market participants in a nondiscriminatory way since no single market actor can reserve the capacity for own use. In a case of congestion all the market actors can readily see the effect, since the energy price will rise in the deficit area. This gives the right incentive to market actors on both sides of the congestion. The price in an area also reflects the value of the electricity for the market actors. Congestion between the areas will generate a congestion rent to the TSOs. According to EU Directive the congestion rent can only be used for transmission investments, reduced tariffs or counter trade. In the market splitting method the central load dispatch has been replaced by market forces. This gives as result a system where generation and load are in balance in the planning phase. During the operating hour there is a decentralised dispatch where the generators follow their schedules. The System Operators will take care of the imbalances occurring after the planning by using the balancing market. Chapter: Congestion Management 15

40 16 Flow based Market Coupling 2.4 LOCATIONAL MARGINAL PRICES (LMP) More physical network models, for example LMP, make it possible to some extent to operate closer to the limit. The LMP takes into account that inter area transactions may have different impact on the flow on congested transmission lines between those areas as well as on loop flows. This impact on the network flow is the reason for possible enhanced utilization of the transmission system. As long as the prices can be different on all nodes, it is in principle possible to find a power flow that gives the overall optimal solution and utilizes the transmission system as good as possible based on the bid curves. This requires however that all technical and economical details to run an optimal power flow are available when the SO is performing central dispatch. There are different types of transactions where some are firm and may have transfer rights while others are more of opportunistic character. It is important to account for these differences when congestion management schemes are developed. In the follow description and discussion, the scheme developed by PJM and MISO will used as an example(pjm Interconnection). The purpose of the scheme (CM Process) was to balance system security with a market sensitive usage of PJM s available transfer capability. The cornerstone of CM Process is the delineation between energy flows associated with serving firm load; and other energy flows associated with commercial activity. Firm flows are defined as those flows that would occur to serve only customer loads and those transactions that have Firm Transmission Rights (FTR). What is left over would be opportunistic transactions that do not have Firm Transmission Rights. This approach allows those with firm FTRs to be secure that their respective transactions will flow; and those without FTRs can still make use of the system as subject to the requirement that there are no actual transfer problems. As long as the actual tie line flows are within the agreed transfer limit, then there is no reason to preclude additional transactions. In an LMP based system the bus prices will reflect which resources are more financially attractive and which resources are not. As long as the Market price shifts keep the flows within the transfer limits, the opportunistic transactions continue. When the normal market prices are not sufficient to control the flows, then the opportunistic transactions would be curtailed. This approach allows SO to let transactions continue as long as the market players are willing to pay the energy costs caused by redispatching the system. The LMP will diverge until the flows are appropriately reduced. Another benefit is that the transfers are based on real time flows and not on predicted flows. This approach increases both reliability (in the case where conditions adversely impact the transfer limits) and commerce (in the case where conditions support additional interchange.) The use of the simplified Interconnection Model provides recognition that in a meshed interconnection transactions themselves are neither good nor bad, it is the effects of the transactions that help or hurt and those effects are a result of more than just the owner area of the point of delivery and point of receipt of the respective transaction. 2.5 CONCLUSION Congestion management methods can be divided into different sub categories. It can be methods for handling congestions between areas (inter area) or it can handle congestions within an area (intra area). Market based or not market based is another subdivision. Finally it should be divided between methods used to allocate capacity up to the capacity limit (allocation methods) and methods used to alleviate the transmission network down to the capacity limit (alleviation methods).

41 It is not possible to state that one method is superior to the others. It is more a question of how it is necessary to prioritize between the general requirements of congestion management methods given specific systems. Giving priority to some specific features will often make it necessary to come up with countermeasures for the requirements not that well covered. Priority based rules, such as first come first served, and pro rata rationing are allocation based and non market based methods. Explicit and implicit auctions are the favored methods by the European Union. They are market and allocation based methods. Market splitting and market coupling are special cases of implicit auctions. In an implicit auction energy and capacity are traded at the same time. In an explicit auction only transmission capacity is traded. Flow based market coupling tries to use both the technical specification of the system and market coupling together. FMC would be discussed in detail in the next chapter. Chapter: Congestion Management 17

42 18 Flow based Market Coupling 3 FLOW BASED MARKET COUPLING The new congestion management guidelines (amending the existing annex to the European Regulation (EC) 1228/2003) entered into force on 1 December These new binding guidelines 3 prescribe that TSOs should implement a flow based allocation system/or a similar mechanism that couples the markets and takes care of the physical flows as soon as possible.(congestion Management Guidelines 2006) At the moment, TSOs calculate the for the market available capacity for each border separately based on quite conservative base cases. After this, the calculated available capacity is attributed to the market through a certain allocation method such as explicit auctions (capacity only) or implicit auctions (capacity and energy actions). New regulatory guidelines of the EU prescribe the implementation of flow based. In short, (FMC) is a new method for Cross Border Congestion Management. It combines commercial energy bids with physical reality to optimize network use with respect to market value. Commercial energy bids and available capacity are evaluated simultaneously in an iterative process which should lead to a more efficient use of transmission capacity with respect to commercial value. Optimization is performed based upon commercial bids and the linear relation between accepted bids and physical flows on flowgates (defined in the PTDF matrix). Theoretical FMC has been proven to be an efficient market design: (Chao, Peck et al. 2000) System of flow based transmission rights enables market based congestion management or efficient energy and transmission markets. Further, once a system of tradable flowgate rights is established, the control of the transmission system is shifted from line owners to the market, in which the transmission charges are determined competitively without excessive complexity or monopoly power abuses. Flow based market coupling is an implicit auction similar to market splitting but performed in opposite order. First each sub market is cleared, and then these markets are coupled. It is a mixture of a flow based modelling and a Decentralised Market Coupling. A flow based modelling considers the physical flows that can be exchanged between different electric systems taking into account the mutual influence of the exchanges. A Decentralised Market Coupling is a method to execute a coordinated market among different markets, using their own market rules in each area. This method is still under development but is likely to become an important solution in Europe. A simplification must be made, considering that a joint system can be operated as a number of single price regions, connected with the other regions by notional transmission circuits. The real flow between different nodes is modelled by flow factors and the limits between nodes are calculated taking into account the influence of the bottleneck capacities on the crossborder inter connectors. Bottleneck capacities (F max and F ref ) and flow factors (PTDFs) need to be estimated and published in advance to inform users and updated before operation of the day ahead market by the TSOs. This information is required by the day ahead markets to describe the state of the simplified transmission model used for flowbased market coupling. Users have the possibility to submit bids and offers to the regional day ahead markets in order to purchase or sell energy in their region. In the same way they also have the option to submit price difference bids in which 3 See guidelines Article 3.2 prescribes that regional coordination shall include, in particular: (a) Use of a common transmission model dealing efficiently with interdependent physical loop-flows and having regard to discrepancies between physical and commercial flows, (b) Allocation and nomination of capacity to deal efficiently with interdependent physical loop-flows

43 they offer to transfer energy between two markets and pay or receive the inter regional congestion rent. This method allows users to participate for crossborder trade through bilateral contract or day ahead access in a non discriminatory manner. Every market will execute the matching process, firstly without exchanges, and will calculate an import/export curve for every hour, that represents the market price with different quantities of imported or exported energy. Every market executes the process with their own rules. A central module with all import/export curves, considering local prices, and price difference bids, optimizes the flows between the regions, subject to the inter regional transmission constraints represented by the simplified transmission model. Every market with the result (imported/exported energy) will execute its matching process considering imported/exported energy and calculates a new import/export curve, sending the new curves to the central module. Iteration is required between the different areas because of block bids and offers. This sequence is repeated until it converges. The described method can coexist with several forms of forward market arrangements. This will be needed for those users that want to hedge cross border price risk (instead of facing the risk in the day ahead market). It should be done independently from the FMC through forward transmission rights, explicit auctions or electricity financial markets. Therefore it is not needed to have a single market in a whole area, as it is needed in the market splitting method. The corridors will always be full of energy from the cheapest market to the most expensive one. The only problem that must be resolved is that a production shift in an area is a meshed network to some extent influence the flow on all connections. Therefore, it will be necessary to establish a base case fairly close to the final schedules for the calculated maximum transfer capacities between market areas to be valid. 3.1 OPERATION OF MARKET UNDER FMC FMC couples the physical flow path modeling with the market coupling as can be seen in Figure 7. Both these mechanisms will be discussed in next. (ETSO and EuroPEX JWG 2004) FIGURE 7. OUTLINE OF FMC A COMBINATION OF FLOW BASED MODELLING THAT TAKES CARE OF THE PHYSICAL FLOW PATH AND MARKET COUPLING TO MAXIMIZE THE COMPETITION AND EFFICIENCY OF ELECTRICITY MARKETS (ETSO AND EUROPEX JWG 2004) Chapter: Flow based market coupling 19

44 20 Flow based Market Coupling 3.2 FLOW BASED MODELING Flow based refers to fact that one takes into account the physical effect of a certain contract on various flow gates as electricity follows Kirchhoff s laws. For example, a commercial transaction from A to B could lead to a physical flow from A to B, B to C and C to A. Bottleneck capacities are the operational limits on variations in physical bottleneck flows. The relation between a commercial exchange and the physical flows through flowgates is expressed in so called power transfer distribution factors (PTDFs). FMC relies on PTDFs to evaluate the impact of each transaction on the transmission overload. However, PTDFs provide only approximate and sometimes even inaccurate results, and are not suitable for the next day line loading forecast. There is need to be concerned with the accuracy of the PTDF approach. There are deficiencies with current PTDF which need to be addressed before it can be implemented for cross country electricity trade (Yan 1999). Also because of asymmetric information with the TSO (they calculate the PTDFs and also earn the congestion rent) there is a possibility of perverse incentives to keep the congestion on the borders. 3.3 MARKET COUPLING Market Coupling refers to a capacity allocation method using implicit auctions in which capacity is allocated implicitly (see Figure 8). Market Coupling may also be used without using a Flow Based capacity calculation method as is in fact the case today in the North West European region. On the other hand, a Flow Based method could also be used in combination with explicit auctions. However, the introduction of a flow based mechanism is often explored in combination with implicit auctioning (e.g. market coupling). Market coupling has obvious benefit of improving net welfare of the whole system. However it is not clear how the customers of the exporting country would react to increased prices. PX area A PX area B Price Import Price Export Independent area price A Independent area price B Volume Volume Area Prices Market Coupling B to A no congestion congestion B 0 A to B Interconnector capacity (used) A Im/export price dependency curve A Independent area Price A Integral system price Independent area Price B Im/export price dependency curve B FIGURE 8. MARKET COUPLING BETWEEN TWO MARKETS LEADING TO INCREASE IN THE NET WELFARE (DE JONG, H. M. AND R. A. HAKVOORT (2007)) 3.4 MARKET PROCESS OF FMC As it already follows from the discussion thus far FMC is a complex mechanism and is fairly technical and not transparent to the policy makers. It is hence desired to create a simplified simulation model of the market to aid in understanding the behavior of the market under various conditions. Figure 9 shows the simplified model of the market operation under FMC. This process would be modeled for a simplified network and limited number of actors to understand the distributive effects and sensitivity of the prices to the PTDFs.

45 FIGURE 9. THE PROCESS OF PRICING UNDER FMC, THIS WILL BE MODELLED FOR A SIMPLIFIED MARKET CASE TO CREATE A SOFTWARE MODEL THAT HELPS UNDERSTAND THE MARKET OPERATION BETTER (ETSO AND EUROPEX JWG 2004) 3.5 DIFFERENCE BETWEEN FMC AND MARKET COUPLING It is important to understand the fundamental differences between MC and FMC, as both of them on the surface seems to be very similar both are implicit market coupling based, and both use a measure of capacity on the interconnector. To be able to appreciate the difference a few variables would need to be defined: Total Transfer Capacity (TTC): is the maximum exchange programme between two areas compatible with operational security standards (stated in the grid codes) applicable at each system if future network conditions, generation and load patterns were perfectly known in advance. Net Transfer Capacity (NTC): NTC is the maximum exchange programme between two areas compatible with security standards applicable in both areas and taking into account the technical uncertainties on future network conditions. o NTC = TTC TRM Available Transfer Capacity (ATC): is the transfer capacity remaining available between two interconnected areas for further commercial activity over and above already committed utilization of the transmission networks. ATC is given by the following equation: o ATC = NTC NTF Transmission Reliability Margin (TRM): is a security margin that copes with uncertainties on the computed TTC values arising from: o Unintended deviations of physical flows during operation due to the physical functioning of load frequency regulation o Emergency exchanges between TSOs to cope with unexpected unbalanced situations in real time o Inaccuracies, e. g. in data collection and measurements Notified Transmission Flow (NTF): can be interpreted as the already occupied part of NTC by the already accepted transfer contracts at the studied time frame. Chapter: Flow based market coupling 21

46 22 Flow based Market Coupling FIGURE 10 TRANSFER CAPACITY DEFINITIONS The above set of capacity parameters are in terms of bilateral exchange programmes between two neighboring areas. In case of MC, the values of NTC and ATCs are calculated for each pair of countries that are being coupled, without any consideration of parallel flows. The ATC values are used in the implicit auctioning algorithm as a constraint to the maximum energy exchange that can take place between these two countries. In the highly meshed interconnected transmission networks in Europe programmed exchanges and physical flows differ often considerably. They would be closely connected to the power flows through the cross borders only in the ideal case of a peninsular system and its neighbor if both were interconnected through a single tie line, in case of market coupling in between Netherlands, Belgium and France there are no parallel paths hence it is a good approximation to peninsular system with the neighbors lying in a straight line. However, in a widely interconnected network like for example the UCTE network the power flow through the cross border tie lines between two neighbor areas A and B may be interpreted as a superposition of a direct flow, which is related to exchanges between A and B and a parallel flow, which is related to all the other exchanges in the meshed network and to the location of generations and loads in the several grids. Therefore there would be a parallel flow even if all the exchanges in the interconnected system were set at zero. With introduction of Germany into the market coupling there would be large implications due to transit and loop flows on the values of ATC and NTCs. Hence it would not be advisable to use NTC based MC for integration of NL, DE, BE and FR markets. The figures provided about capacities for highly meshed systems are limited in scope, in several senses: TTC and NTC values are computed between neighbour areas; these values are the result of assuming that only the transactions between these two areas are modified and the rest ( third parties transactions) remain unaltered. This fact has two consequences: The published values cannot be used for an exact planning of transactions if these do not correspond to generation and to consumption in the pair of control areas for which capacities are defined, i.e. NTCs cannot be combined to derive possibilities of executing transactions according to a given transaction path (contract path). If the pattern of third party transactions differs noticeably from that taken into account in the forecast, TTC values may significantly differ. That may have an important impact upon the NTC value. NTC values between pairs of control areas in meshed network systems are interdependent. For planning and for the sake of simplicity normally only one set of NTC values, that do not reflect NTC

47 interdependencies between several borders, is published. In case of strong NTC interdependencies, better information can be provided by additionally computing values of transfer capacities for groups of areas. i.e., if there is a strong physical coupling between areas A and B regarding exchanges to area C, NTC would be provided from area A to area C, from area B to area C and from areas A + B, considered as a whole, to C. However, during the allocation phases the coupling between the areas has to be respected. Allocation thus may lead to new restrictions as shown in the following Figure 11. FIGURE 11 INTERDEPENDENCIES OF NTC BETWEEN TWO ARAES ( NET.ORG) In the figure it is assumed that in the planning phase the NTC value between areas A and C was assessed to 2000 MW and that independently from this the calculation of NTC between areas B and C lead to a value of 4000 MW. For planning purposes the TSOs thus have given to market participants maximum values, not reflecting the interdependencies between the areas. Indeed the sum of import to area C may be limited to only 5000 MW. Then, at least during the allocations this fact has to be taken into consideration. It is out of the scope of this document to define the criteria for the split of this total value into the capacity for allocation from A to C and from B to C. Finally the NTC values itself do not provide the basis for a co ordinated method of allocating cross border trade over several borders in a meshed network. A vision for a coordinated approach was already presented in ETSO document 4. It relies on the same computation principles as NTC, but the allocation of transfer capacities would be effected on the basis of the consequences in terms of load flows and not directly using the bilateral values of NTC. Therefore, the importance that NTC values actually have in the transaction based concepts of the international trade in Continental Europe will diminish. From the discussion above it can be concluded that NTC values have significant shortcomings while dealing with parallel flows, hence it would not be easy to implement this system for coupling Germany into the current system of MC existing between NL, BE and FR. Even if it can be implemented major changes would need to be introduced into the algorithm to be able to care of the new additional complexity. TSO s would also request higher TRMs to be able to balance the unexpected flows. Hence MC is not a suitable system for coupling North West European region. FMC is able to address these issues by considering the whole network at once and calculating the effects of each transaction on each tie line. The details of FMC and power transfer distribution functions (PTDF s) are presented in section By introducing PTDF based allocation, the transmission capacity available for each border and exchange direction is only specified within the allocation process (and thus not ex ante before the auction as is the case with NTC). This is implemented taking the capacity bids and the physical interrelationships between the exchange directions and the load flows on crucial transmission lines into account. In practical terms, this facilitates short term capacity transfers between borders compared to the NTC model. 3.6 CONCLUSION In short, FMC is a new method for Cross Border Congestion Management. It combines commercial energy bids with physical reality to optimize network use with respect to market value. FMC is similar to by yet 4 Co-ordinated Auctioning of Transmission Capacity in Meshed Networks, Discussion paper, ETSO, November 2000 Chapter: Flow based market coupling 23

48 24 Flow based Market Coupling different from the market splitting approach implemented in the Nordic electricity system, which it tries to imitate (Glachant and Pignon 2002). Market splitting approach has one common pool where generation entities submit offers and load entities submit load bids. Then the market is cleared based on merit order while respecting the system constraints and reliability margins, and the subsequent unit dispatch and scheduling is managed by the independent national system operators. The market is split, or has different prices if there exists any congestion preventing complete convergence of prices across all area. However in case of FMC the bids are submitted to the individual Power Exchanges who calculate the import export curves for each country and submit it to a supra national entity (or an algorithm) that calculates the prices for each market while respecting the capacity available at the tie lines between the countries. Thus it tries to couple the markets; though the results of the algorithm may be same to that of market splitting the approach is different. The other major difference lies in the fact that Market Splitting and Market Coupling is still based on NTC method for capacity calculation and not on flow based method. FMC is also different from nodal pricing which is implemented in PJM, the largest single energy market in the world. Nodal pricing calculates price on each node based on the price and congestion, it is also called Locational Marginal Pricing, and each node has different price. FMC does not give one price for each node, but gives zonal prices, i.e. it considers set of nodes as a zone and assumes them as being copper plates while the zones are connected by means of tie lines. The tie line limits the amount of electricity that can be traded across the regions and so the prices in the two regions may or may not converge depending on congestion of the tie lines. FMC is complex market mechanism to understand and hence its implications on the national welfare are not fully understood. The thesis would aim to answer some of these questions by developing a simulation tool to help understand the system and its sensitivities towards aspirations and actions of involved actors.

49 4 PROBLEM FORMULATION For public authorities (who must approve of the congestion management method to be used) the introduction of a more efficient method such as FMC seems to be an attractive option. As the implementation of the system is now also prescribed in European legislation, the introduction of FMC is currently high on the political agenda (especially in the Central West Region Netherlands, France, Germany, Luxemburg and Belgium). The chosen approach for congestion management (FMC) will impact the possibility and incentive to utilize the transmission network. There are a few but important requirements to congestions management methods as: fair and non discriminatory, economically efficient, transparent, feasible and compatible with different types of trades. The currently used method will to an extent fulfill these requirements, but it is important to realize that systems have different structure and by this technical challenges. From the feasibility point of view, the methods will have to be analyzed. The market design will also be important for which kind of method that best suits a particular system. The functioning of FMC method is quite complex, the sensitivity of the system are unknown and the effects (on national level) are difficult to predict. Various questions remain unanswered at this moment. The main research question can be formulated as follows 4.1 PROBLEM STATEMENT What are the impediments to and implications of implementing Flow based Market Coupling as a Congestion Management Mechanism in the North West European Countries? To answer the above question one needs to understand FMC, and its operation. Once the operation is understood next step is to understand the distributive effects of the welfare gain. To this cause a model needs to be developed for FMC which can enable better understanding of what if scenarios. Thus the research is hinged on development of an decision support model which would enable the policy maker, in case of present research the Ministry of Economic Affairs of The Netherlands, to understand the implications of FMC to Dutch electricity supply. The main research objective can thus be formulated as follows: To develop a decision support model for evaluating the implications of and impediments to implementing Flow based Market Coupling for alleviating the issue of interconnector congestion among European countries. 4.2 SCOPE DEFINITION However the scope of the question is broad and needs to be structured to make it easily tractable. The boundaries of the problem are defined both in geographical sense and temporal sense. The scope covered in the research would be: GEOGRAPHICAL SCOPE The Netherland and immediate continental neighbors (Northwest European region namely Belgium, Germany and France). It should also be stated that Luxembourg is also officially a part of the region and is also participating the regional initiative and pentalateral forum. Luxembourg has two electricity transmission networks that are not interconnected, but are integrated with the networks of the neighboring countries, Germany and Belgium. Hence Luxembourg is not considered as a special prices area and because the network Chapter: Problem Formulation 25

50 26 Flow based Market Coupling is not connected within Luxembourg it does not lead to any parallel flows and can hence be counted out of the market model. FIGURE 12 THE SYSTEM UNDER CONSIDERATION TEMPORAL SCOPE The study is going to study the distributive effects of the benefits earned through the FMC, so it needs to consider both the short term and long term effects. Hence the study would be conducted at two levels. Firstly, from the operational efficiency point of view focusing on day ahead or intraday market, on which much of the literature is focused and secondly, from a long term capital investment perspective. Given the investment cycle is usually years; the second analysis should take into consideration this time scale. The former part of the study would be answered primarily by the outcomes of the model simulations. Once the operational results have been determined it is possible to predict the longer term implications by a mix of expert opinion and the literature survey. 4.3 PERSPECTIVE/PROBLEM OWNER The main problem owner for the given research question is the Dutch Ministry of Economic Affairs. The problem will be analyzed through their lenses. It is also important to state here hat partly the regulator DTe and TSO Tennet also could have considerable interest in outcomes study. By taking the Ministry of Economic Affairs as the problem owner, additional constraints exist for the end result of this thesis: The results should be such that people at the Ministry who deal with market integration but who do not want to bother with detailed technical or economic calculations may gain from the results of the present study. 4.4 SUB QUESTIONS Further the problem statement can be sub divided and grouped under the following classifications to make the research better structured.

51 4.4.1 SYSTEM BEHAVIOR a) How sensitive is the system for the choice of a specific set of PTDFs? b) What is the effect of constraints on the system? Could it lead to a less optimal outcome with respect to optimizing market value? E.g. policy makers may want to reserve a minimum amount of capacity on certain flow gates with respect to the investments they made regarding those flowgates in the past or with respect to security of supply c) What will be the net benefit of future network investments on available interconnector capacity and the Dutch electricity price level? o o o What would be the effect on the price level within Netherlands? Would it lead to more transits and loop flows? How would the actors perceive this change? SECURITY OF SUPPLY What is the risk that FMC would lead to a net reduction of the import capacity? How may FMC interfere with security of supply in the Dutch electricity system? (As less or no capacity at all could be distributed to a certain border if market value is low)? WELFARE EFFECTS What could be the effect of such system on national welfare as the system optimizes regional welfare in terms of market value? What is the effect of the presence of four national systems on the distribution of the welfare for each of the four systems? How many incentives will FMC give to increase electricity production from renewable energy sources? IMPLEMENTATION ISSUES What is the effect on transparency? Would some kind of European central decision making unit be required? (e.g. to decide about the distribution of capacity on the national borders). What about gaming opportunities because of strategic actors? The questions on system and welfare would be answered primarily by the simulation model. The questions with respect to security and implementation would be answered using the expert opinion/interviews with the problem owner/literature survey. 4.5 RESEARCH DESIGN To be able to answer the above mentioned questions it was necessary to first understand FMC from the theoretical perspective. It would be possible to gauge the effects of FMC on the national and regional level if a simulation model with FMC can be set up with the present level of generation, loads and the transmission line typography. The first step in the research would hence be to study FMC and its working within the North West European region context. The study would aim to study the impact that FMC would have on the various actors Chapter: Problem Formulation 27

52 28 Flow based Market Coupling within the four countries defined in the geographical scope of research. The part of research would be based on the Neo classical economics treatment of the market design. The details of the simulation model would be discussed in the subsequent chapter. FIGURE 13 RESEARCH DESIGN The second part that needs to be explored is the institutional aspects of the multi actor environment into which this system would be introduced. The process of implementation of such a market design is further complicated because of presence of large number of actors and stakeholders. This impact of this market design on various actors and interests of the various actors would be studies in this section. The last part of the report would detail in brief the issues associated with implementation of a market design like transparency, legal aspects, etc.

53 Section 2: Simulation Model A step by step description of how the model was created is presented in this section. Details and relevant data is presented in the appendices. Chapter: Problem Formulation 29

54 30 Flow based Market Coupling

55 60.00 Prices (euros) NL BE FR D E P* NL Supply Curve Generation Capacity (GW) y = 3.183x R ² = NL Supply Curve 5 STRUCTURE OF SIMULATION MODEL Flow based allocation is a supra national approach: all bids for energy and/or cross border capacity are optimized by collected by the national PXs, and are then combined together by using an algorithm that takes care of final allocation and dispatch. In the flow based allocation mechanism, the commercial transactions are no longer limited to the interconnections where they are reported, but they are converted into physical power flows by using a simplified representation of the network so that their impacts on third interconnections can be considered thus ensuring overall security. Pri ce (euros/mwh) FIGURE 14 DESIGN OF THE SIMULATION ENVIRONMENT The simulation model for the integrated markets in the four countries of interest was created. None of the commercially available software was able to address all the aspects of modelling. Hence the work was divided into three main parts for both data gathering and modelling: 1. Transmission network module for the four countries (BE, NL, FR and DE). This included simulating the transmission network in the north west European region using specialized software used for solving large transmission network problems (PowerWorld). 2. Electricity demand and supply module: Includes a sub model for calculating the marginal cost of generation. Marginal cost of generation data along with installed capacity data is used to create a representative merit order for each country. From the merit order a linear supply function is extracted. Demand function is also defined. 3. Electricity day ahead market module: This is essentially an optimization module that simulates the implicit flow based market coupling for calculating the electricity dispatch on day ahead basis. The process is outlined in Figure 14. The next sections would elaborate the process of creating each one of the three modules outlined above. Chapter: Structure of Simulation Model 31

56 32 Flow based Market Coupling 5.1 TRANSMISSION NETWORK MODULE The research required an appropriate benchmark system to base the simulation on. Creation of a realistic benchmark system is difficult as utilities are often not willing to disclose details of their own systems because of commercial sensitivity and security reasons. Most TSOs, generators and industrial loads normally refused to share any information whatsoever citing commercial sensitivity or security reasons. The former excuse is debatable as transparency is one of the main pre requisites for an efficient market. The subject becomes more complex because of dealing with four countries and hence most of the required information was inaccessible. Also there was no specialized software available at TU Delft that could be used to simulate such a complex network. After spending considerable time on trying to resolve these operational and informational issues it was considered to look at past research on UCTE region conducted elsewhere. University of Edinburgh has developed a benchmark transmission network for the UCTE region. It was possible to get this model and use it for the current research(qiong and Bialek 2005). With access to the UCTE transmission network model and purchase of PowerWorld the work could proceed. The entire UCTE transmission network model was used as decoupling the four countries of interest from the rest would lead to imbalance in the system and hence the solution could not converge. FIGURE 15 INTERCONNECTED NETWORK OF UCTE ( The approximate model of a European interconnected system was used to study the effects of cross border trades. The system was considered as a realistic representation of the reality for the purpose of the simulation because the results from the model were tested with the real PTDFs values by the researchers and significant correlation was found between the two. More details are attached in the chapter on verification and validation chapter 66 Validation and Verification of model, pg 56 Comparison of simulation results conducted on the test system with the published cross border flows and power transfer distribution factors showed a very good correlation, exceeding 90%..(Qiong and Bialek 2005) Using the model as a research tool which would replicate main physical characteristics of the real network without pretending to be an accurate model, the PTDF s were derived for the four countries.

57 5.1.1 POWER TRANSFER DISTRIBUTION FACTOR Transmission capacities are nowadays allocated at all European borders on the basis of NTCs. The NTC values indicate the maximum commercial power exchange allowed between two countries (more specifically: control zones and blocks). They are fixed prior to allocation (e.g. before bids for an explicit capacity auction are submitted) by the TSOs involved and remain static during the allocation. Electricity transmission flows fan out across all available parallel paths in accordance with the laws of physics. This is illustrated in the next figure, in which all countries are balanced and only a 100 MW commercial exchange from Germany (DE, source) to France (FR, sink) is taking place and the resulting physical crossborder flows are shown. FIGURE 16 PHYSICL FLOWS BECAUSE OF A TRANSACTION OF 100 UNITS OF TRADE BETWEEN FRANCE AND GERMANY ( NET.ORG) PTDF factors describe what physical flow on a given interconnection would be provoked by a requested commercial exchange between two countries or two control areas (or hubs ). These two hubs do not necessarily need to be directly connected. Simply said, the PTDF factors translate a commercial transaction between two hubs into the expected physical flows over the entire network. The PTDF factors that correspond to the situation that is illustrated in the Figure 16, can be seen from the column on the right: the PTDF factor D NL with regard to an exchange D F is 27%, the PTDF D F is 36% etc. Thus it can be easily derived for example, that a 100 MW commercial exchange from Germany to France results in a 27 MW flow from Germany to the Netherlands. In theory the PTDFs are calculated as described below. FIGURE 17 GENERIC REFERENCE CASE FOR CALCULATION OF PTDF The PTDF in theory can be calculated as follows: Let be the flow of the branch j > k in the reference state. Chapter: Structure of Simulation Model 33

58 34 Flow based Market Coupling Let be the flow of the branch j > k after the injection of 1 MW at node I (and compensation at the hub also referred to as the slack node). The expression of the nodal PTDF which means the influence of the injection of 1MW at node i on the branch j > k is: This process is automated while using PowerWorld. Next section would describe the process followed by PowerWorld for calculating PTDF, followed by the section on how to actual perform the process USE OF PTDF Allocation on the basis of Power Transmission Distribution Factors (PTDF) represents a further development of this concept which dynamises the specification of commercially available capacities for each border. This does not occur prior to the allocation, but is part of it, based on an economical optimization, taking parallel flows into account, coordinated for several borders and taking account of the differences in price and costs of power generation between the regions. To this end, so called PTDF coefficients are used that describe how commercial power exchange between regions is distributed on physical power flows at the interfaces (interconnectors) between these and between other regions. Compared to traditional NTC based allocation, this basically facilitates a transfer of commercial transmission capacities between the borders that is not at a ratio of 1:1, but at a ratio of the respective PTDF values Based on the values of PTDFs, transactions are curtailed and increased in order to relieve lines overloads. Due to the phenomenon of loop flows in a meshed network, PTDF calculation requires the full network information CALCULATION OF PTDF IN POWERWORLD The Power Transfer Distribution Factor (PTDF) display is used to calculate the incremental distribution factors associated with power transfers between two different areas or zones. These values provide a linearized approximation of how the flow on the transmission lines and interfaces change in response to transaction between the Seller and the Buyer. These values can then be visualized on the onelines using animated flows (see below for details). The transaction for which the PTDFs are calculated is modeled by scaling the output of all generators on Automatic Generation Control in the source and sink areas in proportion to their relative participation factors,i.e, generators in the source area increase their output, while generators in the sink area decrease their output. FIGURE 18 SCHEMATIC OF THE TRANSMISSION NETWORK MODULE

59 An important aspect to consider in calculating the PTDF is how the losses associated with the transfer are allocated. Simulator assumes that the Seller increases the output of its generators by 100% of the transfer amount, while the Buyer decreases the output of its generators by 100% minus any change in system losses. In other words, the Buyer accounts for the entire change in the system losses. Of course it is possible that a transfer may result in decreased system losses; for that case, the Buyer s generation will be greater than 100% of the transfer HOW TO GET PTDF FROM POWERWORLD Open the *.pwb file using PowerWorld open new file menu. Once the model has been opened perform the following steps to get the PTDF for each set of the country. Repeat the steps with other combinations of countries till all the required data has been collected. Perform an initial Power Flow Solution. In Run Mode, select Tools > Power Transfer Distribution Factors (PTDFs) from the main menu to open the Power Transfer Distribution Factors Dialog. Supply the requested information on the Power Transfer Distribution Factors Dialog and click the Calculate PTDFs button. The distribution factors are calculated and displayed for the element set of your choice in the table at the bottom of the dialog. Figure 20 displays the UCTE model from PowerWorld, while zooming in on the Dutch Belgian border. The circles show the amount of interconnector capacity that is being utilized. FIGURE 19 THE UTILIZATION OF INTERCONNECTOR BETWEEN THE COUNTRIES IN THE NORT WEST EUROPEAN REGION FROM THE MODLE RUN IN POWERWORLD Chapter: Structure of Simulation Model 35

60 36 Flow based Market Coupling Zoom in on Dutch Belgian Interface showing the usage of interconnector capacity and the direction of flow of electricity for winter offpeak base case FIGURE 20 UCTE NETWORK WITH ENHANCED NL BE BORDER

61 TABLE 1 SENSTIVITY MATRIX AND PTDF BASED ON THE UCTE WINTER PEAK MODEL FROM POWERWORLD Sensitivity Matrix NL NL BE NL FR NL DE NL BE FR BE NL DE FR DE NL PTDF Matrix Commercial Exchange BE NL FR NL DE NL FR BE DE BE DE FR BE FR BE NL DE FR DE NL Physical Flow It can be observed that sum of the numbers do not add up to 1,in the PTDF and Sensitivity matrix the reason is because the north west European region is not completely decoupled from rest of UCTE region and some of the flows take place through neighboring countries. In case of a commercial transaction between France and Netherlands almost 14% of flow takes place from France through Switzerland and then into Germany. This is interesting feature of the FMC as not there could be exchanges between Switzerland and Germany or Switzerland and France that would make use of the interconnector between France and Germany however Switzerland is out of the FMC and hence calculation of the rent or marginal cost for usage of interconnector capacity would create additional problems. These PTDF values were used in the market module to calculate the actual flows occurring on the interconnectors given the commercial exchanges which would maximize the market value, or minimize the net cost of electricity generation. 5.2 ELECTRICITY DEMAND AND SUPPLY MODULE Economic model of the system comprised of getting the supply and demand curves for the four countries. Supply curve was based on the real data for cost of generation and the installed generation in every country. The demand curve was derived from the supply curve by substituting the average demand in the country and then obtaining the slope and the intercept. The next sections will describe in detail the procedure to obtain the demand and supply curves. Chapter: Structure of Simulation Model 37

62 Generation Capacity (GW) y = 3.183x R² = NL Sup ply Curve 38 Flow based Market Coupling NL Supply Curve Price (euros/mwh) FIGURE 21 SCHEMATIC OF THE ELECTRICITY DEMAND AND SUPPLY MODULE SUPPLY CURVE Supply curve for a country needs two basic inputs 1. Cost of electricity generation 2. Installed generation capacity However the supply curve obtained this way, as also in real life, is not linear it is a discrete stepwise function. To simplify the calculations linearity assumption is made. The data for cost of electricity generation and capacity is plotted and then a best fit for the supply curve is calculated. The data for the cost and capacity is collected from different sources as variable cost data is not easily available because it is classified. For the actual values turn over to Appendix A COST OF ELECTRICITY GENERATION The sources of electricity generation can be grouped predominantly into two categories fossil (coal, oil and gas) and others (wind, hydro, nuclear and biomass). To convert energy from its original form into electricity many different technologies can be employed Turbines, Internal Combustion engines, nuclear fission based power plants, combined cycle plants and combined heat and power plants. Cost structures of electricity generated from all these resources and technologies vary widely. Hence a comprehensive data base was collected to get as realistic values of the costs involved as possible. Also there is a debate on use of short term variable cost or the long term total cost for the market simulation. However fixed costs are also earned back based on marginal pricing, i.e. equal to only the variable costs(stoft 2002) because of price volatility and peaks in prices. Based on the cost of generation the technologies were classified into following categories: TABLE 2 CATEGORIES OF GENERATION TECHNOLOGIES Energy Technology Hydro Wind Nuclear Hard Coal Condensation Lignite Condensation CHP Natural Gas Natural Gas Combined Cycle Oil CHP Biomass CHP Oil Symbol H W N HC LC CHP G G O CHP B CHP O

63 CHP oil is not present in all countries, and the installed capacity is very low. Also the cost of generation are usually higher, hence it was removed from the list for further analysis. The generation companies do not reveal any data about the marginal cost of generation or the variable costs. It is considered to be sensitive information and is kept secret. Hence data was collected from varied sources: it was ensured that data be from real plants. The secrecy of the plants was not revealed in the reports by not declaring the plant name, location or ownership information thus also respecting the fears of generation companies. Data from various sources was collected for all three cost categories (more can be looked up in the Appendix A1), i.e. 1. Fixed cost (FC): Fixed cost comprised mainly of the cost of capital. It includes debt and equity costs associated with the capital invested in the plant that do not depend on its level of operation. The upfront investment is distributed over the lifetime operation of the plant for being earned back. Nominal discount rate is applied to get the NPV of the investment. 2. Variable Cost (VC): A cost that varies with the level of output. In special case of electricity this is also the generators marginal cost until the full output is reached. This cost is comprised mainly the fuel cost, though it also includes operation and maintenance costs. 3. Total Cost (TC): This is the total cost of generation of one MW of electricity. It includes fixed costs, variable costs, start up costs and no load costs. It is also called total cost of production. 4. Start up Costs: The costs associated with additional fuel usage and extra labor and process required to start generation in a plant. These are usually high for base load coal based power plants and are much smaller for gas based CCGT power plants. 5. No load Costs: The costs associated with just keeping the plant in stand by mode, including the cooling and fuel costs are termed as no load costs. The costs are related as follows: Total Cost = Variable Cost + Fixed Cost + Start up Costs + No load Costs Also in present scenario, with Carbon trading coming in force, emission costs were also included. Hence information was also collected on the emissions related to each generation technology and the cost of carbon credits. Thus giving rise to an additional category of costs that would be referred to as environmental costs (EC) henceforth in the report. Therefore the total costs can be defined as Total Cost = Variable Cost + Fixed Cost + Start up Costs + No load Costs + Environmental Costs Additionally, given the importance that has been attached to the environmental impact from the generation technologies and the emission trading a cost for the emissions has also been included. Variable costs as displayed in the model are the sum of fuel costs, operation and maintenance costs and the CO 2 emission costs. The No Load Cost is defined as the total theoretical heat or fuel input at zero net output multiplied by the performance factor, multiplied by the (Total Fuel Related Cost (TFRC)), plus the No Load Additional Labour Cost(PJM 2006). The dollars per start as determined from start fuel, total fuel related cost, performance factor, electrical costs, start maintenance adder, and additional labour cost, if required above normal station manning levels (PJM 2006). For the model assumption is made that start up costs and no load costs are negligible compared to the rest and are hence assumed to be absent. i.e. TC = VC + FC The generation cost for each of the technologies was broadly based on six different sources: Chapter: Structure of Simulation Model 39

64 40 Flow based Market Coupling 1. Risto TARJANNE, EU POLICY AND CARBON EMISSION TRADING: IMPLICATIONS FOR THE ENERGY MARKET, Lappeenranta University of Technology 2. Vattenfall Annual Report 2006 (Vattenfall 2006) 3. Projected Costs of Generating Electricity 2005 Update (OECD) 4. Paper about game theoretic model of the Northwestern European electricity market (Lise, Linderhof et al. 2006) 5. Federal Energy Regulatory Commission, FERC Form 1, "Annual Report of Major Electric Utilities, Licensees and Others." 6. Hoogwijk, M., D. van Vuuren, et al. (2007). "Exploring the impact on cost and electricity production of high penetration levels of intermittent electricity in OECD Europe and the USA, results for wind energy." Energy 32(8): The Vattenfall annual report for 2006 gave detailed break up of cost for 7 different generation technologies. More details about the advantages, disadvantages and brake up of the costs can be found in the appendix A.2 TABLE 3. TOTAL COST OF GENERATION FOR DIFFERENT TECHNOLOGIES Energy Technology Symbol Installed Capacity Investment Cost Specific Investment Full Load Cost Utilization Time Capacity Factor Economic Lifetime Real Interest Rate Anuity Payment Capital Cost MW million /kw h/a % a % /MWh Hydro H , % 50 5% ( 82,165.10) Wind W , % 25 5% ( 78,047.70) Nuclear N , % 40 5% ( 179,205,345.59) Hard Coal Condensation HC , % 25 5% ( 37,250,040.08) 9.31 Lignite Condensation LC , % 25 5% ( 12,239,298.88) CHP Natural Gas CHP-G , % 40 5% ( 6,060,928.76) 9.35 Natural Gas Combined Cycle G % 25 5% ( 16,460,970.09) 5.14 Oil O , % 25 5% ( 8,301,437.50) CHP Biomass CHP-B , % 30 5% ( 222,150.65) O&M Costs, Total Variable Total Variable Total Cost Cost (with (with Symbol Fuel Price Fuel Cost of Net Efficiency Rate Electricity for given utilization Heat Credit Cost (w/o (Only CHP) emission) Total Cost (w/o emission) Emission Price Emissions Emission Cost emission pricing) emission pricing) /MWh % /MWh /MWh /MWh /MWh /tonne kg CO2/MWh /MWh /MWhe /MWhe H - 100% W - 100% N % HC % LC % CHP-G G % O % CHP-B TABLE 4 FORMULATION OF COST SHEET Variable Units Significance Installed Capacity MW The electricity generation capacity of the plant whose data was used as being representative of plants operating with same fuel type. Investment Cost million Overnight cost of installing the plant under consideration. Specific Investment Cost /kw Ratio of Investment cost to installed capacity Full Load Utilization Time h/a Number of hours the plant is expected to operate in a particular year Capacity Factor % Number of hours of operation divided by total number of hours in a year Economic Lifetime a The number of years the plant is expected to be in operation. Real Interest Rate % The interest rate for the loan that was used to pay for installation of

65 the power plant Annuity Payment It was assumed that the loan was paid back by means of installments paid annually for the period of lifetime of the plant. Depreciation was not considered in the model. Annuity was calculated using PMT function from excel. PMT(rate,nper,pv,fv,type) = PMT(Real Interest Rate, Economic Lifetime, Investment Cost) Capital Cost /MWh The fixed cost associated with generation of 1MWh. It was calculated by dividing the Annuity Payment by the net electricity generation per year. Net electricity generation per year was calculated by multiplying the installed capacity with full load utilization time. Fuel Price /MWh The cost of fuel used in the plant in Euros per MWh. This is the major part in the variable cost for plants based on fossil and nuclear fuels. Net Efficiency Rate % Net electrical efficiency of the plant Fuel Cost of Electricity /MWh Cost of generating one MWh of electricity because of the fuel. It was calculated by dividing the fuel price by net efficiency rate. O&M Costs, for given utilization /MWh The operating and maintenance costs for the plant. It is the second important contributor to the variable cost of electricity generation. Heat Credit (Only CHP) In case of CHP, heat as well as electricity is generated. Hence it is important to consider the profits that are made by selling the heat which is generated as by product of electricity generation. This was collected from the reference data as well. (mainly the OECD report) Total Variable Cost (w/o emission) /MWh Sum of fuel cost of electricity and O&M costs. In case of CHP the heat credit was subtracted from the sum to get total variable cost. Total Cost (w/o emission) /MWh Sum of the total variable cost and capital cost Emission Price /tonne The price of carbon credits in euros per ton of carbon di oxide emission Emissions kg CO 2 /MWh Emission of carbon di oxide because of generation of 1 MWh of electricity using a particular fuel type Emission Cost /MWh Product of emission and emission price Total Variable Cost (with emission pricing) Total Cost (with emission pricing) /MWhe /MWhe Sum of total variable cost and emission cost Sum of capital cost and total variable cost(with emission pricing) The data for the 3 technologies namely CHP Natural Gas, CHP Oil, Oil was based on assumptions and other sources. Oil plant data was based on report on projected cost of generation 2005 (OECD). Rest of the data was based on Vattenfall s annual report and Risto T., EU POLICY AND CARBON EMISSION TRADING: IMPLICATIONS FOR THE ENERGY MARKET, Lappeenranta University of Technology. For CHP plants assumption was made that they share the similar cost structure as the natural gas combined cycle. Detailed assumptions are outlined in the model itself. If the markets are working efficiently is reasonable to make assumption that the market players would bid according to their marginal costs. Though it may sound contradictory as marginal costs would equal to the variable costs and thus it would not be possible to recover the investment cost. However it must be kept in mind that the markets are volatile and the prices are set at the market clearing price(mcp) or equal to the marginal cost of the most expensive generator that was dispatched, and the generators would make profit because their marginal costs were lower than MCP. Also in times of price peaks the generators make profits that substantially offset the investment cost. Hence in the model variable cost of electricity generation was used for obtaining the merit order. In case of scenario with carbon pricing the total variable cost with emission pricing was used. The fixed cost data is important for use in investment decisions. It might be used in further studies on congestion management or investment decisions under FMC system INSTALLED GENERATION CAPACITY The installed generation capacity in based on UCTE system adequacy report: Forecast (UCTE 2007). The data is given for the fossil and renewable sources. However the fossil fuels are not further divided Chapter: Structure of Simulation Model 41

66 42 Flow based Market Coupling into CHP and non CHP based generation. For obtaining this classification data from this aggregated data results from a previous study (Lise, Linderhof et al. 2006) were used. The data is based on statistics of 2001, hence assumption was made that most of plants installed post 2001 are based on CHP and hence it was assumed that the conventional plants capacity stayed at 2001 level. To get CHP capacity difference between current total capacity and 2001 capacity of conventional plants was taken. Other assumptions that were made have been elaborated in the model itself. TABLE 5 INSTALLED GENERATION CAPACITIES IN THE NORTH WEST EUROPEAN REGION Variable Cost Installed Generation Capacity Symbol VC Avg NL BE FR DE /MWh GW GW GW GW Hydro H Wind W Nuclear N Hard Coal Condensation HC Lignite Condensation LC CHP Natural Gas CHP G Natural Gas Combined Cycle G Oil O CHP Biomass CHP B CALCULATION OF SUPPLY CURVE SLOPE The data of capacity and cost was used to generate the merit order curves for each country. The variable cost of electricity generation was used to sort the electricity sources in increasing order of their variable costs and get the merit order. It was assumed to use a straight line for the supply curve approximation hence a linear regression fit was made to the plotted data points. The equation of the curve and the R 2 values were calculated. Based on the R 2 (statistical significance parameter) values it was decide to exclude CHP biomass from the merit order. It had a very high variable cost and effected the equation of straight line fit substantially. TABLE 6 SUPPLY CURVES SLOPE FOR THE FOUR COUNTRIES. Inverse Supply Function y=mx+c y Price /MWh x Quantity Supplied MWh m c NL BE FR DE

67 The intercept value for each curve was set to zero to have consistent equations, without this assumption some of the demand curves had negative values for supply at zero price. This value was then entered into the output matrix for the supply curves. The supply curves for each country and their equation are displayed in Appendix A DEMAND CURVE To complete the economic model demand curve was required. It is debatable whether the demand curve for electricity is inelastic and hence vertical or it is elastic. For the sake of model the demand curve was assumed to be elastic which is more realistic especially while considering the wholesale customers. Two set of data were required to calculate the demand curve: 1. Average electricity demand per country: This data was collected from UCTE System Adequacy Forecast (UCTE 2007). The value for demand was assumed to be equal to the reference load on the 3rd Wednesday of January, 2007 at 11:00. TABLE 7 AVERAGE DEMAND OF ELECTRICITY BASED ON THE UCTE SYSTEM ADEQUACY FORECAST Average Demand GW NL BE FR DE Reliable Available Capacity: The net generation capacity for each country has already been calculated in the supply curve calculations. However for the demand curve calculation reliable available capacity was needed. This data was collected from UCTE System Adequacy Forecast (UCTE 2007). The value for demand was assumed to be equal to the reference reliable capacity available on the 3rd Wednesday of January, 2007 at 11:00. The capacity is given below in the Table 8. The methodology for calculating the reliable available capacity is as follows Reliably available capacity = National generating capacity non usable capacity maintenance and overhauls (nuclear and fossil fuel power stations) outages (nuclear and fossil fuel stations) system services reserve TABLE 8 RELIABLE GENERATION CAPACITY BASED ON THE UCTE SYSTEM ADEQUACY FORECAST Reliably Available Capacity GW NL 17.1 BE 13.5 FR 91.6 DE 83.9 To calculate the slope of demand curve the following procedure was followed. 1. Calculation of P * : The demand curve and supply curve at equilibrium would intersect to give the equilibrium price and quantity. Assuming that the average demand (AD) would be the equilibrium demand (Q * ) and hence the result of intersection of demand and supply. The AD value was Chapter: Structure of Simulation Model 43

68 44 Flow based Market Coupling substituted into the supply function calculated in This gives the equilibrium price (P * ). This is one point on the demand curve. If another point on the demand curve can be calculated it would be possible to get the slope and intercept of the demand function. As a validation the value of P * were also compared to the actual energy spot prices in the four countries. 2. The net installed capacity was assumed to be equal to the demand when price of electricity is equal to zero. This is perhaps not completely true however the demand curve becomes vertical at a point, this was assumed to be the point [refer Figure 25 Principle of MCP (from This gives the second point on the demand curve, which is assumed to be linear in the simulation model 3. With two points known on a linear curve the slope and intercept was calculated. The equation of the line was given in the form : y = mx + c Demand data from UCTE System Adequacy Forecast Reference load on the 3rd Wednesday of January, 2007 at 11:00 FIGURE 22 THE EQUATION OF THE DEMAND FUNCTION FOR THE NORTH WEST EUROPEAN REGION 5.3 TIE LINE CAPACITY The constraints for the market model are based on the capacity of the interconnector. The maximum amount of electricity that can be exported or imported by a country depends on the capacity of the interconnector, this is called the Flow gate capacity (F max ). A part of the maximum allowed flow on the critical branches is already used (i.e. prior to the allocation) by so called already occupied flows that can consist of the following components: natural flows (cross border physical flows that will always occur, even when there is no scheduled commercial exchanges between areas); flows that result from sources/sinks that are located in a single control area commercial exchanges resulting from firm nominations in previous auction rounds flows caused by exchanges between sources and sinks that are not located in the region that is participating in the flow based allocation flows caused by exchanges between areas where the source is located in the region that is participating in the flow based allocation whereas the sink is outside, and vice versa. These values would be summed up to calculate the base case F ref or the capacity that is already used up. It can be assumed to be similar to NTF (notified transmission flow) from the NTC based methods, the main difference being that it considers flow based calculation. The capacity that is allowed to for the implicit market coupling is then equal to the difference of F max and F ref.

69 These values are not available for the countries right now. Hence couple of assumptions were made. First the F max was assumed to be equal to the current NTC value at the borders. Secondly it was assumed that all the flows are allocated on day ahead. Hence F ref was assumed to be zero. It should also be remarked here that NTC values are not the correct measure of how much physical capacity is available. It is also likely to go up with implementation of FMC, as reserve margin could be reduced because of better prediction of flows. However it is tough to define actual values that capacities to get a picture of what the maximum thermal capacities that exist on the borders more data was collected on the actual transmission lines that currently connect the national markets. The data on actual thermal capacities is attached in the appendix for reference A.6. A transmission network consists of a number of transmission elements (e.g. lines and transformers), each one having a definite maximum transport capacity that is mainly determined by thermal limits. NTC values reflect real electricity transport possibilities in the meshed and internationally interconnected transmission systems. Two factors limit transfer capacities at values normally much below the thermal capacities: The network element within a transmission system that is the most sensitive to the load flow, When temporary admissible overloads are taken into account, the security criteria related to voltage or frequency stability lead to constraints that may limit physical load flows at significantly lower values than the thermal capacities of lines. FIGURE 23 TIE LINE CAPACITY FOR THE NORTH WEST EUROPEAN ELECTRICITY NETWORK (BASED ON DATA FROM WEBSITE OF TENNET, RTE, RWE, ELIA AND TSO AUCTION WEBSITE) For the model it was assume that the F max values are equal to the current NTC values on the borders and were entered into the interconnector capacity tab. The Figure 23 shows the values that were used in the model. Also not that the capacity values in two directions are different, this is also because NTC values in both directions depend on the initial load flow conditions set by the base generation and load patterns due to the already existing transactions. If a sensitive network element regarding NTC limits has an initial load in one direction, additional cross border transactions that increase the load flow over this element in the same direction result in tighter NTC limits than cross border transactions in the opposite direction. Chapter: Structure of Simulation Model 45

70 NL BE FR DE P* 46 Flow based Market Coupling 5.4 ELECTRICITY DAY AHEAD MARKET MODULE Prices (euros) FIGURE 24 SCHEMATIC OF THE MARKET MODUEL In the market module implicit flow based day ahead wholesale market four electricity is simulated. The previous sections have already described the data that would be used as input to this module. Essentially this module is a optimization module. Here the electricity dispatch is calculated which respects the constraints while maximizing the market value, or while minimizing the net cost of electricity generation for the whole region. The description of the market model is divided into two parts, first the description of model which calculated prices without any imports and exports and then with flow based market coupling.

71 5.4.1 EQUILIBRIUM WITHOUT MARKET COUPLING FIGURE 25 PRINCIPLE OF MCP (FROM Bid blocks are arranged in bid price ( /MWh) order to form the merit order for the system. The energy market is then dispatched by traversing the energy stack, climbing up/down the stacked bid blocks until load demand is met. The market is not coupled hence both the supply and demand curves are local. It must be mentioned here that in real life the demand curves move up and down with time, but the dynamics of the system are not being studied by the current research. Here an average demand function was set up and assumed to be same over the period of time. The supply function that was calculated in the sections above and the demand function, also calculated above, would intersect at equilibrium and the values would give equilibrium price or MCP, P * and the market clearing quantity, Q *. Next step would be to calculate the coupling and optimization WELFARE AND NET COST OF GENERATION CALCULATIONS FIGURE 26 THE DEFINITION OF CONSUMER SURPLUS AND PRODUCER SURPLUS Consumer surplus is the difference between the price consumers are willing to pay (or reservation price) and the actual price. If someone is willing to pay more than the actual price, their benefit in a transaction is how much they saved when they didn't pay that price. It is calculated as the area of triangle ABC from Figure 26 The definition of consumer surplus and producer surplus The producer surplus is the amount that producers benefit by selling at a market price that is higher than they would be willing to sell for. Note that producer surplus flows through to the owners of the factors of Chapter: Structure of Simulation Model 47

72 48 Flow based Market Coupling production, unlike economic profit which is zero under perfect competition. It is calculated as the area of triangle EBC from the Figure 26. Combined, the consumer surplus, the producer surplus, and the government surplus (if present) make up the social surplus or the total surplus or net surplus. The net welfare is calculated as the sum of areas of triangle ABC and EBC. Total cost of generation is calculated from the Figure 26 as the area of triangle CBE. This is done for each country. Each of the above four values are calculated independently for the four countries in north west European region. Finally to get the regional values these are summed up together. The formulations are attached in the appendix. AFTER MARKET COUPLING The optimization module simulates merit order based economic dispatch. Bid blocks are arranged in bidprice ( /MWh) order to form the merit order for the system. The energy market is then dispatched by traversing the energy stack, climbing up/down the stacked bid blocks until load demand is met as shown in Figure 27. In this case the markets are coupled hence the value of demand and supply functions would have to optimize over the whole region under the given constraints. FIGURE 27 SIMPLE STYLIZED ELECTRICITY MARKET The central dispatch process should aim to maximize the value of spot market trading i.e. to maximize the value of dispatched load based on dispatch bids less the cost of dispatched generation based on dispatch offers, subject to: dispatch offer and dispatch bid quantity constraints; constraints due to availability and commitment; non scheduled load requirements in each zone; constraints due to power system reserve requirements determined; intra zonal network constraints and intra zonal losses; inter zonal network constraints and inter zonal losses; constraints consistent with registered bid & offer data; current levels of dispatched generation and load; and procedures to take account of constraints imposed by ancillary services requirements. The above statements establish the framework within which more detailed rules on specific topics may be defined. In the current module, given the non availability of data for detailed constraints, only the following constraints were modeled:

73 dispatch offer and dispatch bid quantity constraints; intra zonal network constraint; EFFECT OF IMPORT AND EXPORT ON PRICES AND QUANTITY DEMANDED FIGURE 28 EFFECT OF IMPORT AND EXPORT ON PRICES AND QUANTITY IN THE COUNTRY It is important to calculate the effect of import and export on the prices and quantities demanded and supplied in the country. The calculations were made as follows, the diagram of the effect is shown above in Figure 28 Effect of import and export on prices and quantity in the country Assume that the demand function is defined as follows The supply function is defined as follows However the prices would equalize, i.e. Also since Replacing value of Q s from above in the equations And Chapter: Structure of Simulation Model 49

74 50 Flow based Market Coupling This was programmed into the MS Excel model. The input variable for the optimization was set be Q. Since there are four countries there were also all four Q s WELFARE CALCULATIONS FIGURE 29 WELFARE CALCULATIONS AFTER INTER COUNTRY TRADE Consumer Surplus = Area ( AIK) Producer Surplus = Area ( KJD) Net Welfare = Area ( AIK) + Area ( KJD) Cost of Generation = Area ( DJG) These values for each country were added up to give the net regional values for the North West European region SETTING UP THE OPTIMIZATION USING SOLVER Since the main aim of the simulation program is to study the average static behavior of the markets bring out sensitivities it was considered sufficient to consider only the above two constraints. For setting up the optimization problem Solver feature of MS Excel was employed. The optimization variable was the net import/export into a country. If there are imports, or exports from a country it leads to change in both supply and demand. This was modeled using the equations that calculate the change in demand and supply in a country that import/exports some known amount of electricity from outside. This has already been discussed above in the section, the equations were entered into the excel sheet and the solver was set up as follows. It needed the reference to the (Figure 30 and Figure 31) objective function net cost of generation, what the objective was minimization of the objective function, and constraints

75 FIGURE 30 OPTIMIZATION USING MS EXCEL SOLVER FIGURE 31 OPTIMIZATION PARAMETERS CHOICE OF OBJECTIVE FUNCTION Congestion management methods can have different goals. Before the operating hour, it be hours, days, weeks or months, the goal may be to optimally allocate available capacity among market parties. In real time the goal may be to deal with congestions occurring during operation. In the current simulation the focus was solely on the day ahead markets hence it was required to decide on an objective function that was suitable for this situation and was practical with the assumptions made and the data available. There is a debate on what should be used as an objective function maximization of market value, minimization of cost of generation, maximization of traded volume, etc. For the current study it was decided to use minimization of cost of generation as the objective function mainly because that is the ultimate objective of coupling markets. Also there are operational and mathematical issues with using maximization of market value with approximate linear demand and supply curves this method works well for explicit auction and also implicit auctions but when actual merit order is given, but not an approximate supply curve. In case of approximate linear supply curves with optimization variable are the net imports into the four countries the optimization problem that is set up can have multiple solutions. The solution also depends on the initial values of trade what are entered into the model. Basically it s a 8 variable problem with only five governing equations, and can hence have any number of solutions. This issue can be possibly resolved by introducing more constraints and intelligent actors into the simulation. However as the aim of research was to understand the working of the market and distinguish it from the effects that might happen due to idiosyncrasy of the actors (bids and offers) it was decided to use another objective function. The In case of approximate linear supply curve Chapter: Structure of Simulation Model 51

76 52 Flow based Market Coupling minimization of total cost of generation leads to unique solutions and also to higher net welfare for each country. However in real implementation of the system it is possible to use net welfare or market value maximization when actual bids and offers are available. It is possible to use quadratic programming to solve such problems. FIGURE 32 EFFECT OF AN INELASTIC DEMAND FOR DECIDING THE OBJECTIVE FUNCTION Other reason for using the minimization of the total cost is because of the nature of the demand curve. The demand curve that was calculated from the model was inelastic. (slope was almost a 65 for Netherlands, meaning that demand would change by 1GWh for a change of 65 in the electricity price). Under such situation the demand supply graph would look as shown in figure. If the demand function is assumed to be perfectly inelastic then the consumer surplus would tend towards infinity. More inelastic demand makes the Net Welfare completely dependent only on consumer surplus. Producer surplus becomes less significant if the slope of demand function is high. In PJM, with LMP, the optimization function also employs an perfectly inelastic demand function and minimization of cost of generation as objective function. If net welfare is completely dependent on consumer surplus, another way of saying maximization of net welfare, hence consumer surplus is to minimize the cost of generation of electricity. Also if the aim of integration of markets is to achieve a copper plate then the final aim of markets would be to minimize the net cost of generation. The objective function can be formulated as follows: : Where, dq si is the electricity demand in a country fulfilled by local generators. P i is the price of electricity after coupling in the national market CONSTRAINTS The constraints were three: Net trade should equal to zero

77 All electricity that is generated is also consumed Tie line capacities are respected Using these inputs to the Microsoft Excel Solver it was possible to obtain a solution to the market optimization. The results of this would be discussed in the section 7.1Base Case. 5.5 MODEL OUTPUTS The model gives the following outputs OUTPUTS FOR THE NON COUPLED MARKETS The following results were calculated for the national markets, assuming there was no coupling. TABLE 9 OUTPUT LIST FOR MODEL FROM NON COUPLED MARKETS NL BE FR DE Price and Quantitiy Surplus GW /MWh /hr /hr /hr /hr Consumer Producer Cost of Q* P* Surplus Surplus Net Welfare generation Net Welfare Total Cost of Generation OUTPUTS FOR THE COUPLED MARKETS Following parameter were calculated from the model after FMC. TABLE 10 OUTPUTS FROM THE MODEL AFTER FMC NL BE FR DE Price and Quantity Surplus GW GW /MWh GWh /hr /hr /hr /hr Net Consumer Producer Net Welfare Cost of Qd Qs P' Exchange Welfare(FMC) Welfare(FMC) (FMC) generation Chapter: Structure of Simulation Model 53

78 54 Flow based Market Coupling OVERALL RESULTS The comparison was made between the results of market after FMC and the independent market case. In scenarios comparison was also made between the results from the scenario and the results of FMC if applied in present case. Two main values of interest were the following, which were calculated from the model. 1. COST OF GENERATION: REGIONAL CHANGE 2. NET WELFARE: REGIONAL CHANGE COMPARATIVE RESULTS The following result were calculated to be able to make decisions on the effectiveness of FMC under different scenarios. TABLE 11 COMPARATIVE RESULTS FROM THE MODEL BETWEEN FMC CASE AND THE BASE CASE NL BE FR DE /MWh Price % Price % Consumer Surplus % Producer Surplus % Net Welfare % Demand Imported % Installed Capacity Used CAPACITY UTILIZATION RESULTS It was important to also see what the results were from the utilization perspective of the tie lines. It would help in deciding which tie lines need further investments to relieve congestion. TABLE 12 CAPACITY UTILIZATION RESULTS FROM THE SIMULATION MODEL Reverse Capacity Forward Capacity Reverse Flow Forward Flow Reverse Unused Capacity Forward Unused Capacity Reverse Capacity Utilizatio n Forward Capacity Utilizatio n GW GW GW GW GW GW BE FR BE NL DE FR DE NL 5.6 GRAPHS The following graphs were also plotted to give easier insights and comparisons into the market operation and comparison. Price of Electricity o Electricity Prices o Change in Prices Electricity Demand o Consumption per Country o Imports and Exports

79 Welfare Data o Consumer Surplus o Producer Surplus o Net Welfare o Distribution of Welfare Interconnector Capacity Utilization o Actual Flows o % Utilization of capacity Chapter: Structure of Simulation Model 55

80 56 Flow based Market Coupling 6 VALIDATION AND VERIFICATION OF MODEL As it is a simulation model and all simulation models have the property of garbage in garbage out. Hence there was a need to carefully and considerately ensure quality of the input data. It was tough to ensure this as most of data was not available directly from the owners, but was picked up from other studies, and was sometimes dated. It was tried to the utmost degree to have data which reflects reality, so that conclusions can be made on basis of this data. In this chapter you would find detailed verification of model and validation of the intermediate results. It is not possible to validate the final results as the market in this form does not exist yet. 6.1 VALIDATION OF TRANSMISSION MODEL The model has already been verified in detail by the authors (Qiong and Bialek 2005). After studying the report it was considered sufficient to take the model as being correct. 6.2 QUALITY OF DATA USED FOR SUPPLY CURVE This was one of the most important data for the economic calculations as all the rest of calculations would hinge on it. Any issue with data would reflect in the results of from the model. Hence below the supply curve calculated from the simulation model and actual merit order curves for the four countries are given below. Merit order curve based on the cost data from current model Merit order based on the actual data from the past (DG Competition report on energy sector inquiry (SEC(2006)1724, 10 January 2007) NL Supply Curve Price (euros/mwh) y = 3.183x R² = Generation Capacity (GW) NL Supply Curve Linear (NL Supply Curve) Price (euros/mwh) BE Supply Curve y = 4.217x R² = BE Supply Curve Generation Capacity (GW) Linear (BE Supply Curve)

81 Price (euros/mwh) FR Supply Curve y = 0.392x R² = FR Supply Curve Generation Capacity (GW) Linear (FR Supply Curve) Price (euros/mwh) DE Supply Curve y = 0.452x R² = FIGURE 33 MERIT ORDER CURVES FOR THE NORTH WEST EUROPEAN REGION, MODEL DATA AND COMPARISION WITH ACTUAL DATA FROM THE PAST As can be clearly seen both the graphs are very similar considering both the prices and the installed generations. Hence it would be acceptable to use the data that has been collected from the sources and still be able to get realistic results. 6.3 QUALITY OF DATA USED FOR AVERAGE DEMAND Second most important data was the data for average demand in a country. There were many sources to gather this data from Eurostat website, past studies and UCTE system adequacy report. For the model the values were picked up from the UCTE system adequacy report This was the most recent data. However the values were not averages but were snapshots in time. To see how well these values reflect reality they were compared to the Load Duration Curves Calculated for the last three years for the north west European region by London Economics report for the DG Trent. (DG Competition report on energy sector inquiry (SEC(2006)1724, 10 January 2007) Chapter: Validation and Verification of model Generation Capacity (GW) DE Supply Curve Linear (DE Supply Curve) 57

82 58 Flow based Market Coupling TABLE 13 COMPARISION OF AVERAGE DEMAND OF ELECTRICITY USED IN THE MODEL AND HISTORICAL DATA Average Demand used in model (from UCTE one day value) Average Demand based on the load duration curve NL Method of calculation of average demand based on load duration curve = average of values at 3000 and 6000 hours FIGURE 34 LOAD DURATION CURVE FOR NETHERLANDS BE FIGURE 35 LOAD DURATION CURVE BELGIUM FR FIGURE 36 LOAD DURATION CURVE FOR FRANCE

83 DE FIGURE 37 LOAD DURATION CURVE FOR GERMANY As is clear from the values, the values used in the model are higher than what is reflected from the past load duration curves. This could be attributed to two factors The values in the study are older, only till If we consider the annual increase there would be some convergence of the values. The value snapshot considered by UCTE, i.e. Referecne value on the 3rd Wednesday of January, 2007 at 11:00, might not reflect the reality as well. However the values still are reflective of the trend, and have been achieved at the same instance of time in the past. Hence it is much better reflective of reality than the average values reflected in the data from duration curves. 6.4 PRICE OF ELECTRICITY WITHOUT IMPORTS AND EXPORTS To test the correctness of the supply curve and the demand data it was decided to compare the results of price calculated from the simulation model without inter country trades and the past prices from the day ahead markets in APX, BelPEX, PowerNext and EEX. This data was tough to gather too. However it was collected for 24 hours and for the past two years from the respective power exchanges. The data analysis of the data yielded the statistics given in the Figure 38 Comparision of wholesale price of electricity based on the model and actual past data from the power exchanges for year The software SAS and MS Excel were utilized for the analysis. The results can be found in the appendix A.4. FIGURE 38 COMPARISION OF WHOLESALE PRICE OF ELECTRICITY BASED ON THE MODEL AND ACTUAL PAST DATA FROM THE POWER EXCHANGES FOR YEAR As can be seen for NL and BE the prices calculated from the model are much higher than the actual day ahead prices. However for Germany and France it is much closer to the reality. Other observation is the fact that mean of prices in Germany is actually higher than the mean in Netherlands and Belgium, which might not be true. This could be attributed to the fact that a lot of values from the data set for EEX were missing for some reason. The prices in NL and BE are expected to be higher as all the generation is assumed to be local, and that would mean it would be expensive. In real life prices in NL and BE are lower because of cheaper imports from Germany and France. Chapter: Validation and Verification of model 59

84 60 Flow based Market Coupling To get better perspective on the prices it was decided to compare the prices to the futures prices, which showed better correlation. 6.5 VALIDATION OF PTDF VALUES FROM POWER WORLD The basic tool for congestion management in FMC is Power Transfer Distribution Factor (PTDF) which assesses sensitivity of a given line flow to the changes in nodal generations and demands. PTDF factors are used to measure the impacts of a commercial transaction on the physical power flows in the critical transmission lines/ or interfaces. Based on the values of PTDFs, transactions are curtailed and increased in order to relieve lines overloads. Due to the phenomenon of loop flows in a meshed network, PTDF calculation requires the full network information. This next section is based on the paper by (Zhou and Bialek 2005) TABLE 14 PTDF VALUE COMPARISON FROM POWER WORLD MODEL TO ACTAUL PUBLISHED VALUES In order to test the applicability of the UCTE network model under various dispatch conditions, the published PTDF values were compared with the calculated ones. Although not all PTDF values are publicly available, some of the PTDFs in the UCTE network were obtained from European Commission reports 5 (Haubrich and Fritz 1999). As our study was focused on the cross border congestions, we were interested only in the values of PTDFs on cross country interfaces rather than individual PTDFs for all internal lines within each country. Tables II IV give the PTDF factors (in percent) on the crossborder interfaces for the transactions from Belgium to Italy 6, North France to the Netherlands and France to Italy 7, respectively. The correlation factors are 97%, 91% and 95%, proving very good agreement between the expected and actually calculated values. 6.6 CONCLUSIONS The conclusions from the validation and verification yielded confidence into the data. The data from the studies were lower for the marginal costs and average prices seen on the spot markets. Main reason for this is that the plants right now in operation have been in operation for long and have already paid for their investments. It is profitable to just bid on the marginal costs in the market. However the data which was Net Transfer Capacities (NTC) and Available Transfer Capacities (ATC) in the Internal Market of Electricity in Europe (IEM) Information for User, ETSO,

85 collected was mostly for newly installed or planned plants in the cost sheet. Hence the supply curve was shifted towards higher costs in the model. This was the main difference. Also the demands have risen in the present case higher than the once in the published reports from 2004 and Considering these facts it is possible to justify the values that were considered in the model. Also counting the fact that the model is built to give approximate values and is mainly built for the purpose of understanding the behavior of FMC it was justified to use the data. Chapter: Validation and Verification of model 61

86 62 Flow based Market Coupling

87 Section 3: Results Here the results from the model are presented and conclusion are drawn Chapter: Validation and Verification of model 63

88 64 Flow based Market Coupling

89 7 RESULTS This chapter is divided into two main sections. First section would detail the results from the base case. The second part would detail the results from the scenarios. 7.1 BASE CASE The base case was run with the values which have already been discusses in Chapter 5(Structure of Simulation Model). The model is run first without coupling, i.e. all the markets calculate MCP without any imports or exports. The results of the calculations are given below in the table. FIGURE 39 MCP, EQUILIBRIUM DEMAND AND SURPLUS CALCULATION WITHOUT MARKET COUPLING The prices in Belgium are the highest, followed by Netherlands. France with its huge installation of nuclear energy has the cheapest energy prices. Germany is also substantially lower priced compared to NL and BE because of coal installations. The second step was to run the solver in excel to minimize the net cost of generation by varying the imports and exports into each country PRICE OF ELECTRICITY Prices (euros/mwh) ELECTRICITY PRICES PRICES From the end consumer viewpoint the most important indicator is the price of electricity. It must be considered that this graph only gives the price of wholesale electricity; however it contributes substantially to the final electricity price. As can be seen after flow based market coupling Dutch and Belgian customers would earn the benefit. NL BE FR DE P* P' Chapter: Results 65

90 66 Flow based Market Coupling 2.00 (2.00) (4.00) (6.00) (8.00) (10.00) Prices Change (euros/mwh) NL BE FR DE Price CHANGE IN PRICES As already said Dutch and Belgium see substantial decrease in electricity prices. The biggest gainer is Netherlands with almost 8, Belgium also has reduced prices but not as much reduction as seen in Netherlands 6. It must also be stated here that the benefit driven by countries shows a very strong dependence on the slope of calculated supply curve. Hence it is very important to get the supply curves as accurate as possible. This was the reason behind creating a costing model for all the technologies and obtaining real world data to populate the matrices ELECTRICITY DEMAND Electricity Demand (GW) NL BE FR DE Q* Qd CONSUMPTION PER COUNTRY The consumption does not change so much in the countries. However as expected, the countries in which prices went up namely Germany and France, the consumption decreased. This is true because of the assumption of elastic demand curve. Netherlands and Belgium saw more consumption because of falling prices. 5.00% 0.00% 5.00% 10.00% 15.00% 20.00% Import Export ( ve means import) NL BE FR DE % Demand Imported IMPORTS AND EXPORTS The high price countries import from the low electricity priced countries. Netherlands imports almost 17%(this number is comparable to what Netherlands imports currently at about 20%) of its net consumption. Belgium is able to import only 12%. However current values are based mostly on explicit contacts, trading on Power exchanges is still limited. However if all electricity is traded over the power exchange and calculations are done according to implicit flow based coupling method with objective function being the total cost of generation then the scenario might change to what is shown in the figure.

91 7.1.3 WELFARE DATA Millions $10.00 $9.00 $8.00 $7.00 $6.00 $5.00 $4.00 $3.00 $2.00 $1.00 $0.00 Consumer Surplus (Meuros/hour) NL BE FR DE Consumer Surplus Consumer Welfare(FMC) CONSUMER SURPLUS The consumer surplus has a marginal increase in importing countries mainly because of reduced prices and also increased demand. Change in consumer surplus in exporting countries is relatively small because of their large size and large spare installations for generation. Millions PRODUCER SURPLUS Producer surplus did see a substantial reduction in the quantity. This was visible in both the importing countries and exporting countries. Importing countries would hurt the local generating companies. At the same time the exporting countries would benefit their local generators as they would now be able to sell outside of their territory and earn benefit NET WELFARE Net Welfare for all countries remained almost the same. The benefit earned by the consumers was countered by the loss to the generators in the importing countries and was reversed in case of exporting countries. In general overall for the local economy the benefit is almost negligible. Chapter: Results Producer Surplus (Meuros/hour) NL BE FR DE Producer Surplus Producer Welfare(FMC) Millions Net Welfare (Meuros/hour) NL BE FR DE Net Welfare Net Welfare (FMC) 67

92 68 Flow based Market Coupling 10.00% 5.00% 0.00% 5.00% 10.00% 15.00% 20.00% 25.00% 30.00% 35.00% Distribution of Welfare NL BE FR DE % Consumer Surplus % Producer Surplus % Net Welfare DISTRIBUTION OF WELFARE Looking at percentage is interesting when considering welfare. This would decide the amount of opposition that the proposal would face from various actors. As can be clearly seen that the Producers in Netherlands would oppose this move, and so would the producers in Belgium. However the remaining actors would be more or less neutral to the change INTERCONNECTOR CAPACITY UTILIZATION % % 50.00% 0.00% 50.00% % % % Utilization BE FR BE NL DE FR DE NL % UTILIZATION OF CAPACITY As already said, the utilization on the BE FR and DE NL border is the highest. These would be the tie lines that would need to be reinforced in order to attain the copper plate or a successful competitive and liquid market in the north west European region. However it must be noticed that this utilization % is of the physical available capacity and not the NTC values. Net Utilization Utilization of Interconnector Capacity BE FR BE NL DE FR DE NL Forward Unused Capacity Forward Flow Reverse Unused Capacity Reverse Flow ACTUAL FLOWS Considering the tie line capacity utilization factor is important. As can be seen BE FR and DE NL are completely used. This is what would be expected as well. Only when two of the four interconnectors have been fully utilized there could be no additional flow from any direction possible and hence it would lead to a closed constraint.

93 7.1.5 CONCLUSIONS FROM BASE CASE On the whole FMC is effective in bringing down the net cost of generation for the whole north west European region. In terms of net welfare every country gained except France, where there was no change observed. From the perspective of energy prices, they went down significantly for Netherlands and Belgium, though there was a very insignificant increase in the prices in Germany and France. The interconnectors on BE FR and DE NL borders were completely utilized to their limits. Overall it seems to be an efficient system for optimization of market value and better usage of the interconnector capacity. Hence it can be concluded that FMC would lead to regional gains if implemented in the North West European region. The detailed numbers from the analysis are attached in the appendix A.9 and A.10 Next step was to compare FMC with NTC based Market coupling method which is already implemented between NL, BE and FR. However there was issue of defining the contract path with addition of Germany. Germany offers a parallel or loop flow, which can not be defined using a contract path, as the flows resulting from execution of a contract between France and Netherlands would pass through Germany and the German TSO should be paid for its marginal value. This makes definition of contract path complex. As there are no documents of how the system would define contract path if not basing them on PTDFs it was decided to base them on assumptions. For testing if FMC is better than Market Coupling, three different contract paths were defined clockwise, anti clockwise and 50% division. The results of the analysis would be discussed below. 7.2 CURRENT SITUATION VS FMC As already discussed in chapter 2.2 Trilateral Market Coupling, the current situation in the North West European region already has coupled markets. In order to be able to comment on the additional benefits that FMC entails it was necessary to make comparison with the current system. Hence a modification was made to the model with following parameter values. FIGURE 40 TRADED VOLUMES FOR FMC AND TRILATERAL MARKET COUPLING COMPARISION Chapter: Results 69

94 70 Flow based Market Coupling The explicit trades as mentioned in the picture above lead to already some change in the demand and prices in the countries as compared to the base case of no coupling at all. A new optimization problem was now set for the trilateral market coupling countries NL, BE and FR. The optimization objective function was set to be minimization of generation cost and not net welfare (present system employs Net welfare, for justification see chapter of methodology). The results from the optimization were plotted against the results from the base case without any coupling and FMC to clearly delineate the effects Prices (euros/mwh) Prices Change Compared to no coupling (euros/mwh) Electricity Demand (GW) (2.00) NL BE FR DE (4.00) (6.00) (8.00) NL BE FR DE P*(No Coupling) P'(MC) P'(FMC) (10.00) Price(MC) Price(FMC) 0.00 NL BE FR DE Q*(No Coupling) Qd(MC) Qd(FMC) Import-Export (-ve means import) $10.00 Consumer Surplus (euros/hour) 1.40 Producer Surplus (euros/hour) 5.00% 0.00% $9.00 $8.00 $ NL BE FR DE $ % % $5.00 $4.00 $ $ % % % Demand Imported(MC) % Demand Imported(FMC) $1.00 $0.00 NL BE FR DE Consumer Surplus(No Coupling) Consumer Surplus(MC) - NL BE FR DE Producer Surplus(No Coupling) Producer Surplus(MC) Producer Surplus(FMC)

95 0.14% Distribution of Welfare Utilization of Interconnector Capacity % Utilization 0.12% 0.10% 0.08% % % 50.00% 0.06% 0.04% 0.02% (1.00) (2.00) BE-FR BE-NL DE-FR DE-NL 0.00% % % BE-FR BE-NL DE-FR DE-NL 0.00% (3.00) % NL BE FR DE % Net Welfare(MC) % Net Welfare(FMC) Forward Unused Capacity (MC) Forward Flow(MC) Reverse Unused Capacity (MC) Reverse Flow(MC) Net Utilization (MC) Net Utilization(FMC) FIGURE 41 RESULTS FROM SIMULATION MODEL FOR CURRENT STATE OF TRADE IN NORTH WEST EUROPEAN REGION Conclusions As can be clearly seen above FMC is more beneficial as compared to present implementation of trilateral market coupling between NL, BE, FR and explicit auctions between NL, DE and DE, FR. Clearly there are benefits for Netherlands there is almost 2 per MWh further drop in prices with FMC. For Belgium there is almost no difference. Germany sees higher prices with FMC as it is able to sell more in the market and substitute the trades which were otherwise going from France. There is also a net gain the welfare of the region. The percentages are small however it should be considered that the initial values of consumer surplus were high due to the inelastic nature of the demand function. It can be concluded on basis of the above mentioned results the FMC would indeed lead to gains both in national and regional levels. Though it should also be kept in mind that there was a very insignificant drop in welfare within France, mainly because Germany replaced some of its export benefits with introduction of FMC. 7.3 COMPARISON OF FMC WITH MARKET COUPLING The next step was to compare the claim of FMC with Market coupling The following situations were simulated. Chapter: Results 71

96 72 Flow based Market Coupling FIGURE 42 FLOW DIRECTION FOR COMPARISION OF FMC WITH NTC BASED MARKET COUPLING The results are attached in the appendix A.11 The results can be discussed from perspective of any of the involved countries, and actors. As already stated the problem owner being the Ministry of Economic Affairs this report would aim to analyze the results from their perspective.

97 1. Consumer Perspective: Wholesale MCP after inter country trading The end consumers, in this case the domestic users, care only about the price to electricity that is charge to them. From the results of the analysis the Dutch customers would pay least in case of NTC based market coupling with equally distributed assumption on the contract paths( 8.18 /MWh), followed closely by the results from Flow based market coupling based on actual PTDF values ( 8.21 /MWh). However the difference is not significant only at the second decimal value ( 0.03) which could be neglected. Basically both FMC and distributed contract Market coupling gave same results on Dutch market prices. However considering prices in Belgium do not get benefit in case of distributed contract based market coupling (a drop of less than 2) which dropped substantially with FMC (almost a drop of 6). The results from anti clockwise and clockwise Market coupling are much different from the other two cases. As can be seen that prices fall by about 4 in Netherlands FMC Prices Change (euros/mwh) 2.00 NL BE FR DE (2.00) (4.00) (6.00) (8.00) (10.00) Price MC Anti Clockwise Prices Change (euros) (1.00) NL BE FR DE (2.00) (3.00) (4.00) (5.00) Price MC Clockwise Prices Change (euros) (0.50) NL BE FR DE (1.00) (1.50) (2.00) (2.50) (3.00) (3.50) Price MC Distributed Prices Change (euros) 2.00 NL BE FR DE (2.00) (4.00) (6.00) (8.00) (10.00) Price 2. Effects on Welfare Distribution: Welfare distributions can be predicted on basis of above graphs, the importing countries saw the consumer surplus going up while at the same time exporting countries saw consumer surplus going down because of slightly higher prices. Also to be noted is the fact in general it is observed that the Net Welfare change for all countries is always positive but is only a small percentage. More information and graphs can be looked up in the appendix A.11.3 Chapter: Results 73

98 74 Flow based Market Coupling 3. Dependence on imports: Imports and exports for each of the partners involved in market coupling after equilibrium To an extent the amount of imports decide the level of dependence that the country has on its trade partners. From the perspective of the supra national aspirations this is good. However thinking of the energy independence and self sufficiency too much imports create a sense of insecurity in the local market. In the long term they may create incentives for the generators to move the generation plants in the countries with lower cost of generation, thus causing further erosion of spare capacity and could lead to higher volatility in the markets. In case of FMC Holland ends up importing almost 15% of its total demand from outside. This is actual in line with the present imports which stand at around 15 20%. Other interesting thing to notice is the fact that France become the sole exporter in case of MC. This should not be very surprising given the fact that electricity is the cheapest in this country and the objective function for the current simulation is minimization of the oveall cost of generation. FMC 5.00% 0.00% 5.00% 10.00% 15.00% 20.00% Import Export ( ve means import) MC Anti Clockwise NL BE FR DE % Demand Import Export ( ve means import) 8.00% 6.00% 4.00% 2.00% 0.00% 2.00% 4.00% 6.00% 8.00% 10.00% NL BE FR DE % Demand MC Clockwise 6.00% 4.00% 2.00% 0.00% 2.00% 4.00% 6.00% 8.00% Import Export ( ve means import) MC Distributed 5.00% 0.00% 5.00% 10.00% 15.00% 20.00% NL BE FR DE % Demand Import Export ( ve means import) 10.00% NL BE FR DE % Demand 4. Effect on local demand It was also noted that demand changes only slightly for each case this could be attributed to the fact that demand function is almost inelastic. The slope of demand function in NL and BE is of the order of 1000, hence there is not much change in the demand with increasing or decreasing prices. This is in line with the reality too. For details refer to appendix

99 5. Tie Line capacity utilization % utilization of the interconnector utilization FMC MC Clockwise To make decisions about investment into grid expansion it is very important for the government to look at the actual picture of congestion. The results from the analysis of capacity usage under different market methods were largely different. This probably also shows that there would exist gaming opportunities for the actors with large market power both generators and large consumers. In case of FMC the most congested borders are BE FR and DE NL. Is should be noted that France sort of determines which tie lines would get congested because of low prices. In case of MC clockwise the BE FR tie line gets congested because France is exporting to both Belgium and Netherlands. It is also trying to export to Germany, thus the DE FR tie line also gets congested. Similar is the case with anti clockwise flow but with reversed signs. MC distributed leads to appearance of three congested lines, i.e. all tie lines except the DE NL tie line. For details refer appendix A % % 50.00% 0.00% 50.00% % % MC Anti Clockwise % % 50.00% 0.00% 50.00% % % % Utilization BE FR BE NL DE FR DE NL Net Utilization % Utilization BE FR BE NL DE FR DE NL Net Utilization 80.00% 60.00% 40.00% 20.00% 0.00% 20.00% 40.00% 60.00% 80.00% % % MC Distributed % % 50.00% 0.00% 50.00% % % % Utilization BE FR BE NL DE FR DE NL Net Utilization % Utilization BE FR BE NL DE FR DE NL Net Utilization CONCLUSIONS In the highly meshed interconnected transmission networks in Europe programmed exchanges and physical flows differ often considerably. They would be closely connected to the power flows through the cross borders only in the ideal case of a peninsular system and its neighbor if both were interconnected through a single tie line, in case of market coupling in between Netherlands, Belgium and France there are no parallel paths hence it is a good approximation to peninsular system with the neighbors lying in a straight line. However, in a widely interconnected network like for example the UCTE network the power flow through the cross border tie lines between two neighbor areas A and B may be interpreted as a superposition of a direct flow, which is related to exchanges between A and B and a parallel flow, which is related to all the other exchanges in the meshed network and to the location of generations and loads in the several grids. Therefore there would be a parallel flow even if all the exchanges in the interconnected system were set at zero. With introduction of Germany into the market coupling there would be large implications due to transit and loop flows on the values of ATC and NTCs. These transit flows cannot be handled by the NTC based MC. Chapter: Results 75

100 76 Flow based Market Coupling FMC leads to substantial drop in prices for both NL and BE. The drop is almost twice as much as achieved by using MC (clockwise and anti clockwise). MC with distributed also did perform well if considering only the prices in Netherlands, however from a regional perspective FMC would be more suitable. FMC also makes best possible use of the interconnector capacity, the interconnector % utilization ratios in other cases is not really the physical limit usage but some approximate flow based on the commercial exchange that does not take care of parallel flows occurring due to transactions between other countries. Hence it is possible to conclude that FMC comes across as an ideal method for managing inter country electricity trade and congestion management. It leads to the best use of actual interconnector capacity and hence also optimal trade volumes. Next part of the study would focus on the different scenarios that has either been planned in the future or are expected to happen in the near future. Effect of FMC on these scenarios would be discussed in the next section. 7.4 SCENARIOS Scenarios were developed in consultation with the Ministry of Economic Affairs and suggestions from experts in the field. The scenarios were mainly on policy and investment options that can be employed. Increase of generation capacity, increase of transmission capacity, effect of carbon markets and the effect of loop flows because of wind generation in Germany have been investigated INCREASE GENERATION CAPACITY IN NETHERLANDS There is presently no coal production in the Netherlands. All of their coal mining operations were closed down following the discovery of large indigenous reserves of natural gas in the early 1970s. Most of the coal reserves are in the southern and east central parts of the country, and recoverable reserves are estimated (as of January 2005) at about 500 million short tons. The Netherlands currently ranks as the world's 32nd greatest coal consumer (and 11th in the EU), accounting for about 0.3% of the world's (and about 1.9% of the EU's) total annual coal consumption (Source: EIA/DOE). With the volatility that has been seen in the Natural Gas and Oil markets recently there is a renewed interest in Netherlands in investments in coal based power plants. Two different scenarios were tested in the current research first with creation of 5000 MW additional capacity and second with additional capacity of 10 GW. The results would be discussed next.

101 INSTALLATION OF 5000 MW COAL BASED POWER PLANTS Price (euros/mwh) NL Supply Curve y = 2.439x R² = NL Supply Curve Generation Capacity (GW) Linear (NL Supply Curve) FIGURE 43 NEW SUPPLY CURVE WITH ADDITIONAL 5 GW OF HARD COAL BASED POWER PLANT The change induced in the model due to increased generation capacity is outlined above, the supply curve of NL changes and the slope becomes flatter. The market results from this change are presented in the figure below: Prices (euros/mwh) 10.00% Import-Export (-ve means import) % % Tie-Line Utilization % 0.00% -5.00% NL BE FR DE % 50.00% % % % % 0.00% % % BE-FR BE-NL DE-FR DE-NL % % NL BE FR DE P* (w/o Coupling) P' (FMC Base) P* P' % Demand Imported (FMC Base) % Demand Imported Net Utilization (FMC Base) Net Utilization Figure 44 Price change in NL, Imports and % tie line utilization for scenario with installation of 5000 MW coal based generation in NL The prices in the Dutch electricity market would drop substantially with installation of extra 5000MW, even in absence of any market coupling prices drop to almost 40 per MWh. However there are still substantial benefits from FMC. If FMC was introduced in the present sate the prices would drop form to Chapter: Results 77

102 78 Flow based Market Coupling per MWh. Then if government decides to invest in 5000 MW coal based power plants, then the prices would drop further to NL is also the country that derives most benefit in from FMC with the highest price drop of 6.48 per MWh, followed by Belgium 5.55 per MWh. Also Belgium begins to import more from other countries; it now imports almost 23% of its demand from outside, while Netherlands has shown an opposite trend with imports now standing at around 7%. There is also reversal of flow on the Dutch Belgian flow. In FMC base case the flow was mostly from Belgium to Netherlands due to cheaper imports coming from France, however now the flows are from Netherlands to Belgium, as electricity in Netherlands is being supplied more locally. It goes on to prove that NL becomes much more independent of the imports it s also shown in the figure above. However it must to noted that still Netherlands is a net importer despite investing so largely in the coal based plants. The detailed results for the analysis are given in the appendix A.12 One important aspect that should be kept in mind is that installation of coal based power plants though brings down the prices they would also push the renewable energy technologies out of business for most of the time in a year. This should be considered by the ministry before taking such a step. The renewable technologies would need government support to be able to remain competitive and be able to survive this might offset some of the benefits of installing more coal based power plants INSTALLATION OF 10,000 MW COAL BASED POWER PLANTS Price (euros/mwh) NL Supply Curve y = 1.951x R² = Generation Capacity (GW) NL Supply Curve Linear (NL Supply Curve) FIGURE 45 NEW SUPPLY CURVE WITH ADDITIONAL 10,000 MW OF HARD COAL BASED POWER PLANTS The change induced in the model due to increased generation capacity is outlined above, the supply curve of NL changes and the slope becomes flatter. The market results from this change are presented in the figure below:

103 Prices (euros/mwh) Import-Export (-ve means import) % Utilization % % % % % -5.00% NL BE FR DE 50.00% % 0.00% % % % BE-FR BE-NL DE-FR DE-NL - NL BE FR DE % % % % P* (w/o Coupling) P' (FMC Base) P* P' % Demand Imported (FMC Base) % Demand Imported Net Utilization (FMC Base) Net Utilization FIGURE 46 PRICE CHANGE IN NL, IMPORTS AND % TIE LINE UTILIZATION FOR SCENARIO WITH INSTALLATION OF 10,000 MW HARD COAL BASED GENERATION IN NL The most striking observation from this scenario is that Netherlands becomes a net exporter. Belgium is the only country which is importing now from both Netherlands and also France and a small amount of exports from Germany. Belgium prices drop down and also it is importing now almost quarter of its demand from other countries. The Netherlands ends up exporting electricity amounting to 5% of its local demand. Also the producer surplus in the Netherlands increases mainly attributed to the fact that the prices go up and electricity is being exported to other countries. Another interesting observation is that Germany is almost out of the loop. This could mainly be because of presence of low cost countries in the neighborhood of Belgium, and no direct connection between Germany and Belgium. This could however change if there were to be integration of networks in Luxemburg, allowing Germany to trade electricity directly to Belgium over that tieline. Presently the flows from Germany to Belgium have to pass through either France or Netherlands and hence have lower value. Also worth noting is the local prices of electricity in Netherlands drops down to per MWh, which is marginally lower than the local prices in Germany which stand at per MWh. The detailed results for the analysis are given in the appendix A.13 However it should also be noted that if there was no market coupling and Netherlands was to install 10,000 MW coal based power plants than the local prices will drop down to per MWh, and actually there is increase in prices due to coupling as there is exports to Belgium leading to a small increase in prices. Hence there is no clear cut benefit with installation of 10,000MW and FMC together if only considering the national interests of Netherlands. Just going for installation of 10,000MW would give the benefits. Conclusion Installation of additional coal based capacity in Netherlands has strong impact on the local prices and status of Netherlands as being an net importer of electricity. However it should be noted that installation of large capacity brings down the benefit that Netherlands was deriving out of FMC, actually with installation of +10,000 MW of coal based power plants Netherlands became a net exporter. However one thing that was not considered here was the carbon prices. The price benefit that has been calculated was based on variable cost of electricity generation but without any carbon prices. It is possible that with inclusion of carbon prices (and assuming that it would rise from the present levels of less than a euro per ton of CO 2 emission) the Chapter: Results 79

104 80 Flow based Market Coupling variable cost would go up. Also from the purely environmental and health perspective government needs to investigate the implications such an huge change in generation mix might have on the country INCREASED TIE LINE CAPACITY Tennet is working on a new direct Dutch German 1 2 GW power interconnnector from Doetinchem to Wesel, planned to come onstream 2012, which could physically complete an integrated north west European power market. It was considered important to look at the effect FMC would have on the electricity market landscape after installation of the new transmission capacity. Two scenarios were modeled to study the effects first with installation of 2000 MW and the second with installation of 5000 MW on the DE NL borders. The section below first describes the process of setting up the model, which was similar for both cases. The second part details the results. The new transmission line would have effect on two aspects of the model first it could induce changes in the PTDFs as there is now an additional path available for the current to pass through and second that the interconnector capacity now is more hence there could be more exchange of electricity between the countries, leading to better convergence of the prices. Both these aspects were taken into consideration for studying the effect on new transmission line on the electricity markets in north west European region. The new line was assumed to be installed between Gronau in Germany and Hengelo in Netherlands (referred to as Gronau Hengelo, from here on). The line is parallel to already existing transmission line between the two stations. The technical details for the transmission line that were entered into Power World model are given below. TABLE 15 PARAMETERS FOR THE NEW TRANSMISSION LINE BETWEEN DE NL TieLine Records Series Reactance Meter MW Meter Mvar Status Lim MVA MW Loss Mvar Loss Closed Closed The values of series reactance were picked to be similar to the ones of existing lines of similar capacities from within the model. FIGURE 47 THE DIALOG IN POWERWORLD TO SET UP THE VALUES OF PARAMETERS FOR THE TRANSMISSION LINE

105 After installing a new transmission line as described above, the PowerWorld model is run to get the new values of the PTDF s. The new values compared to the previous values are shown below. TABLE 16 COMPARISION OF THE PTDF VALUES BEFORE AND AFTER INSTALLATION OF 2000 MW TRANSMISSION LINE BETWEEN DE NL Base Case With 2000 MW transmisison line between DE NL NL NL BE NL FR NL DE NL NL NL BE NL FR NL DE NL BE FR BE NL DE FR DE NL The first observation is that there is no significant variation between the values. There is only a very minor change in the values at the DE NL border. However even that is at the third decimal point. It can thus be assumed that installation of a 2000 MW line between DE NL is an insignificant event considering the whole network. It can safely be concluded that PTDF values are not very sensitive to such changes. However it is still not clear what effect this small change in PTDF can have on the outcomes of the market module. Next step was to increase the tie line capacity in the market module by 2000 MW, followed by the calculations of the final results. The line was assumed to be bi directional; hence the capacity was increased in both directions. FIGURE 48 NEW TIE LINE CAPACITY VALUES Chapter: Results 81

106 82 Flow based Market Coupling BETWEEN GERMANY AND NETHERLANDS (+2000 MW) Prices (euros/mwh) Import-Export (-ve means import) % Utilization % % % 5.00% % % 50.00% % % % NL BE FR DE 0.00% % BE-FR BE-NL DE-FR DE-NL NL BE FR DE P* (w/o Coupling) P' (FMC Base) P* P' % % % % Demand Imported (FMC Base) % Demand Imported % % Net Utilization (FMC Base) Net Utilization FIGURE 49 RESULTS FROM FMC AFTER INSTALLATION OF 2000 MW TRANSMISSION LINE BETWEEN DE NL The change in prices in the Netherlands is high it drops by 4.62 compared to case in which FMC is implemented in present situation. It shows that there is sensitivity of the FMC market results on the transmission capacity. Also Germany becomes the only exporter, despite lower prices in France. The explanation to this also lies in how FMC operates, Netherlands only has an interconnector capacity of 1350 MW with Belgium, that so that means maximum amount of electricity that can come from France to Netherlands through Belgium is 1350 MW. Additionally the interconnector between France and Germany allows only 1700MW to flow from France to Germany. Hence there isn t a possibility of large imports into Netherlands from France, because of congestion on BE FR tie line. When the interconnector on Germany and Netherlands border gets more capacity installed then Germany is able to deliver large amounts of electricity and 10% of that flows through the parallel path, DE > FR > BE > NL. Hence looking at the whole system it is more beneficial to dispatch the generators in Germany and not France. This should be kept in mind the FMC results have strong dependence on topology of the system. Any change on one tie line can lead to completely different results, which are not intuitive to the sense of dispatching the cheapest generators. The increased transmission capacity allows more flow from Germany to both Belgium and Netherlands which allow these two countries to get the benefit. The prices in Belgium also drop by almost 8, compared to the drop of 6, in the base case. The net welfare for all countries is positive. For France the change is almost zero. The details of the results can be checked from the appendix A.14 Also still there is congestion on the DE NL and BE FR border. That shows that there could be additional benefits with increasing the capacity of the DE NL border further. A case was also run with increasing BE FR tie line capacity, as that is also congested, in both the base case and current case.

107 Prices Change (euros) FMC Base FMC Scenario Prices Change (euros) FMC Base FMC Scenario % Utilization % % (1.00) (2.00) NL BE FR DE (1.00) (2.00) (3.00) NL BE FR DE 50.00% 0.00% % BE-FR BE-NL DE-FR DE-NL (3.00) (4.00) % (4.00) (5.00) % (5.00) (6.00) Price Price Net Utilization (FMC Base) Net Utilization FIGURE 50 PRICE CHANGE AFTER INSTALLATION OF +2000MW ON NL DE TIE LINE FIGURE 51 PRICE CHANGE WITH INSTALLATION OF +2000MW ON BE FR TIE LINE FIGURE 52 UTILIZATION OF FLOW LINE WITH MW ON BE FR LINE, COMAPRE TO FIGURE 45 The detailed results are attached in the appendix A.15. It can be seen that clearly the benefit now shifts to Belgium as an importer and to France as an exporter. France becomes the sole exporter and Germany is also importing small amount of capacity from France BETWEEN GERMANY AND NETHERLANDS (+5000 MW) The case was setup using the same steps as above. The results are discussed below. TABLE 17 COMPARISION OF THE PTDF VALUES BEFORE AND AFTER INSTALLATIONOF 5000 MW TRANSMISSION LINE BETWEEN DE NL Base Case With 5000 MW transmission line between DE NL NL NL BE NL FR NL DE NL NL NL BE NL FR NL DE NL BE FR BE NL DE FR DE NL Again the same observation is made about the effect of a new transmission line on the PTDFs, it is very small, affecting only the third decimal value. This time the value of interconnector transmission capacity between DE NL was increased by 5000 MW, taking it up to 7500 MWs. The results of the FMC would be discussed next. Chapter: Results 83

108 84 Flow based Market Coupling Prices (euros/mwh) Import-Export (-ve means import) % Utilization NL BE FR DE P* (w/o Coupling) P' (FMC Base) P* P' 15.00% 10.00% 5.00% 0.00% -5.00% NL BE FR DE % % % % % % Demand Imported (FMC Base) % Demand Imported % % 50.00% 0.00% BE-FR BE-NL DE-FR DE-NL % % % Net Utilization (FMC Base) Net Utilization FIGURE 53 RESULTS FROM FMC AFTER INSTALLATION OF 5000 MW TRANSMISSION LINE BETWEEN DE NL As can be seen the results are identical to the results of 2000 MW case, prices change and imports into Netherlands are almost the same. This shows that congestion is now no longer on the NL DE border. The only congestion now is on BE FR line. As can be seen in the first case two of the interconnectos are congested BE FR and DE NL. In the case of increased capacity on the DE NL border further, the congestion shifts to BE NL border. More details on the model can be found in the appendix A.16 The case was also run with increasing the border capacity on DE NL border by 10,000 MW, however the results resembled the results from additional 5000 MW case, as the congestion is not on that border anymore. See Appendix A DISCUSSION ON SENSITIVITY OF PTDFS As seen from the two cases, first with installation of +2000MW and then with +5000MW, the PTDFs do not change substantially. The effect is seen only at the third decimal place. This could also be expected mainly because the whole of UCTE network is considered here and an additional 1000 s of MWs only have a minor impact the topology of the network. Hence it could be assumed that there would not be major changes in the PTDFs with small changes in the network. Next it is also important to see what small differences can have on the market results. To check the impact of PTDF on the solutions case with PTDFs of 5000 MW, and tie line capacity on NL DE border of additional 2000 MWs. Now the results could be compared to the results of model with the PTDFs from 2000 MW and additional capacity of 2000 MW on DE NL border (A.14). There was no remarkable change with between the two.

109 3.00 Prices Change (euros) FMC Base - FMC Scenario % % Utilization % 50.00% (1.00) NL BE FR DE 0.00% (2.00) (3.00) % BE-FR BE-NL DE-FR DE-NL (4.00) % (5.00) Price % Net Utilization (FMC Base) Net Utilization FIGURE 54 PRICE CHANGE AND % UTILIZATION FROM +2000MW ON DE NL BORDER AND PTDF'S ALSO FROM +2000MW CASE 3.00 Prices Change (euros/mwh) FMC Base - FMC Scenario % Utilization % % % (1.00) NL BE FR DE 0.00% (2.00) % BE-FR BE-NL DE-FR DE-NL (3.00) % (4.00) % (5.00) Price Net Utilization (FMC Base) Net Utilization FIGURE 55 PRICE CHANGE AND % UTILIZATION OF INTERCONNECTOR FROM +2000MW CAPACITY ON DE NL BORDER AND PTDFS FROM +5000MW The above two figure show that there is no difference between the results. It is possible that under extreme change of values the results would change as seen in the comparison of FMC with MC in section 7.3. However there is no significant dependence on the PTDF s for minor changes as seen here. Hence though PTDF s play a very important role in deciding the market outcome they are also not very sensitive to small changes in transmission capacity, neither do small changes in PTDF s lead to large changes in the market outcomes. Conclusion Tie line capacity has strong effect on the prices in a country. As was seen with installation of 2000MW on the DE NL border and BE FR border, the prices change is higher with installation of more capacity on the congested tie lines. However it was also noted that installation of additional 5000MW has very small additional benefit for the region. The tie line is not fully utilized and hence is not recommended. PTDFs play an important role in the market outcome from the FMC. However calculation of PTDFs depend only on the topology of the network. UCTE is composed of the highly meshed system of 24 European countries consisting of some 200,000 km of 400 and 220 kv lines. Installation of a new tie line between countries like NL and DE, have small impact on the topology of the system as a whole. Hence there is only a minor change on the PTDFs with installation of new tie line. Secondly, though the markets do depend on the PTDFs, the dependence is not as strong so as to get affected by change on PTDFs in its third decimal. Hence it Chapter: Results 85

110 86 Flow based Market Coupling is possible to be certain that the TSOs won t be able to guide the market results as much. Markets could still be efficient, though they might be owned by TSOs, as in case of NL. Next part of scenario analysis would focus on studying the effect of Carbon Market which has been evolving over the recent past. As already discusses in the development of model the cost sheet included carbon pricing into it. However it was decided in the model thus far not to consider them. In the next section carbon prices and its effect on the market results would be studied EFFECT OF CARBON MARKET The European Union Emission Trading Scheme (or EU ETS) is the largest multi national, greenhouse gas emissions trading scheme in the world and was created in conjunction with the Kyoto Protocol. It commenced operation in January 2005 with all 25 (now 27) member states of the European Union participating in it. It contains the world's only mandatory carbon trading program. The program caps the amount of carbon dioxide that can be emitted from large installations, such as power plants and carbon intensive factories and covers almost half of the EU's Carbon Dioxide emissions FIGURE 56 CONTRIBUTION OF ELECTRICITY AND HEAT GENERATION TOWRDS TOTAL CARBON EMISSIONS ( As shown above in the Figure 57 electricity and heat generation accounts for 24% of total emissions within EU (EEA 2006) and also the fact the prices of Carbon have been very volatile over the past two years since introduction of trading. The prices of Carbon credits since its introduction have been given below in Figure 57. The prices started at around 8 per tonne of emission, it increased within the next 18 months to a high of more than 30 per tonne and then suddenly dropped in April 2006, since then it has been tumbling down to below 1 recently. FIGURE 57 DEVELOPMENT OF CARBON PRICES IN EU, INCLUDING THE CRASH IN APRIL 2006 (SOURCE: POINT CARBON) Carbon prices add to the variable cost of generation of electricity based on fossil fuels. Considering the importance coal and natural gas play in the energy portfolios of the north west European countries it was

111 important to study the effect that carbon prices can have on the merit orders and hence the energy export and import. On basis of the values for carbon credits given four different scenarios were defined. 15 /ton of CO 2 20 /ton of CO 2 30 /ton of CO 2 50 /ton of CO 2 The details of the results are attached in the appendix A.19. Some of the interesting observations would be discussed below EFFECT OF CARBON PRICING ON MERIT ORDER The merit order did not change with increasing the prices from 0.00 till 20. Till then the merit order was same as the base case. Only when the prices for carbon rise to 30 there is a change in the merit order. CHP Gas becomes cheaper compared to Lignite coal and hard coal based power plants. At the extreme end of the spectrum with the prices rising up to 50 the Gas based plants become cheaper than the Lignite based power plants. This is mainly an issue for Germany as they have GW installed capacity based on Lignite. This also reflects in the electricity prices in Germany after introduction of carbon pricing. TABLE 18 IMPACT OF CARBON PRICING ON THE MERIT ORDER BASED ON FUEL AND ENERGY TECHNOLOGY TYPE (THE ITEMS MARKED WITH OLIVE ARE THE ONCE THAT MOVED) Environmental Preference Base Case Price of Carbon Emission ( /tonne) Merit Order H H H H H H W W W W W W N N N N N N CHP B HC HC HC CHP G CHP G O LC LC LC HC HC CHP G CHP G CHP G CHP G LC G G G G G G O HC O O O O LC LC CHP B CHP B CHP B CHP B CHP B Key Hydro Wind Nuclear CHP Natural Gas Hard Coal Condensation Lignite Condensation Natural Gas Combined Cycle Oil CHP Biomass H W N CHP G HC LC G O CHP B This is also reflected in the changing equation of the supply curve. The demand curve was assumed to be the same as before, to ensure that the results can be compared to each other. The demand curve changes are shown in Figure 58. Chapter: Results 87

112 88 Flow based Market Coupling

113 FIGURE 58 EFFECT OF CARBON PRICING ON THE MERIT ORDER WITHIN NORTH WEST EUROPEAN REGION The biggest change is observed within Netherlands and Germany. France has large installations of nuclear and is predominantly unaffected by the prices. It is interesting to see the operating point of the market as well the average demand for the north west European region. Chapter: Results 89

114 90 Flow based Market Coupling TABLE 19 AVERAGE DEMAND IN NORTH WEST EUROPEAN REGION (UCTE SYSTEM ADEQUACY REPORT ) Average Demand GW NL BE FR DE The average demand within France can be satisfied by its installed capacity of Hydro, Wind and Nuclear and is thus almost not affected by the carbon pricing. However the results of FMC could shed more light into the details. Prices Change (euros/mwh) FMC Base - FMC Scenario % Import-Export (-ve means import) % Utilization % -5.00% NL BE FR DE % % 50.00% 0.00% % % BE-FR BE-NL DE-FR DE-NL % % % - NL BE FR DE Price % % Demand Imported (FMC Base) % Demand Imported Net Utilization (FMC Base) Net Utilization CO 2 Price 15 per ton Prices Change (euros/mwh) FMC Base - FMC Scenario % Import-Export (-ve means import) % Utilization % % -5.00% % NL BE FR DE % 50.00% 0.00% % BE-FR BE-NL DE-FR DE-NL % % % - NL BE FR DE Price % % Demand Imported (FMC Base) % Demand Imported Net Utilization (FMC Base) Net Utilization CO 2 Price 20 per ton

115 Prices Change (euros/mwh) FMC Base - FMC Scenario % Import-Export (-ve means import) % Utilization % % % % -5.00% % NL BE FR DE 50.00% 0.00% % % BE-FR BE-NL DE-FR DE-NL % % - NL BE FR DE Price % % Demand Imported (FMC Base) % Demand Imported Net Utilization (FMC Base) Net Utilization CO 2 Price 30 per ton Prices Change (euros/mwh) FMC Base - FMC Scenario % 0.00% -5.00% Import-Export (-ve means import) NL BE FR DE % % 50.00% 0.00% % Utilization % % % % % BE-FR BE-NL DE-FR DE-NL - NL BE FR DE Price % % Demand Imported (FMC Base) % Demand Imported Net Utilization (FMC Base) Net Utilization CO 2 Price 50 per ton FIGURE 59 RESULTS FROM FMC AFTER CONSIDERING CARBON PRICING As expected, after implementation of Carbon pricing in the FMC model, the price of electricity in the markets would go up. Belgium sees the highest price change for prices increasing to 15 and 20 per ton, compared to the FMC base case with no carbon pricing. The second most affected country is Netherlands. With 15 per ton, there is an increase of around 4.50 per MWh in Netherlands, and an increase of 5 in Belgium. In general every country sees a price increase of around 4 5 per MWh. Still Germany is exporting and the congested tie lines are NL DE and BE FR. With the price of carbon going up to 20 per ton of CO 2 the pattern of results is similar. Though now the prices increase in the region is higher around 5.50 to The highest increase in again seen in Belgium Chapter: Results 91

116 92 Flow based Market Coupling and followed by Netherlands at It can be seen that because these countries depend strongly on the imports they see slightly higher increases in prices as compared to the exporters. Germany is still exporting. When the prices hit 30, the scenarios begin to change. Germany sees a very high increase in electricity price of around 1 per MWh, followed by Belgium at around 9 per MWh. The lowest price change is now seen in Netherlands at around 7 per ton. Though still Germany is exporting, mainly the reason for that is good interconnector that it has with Netherlands and higher flow capacity from DE > FR(5000MW), as compared to FR > DE (1700MW). Hence Germany has a good access to the high price areas (NL and BE, as compared to France. With further increase in price 50 per ton of CO 2, the picture looks much different. Now the German prices shoot up by 18 per MWh, followed by France at around 14. Netherlands sees a price increase of around 8 per MWh, which is the lowest among all countries. The detailed results for each case from every perspective can be checked in the appendix A.19 Conclusion It can be concluded from the above results that CO 2 pricing though has strong effects on regional price levels. Netherlands is amongst the least affected countries when it comes to price changes induced by carbon pricing. With lower prices (around 15) there was an increase of around 4, and with higher prices ( 50) there was a change of around 8. However other countries most importantly Belgium saw bigger price changes and Germany gets badly affected when the prices go high DECOUPLING OF GERMAN SYSTEM INTO NORTH AND SOUTH BLOCK Germany was required to be split up into north and south zones in order to study the effect that loop flows have on the north west European electricity system, that occur due to wind generation in North of Germany and consumption in the southern industrial demand area within Germany. The PowerWorld model for the UCTE was used for this purpose. Two new areas were defined out of the current single area. The division was made along the line shown in Figure 60. The new areas were then used to calculate new set of PTDF s for the region. Figure 61 shows the flows after splitting of Germany. These new PTDF values would be used to understand the effect that these loop flows have on the interconnector congestion in the North West European region.

117 FIGURE 60 THE DIVISIONOF GERMANY INTO NORTH AND SOUTH BLOCK FIGURE 61 FLOWS AFTER SPLITTING GERMANY INTO NORTH AND SOUTH BLOCKS Chapter: Results 93

118 94 Flow based Market Coupling EFFECT OF WIND GENERATION IN GERMANY (15,000 MW FROM N S) Current installation of wind generation in Germany is around 21 GW, it is expected to rise to 29 GW by the year 2010 ( Wind is highly intermittent and German laws enforce that all wind generation should be accepted into the grid. Figure 62 shows the share of wind power in the net generation within Germany. FIGURE 62 ANNUAL SHARE OF DAILY WIND POWER IN RESPECTIVE DAILY PEAK DEMAND ON E ON GRID (GERMANY) With the fast extension of wind capacities especially in Northern Germany congestion management has become a serious issue in the North Western European electricity grid. Due to power distribution through the entire European integrated network (UCTE grid) according to relative line impedances, Germany s neighbors to the North West, namely Benelux countries, are affected by unintended but inevitable cross border flows congesting their grids. With the intended expansion of offshore wind capacities in the German North Sea, this problem is bound to aggravate. TABLE 20 PTDF VALUES AFTER SPLITTING GERMANY INTO NORTH AND SOUTH BLOCK Interface % PTDF BE FR 9.4 BE NL 9.4 DE S FR 0.90 DE N NL 9.4 DE N DE S 77.4 As can be seen from the table above 9.4% of all electricity that is supposed to go from DE N to DE S actually takes its path through NL. Hence if 20,000 MW of electricity has to go from DE N to DE S during a storm in the northern part and operation of the wind generation, 9.4% of this would pass through the DE N NL tie line MW is the capacity, NTC value of this interface. Therefore 1880 MW of this 2500 MW would already be used up because of German wind energy. This would leave only 620 MW for making trade with other countries (in the direction DE > NL). Also there would be less capacity on the BE NL border and also on BE FR border and also on FR DE border. BE NL border has only a capacity of 1350 MW, it cannot transfer 1880 MW through. Hence BE NL interconnector would already be congested. Hence there is no possibility of making

119 any trade with any other country. The prices in the countries would stay equal to the prices that would occur without coupling of networks. Hence it is concluded that there is significant effect that loop flows from a large market (like Germany and France) could have on interconnector usage between smaller market like Netherlands and Belgium. Loop flows because of German wind power are enough to congest the interconnector within other countries. Next section would discuss the effect of FMC on loop flows because of wind FMC AND LOOP FLOWS The relatively large loop flows through the Netherlands and Belgium and France as shown in the previous discussion due to wind generation within Germany and the law within Germany to consider these flows as local clearly ignores the physics of the interconnected network. FMC in present state is not able to deal with the loop flows. FIGURE 63 THE DIFFERENCE BETWEEN TRANSIT AND LOOP FLOWS (GREEN ARROWS MARK THE TRANSIT FLOWS, GREY ARROWS MARK THE LOOP FLOWS) As can be seen from the Figure 63, the green arrows mark the transit flows that occur due to commercial exchange of electricity between Germany and France (from Germany to France).Please note that there would be direct flow over the DE FR interconnector as well, which is not show in the image above. These flows that occur in the interconnectors between DE NL, NL BE and BE FR because of commercial flow from DE >FR are termed as transit flows. These flows are all accounted for by using the PTDFs approach during market coupling. The loop flows on the other hand the flows that happen in the whole system due to flows which are assumed to be local. As can be seen in the picture above, the flow from Germany north to south would also utilize the networks in NL, BE and FR. However there was no commercial exchange, and hence it was not considered in the calculation during FMC. Hence FMC would not resolve the issues of loop flows from Germany. Also interesting is the fact that any other country in North West European region shares its border with every other country, unlike Germany; hence the issue of loop flows in case of other countries is much smaller. The only possible solution to this issue is either to consider Germany as two price areas, two independent markets. If that is the case, then the FMC would be between 5 market areas, and all the flows which are now considered local to Germany would be from Germany north to south. These flows can then be taken in the FMC algorithm. The markets would become more efficient and transparent in that situation. This is a suggestion and might not be acceptable to German policy makers. This however is an important issue to be discussed with the German TSOs and PXs in case there can be some other arrangement that can be worked out. Chapter: Results 95

120 96 Flow based Market Coupling 8 ACTOR ANALYSIS Till now the report focused on understanding FMC and its distributive effects on the involved countries. However, it is important at this point to realize the fact that FMC would have to be implemented in a multiactor environment many of whom might have conflicting interests. The last sections showed that FMC is indeed an economically efficient method for cross border congestion management it makes most optimal use of the available transmission capacity while respecting the physical limits of the lines. This section would focus on researching the actors that would be involved in the process. First step would be to look at the context in which the present system would be implemented i.e. very high level understanding of the market place. Consequently, the section would detail the various actors in the electricity market arena. After the listing is complete, their interests and positions would be listed. Based on their positions and needs a framework for analysis of assessing the acceptability of FMC would be outlined. Finally the chapter would end with detailing the FMC s effect on the positions of the involved actors. 8.1 GENERAL OVERVIEW The four countries in focus here are significantly different from other in terms of size, population and electricity demand and supply characteristics. The table below details some of the important features of the markets under consideration. The smaller countries NL and BE are also the importers. Net imports for NL were the highest, underlining the importance imports have for the country. France is the largest exporter. Germany consumes the highest amount of electricity amongst the four and also produces the highest amount per year. However the difference in generation between DE and FR are not huge. TABLE 21 GENERAL CHARACTERISTICS OF THE NORTH WEST EUROPEAN ELECTRICITY MARKET Area Population Installed Generation Annual Electricity Generation Growth Forecast sq km GW TWh % NL 41,526 16,570, BE 30,528 10,392, FR 643,427 63,713, DE 357,021 82,400, Annual Electricity Consumption Annual Electricity Generation Electricity Imports Electricity Exports TWh TWh TWh TWh NL BE FR

121 DE STRUCTURE OF THE ELECTRICITY MARKETS FIGURE 64 THE ELECTRICITY VALUE CHAIN The electricity value chain can be described under the following headers: 1. Generation: Refers to generation of electricity from coal, gas, nuclear, hydro or other sources 2. Electricity Trading: Generators sell energy via market places to sales companies or major resellers. At places where the generation companies are not yet unbundled some generation units go directly to the sales of the generationn company. The price paid by the sales units is determined in the market place. 3. Sales: Private persons and industries sign delivery agreements with the different actors at the sales stage. Prices are regulated via various kinds of agreements between sales companies and customers. 4. Transmission: The national grids are large high voltage networks that transport large volumes of electricity over long distances. National grids are for most part owned by state. 5. Distribution: Regional and local electricity networks transport electricity to the end customer. There are some large companies coexisting with regional and local network companies. The networks are operated by separate legal entities, regardless if they are owned by major power utilities, municipalities or other players. 6. Market Places: Electricity is traded on electricity exchanges or in the OTC market. Actors such as producers, resellers and very large industrial companies trade on the electricity exchanges. Some major actors are both producers and suppliers, and act as both buyer and seller on the electricity exchange. In the OTC market which is outside the exchange, actors trade directly with one another. Pricing in the OTC market often reflects expectations regarding the electricity exchange s wholesale price. The import export between national borders take place depending on difference in the wholesale prices between countries. 7. Regulators: Electricity markets have inherent issues of monopoly and economies of scale there is a requirement of monitoring and control from an external agency. Each country has an agency committed to making energy markets work as effectively as possible by implementing various regulatory instruments. This entails safeguarding access to networks, maintaining sufficient transparency (access to essential information) and protecting consumers against potential malpractices resulting from the (inherent) dominant position of providers. In the restructured environment, there are increasing market actor demands for better power system security and quality. On the other hand, market actors also demand an increase in the power system limits and transfer capacities. These demands work in opposite directions. Considering that the research deals with four Chapter: Actor Analysis 97

122 98 Flow based Market Coupling countries a detailed list of actors would be prepared. The next section would present the list of all the actors NL, BE, FR and DE. 8.3 LIST OF ACTORS TABLE 22 LIST OF ALL RELEVANT ACTORS FOR THE NORTH WEST EUROPEAN REGION Regulator NL Dienst uitvoering en toezicht Energie (DTe) BE Commission de Regulation de l Electricite et du Gas (CREG) Power Exchnage APX BelPex Generation Large Scale Generation: Electrabel SA Delta Electrabel E.ON Essent Nuon Trade Generation and Supply Generation and Supply APX Belpex Imports/Exports Imports/Exports Long Term Contracts Long Term Contracts Transmission Tennet ELIA Distribution Cogas Private companies: 1 Delta Municipality companies: 4 Eneco Mixed Inter municipality companies: 15 Essent Pure Inter municipality companies: 8 NRE Network Nuon ONS Rendo Westland Energie Retail, B2B Cogas Many Delta Eneco Energiebedrijf.com Essent Intergas NRE Network Nuon ONS Rendo Westland Energie

123 FR Regulator Commission de Régulation de l Electrité Bundesnetzagentur für Elektrizität, Gas, Telekommunikation, Post und Eisenbahnen (BNetzA) DE Power Exchnage Powernext European Energy Exchange (EEX) Generation Electricite de France Berliner Kraft und Licht AG Energie du Rhône E.ON Energie AG Société Nationale d Electricité et de Thermique ENBW Energie Versorgung Schwaben Société Hydroélectrique du Midi (SHEM) Hamburgische Elec werke AG Electrabel (Suez Group) Neckarwerke Suttgart AG (NWS) RWE Energie AG STEAG AG Vattenfal Europe AG Trade Generation and Supply Generation and Supply Powernext EEX Imports/Exports Imports/Exports Long Term Contracts Long Term Contracts Transmission Reseau de Transport Electricite (RTE) EnBW Transportnetze AG E.ON Netz GmbH RWE Transportnetz Strom GmbH Vattenfall Europe Transmission GmbH Distribution Electricité de France (EDF) More than 900 distribution companies Electricité de Strasbourg (ES) Gaz et Electricité de Grenoble (GEG) Usine d Electricité de Metz (UEM) Vialis (Colmar) Retail, B2B Energie du Rhône sales RWE PLUS Citiworks ENDESA France Electrabel (via SHEM) E.ON Sales & Trading CNR (Main shareholder Electrabel 49,97) EDF Trading SNET (Main share holder Endesa 65%) Electrabel Deutschland Total (via Gas and Power North Europe) EnBW Ges. für Stromhandel Vattenfall Europe Endesa Trading Enercity Trade EOS Avenis GETEC Energie MVV Energiehandel NUON Energy Trade & Wholesale RWE Trading Trianel European Energy Trading More details on the actors can be found in attached appendix A.20. It is true that most of these actors have different interests and goals. However it is possible to consider them under groups generators, consumers, traders, regulator, system operators and power exchanges. Their class interests are similar, though it would be in favor of their own citizens or industries. The next section would look at the interests of these groups. Chapter: Actor Analysis 99

124 100 Flow based Market Coupling 8.4 INTERESTS GENERATORS Generators wish to maximize their profits. It is possible to achieve in two ways by volume or by price. Price is usually decided by the market (assuming little or no market power with the generators), and hence cannot be changed at will by the generator. Hence most important measure is how much of the installed capacity is being utilized at any given instance. Another parameter that can be monitored is the producer surplus CONSUMERS Consumers can also be categorized further into domestic and industrial. However it can be said that all consumers want the price of electricity to be low. This is also reflected in the consumer surplus, but considering that perception also has a role to play in the consumer satisfaction it was decided to consider the change in price as the indicator for consumers preference towards a system REGULATOR Regulator is a government agency committed to making energy markets work as effectively as possible. This entails safeguarding access to networks, maintaining sufficient transparency (access to essential information) and protecting consumers against potential malpractices resulting from the (inherent) dominant position of providers. Consumer surplus is important from their perspective as this reflects how beneficial the energy markets are to the end consumer GOVERNMENT The government has an high level goal of ensuring goodwill of people. They have to listen to both producers and consumers at the same time. At times these two goals contradict each other. However it was decided to consider net welfare as a measure of acceptance for a market method. Another issue that government is concerned with is the correct investment signals. When a flow based transmission model is used for regional capacity allocation purposes, the market will choose the most economically efficient cross border trades by itself. The flow based method will reveal, in a transparent way, the location of the limiting constraint. This facilitates an efficient investment foundation for the transmission infrastructures in the entire region TRADERS From the market actor s point of view, it is desirable to be able to trade with any other actor in any system and to rapidly react to changed market conditions. In cases with limited transmission capacity, knowledge about the actual transfer capacity and consequences for own decisions are required. Such information should be simple enough to be understood without too detailed knowledge about the dependence and interactions between the transmission corridors TSO The TSOs define the available transfer capacity between regions before the market operators (MO) calculate the balance prices and exchange volumes. The given transfer capacities will therefore impact the output from the activities of the market participants. The TSOs are constantly under pressure to give as much transfer capacity as possible. Since the transfer limit calculations directly affect the market actors, the TSOs are to a larger extent requested to justify any limits. Different flow conditions may cause different system phenomena to be limiting even on the same set of lines. The different transfer limits will vary and for an overall system optimization the set of limits cannot be calculated accurately in advance. With heavy power transfer between neighboring control areas, the consequences of incidents are more wide spread. System wide information is needed for transfer limit calculation and information exchange between SOs are very important for the quality of the results.

125 8.4.7 POWER EXCHANGES The power exchanges play an increasing important role in the electricity markets post liberalization. True competition can only be introduces if all market participants trade energy via an implicit market. The main interest of PX is to maximize their profits by maximizing market participation and hence fostering liquidity. If the new system is likely to increase the inter country trade taking place via the power exchanges then it can be said that the PX would support it. The secondary aims of power exchanges are increasing competition by creating price transparency and implementing the European single electricity market. supporting the liberalization of the different European Electricity Systems dealing with the issue of international trading, with special emphasis on providing a market solution to the congestion problems. TABLE 23 INTERESTS OF THE ACTORS AND THE QUANTIFYING/QUALIFYING PARAMETER Actors Interest Importance Quantifying Parameter Generators Profits Minimization of cost of generation % change in producer surplus % of installed capacity utilized Consumers The electricity price % Change in electricity price Regulator For a more fair and efficient system To ensure fair operation of market % Change in Consumer surplus Government Welfare of the citizens % Change in Net Welfare Traders Profit To ensure that arbitrage possibilities do not exist Complexity of system Also they like complexity More risk, more opportunity to arbitrage and earn profits Power Exchanges Profits through maximization of traded volume Design of market algorithms Liquidity Facilitate the whole integration process TSO System Security Ensure reliability of system Security margin Method of calculation 8.5 POSITION OF ACTORS The TSO, regulators, government, traders and consumers have already been covered in detail in the sections on interests and market structure. The other interesting players that need to be further detiled are PXs and the strategic generation companies. Chapter: Actor Analysis 101

126 102 Flow based Market Coupling POWER EXCHANGES Currently there are 4 power exchanges in North West European region one in each country. The largest power exchange is in Germany (EEX). It handles both spot and futures market for electricity. Belpex is the newest one of the lot. The start of the day ahead trade of electricity on Belpex started from 21st of November The Powernext, Belpex and APX are participants in setting up market coupling between the three countries. Integration of EEX would be crucial for success of FMC. The figure below details the current trades and prices on the four exchanges. APX NL Spot Price Yearly Average ( /MWh):: 57.3 Electricity Volumes [spot] (TWh): 19 NL Belpex Spot Price Yearly Average ( /MWh): 46.0 Electricity Volumes [spot] (TWh): 1 BE DE EEX Spot Price Yearly Average ( /MWh): 50.8 Electricity Volumes [forward] (TWh): 1044 Electricity Volumes [spot] (TWh): 89 Powernext Spot Price Yearly Average ( /MWh): 47.2 Electricity Volumes [forward] (TWh): 83 Electricity Volumes [spot] (TWh): 30 FR FIGURE 65 PRESENT POSITION OF THE POWER EXCHANGES GENERATORS The North West electricity market represents 1,1 million GWh of electricity consumption, 42% of the EU 25 electricity market 8. There are many generation companies in operation in the region. The detailed list and statistics could be looked up in the previous section on the list of actors or the appendix A.20. For the regulator and the government of Netherlands it is important to understand who the possible monopolists would be in the regional market after integration. Hence The most important players in the market were identified and data was collected. TABLE 24 SHARE OF GENERAION COMPANIES IN THE TOTAL GENERATION WITHIN NORTH WEST EUROPEAN REGION Country Firms capacity % share in national market % share in four countries MWe Netherlands E.ON Energie AG % 0% Electrabel SA % 2% Essent Energy Production BV % 1% Nuon (Reliant) % 1% Belgium Electrabel SA % 6% 8

127 France Comp Nationale Du Rhone % 1% Electricite de France % 38% Germany Berliner Kraft und Licht AG % 0% E.ON Energie AG % 11% ENBW Energie Versorgung % 4% Hamburgische Elec werke AG % 2% Neckarwerke Suttgart AG (NWS) % 1% RWE Energie AG % 10% STEAG AG % 2% Vattenfal Europe AG % 4% FIGURE 66 MAIN STRATEGIC ACTORS FOR THE NORTY WEST EUROPEAN REGION (FROM VATTENFALL ANNUAL REPORT 2006) Looking at the above figures it is visible that EDF, E.ON, RWE, Electrabel and EnBW are the biggest players even within the regional markets. Hence is was important to look into the detail of what their present situation was within the region. The next section would detail the information on these companies. Essent was also considered in the analysis being the big local company to see how it compares with the bigger European giants TSOS Using the PTDF capacity model increases planning certainty for TSOs and creates the possibility of distributing cross border capacities following the submission of bids, taking into account the willingness to pay for power in the various zones and the load flows resulting from the market result. However at the same time they would be forced by the regulator and power exchanges to reduce the reliability margin. Hence they could be both for and against FMC. Chapter: Actor Analysis 103

128 104 Flow based Market Coupling 8.6 STRATEGIC GENERATORS TABLE 25 DETAILS ABOUT THE MOST STRATEGIC PLAYERS IN THE NORTH WEST EUROPEAN MARKET E.ON RWE EnBW Country Germany Germany Germany Listing info Listed (free float: 85%) Listed (free float: 67%) Listed (EDF owns 45.01%) Electricity sales 2005, 404 (of which, Europe 367) TWh Number of customers, millions Primary products Primary markets Strategies Electricity: 22 Gas: 8 Electricity: 20 Gas: 10 Electricity: 5 Gas: 0.4 Electricity, gas (upstream, downstream) Central and Eastern Europe, UK, Nordic countries, Italy To be the world's leading power and gas company Integrate and strengthen electricity and gas operations Expand in gas production Organic growth Focus on new markets, such as Russia and Italy Electricity, gas. (Water operations are being divested) Germany, UK, Central and Eastern Europe Focus on electricity and gas Focus on core markets Divestment of water operations in the USA and the UK Organic growth Electricity, gas, (heat, waste, water) Germany, Central and Eastern Europe Focus on core business Strengthen the German operations. Ambition to be number three in German energy market Advance positions in Central and Eastern Europe Focus on value creation EDF Essent Country France Netherlands Listing info Listed in 2005 (85% owned by French state) Unlisted Electricity sales 2005, TWh Number of customers, Electricity: 40 (Of which, Europe 37) Electricity: 2.5 Gas: 1.9 millions Primary products Electricity, gas Electricity, gas, heat Primary markets Strategies France, UK, Germany, Italy, Eastern Europe (Asia, USA and Africa) Strengthen positions in Western and Central Europe Divest non core businesses (e.g., Brazilian assets) Improve productivity and reduce costs Invest in gas assets in order to be able to offer customers both electricity and gas Netherlands, Germany, Belgium Aspiration to become a leading utility in northwestern Europe Safeguard the company's independence and handle demands on transmission grid undbundling. (In early 2007, plans were announced for a merger with the Dutch company Nuon) Look for collaboration with other companies in northwestern Europe, primarily in the Netherlands It would be important for DTE and the Dutch ministry of economic affairs to closely monitor the behaviour of EDF, E.ON, RWE, Electrabel and EnBW. The local company Essent is holding almost 20% of the local market though from a regional perspective it is a very small player.

129 8.7 QUESTIONS RAISED BY THE ACTORS Also during discussions many questions were raised by the involved parties. Which though have not been answered in the report are still worth mentioning here REGULATOR How would you calculate the transmission capacities, when 7 different TSO s are involved? What about the aspects of information exchange & Transparency? The sensitivity of the system is very important TSO How would it be able to work out one technical model when 7 TSO s with different standards on information have to collaborate? Would it lead to a more conservative estimate than the present margins? What information is secret and what is open? One solution is creation of a supranational body that controls the system. Going to one supra national TSO might jeopardize their existence TSOs always want to maximize the reserve to be safe. Negotiations may be get one sided if system is very sensitive to the information that TSO s have monopoly over. Wish to have more power on allocation of capacity. Should the importance of Power Exchange be more or less? Issues of uncertainty about the future POWER EXCHANGE Currently PXs are partly owned by the TSOs, would it be still feasible under FMC? Is there a need of a central auction office as suggested in OMC? If there is a central office, what happens to their existence? 8.8 RESULTS FROM FMC The final section of the actor analysis chapter would populate the interests matrix that was composed previously with results from the FMC base case. The FMC base case was already discussed in detail in the chapters on methodology and results. TABLE 26 ACTORS PERCEPTION OF FMC Actors Quantifying Parameter Netherlands Belgium France Germany Generators % change in producer surplus 29.18% 20.86% 3.52% 6.30% % of installed capacity utilized 62.06% 68.64% 69.43% 62.24% Consumers % Change in electricity price 15.85% 11.04% 1.74% 3.10% Regulator % Change in Consumer surplus 1.56% 1.58% 0.53% 0.83% Government % Change in Net Welfare 0.12% 0.09% 0.00% 0.01% Traders Complexity of system Power Exchanges Liquidity TSO Security margin Method of calculation As can be seen from the table, the Dutch consumers, regulators and government should be pro FMC. The generators would lose some business due to cheaper imports from France and Germany. Belgium should also be supportive of the idea as it has gained in almost every front. In France there was a small increase in the prices (1.74%) but it was compensated by additional income from exports. The net welfare gains in France remained almost zero. German generators would be supportive, and not the consumers though there would Chapter: Actor Analysis 105

130 106 Flow based Market Coupling be an microscopic positive change in the net welfare. The net welfare does not show such large change mainly because the gain to the consumers is compensated by loss to the producers and vice versa. At the same time it must be noted that FMC of ETSO Europex proposal, offer more security to the producers and consumers since the implicit auction mechanism will ensure congested lines to be fully utilized. On the other hand, explicit auctions of options of international exchange flows do not offer the same security to TSOs, since the owner of the options can decide at the last moment not to use it. Hence from this perspective FMC should be favored by all producers, consumers and TSOs. 8.9 CONCLUSIONS In the conclusion it can be said that the negotiation arena is very complicated with large number of actors. Multiple actors with varied interests and the system under consideration is complex hence it would be a problem to get everyone to agree on any issue. The most important consideration identified was that of strategic behavior from large generation companies. The large French and German generators might be able to have monopoly position in the North West European region. It is important to check the behavior. Other important aspect is of co operation between the 7 TSO s, each with own system codes and standards. The success of the implementation rests almost entirely on the success of the integration of the technical models as PTDFs play a very important role in the convergence and results of the market. Also the technical model is required to calculate F max and F ref the most important constraints governing the market outcome.

131 Section 4: Market Design Issues The final section of the report sums up the research by identifying and delineating the risks and issues associated with implementation of a new market design in a multi actor setting. The main focus would hinge around the institutional and informational aspects Chapter: Actor Analysis 107

132 108 Flow based Market Coupling

133 9 MARKET DESIGN ISSUES Although, ETSO and EuroPEX agreed on FMC as one way of cross border congestion management, the proposal provides a list of still outstanding issues, such as the development of a simplified transmission model and its consequences, the development of the coordinating algorithms, the definition of performance measure etc. There is need to dig further into what we do not know and what we need to know aspect of market design like FMC. Having answered some of the economical and technical questions in the previous chapters now the focus is shifted towards looking more at the issues that need to be addressed in case government decides to go ahead with FMC. 9.1 LEGAL AND ORGANIZATIONAL ISSUES Implementing any sound market based mechanism in coupling separate markets raises some crucial organizational and legal issues whose answers are necessary prerequisites 9 : Compatible legal framework must be enforced in each market to be coupled in order to ensure fair conditions to all players (11 th Florence Forum requirement for Switzerland) China walls must be installed in every remaining vertically integrated firm (especially with large German producers and French behemoth EDF) Non discriminatory behavior of TSOs and MOs(PXs) must be guaranteed and externally audited The discrepancies of the current transparency of each market to be coupled must be reduced To avoid the apparition of gaming opportunities for the better informed market actors, the market transparency must be levelised in the whole integrated market. For more refer to chapter on actor analysis. To improve the use of interconnectors and reveal market prices (or one common price, when the interconnectors are not congested) reflecting the actual situation of the electrical system, the publication of information has to be improved and harmonized as regards notably: Access grid data: Forecasted interconnection capacities, forecasted and real time load level Generation data: Forecasted availability of generation units, actual generation by market time unit 9.2 TECHNICAL ISSUES Implementing the ETSO/EUROPEX model raises some important technical issues that will take time to be fully solved: The insufficient coordination between TSOs The global approach model to determine the capacities requires a very strong operational coordination and a very detailed information exchange process between all the involved TSOs (technical systems actually exist but are they fully used?) The determination of the PTDFs matrix What is the optimal zonal segmentation of the regional network? The interaction between TSOs and PXs 9 Chapter: Market design issues 109

134 110 Flow based Market Coupling It is not yet demonstrated how it can work in order to guarantee simultaneously efficient trade and operational safety of the grid The determination and the operational working of the PXs clearing algorithm Harmonization of the respective time schedules for bidding, harmonization of the respective PX s legal statute, etc. If more than two PXs are involved, the complexity of the clearing algorithm is strongly increased (cf. Belpex project with 3 PX) The current unequal level of liquidity of the existing PXs In some regions, almost all the market players expressed their strong concerns about the potential impact on the wholesale PXs prices of allowing PXs to manage suddenly a large amount of interconnection capacities The most recent ETSO paper (Regional Flow based allocations State of play, March 2007, has outlined the following issues in its paper: 9.3 MARKET RELATED ISSUES TRANSPARENCY TOWARDS THE MARKET In an NTC based allocation mechanism, an Available Transmission Capacity (ATC) is given to the market, and the market actors send their bids for parts of this capacity to the allocation entity. In case of flow based allocation, there is no such thing as an ATC between two control areas. What is available is the maximum allowable flow on certain branches (F max ) and an estimate of the flow that is already present at those branches prior to the allocation (F ref ). Although evolving from an NTC based market coupling to flow based market coupling is rather smooth for market participants (they are still buying energy from and selling energy to their local power exchanges), this is a completely different situation when evolving from NTC based explicit auctions towards flow based explicit auctions. In fact, the main challenge of flow based allocation is to find the proper balance between safeguarding the network security on the one hand and facilitating the market, by providing a transparent allocation mechanism, on the other hand ECONOMIC SIGNALS TO MARKET PARTICIPANTS AND SHARING OF CONGESTION INCOME Market actors with commercial exchanges between two adjacent control areas with uncongested tie lines, can contribute to congestions somewhere else in the grid. One of the main advantages of flow based allocation in a regional setting, is that this effect is taken into account during the allocation phase. In this way, low priced bids between two control areas of which the interconnections are not congested have to compete with, amongst others, the high priced bids between two control areas with congested interconnections, according to their contribution to the congestion LIABILITIES OF TSOS AND POSITION OF INDIVIDUAL REGULATORY AUTHORITIES In principle (given the physical laws of the system) any commercial transaction will use capacity on each interconnection of the interconnected system. Virtually, one could imagine that if a certain TSO offers 0 MW (or an unrealistically low capacity) on some interconnections in order to cope with congestions in its Control Area or on its interconnections, all additional transactions in the whole region and all capacities would be blocked simultaneously. Appropriate revenue distribution methods among TSOs and proper political, regulatory, and TSO coordination should prevent such situations to occur. 9.4 TECHNICAL IMPLEMENTATION ISSUES REGIONAL BASE CASE To calculate the flow based parameters (PTDF, F max, F ref ) as accurately as possible, a common model of the grid, with predicted generation and load should be prepared. A part of the maximum allowed flow on the

135 critical branches is already used (i.e. prior to the allocation) by so called already occupied flows that can consist of the following components: natural flows (cross border physical flows that will always occur, even when there is no scheduled commercial exchanges between areas); flows that result from sources/sinks that are located in a single control area commercial exchanges resulting from firm nominations in previous auction rounds flows caused by exchanges between sources and sinks that are not located in the region that is participating in the flow based allocation flows caused by exchanges between areas where the source is located in the region that is participating in the flow based allocation whereas the sink is outside, and vice versa. The base case should be representative for the situation/day/time at which the actual allocation should take place. It is evident that there is a lot of uncertainty involved in creating such a base case. To quantify this uncertainty, a thorough testing should be applied PHASE SHIFTING TRANSFORMERS Some of the phase shifters installed at his moment, or to be installed in the (near) future, are applied in interconnections. Phase shifting transformers offer the possibility to control the active power flow through the interconnection in which it is installed. Naturally this has an impact on the flows in the vicinity of the phase shifting transformer too. Therefore, the flow based parameters (PTDF, F ref ) are influenced by the operation of phase shifters. In the 'classical' formulation of the flow based allocation, the allocated quantity is the only control variable to optimize the market value. A limited range of tap positions of phase shifters could be used as an additional control variable NETWORK REPRESENTATION Currently a control area is represented by a single node that is connected to neighboring areas by a single cross border transmission link. It is evident that this is only a rough approximation of the physical reality. The translation of the grid in the representation that will be used during the flow based allocation is a key issue for the TSOs. The grid representation must be such that the network security is guaranteed, even in the case of contingencies, but still results in a transparent mechanism for the market actors. Thorough testing should reveal for the TSOs whether a chosen grid representation in the flow based allocation is sufficient to safeguard the network security GENERATION REPRESENTATION The exact distribution of generation within an area is not known exactly ex ante. This means that TSOs or auction offices cannot know which power plant will be used to serve a requested capacity towards another area. This is one of the major sources of uncertainty when estimating and assessing future grid situations DEFINITION OF FLOW GATE CAPACITY (F MAX ) This issue is closely related with the network representation that has been chosen. In case that the network is represented as a single node that is connected to neighboring areas by a single cross border transmission link the maximum allowed flow between two price areas is referred to as border capacity (BC) or flow gate capacity. Such a border capacity is an aggregated value and has no direct relationship with the physical capacity of a transmission line. Hence computation of the F max is an issue. In present case the border capacity is the aggregated value of the cross border flows when maximum exchange between two neighboring, taking the n 1 security into account, is applied. In case that a more refined model, that allows multiple nodes per TSO or control area, is implemented, the F max on the critical branches interconnecting those nodes can be made explicit therefore adding more accuracy and transparency to the system. Chapter: Market design issues 111

136 112 Flow based Market Coupling TYPE OF BIDS In a market coupling environment, bids for the purchase or sale of energy refer to the market participant s price area. In an explicit auction environment, in case of bilateral bids, the market participant specifies both the source and the sink. In case of a flow based allocation mechanism, in a regional setting, it is even possible that market actors specify a source and sink in, electrically, non neighboring countries. 9.5 POSSIBLE SOLUTIONS The legal and organizational issues need to be resolved by dialogue between the involved countries. One possible solution could be formation of a supra national entity (similar to the one proposed by Germany s Open Market Coupling market design) that serves as auctioning office. Also to sort out the need of sharing of information and transparency between TSO s there can be two solutions: Setting up of a supra national entity. It would be an effective solution however may not be acceptable to the TSO s. A supra national entity would diminish their importance and might in the long run jeopardize their existence. However it is possible to implement this with clear cut discussion on duties and responsibilities at national and regional level. Technology: Information technology lowers transaction cost. Recent rapid technological advances should be able to settle the issue of complexity. In the past few years, rapid developments of advanced metering, two way communications and Internet based information technologies have clearly set the trend for lowering market transaction costs. Right technology solutions might be able to alleviate, if not completely solve the issue of information sharing and transparency. In conclusion it can be said that the shift from current system to the new system would not be easy, however considering the benefits it should be worth it. Further integration plans with UK and Norway would also be bolstered from this experience with integration with German market.

137 10 CONCLUSIONS What are the impediments to and implications of implementing Flow based Market Coupling as a Congestion Management Mechanism in the North West European Countries? Current research was undertaken to provide answer to the above mentioned question. The basic underlying mechanism of operation of FMC was studied using a neo classical economic approach. Once the operation was understood a model was created to study its impacts on national and regional level. Then model was used to understand the effects on national and regional levels for different what if scenarios. The research hinged around development of decision support model which would enable the policy maker, in case of present research the Ministry of Economic Affairs of The Netherlands, to understand the implications of FMC to Dutch electricity supply. A decision support model for evaluating the implications of implementing Flow based Market Coupling for alleviating the issue of interconnector congestion among European countries was created as the outcome of research efforts. Second part of the question was on understanding impediments to implementing FMC. This was studied using the actor analysis and literature survey on market design. The conclusions are divided into sub sections to discuss each part individually: 10.1 SYSTEM BEHAVIOR How sensitive is the system for the choice of a specific set of PTDFs? PTDFs play an important role in the market outcome from the FMC. However calculation of PTDFs depends only on the topology of the network. UCTE is composed of the highly meshed system of 24 European countries consisting of some 200,000 km of 400 and 220 kv lines. Installation of a new tie line between countries like NL and DE, have small impact on the topology of the system as a whole. Hence there is only a minor change on the PTDFs with installation of new tie line. Secondly, though the markets do depend on the PTDFs, the dependence is not as strong so as to get affected by change on PTDFs in its third decimal. Hence it is possible to be certain that the TSOs won t be able to guide the market results. Markets could still be efficient, though PXs might be owned by TSOs, as in case of NL. What is the effect of constraints on the system? Could it lead to a less optimal outcome with respect to optimizing market value? F max has an important role to play in the market coupling algorithm. Congestion of one line has impact on the whole region. DE NL tie line constraint also has impact on the welfare gains within Belgium and congestion on BE FR tie line has impact on welfare gain in NL. Though this issue was not represented in the simulations that well, mainly because of presence of one low price country next to a high priced country (DE for NL and FR for BE). However it becomes more and more important to consider regional perspective for planning further investment in transmission capacity (section 7.4.2). The system could lead to less optimal outcomes compared to the hypothetical case of implementing FMC in the current situation only with very extreme low values for tie line capacities. With the current values, the outcomes lead to benefit for all countries, in terms of welfare. The prices were lowered significantly in NL and BE because of imports. There was a marginal increase in prices in FR and DE markets because of exports. Chapter: Conclusions 113

138 114 Flow based Market Coupling Policy makers may want to reserve a minimum amount of capacity on certain flow gates with respect to the investments they made regarding those flowgates in the past or with respect to security of supply. It should be mentioned here that if countries reserve huge capacity on a certain flow gate this might prevent other countries from importing as well (refer A.11.2). While at the same time drop of prices in their own countries might not drop as much, as there is no netting of flows on a reserved capacity. So reserving capacity on interconnector is not a good idea if the markets are working well, it disables the markets from performing well. What will be the net benefit of future network investments on available interconnector capacity and the Dutch electricity price level? Tie line capacity has strong effect on the prices in a country. As was seen with installation of 2000MW on the DE NL border and BE FR border, the prices change is higher with installation of more capacity on the congested tie lines. However it was also noted that installation of additional 5000MW has very small additional benefit for the region. The tie line is not fully utilized and hence is not recommended. o What would be the effect on the price level within Netherlands? The prices drop till the capacity is increase to a little above 2000MW on the Dutch German border after which it does not have any significant effect anymore. o Would it lead to more transits and loop flows? Loop flows were not modeled implicitly in the model, so it is not possible to predict effect on them. It is not possible to resolve individual transit flows through the interconnectors, as the simulation only gives the net sum of all flows over the interconnector. However the net flows occurring over the interconnectors increase with increasing interconnector capacity hinting at higher transit flows. However it must be noted again that transit flows are not necessary eating up the capacity with FMC they also ease up the congestion if they are flowing in opposite direction to the congestion. o How would the actors perceive this change? The effect of net capacity is different of different actors. The generators in exporting countries like FR and DE would be for it, and so would be the consumers in importing countries. Consumers in FR and DE should have only marginal interest in the issue as the price change for them is minimal. The generators in NL and BE would be concerned because of dropping prices and hence they might not be dispatched SECURITY OF SUPPLY What is the risk that FMC would lead to a net reduction of the import capacity? Import capacity as concept does not exist in FMC. F max is defined for each tie line, but it is possible to the transactions to net over this line and lead to higher imports than what F max value is. It is not possible to answer this question on basis of the current research, as it requires also consideration of F ref.. The data is required to answer these questions further; it depends on the long term (monthly and yearly) contracts and consideration of other natural flows that occur in the system. Further work with a more detailed model and Monte Carlo simulation might be able to shed some light on this issue. (section 9.4.1) How may FMC interfere with security of supply in the Dutch electricity system?

139 Security can be defined in terms of availability and reliability. The reliability aspect cannot be answered from the present research it depends on information sharing between PXs and TSOs in North West European region. However it is possible to talk about the availability aspect. As seen from all simulation results, Netherlands becomes dependant on the foreign imports. Hence in the long term high priced local generators would either have to shut down or move their generation capacity to other countries FR and DE. However this risk of dependence can be mitigated by diluting dependence on one particular country. Interconnections with Norway and UK would mitigate the effect of market power from a single large exporting country WELFARE EFFECTS What could be the effect of such system on national welfare as the system optimizes regional welfare in terms of market value? It was seen in all the run cases that Net Welfare for each country goes up (except for marginal drop in France). However it must also be mentioned here that the objective function was minimization of total cost of generation and not net welfare. It is possible that results would change with a different objective function. Refer to section 7.2 Current Situation vs FMC. What is the effect of the presence of four national systems on the distribution of the welfare for each of the four systems? The effect on not welfare was not very pronounced, because net welfare was sum of increasing producer surplus and decreasing consumer surplus or vice versa. However there were substantial changes in producer and consumer surplus. The producers in exporting countries derived surplus gains from FMC and so did the consumers in importing countries compared to the current situation. How many incentives will FMC give to increase electricity production from renewable energy sources? This is an important question to be answered; till date renewable technologies are expensive when compared to conventional methods of electricity generation. Even with extremely high carbon prices ( 50 per ton which is extreme compared to the current carbon prices which are hovering at less than a euro per ton) most of demand was satisfied by conventional sources. Though it does lower the dependence on polluting technologies like Lignite based condensation plants, effect on dispatch of renewable sources is very marginal. Hydro is almost always dispatched, but it is not present within Netherlands. Wind is intermittent and cannot be described on this high level model. Bio mass based technologies are too expensive and would require support of government to be able to make profitable business IMPLEMENTATION ISSUES What is the effect on transparency? The FMC is based on PTDFs and the maximum allowable flow on certain branches (F max ) and an estimate of the flow that is already present at those branches prior to the allocation (F ref ). Although evolving from an NTC based market coupling to flow based market coupling is rather smooth for market participants (they are still buying energy from and selling energy to their local power exchanges), this is a completely different situation when evolving from NTC based explicit auctions towards flow based explicit auctions, especially between Germany and Netherlands, and also on Germany and France border. In fact, the main challenge of flow based allocation is to find the Chapter: Conclusions 115

140 116 Flow based Market Coupling proper balance between safeguarding the network security on the one hand and facilitating the market, by providing a transparent allocation mechanism, on the other hand. The PTDFs need to be calculated by close cooperation of involved TSOs. PTDFs are not very sensitive to changes of an order of magnitude 1000MWs, hence it is not possible to impact them significantly single handedly for any TSO. PTDF s play an important role in the market outcomes and hence should be closely guarded by the respective regulating authorities. From transparency point of view the authorities should request publication of the following parameters: a. Optimization Parameters: Publication of the optimization algorithm details namely the objective function, constraints and process for resolution of non convergence of the optimization process. b. Prior to trade publications: Publication of PTDF and border capacities for the day ahead trade c. Post trade publications: Publications of market outcomes from the day ahead market. Would some kind of European central decision making unit be required? (e.g. to decide about the distribution of capacity on the national borders). It is important to have higher degree of co operation between the involved actors. Hence it is a good option to form a central agency to take care of issues on the regional level. A central arena to address only the issue of FMC can be set up to be able to address the issues on national level. For decision making in complex networks it is important to have a guiding authority to direct the negotiations. Overwhelmed with the complexity of the problem relative to the current understanding and success of the Market Coupling, the actors might fall into defensive positions and might not coordinate well. There is no common hierarchy to define an authority to set overall direction and resolve differences in such situation. The first step of the agency would be to study in detail the impact and implications which could be better performed with access to real data. Subsequently follow up with the ideas to address each issue that pops up from the analysis like distribution of congestion rent, payment to the actors that make loss by partially taxing those who get the benefit. This way it would be possible to earn acceptance from all actors towards FMC. It can be seen that theoretically FMC is a better system at utilizing the system, and hence obstacles to its implementation need to be resolved through negotiation and collaboration a central agency solely devoted to this end would be an intelligent move to speed up the process in a neutral and comprehensive way. What about gaming opportunities because of strategic actors? There are large generation companies in France and Germany, which also own part of transmission lines. They own significant % of generation while considering regional generation in North West European region (namely Electricite de France, E.ON Energie AG, RWE Energie AG). These companies could abuse this market power in a coupled environment. It is important for the regulatory and competition agencies to observe their behavior. It is also imperative to decouple these companies from the transmission ownership to level the playing field.

141 11 LIMITATIONS AND RESEARCH RECOMMENDATIONS The final chapter would discuss the limitations of the current study and possible areas that can be researched further LIMITATIONS OF THE STUDY This study focused on the impact of FMC on the North West European considering the effect on electricity prices, welfare and use of installed generation and tie line capacity. It aspired to offer answers to the effect that FMC would have on the various actors and on national and regional level. It also aimed to understand the network of actors that would be affected by implementation of this system. It aimed to give study whether FMC would deliver what it promised and what the possible caveats may be. However there were issues of information. It is not possible to get detailed information on many aspects of the mode as the information is not public. Cost of generation data required for obtaining the supply curves is considered as classified, as their bidding behavior depends on it. Information of F max and F ref is not calculated or published yet. The PTDFs are also not available from the public source. Even the data on demand and supply that is available is highly volatile within every time frame be it within a day, varies from season to season and from country to country. Hence, during the deployment of this study required making several assumptions. Second issue was the available time at hand; in the short time available for the study it was only possible to cover certain aspects and not all. These assumptions led to several limitations, and have been mentioned within the model methodology section. The whole analysis was based on average behavior of the system. It was important as the general observations cannot be based on one set of demand and supply bids and offers. The results are variable and volatile using such analysis could not have provided the insights that were required. The model does not give exact results for the markets. The objective function that was used was minimization of generation cost and not maximization of net welfare though both aim towards a similar aim, i.e. maximization of consumer surplus. It is possible that real results may vary from those of the model. It is also important to note that there are two parameters that need to be considered before making decisions mean behavior and variance. Though mean behavior was studied in the report and it gives insights into operation of the markets, there were no discussions of the range or variance. Basing the policy on mean behavior is not recommended, as depicted in the quote below Consider the case of the statistician who drowns while fording a river that he calculates is, on average, three feet deep. If he were alive to tell the tale, he would expound on the flaw of averages, Sam Savage, Consulting professor at Stanford University Though behavior of system could be understood by assuming averages, it is also true that FMC can give very different solutions for different values as seen in the analysis. Hence there is a need to further understand these effects for all possible variations in of the variables that affect the market output RESEARCH RECOMMENDATIONS Since the scope of a project is always limited, several proposals can be made for the future continuation of this project. Research further with real data from the agencies. Dynamic simulation of the system to obtain the median and mean results from the markets. Chapter: limitations and Research recommendations 117

142 118 Flow based Market Coupling Study of the variance and volatility which occur in the results for variation in the input variables. Investment analysis using the cost data that is available to study how investment decisions on future generation capacity would be affected by FMC. A detailed actor analysis

143 A APPENDIX Chapter: Appendix 119

144 120 Flow based Market Coupling

145 A.1 COST OF ELECTRICITY GENERATION MODEL FIGURE 67 THE COMMON DATA FOR COST OF ELECTRICITY FIGURE 68 FIXED COST DATA FOR ELECTRICITY GENERATION

146 2 Flow based Market Coupling FIGURE 69 VARIABLE COST DATA FOR ELECTRICITY GENERATION ALONG WITH EMISSION COSTS

147 TU Delft Flow based Market Coupling 3 FIGURE 70 FORMULATION FOR THE CAPITAL COST CALCULATIONS

148 4 Flow based Market Coupling FIGURE 71 FORMULAITON OF THE VARIABLE AND TOTAL COST

149 A.2 DATA SOURCES AND ASSUMPTIONS FOR COST OF ELECTRICITY GENERATION MODEL TABLE 27. TOTAL COST OF GENERATION FOR DIFFERENT TECHNOLOGIES BASED ON VATTENFALL Total cents/kwh Advantages Disadvantages cost, (4,4 6,6) Hard coal: BLignite: Nuclear power Hydro power Coal condensing (hard coal, lignite) Good fuel availability No CO2 emissions in electricity generation Long construction time and high technological complexity results in high cost of capital Safety concerns Final storage of spent nuclear fuel is an unsolved issue in many countries High efficiency No emissions to air or water Water (and thus energy) can be stored Easy to regulate generation Highly dependent on water supply Major intrusion on nature that changes landscape Few exploitation objects High investment cost Good fuel availability from politically stable regions and an effective world market for hard coal. The world's coal reserves are expected to last at least 250 years Relatively long construction time High environmental impact, mainly through CO2 emissions. Lignite CO2 emissions are roughly 30% higher than for hard coal Cost example* Fixed cost, mainly cost of capital, cents/kwh Variable cost (mainly fuel, including CO 2 emission allowances, cents/kwh) Generation capacity: 1,600 MW Annual generation: 12 TWh (incl. tax on nuclear capacity of 0.4) 0.5 (incl. cost for final storage of 0.2) Generation capacity: Small scale: up to 10 MW (in EU) Large scale: up to 440 MW Normal generation: approx. 4,000 hrs/year incl. tax, operation and maintenance 0 (No value has been assigned to the so called water value) Generation capacity: 700 MW Annual generation Hard coal: 4.2 TWh Lignite: 5.25 TWh Hard coal: (incl. operation and maintenance) Lignite: (incl. operation and maintenance) Hard coal: Lignite: Total cost, cents/kwh Advantages Natural gas combined cycle Lower environmental impact and higher efficiency than coal and oil Low investment cost Combined heat and power from biofuels Low environmental impact from large, modern facilities CO2 neutral Wind power No emissions to the air or water

150 6 Flow based Market Coupling Disadvantages Concerns over fuel availability (the largest reserves are in politically unstable regions) High and unpredictable fuel cost (major price fluctuations) Environmental impact (emissions of CO 2, among other things) The market for biofuels is still undeveloped in many countries, and conflicts with other uses for the fuel can arise Particle emissions from poorly equipped facilities Higher need for balancing power due to the unpredictability of wind based electricity generation Stability of the electricity grid can be adversely affected with a high share of wind power in the system Requires subsidies Landscape aesthetics Cost example* Generation capacity: 400 MW Annual generation: 2.4 TWh Fixed cost, mainly cost of capital, cents/kwh Variable cost (mainly fuel, including CO 2 emission allowances, cents/kwh) Generation capacity: 30 MW electricity and 80 MW heat Annual generation: 0.14 TWh electricity and 0.36 TWh heat Generation capacity: 110 MW Annual generation: 0.33 TWh Pertains to an offshore wind power farm (including maintenance costs of approx. 1.7) Data Sources Cost of electricity technologies Nuclear EU POLICY AND CARBON EMISSION TRADING: Natural Gas Combined Cycle IMPLICATIONS FOR THE ENERGY MARKET Hard Coal Condensation Lignite Condensation Risto TARJANNE Wind Lappeenranta University of Technology Hydro Oil CHP Natural Gas Federal Energy Regulatory Commission, FERC Form 1, "Annual Report of Major Electric Utilities, Licensees and Others." Projected costs of generating electricity 2005 update book CHP Biomass Hoogwijk, M., D. van Vuuren, et al. (2007). "Exploring the impact on cost and electricity production of high penetration levels of intermittent electricity in OECD Europe and the USA, results for wind energy." Energy 32(8):

151 TU Delft Flow based Market Coupling 7 Installed Generation Capacity Own calculations based on UCTE SYSTEM ADEQUACY: FORECAST CO2 Emission Data Projected costs of generating electricity 2005 update book Comparative Data for Validation Assumptions Wietze et al.(2006) Cost 1. Fixed cost os all CHP plants is same 2. Variable cost from CHP plants is same as the variable cost of normal plants based on the same fuel type 3. Variable cost for CHP-G, CHP-O and O are based on proportional increase in the data for 2000 based on W. Lise et al, Energy Policy 34 (2006) 4. Oil prices have almost doubled since 2000, hence the data from W. Lise et al, Energy Policy (2006) was doubled to get the variable cost data for oil based plants 5. The variable cost range and the fixed cost range for oil are then back calculated from the Total Cost data and the variable cost data 6. CHP natural gas and CHP oil have similar fixed cost structure as the natural gas combined cycle. Capacity 1. Based on installed generation in each country on the 3rd Wednesday for the month of July in year The distribution into CHP is not mentioned in the UCTE adequacy database, hence it is based on the W. Lise et al, Energy Policy 34 (2006). Assuming that more growth has taken place in CHP generation since it was assumed that the data for conventional plants stayed at the same level as Using this assumption the remaining data could be calculated 3. Oil and mixed-oil and gas plants were assumed to have cost profile of the fuel oil based plants 4. The non-attributable sources of generation under the fossil fuel geeneraion were added to the Hard Cola Condensation plants capacity. 5. Renewable sources other than wind were assumed to be the CHP-Biomass based plants. 6. CHP-Oil data was used from W. Lise et al, Energy Policy 34 (2006) 7. All Gas based plants, i.e. combine, closed and open, were assumed to follow the cost structure of a combined cycle plant.

152 8 Flow based Market Coupling A.3 NATIONAL SUPPLY CURVES FIGURE 72 DATA POINTS FOR THE MERIT CURVE (SUPPLY CURVE) NL Supply Curve y = 3.183x R² = Price (euros/mwh) NL Supply Curve Generation Capacity (GW) Linear (NL Supply Curve) FIGURE 73 SUPPLY CURVE FOR NETHERLANDS

153 TU Delft Flow based Market Coupling 9 BE Supply Curve Price (euros/mwh) y = 4.217x R² = Generation Capacity (GW) BE Supply Curve Linear (BE Supply Curve) FIGURE 74 ELECTRICITY SUPPLY CURVE FOR BELGIUM FR Supply Curve Price (euros/mwh) y = 0.392x R² = Generation Capacity (GW) FR Supply Curve Linear (FR Supply Curve) FIGURE 75 ELECTRICITY SUPPLY CURVE FOR FRANCE

154 10 Flow based Market Coupling DE Supply Curve Price (euros/mwh) Generation Capacity (GW) DE Supply Curve Linear (DE Supply Curve) y = 0.452x R² = FIGURE 76 ELECTRICITY SUPPLY CURVE FOR GERMANY FIGURE 77 EQUATION OF THE SUPPLY CURVES (LINEAR REGRESSION ON THE MERIT ORDER CURVES)

155 TU Delft Flow based Market Coupling 11 A.4 AVERAGE HISTORICAL MARKET PRICE OF ELECTRICITY FIGURE 78 DISTRIBUTION OF MCP FOR NETHERLANDS APX, YEAR FIGURE 79 DISTRIBUTION OF MCP FOR BELGIUM BELPEX, YEAR

156 12 Flow based Market Coupling FIGURE 80 DISTRIBUTION OF MCP FOR FRANCE POWERNEXT, YEAR FIGURE 81 DISTRIBUTION OF MCP FOR GERMANY EEX, SEPTEMBER 06 JUN. 07

157 TU Delft Flow based Market Coupling 13 TABLE 28 CONSOLIDATED MCP DATA FOR CENTRE WEST EUROPEAN REGION FOR N Mean Std Dev Range Median Mode Q3 Q1 Samples /MWh /MWh /MWh /MWh /MWh /MWh NL BE FR DE

158 14 Flow based Market Coupling A.5 DEMAND WORKSHEET FROM MARKET MODULE FIGURE 82 DEMAND WORK SHEET VALUES FIGURE 83 DEMAND WORKSHEET FORMULATION

159 TU Delft Flow based Market Coupling 15 A.6 TIE LINE CAPACITIES Reference actual capacities of the tie lines without reliability margins (n 1) and (n 2) corrections Germany Netherlands Tie lines UN Capacity [MVA] Net Tie Line Capacity kv MVA MVA Conneforde Meeden Diele Meeden Gronau Hengelo 380 1,300 Gronau Hengelo 380 1,300 Rommerskirchen Maasbracht Siersdorf Maasbracht Capacity restricted by phase shifting transformer in Gronau Belgium Netherlands Tie lines UN Capacity [MVA] Net Tie Line Capacity kv MVA MVA Zandvliet Borssele Zandvliet Geertruidenberg Meerhout Maasbracht Lixhe Maasbracht ,566 Capacity restricted by phase shifting transformer in Borssele Maximum Allocable Capacity Long term contracts 3600 MW 1500 MW France Germany Tie lines UN Capacity Net Tie Line Capacity kv MVA MVA Vigy Uchtelfangen (RWE Net) Vigy Uchtelfangen (RWE Net) Muhlbach Eichstetten (EnBW) Vogelgrun Eichstetten (EnBW)

160 16 Flow based Market Coupling France Belgium Tie lines UN Capacity Net Tie Line Capacity kv MVA MVA Chooz Jamiolle Lonny Achene Avelin Avelgem Moulaine Aubange Limited by 220/150 kv transformer in Jamiolle, on 150 kv side TABLE 29 TIE LINE CAPACITY VALUES Reverse Capacity GW Forward Capacity GW BE FR BE NL DE FR DE NL Data Sources 1. Website of Tennet, RTE, RWE, Elia and TSO auction website

161 TU Delft Flow based Market Coupling 17 A.7 POWER TRANSFER DISTRIBUTION FACTORS FOR NORTH WEST EUROPEAN REGION PTDFs Sensitivity Matrix NL NL BE NL FR NL DE NL BE FR BE NL DE FR DE NL PTDF Matrix Commercial Exchange BE NL FR NL DE NL FR BE DE BE DE FR BE FR BE NL DE FR DE NL Physical Flow Data Sources Own calculation using PowerWorld using the model from (Qiong and Bialek 2005)

162 18 Flow based Market Coupling A.8 ELECTRICITY DAY AHEAD MARKET MODULE A.8.1 CONSTRAINTS Flow Based Coupling Optimization Constraints Constraint 1: Maximum flow through the tie line is less than the capacity Backward Capacity Net Calculated Flow Forward Capacity GW GW GW BE FR 1.7 <= 1.70 <= 2.5 BE NL 1.35 <= 0.77 <= 1.35 DE FR 1.7 <= 0.49 <= 5 DE NL 2.5 <= 2.50 <= 2.5 Constraint 2: Net Exchange should equal to zero. As there is no storage of electricity Net Exchange Average Demand Installed Generation Absolute Exchange GWh GW GW GW NL BE FR DE Net Exchange Net Exports/Imports A.8.2 OBJECTIVE FUNCTION Minimization of Overall Generation Costs

163 A.9 ELECTRICITY DAY AHEAD MARKET RESULTS Equilibrium without Market Coupling Price and Quantitiy Surplus GW /MWh /hr /hr /hr Q* P* Consumer Surplus Producer Surplus Net Welfare Cost of generation NL ,186, , ,608, , BE ,131, , ,465, , FR ,640, ,229, ,870, ,229, DE ,884, ,237, ,122, ,237, Net Welfare 10,067, ,224, Market Equilibrium under flow based constraints after Market Coupling Price and Quantitiy Surplus GW GW /hr /hr /hr Consumer Producer Net Welfare Qd Qs P' P' Welfare(FMC) Welfare(FMC) (FMC) Cost of generation NL ,253, , ,612, , BE ,234, , ,475, , FR ,624, ,246, ,870, ,246, DE ,816, ,307, ,124, ,307, Net Welfare (FMC) 10,082, ,153, Total Cost of Generation (FMC)

164 20 Flow based Market Coupling A.10 CAPACITY UTILIZATION RESULTS Reverse Capacity Forward Capacity Reverse Flow Forward Flow Reverse Unused Capacity Forward Unused Capacity GW GW GW GW GW GW Reverse Capacity Utilization Forward Capacity Utilization Net Utilization BE FR (1.70) % 0% % BE NL (0.77) (0.58) % 0% 56.93% DE FR (1.70) % 10% 9.76% DE NL (2.50) (0.00) 0% 100% %

165 TU Delft Flow based Market Coupling 21 A.11 COMPARISON OF MARKET COUPLING WITH FLOW BASED MARKET COUPLING A.11.1 BASE CASE FLOW BASED MARKET COUPLING

166 22 Flow based Market Coupling A.11.2 BASE CASE FLOW BASED MARKET COUPLING WITH NL RESERVING 1500MW ON DE NL BORDER

167 TU Delft Flow based Market Coupling 23 A.11.3 A MARKET COUPLING CONTRACT PATH: CLOCKWISE

168 24 Flow based Market Coupling A CONTRACT PATH: ANTI CLOCKWISE

169 TU Delft Flow based Market Coupling 25 A CONTRACT PATH: DISTRIBUTION (50 CLOCKWISE 50 ANTI CLOCKWISE)

170 26 Flow based Market Coupling A COMPARATIVE CAPACITY UTILIZATION RESULTS Base Case Flow based market coupling Contract path: Anti Clockwise Contract path: Clockwise Contract Path: Equal distribution

171 TU Delft Flow based Market Coupling 27 A.12 RESULT OF FMC AFTER INSTALLATION OF 5000MW OF COAL BASED POWER PLANTS IN NL

172 28 Flow based Market Coupling A.13 RESULT OF FMC AFTER INSTALLATION OF 10,000 MW OF COAL BASED POWER PLANTS IN NL

173 TU Delft Flow based Market Coupling 29 A.14 RESULT OF FMC AFTER INSTALLATION OF 2000 MW TRANSMISSION CAPACITY BETWEEN DE NL

174 30 Flow based Market Coupling A.15 INCREASE TIE LINE CAPACITY BETWEEN FR BE +2000MW

175 TU Delft Flow based Market Coupling 31 A.16 RESULT OF FMC AFTER INSTALLATION OF 5000 MW TRANSMISSION CAPACITY BETWEEN DE NL

176 32 Flow based Market Coupling A.17 RESULT OF FMC AFTER INSTALLATION OF 10,000 MW TRANSMISSION CAPACITY BETWEEN DE NL

177 TU Delft Flow based Market Coupling 33 A.18 RESULT OF FMC AFTER INSTALLATION OF 2000 MW TRANSMISSION CAPACITY BETWEEN DE NL WITH PTDF FROM 5000 MW CASE

178 34 Flow based Market Coupling A.19 EFFECT OF CARBON MARKET A.18.1 FMC RESULTS WITH EMISSION PRICE OF 15 /TON OF CO 2

179 TU Delft Flow based Market Coupling 35 A.18.2 FMC RESULTS WITH EMISSION PRICE OF 20 /TON OF CO 2

180 36 Flow based Market Coupling A.18.3 FMC RESULTS WITH EMISSION PRICE OF 30 /TON OF CO 2

181 TU Delft Flow based Market Coupling 37 A.18.4 FMC RESULTS WITH EMISSION PRICE OF 50 /TON OF CO 2

182 A.20 LIST OF ACTORS A.19.1 NETHERLANDS Transmission system operator o TenneT is since 1998 acting as TSO for the Netherlands. It is responsible for providing system services for the whole system and transmission services on the 220 and 280 kv level in the whole country. Transmission services on lower voltage levels are provided by some 20 network operators. In 2004 TenneT has taken over the transmission service of a regional network company and its aim is to become the transmission operator for all voltages above 110kV. During 2005 the operational activities of the regional control centre of the acquired network were integrated in the national control centre of TenneT. Main generators o The main generators in Netherlands market are E.ON Energie AG, Electrabel SA, Essent Energy Production BV, Nuon (Reliant). Distributors o There are a large number of suppliers of electrical energy. However, some merging has taken place among the distributors. The biggest suppliers on the moment are NUON, Essent and Eneco Power Exchange o The Amsterdam based APX (Power NL) is the first independent internet based power exchange in continental Europe. Since May 1999 when it began, APX s Day Ahead Market has provided its members with standardised products and flexible block orders to sell and purchase electricity in the Netherlands. APX also offers a continuous trading facility for Intraday and Strips Markets in the Netherlands. APX is the central counterparty in all electricity trades, retaining anonymity for all members throughout. APX also offers benchmark data and provides industry indices. Main traders & other players o A very important player is the Amsterdam Power Exchange, which operates a day ahead market and an adjustment market. Furthermore, some brokers are active and all producers and suppliers may act as trader, directly or via the APX. Regulator o Dienst uitvoering en toezicht Energie (DTe)

183 TU Delft Flow based Market Coupling 39 A.19.2 BELGIUM Transmission system operator: o The high voltage network is operated by ELIA, an independent public limited company founded in June 2001 in order to comply with the federal requirements of independence. ELIA former shareholders (Electrabel and SPE) reached an agreement with the Federal Government on the shareholder structure. Main generators o Electrabel (Net generation capacity: about MW in Belgium) o SPE (Gen. capacity: MW in Belgium) o RWE (from 2005 with 50% of 385 MW at BASF Antwerp) o Essent (from 2006 with 120 MW at INEOS Antwerp) Distributors o The former distribution companies (mainly inter municipal companies) have been appointed as operator of the distribution network for their respective territory. In order to comply with the regional legal requirements, they transferred their sales activities to another company, when their customers will become eligible. o Number of distribution system operators Private companies 1 Municipality companies 4 Mixed Inter municipality companies(a) 15 Pure Inter municipality companies 8 Power Exchange o Belpex (Belgian Power Exchange) is a day ahead spot exchange for the delivery and off take of electricity on the Belgian hub. Belpex is coupled with its two neighbours, APX in the Netherlands and Powernext in France. The partners of Belpex are APX (Dutch Power Exchange), Powernext (French Power Exchange), RTE (French Transmission System Operator) and TenneT (Dutch Transmission System Operator), under the leadership of Elia System Operator. Endowed with a 3 million euro equity capital, Belpex shares are owned initially by APX (10%), Elia (60%), Powernext (10%), RTE (10%) and TenneT (10%). Belpex is the Belgian Power Exchange for anonymous, cleared trading in day ahead electricity, providing the market with a transparent reference price. Main traders & other players o Many Regulator o Commission de Regulation de l Electricite et du Gas (CREG)

184 40 Flow based Market Coupling A.19.3 FRANCE TSO o o RTE The single French Transmission System Operator RTE is responsible for the operation and maintenance of the French EHV and HV networks between 400kV and 63kV. Main generators o The four main generating companies in France are: Electricité de France (EDF): Turnover 34400M, installed capacity 102GW, generation 482TWh, employees , Energie du Rhône (former CNR): Turnover 459 M, installed capacity 3GW, generation 12,5TWh, employees 1060, Société Nationale d Electricité et de Thermique (SNET): Turnover 810M, installed capacity 2,5GW, generation 8,7TWh, employees 1074, Société Hydroélectrique du Midi (SHEM): Turnover 75M, installed capacity 0,8GW, generation 2 TWh, employees 220, Number of distributors o Based on the French Act of 15 June 1906, the public distribution of electricity has been placed under the responsibility of local territorial administrations (councils, regions, or public joint ventures). The management of these local public utilities has been either given to a public operator (today EDF for 93% of the local councils) or are managed directly by the local councils through communal subsidiaries, which may work together within syndicates. o o Electricité de France (EDF) Turnover electricity 34400M, sales in distribution ~316 TWh, employees [80000 in the distribution sector], customers Beside EDF, there are in France 22 municipal distribution companies. The 4 main distributors are: Electricité de Strasbourg (ES) Turnover electricity ~415M, sales electricity ~6 TWh, employees 1363, customers living in 376 communal entities Gaz et Electricité de Grenoble (GEG) Turnover electricity 59M, sales electricity 8 TWh, employees 448 Usine d Electricité de Metz (UEM) Turnover electricity 135M, sales electricity ~1,7TWh, employees 500, customers , Vialis (Colmar) Turnover electricity 53M, sales electricity 0,4TWh, employees 259, customers 36650, Power Exchange o The introduction of a power exchange (Powernext) in France is a direct response to the opening up of the European electricity markets. The 1996 European Directive, applied into the French Law on the 10th of February 2000, create an opportunity to launch a feasibility study of the French electricity organised market for partners including BNP Paribas, ELIA (Belgium Transmission System Operator), EDF, Electrabel, RTE (French Transmission System Operator), Société Générale and TotalFinaElf under the leadership of Euronext. The conclusions of the one year study led to the incorporation of Powernext SA on 30/07/01.

185 TU Delft Flow based Market Coupling 41 Main traders & other players (exchanges etc.) o The other actors selling to eligible customers on the French market are: Energie du Rhône sales on the French market outside EDF 1,5TWh RWE PLUS sales on the French market 1TWh ENDESA France sales on the French market 0,6TWh Electrabel (via SHEM) sales on the French Market 2TWh CNR (Main shareholder Electrabel 49,97%) sales on the French Market 12,5TWh SNET (Main share holder Endesa 65%) sales on the French market 8,7TWh Total (via Gas and Power North Europe) sales on the French market 9,5TWh HEW (now Vattenfall Europe) sales on the French market 3TWh Regulator o Commission de Régulation de l Electrité

186 42 Flow based Market Coupling A.19.4 GERMANY Transmission system operator o 4 German Transmission System Operators bear responsibility for their control areas and the German system o (EnBW Transportnetz AG o E.ON Netz GmbH o RWE Transportnetz Strom GmbH o Vattenfall Europe Transmission GmbH Main generators o Main generating companies are (A growing number of IPPs enter into the German market) RWE E.ON Vattenfall EnBW Distributors o In Germany, there exist more than 900 Distribution System Operators. Power Exchange o EEX o The European Energy Exchange AG (EEX) currently operates spot and derivatives markets in power, CO2 emission allowances, coal and gas. It is rapidly expanding its range of services with the aim of establishing EEX as Europe s leading, multi product energy exchange. EEX is a product of the recent liberalisation in energy markets and is already Europe s biggest energy exchange in terms of both membership, with over 170 trading members from 19 countries, and growth in trading volume. It is fully regulated and can already claim to be truly pan European, with the majority of its members based outside Germany. Traders confidence in EEX and its clearing system is underlined by the rapid growth in the volume of open interest in power futures more than 300 Terawatt hours in December 2006, valued at around 17bn. Main traders & other players o Citiworks, E.ON Sales & Trading, EDF Trading, Electrabel Deutschland, EnBW Ges. für Stromhandel, Endesa Trading, Enercity Trade, EOS Avenis, GETEC Energie, MVV Energiehandel, NUON Energy Trade & Wholesale, RWE Trading, Trianel European Energy Trading, Vattenfall Trading Services Regulator o Bundesnetzagentur für Elektrizität, Gas, Telekommunikation, Post und Eisenbahnen (BNetzA)

187 TU Delft Flow based Market Coupling 43 BIBLIOGRAPHY Averch, H. and L. L. Johnson (1962). "Behavior of the firm under regulatory constraint." The American Economic Review 52: Bhattacharya, K., M. H. J. Bollen, et al. (2001). Operation of restructured power systems. Boston, Kluwer Academic Publishers. Borenstein, S. (2001). The trouble with electricity markets (and some solutions), University of California at Berkeley. Budhraja, V. (2003). "Harmonizing Electricity Markets with the Physics of Electricity." The Electricity Journal 16(3): Chao, H. P., S. Peck, et al. (2000). "Flow gate Transmission Rights and Congestion Management." The Electricity Journal 13(8): Christie, R. D., B. F. Wollenberg, et al. (February 2000). Transmission management in the deregulated environment. Proceedings of the IEEE. Congestion Management Guidelines (2006). amending the Annex to Regulation (EC) No 1228/2003 on conditions for access to the network for cross border exchanges in electricity. Conscentec and Frontier Economics (June 2004) Analysis of Cross Border Congestion Management Methods for the EU Internal Electricity Market. Volume, 138 DOI: Crampes, C. and J. J. Laffont (2001). "Transport pricing in the electricity industry." Oxford Review of Economic Policy 17(3): Crew, M. A. and P. R. Kleindorfer (1985). "Governance Structures for Natural Monopoly." Journal of Behavioral Economics 14(0): Day, C. J., B. F. Hobbs, et al. (2002). "Oligopolistic Competition in Power Networks: A Conjectured Supply Function Approach." IEEE Trans. Power Systems 17(3): De Jong, H. M. and R. A. Hakvoort (2007). "Congestion Management in Europe: Taking the Next Step. 9th IAEE European Energy Conference Energy Markets and Sustainability in a Larger Europe, June 10 12, Florence, Italy. De Vries, L. J. and R. A. Hakvoort (2000b). "An Economic Assessment of Congestion Management Methods for Electricity Transmission Networks." Journal of Network Industries 3(4): Directive 96/92/EC Directive of the European Parliament and of the Council of 19 December 1996 concerning common rules for the internal market in electricity., Official Journal of the European Union. L 27:

188 44 Flow based Market Coupling Directive 2003/54/EC European Parliament and of the Council of 26 June 2003 concerning common rules for the internal market in electricity and repealing Directive 96/92, (OJ 2003 L 176/37). EEA (2006). Greenhouse gas emission trends and projections in Europe Copenhagen, European Environment Agency. ENERGY SECTOR INQUIRY (2006). EUROPEAN COMMISSION, Competition DG; Energy, Basic Industries, Chemicals and Pharmaceuticals; Energy, Water. ETSO (May 2006). An Overview of Current Cross Border Congestion Management Methods in Europe: 31. ETSO and EuroPEX JWG (2004). Flow based Market Coupling A joint ETSO EuroPEX Proposal for Cross Border Congestion Management and Integration of Electricity Markets in Europe. Gheorghe, A. V., M. Masera, et al. (2006). Critical Infrastructures at Risk. Dordrecht, Springer. Glachant, J. M. and V. Pignon (2002). Nordic electricity congestion's arrangement as a model for Europe : physical constraints or operators' opportunism? [Cambridge, Mass.], Massachusetts Institute of Technology, Center for Energy and Environmental Policy Research. Harvey, S. M., W. W. Hogan, et al. (1996) Transmission Capacity Reservations and Transmission Congestion Contracts. Volume, DOI: Haubrich, H. J. and W. Fritz (1999). Study on Cross Border Electricity Transmision Tariffs.. Aachen, Germany. Hogan, W. W. (1992). "Contract networks for electric power transmission." Journal of Regulatory Economics 4: Hunt, S. (2002). Making competition work in electricity. New York, John Wiley & Sons, Inc. Kirschen, D. S. and G. Strbac (2004). Fundamentals of power system economics London, Wiley, 2004 Knops, H. P. A., L. J. De Vries, et al. (2001). "Congestion management in the European electricity system: an evaluation of the alternatives." Journal of Network Industries 2(3 4): Lise, W., V. Linderhof, et al. (2006). "A game theoretic model of the Northwestern European electricity market market power and the environment." Energy Policy 34(15): OECD "Projected Costs of Generating Electricity 2005 Update ": 232. PJM (2006) Cost Development Guidelines. PJM Manual Volume, DOI: Qiong, Z. and J. W. Bialek (2005). "Approximate model of European interconnected system as a benchmark system to study effects of cross border trades." Power Systems, IEEE Transactions on 20(2): Regulation (EC) No. 1228/2003 (2003). "of the European Parliament and of the Council of 26 June 2003 on conditions for access to the network for cross border exchanges of electricity on conditions for access to the network for cross border exchanges in electricity." THE EUROPEAN PARLIAMENT AND THE COUNCIL OF THE EUROPEAN UNION L 176/1. Samuelson, P. A. and W. D. Nordhaus (2005). Economics. New Delhi, Tata McGraw Hill. Stoft, S. (2002). Power System Economics: Designing Markets for Electricity Piscataway (NJ), IEEE Press. UCTE (2007). UCTE System Adequacy Forecast Brussels, UCTE.

189 TU Delft Flow based Market Coupling 45 Vattenfall (2006). ANNUAL REPORT Stockholm: 122. Vries, L. J. d. (2004). Securing the public interest in electricity generation markets : the myths of the invisible hand and the copper plate Faculty of Technology Policy and Management. Delft, Delft University of Technology. Doctorate: 353. Yan, H. H. (1999). PTDF and TLR from a power marketer's perspective. Zhou, Q. and J. W. Bialek (2005). "Approximate Model of European Interconnected System as a Benchmark System to Study Effects of Cross Border Trades." IEEE TRANSACTIONS ON POWER SYSTEMS5 20(2): 7.

190 46 Flow based Market Coupling WEBSITES general reading on subjects of interest net.org Updates on TSOs and policy Directives and regulations EuroPEX information General information on latest research on electricity markets Nord pool market; /Special_topics/Open_Market_Coupling_32p.html Economic and legal analysis of congestion management methods in EU. reports/vf_com/2006/filter.asp?filename=page_013.html Vattenfall's Annual Report 2006 What does new electricity generation cost? Vattenfall 2005 The European electricity market EEDRB Netherlands Installed Capacity of Electrical Plants AL EUROPA Eurostat Environment and energy [ :: Home [ :: Publications : Library :

191 TU Delft Flow based Market Coupling 47 APX Group 31&cld2=132&cld3= 1&graph=Day&graphtype=spot Powernext.fr EUROPA DG Competition Dr François Boisseleau thesis on The role of power exchanges for the creation of a single European electricity market: market design and market regulation