Costly Remedial Actions Coordination

(1)

IN

DEGREE PROJECT ELECTRICAL ENGINEERING, SECOND CYCLE, 30 CREDITS

,

STOCKHOLM SWEDEN 2019

Costly Remedial Actions

Coordination

QUENTIN WILLAIME

KTH ROYAL INSTITUTE OF TECHNOLOGY

(2)

KTH Royal Institute of Technology

Master Thesis

Costly Remedial Actions Coordination

Author:

Quentin Willaime

Supervisor:

Dr. Mohammad Reza Hesamzadeh

Examiner:

Dr. Mohammad Reza Hesamzadeh

A thesis submitted in fulfilment of the requirements for the degree of Master Thesis

in the

Electricity Market Research Group

Department of Electric Power and Energy Systems

(3)

Declaration of Authorship

I, Quentin Willaime, declare that this thesis titled, ’Costly Remedial Actions Coordi-nation’ and the work presented in it are my own. I confirm that:

This work was done wholly or mainly while in candidature for a research degree at this University.

Where any part of this thesis has previously been submitted for a degree or any other qualification at this University or any other institution, this has been clearly stated.

Where I have consulted the published work of others, this is always clearly at-tributed.

Where I have quoted from the work of others, the source is always given. With the exception of such quotations, this thesis is entirely my own work.

I have acknowledged all main sources of help.

Where the thesis is based on work done by myself jointly with others, I have made clear exactly what was done by others and what I have contributed myself.

Signed:

Date:

i

(4)

Abstract

The energy sector is evolving and the role of the Transmission System Operator (TSO) must adapt to these new changes. In fact, the energy transition modifies both the nature of the production sources by integrating intermittent energies and the classical structure of the grid, going from a centralized topology to a distributed topology.

In this particular context, the TSO has to balance production and demand while en-suring that the system is operated in a safe and optimal manner. These new challenges can not be achieved by independently, it is necessary to introduce a strong coordination and collaboration between European members. This desire to co-ordinate the operating methods of the system on a European scale as well as to harmonize the market in or-der to favor exchanges, is pushed by the European Commission in the form of Network Codes. In order to maximize the exchanges between bidding zones, it is necessary to calculate the available exchange capacity and assess potential risks for the grid by con-ducting joint analyzes between neighboring countries. If there is a risk of congestion on a network element because of an excessive exchange between TSO, it is possible to manage this overload by activating measures called Remedial Actions.The legal framework re-quires TSOs to establish a detailed coordination methodology for Capacity Calculation, Common Security Analyzes and activation of Remedial Actions.

In addition, the activation of RA often leads to a significant cost for the TSO concerned. These congestion management costs may arise especially in particular situations of high demand, poorly estimated production or modified topology due to a fault. Given that these risks of congestions are sometimes induced by neighboring bidding zones, it is necessary to also foresee a methodology for sharing these costs so as not to penalize the TSOs polluted by their neighbors.

(5)

Abstrakt

(6)

Acknowledgements

I would like to acknowledge my examiner at KTH Mohammad REZA HESAMZADEH for answering my questions and supervising my thesis.

I would like to express my gratitude to my supervisor at RTE Valentine WOILLEZ for her support, her valuable guidance and her constructive suggestions during the whole duration of the project.

I would like to thank the whole operation direction team for their integration within the group, working with them was a real pleasure.

I am also grateful to my colleagues Guénolé AUMONT, Benjamin BAA-PUYOULET, Ryan IGHILAHRIZ, Emmanuel JAMES and Élise REN who worked with me on this thesis subject. I sincerely thank them for their time, advice and support.

I would like also to thank the Research and Development Department, and particularly the team of Luc DI GALLO and S´ebastien MURGEY for presenting me their projects and for allowing me to approach some issues from another angle.

Finally, I am thankful to all the employees I have met at RTE for their welcome and their help.

(7)

List of Figures

1.1 Counter Trading . . . 8 1.2 Cross-border Redispatching . . . 9 1.3 Internal Redispatching . . . 10 1.4 External Redispatching . . . 10 2.1 Internal . . . 24 2.2 Loop . . . 25

2.3 Import and Export . . . 25

2.4 Transit. . . 25

2.5 Power Flows Coloring calculation process . . . 28

2.6 Balancing of zones example . . . 29

2.7 Balancing of zones example : Balanced Model . . . 30

2.8 Balancing of zones example : Compensating Model . . . 31

2.9 Balancing of zones example : decomposition of the grid model. . . 31

2.10 Example of Power Flow Decomposition . . . 45

2.11 Vertical Shift Netting . . . 46

2.12 Total Power Flows before and after Proportional Netting . . . 47

2.13 Final Power Flows Decomposition after Proportional Netting . . . 47

2.14 Total Power Flows before and after Proportional Netting per Category . . 48

2.15 Final Power Flows Decomposition after Proportional Netting per Category 49 2.16 Final Power Flows after Equal Netting per Category . . . 50

2.17 Principal prioritization scenarios . . . 54

2.18 Situation before activating Remedial Actions . . . 55

2.19 Situation after activating Remedial Actions . . . 56

2.20 Individual Utility Function . . . 64

2.21 IVBM calculation process . . . 71

3.1 German TSOs. . . 73

3.2 2017 German costs of Remedial Actions . . . 75

3.3 2017 Total German costs of Remedial Actions per TSO . . . 76

3.4 Most congested lines in Germany . . . 80

3.5 Power Flow Decompostion in the 4 different areas . . . 81

3.6 Power Flows in Zone 3 . . . 82

3.7 Prioritization scenarios . . . 84

3.8 Penalized Flows in the 4 different scenarios . . . 86

3.9 Estimation of the German Remedial Actions costs sharing . . . 87

3.10 Estimation of the German Remedial Actions costs sharing per bidding zone 88 3.11 Influence of the Threshold value for scenarios 1 and 2 . . . 89

(11)

List of Figures viii

3.12 Influence of the Threshold value for scenarios 3 and 4 . . . 90 B.1 Day-ahead Capacity Calculation, Coordinated Security Analysis and

Re-medial Actions activation . . . 103 B.2 Intra-day Capacity Calculation, Coordinated Security Analysis and

Re-medial Actions activation . . . 104 B.3 Close to real time Capacity Calculation, Coordinated Security Analysis

(12)

List of Tables

1.1 Different types of Redispatching . . . 11 2.1 Flow Types . . . 24 2.2 Power Flow identification with PFP on a line located in A . . . 42 2.3 Power Flow identification with PFP on a tie-line between A and B . . . . 43 2.4 Example of Power Flow Decomposition . . . 45 3.1 Country Acronyms . . . 79

(13)

Abbreviations

ACER Agency for the Cooperation of Energy Regulators ANE Aggregated Netted External

C2RT Close To Real Time

CACM Capacity Allocation and Congestion Management CC Capacity Calculation

CCR Capacity Calculation Region CGM Common Grid Model

CNEC Critical Network Element and Contingency CNE Critical Network Element

CORE Central Western and Central Eastern Europe Capacity Calculation Region CRAC Contingency list, Remedial action and Additional Constraints

CSA Coordinated Security Analysis CT Countertrading

D2CF Two-Days ahead Congestion Forecast DACF Day Ahead Congestion Forecast ECT External Commercial Trade

ENTSOE European Network of Transmission System Operators for Electricity FAP Fast Activation Process

FLD Full Line Decomposition GSK Generation Shift Key ICT Iternal Commercial Trade ID Intra-Day

IGM Individual Grid Model

IOBM Individual-Optimization Based Mapping IVBM Individual Volume Based Mapping

(14)

Abbreviations xi

LSK Load Shift Key

minRAM minimal Remaining Available Margin NRA National Regulatory Agency

PEX Power Exchange PFC Power Flow Coloring PFP Power Flow Partitionning PST Phase Shift Transformer

PTDF Power Transfer Distribution Factor RAO Remedial Action Optimizer

RA Remedial Action RD Redispatching

RSC Regional Security Coordinator RTE Réseau de Transport de l’ Électricité RT Real Time

SA Security Analysis

(15)

Chapter 1

Introduction

1.1 Background

1.1.1 Global presentation of the power grid organization

All around the world, the electricity supply chain has been built on the same model, which is constituted of three different parts: production, transmission and distribution. These distinct parts correspond to characteristic voltage ranges and particular actors. At first, the electric power is generated in the production section by the power plants often located far from the cities and the consumption areas. The power sources of these power plants are various: coal, oil, gas, nuclear, hydraulic, solar, wind etc The power generators output voltage is between 11kV and 24kV. Before injecting this power in the transmission system, the voltage is raised up to a range of 63kV-400kV in order to reduce the losses. Indeed, reducing the Ohmic heating in long distance overhead-lines decreases the losses. Finally, the voltage will be reduced before the distribution system in order to supply the consumption demands. This classic radial description is a good simple model to understand the global organization of the power grid.

For both the production and the distribution sides, the electricity market is nowadays open to competition between actors. Indeed, various companies now supply the electric production and new distribution offers continue to appear on the market. However, the Transmission System has some particularities that prevent to expose it to competition. In fact, the importance in terms of infrastructure, territory planning and costs of the

(16)

Chapter 1. Introduction 2

Transmission System is such that a monopoly is the most efficient solution. Electric transmission networks are known to be a ”natural monopoly” as it is neither sustainable nor desirable to have different competing system operators in one area. [1] In most of the European countries, the Transmission System is managed by only one Transmission System Operator (TSO).

In France, the Transmission System is managed by RTE (R´eseau de Transport de l’Electricit´e) which operates all over the metropolitan territory. The French power grid is composed of more than 100 000 km of electric lines and 2 700 substations [2], and has the particularity to be strongly meshed. This particularity ensure a good security of the system in terms of reliability, stability and selectivity. One of the main goal of a TSO is to maintain the balance between the production and the consumption at any moment. The means of production dispatching should not only guarantee to respond to the demand, this dispatching should also be the most economic and ecological as possi-ble. Managing to establish the optimal dispatch given all the underlying constraints is a real challenge.

1.1.2 Evolution of the electricity power sector

(17)

accentuated particularly with the inevitable development of electric vehicles. All over the world the consumption of electricity is increasing faster than any other kind of energy. Moreover electricity is involved in almost every aspect of the economic activity. Indeed digitization of activities and economies is built on the reliability of electric supply and just small interruptions can lead to enormous economic losses.

Moreover, not only the nature of the production and the consumption evolves but also the topology of the network and the role of the stakeholders. Indeed, the historic radial topology of the electric power grid is challenged by the integration of distributed energy power plants. Nowadays it is more and more common to install solar panels on one’s own house or to construct wind farms close to residential homes. The power plants are not necessarily installed far away from the nodes of consumption and can be directly integrated on the distribution system. The integration of electric vehicles represents a considerable capacity for electricity storage, since the consumer’s behavior can have a huge impact on the power grid whether he is charging or discharging his car. The average consumer can be considered as an actor of the production of electricity and its impact should be taken into account of the role of the TSO.

(18)

1.1.3 European role of the TSOs

In each European country, the respective national transmission network is operated by its own TSO (with the exception of Germany, which national transmission grid is operated by four different TSOs) but this doesn’t mean that they are totally independent. In contrary, all national power grids are interconnected and constitute one huge meshed European electric network. These interconnections enable TSOs to trade and exchange power so that each one of them can ensure their supply in the most efficient way possible. However, the cross-border transmission capacity of the interconnection is limited and the power exchanges should take into consideration the security constraints of each TSOs. It is thus necessary that every European TSO collaborates together and coordinates their actions in order to maintain the stability of the whole power grid. The goal is to manage to create a pan European power grid which optimizes all the means of production, with an integrated electricity market and a high inter-connectivity in order to enhance power exchanges between countries. The role of each European TSO is to improve this coordination and to work towards an optimized and unified power grid. Nonetheless, creating a pan European electric network is not an easy task. Given the multiplicity of the actors involved, the specifications of their network, their operating doctrines and their personal interests, the implementation of such a European power grid operation requires a lot of work for all the concerned TSOs. European coordination is one of the key aspects to tackle the new upcoming challenges of the electricity sector. The regulation and governance of European power grid should improve the supranational cooperation and coordination of TSOs. [5]

1.1.4 Problematic and report organization

(19)

analyses which is the coordination of remedial actions and their costs sharing. We will first explain more in detail the issue of remedial actions coordination by taking a closer look at its implementation, its regulatory framework and the differences between each region of Europe. Then we will study more precisely the remedial action coordination and costs sharing for a particular region on which I have worked. I will present the European working group’s discussions and my propositions to the different issues raised. Finally, this report concludes with a study case on a particular example of the studied European region.

1.2 European coordination of Remedial Actions

1.2.1 Congestion management: Remedial Actions

1.2.1.1 Risks of congestion

(20)

the hypotheses known two days before, which can of course be different from the real time situation. In fact, in real time the production of renewable sources can vary from the forecasted values. The consumption may also be different from the forecast and the productions dispatch not the same as anticipated. Moreover, the bidding zones in the zonal market model are seen as single nodes without any internal transmission capacity limit. All these particularities induce risk of congestion on the European network that needs to be controlled and mitigated.

1.2.1.2 Costly and non-costly Remedial Actions

(21)

However, in order to modify the production dispatch it is needed to start, increase, stop or decrease other power plants that are often more expensive as they were not scheduled during the optimal dispatch calculation. According to the location and the depth of the congestion, different power plants production will be modified which induces costs for the Remedial Actions borne by the concerned Transmission System Operator. The costs iduced by the congestion management is in a first time borne by the TSOs, but these additional costs are directly reflected on the consumers. Indeed, the congestion costs are integrated into the TSOs’ tarification resulting logicaly at the end to a higher price of electricty paid by the final consumer. Concerning the suppliers, the are rewarded in the same way as the balancing market : they receive an income for upward offers and they pay back for downward offers. (to compensate the fuel saved).

In contrary to primary and secondary control, the Remedial Actions are not activated automaticaly. Indeed, these mechanims are activated in order to relieve network conges-tion or sometimes voltage ajustement after an automated manipulaconges-tion. However, even if Remedial Actions are acivated by an operator, they are ordinary solutions compared to exceptional rolling load shedding. Non-costly Remedial Actions have the particularity to be quickly activated. An operator only needs a few minutes to activate the topological Remedial Actions which are applied immediatly. This is not the case for costly Remedial Actions as it involves to modify the production dispatch. The Mobilisation Leadtime of the Offer as well as the Minimum Usage Period are important factors. Indeed the time of activation can be around 15 minutes for a hydro power plant, and a couple of hours for a thermic power plant . The response time of a Remedial Action is important and should be used wisely in order to optimize the coordination of Remedial Actions. Therefore the optimal set of Remedial Actions is not computed considering one particular timestamp but for a whole day.

1.2.2 Counter Trading and Redispatching

(22)

1.2.2.1 Counter Trading

Counter Trading principle is illustrated in Figure 1.1, constituted of two bidding zones A and B. Considering that bidding zone A is exporting towards B and that a transmission line located in B is congested, the production dispatch needs to be modified in order to reduce the power flow between the two bidding zones.

Figure 1.1: Counter Trading

The net position of a bidding zone corresponds to the difference between its production and its consumption. By reducing the net position of bidding zone A and increasing the one of B, the power exchange between the two countries is reduced and the congestion can be solved. For Counter Trading, the involved power plants or consumption curtail-ments are not directly chosen by the TSOs. In fact the dispatching modification is lead by the market rules in the concerned bidding zone. For most of the European bidding zones, Countertrading measures are activated on the balancing market.In France Coun-tertrading mechanism is operated in the balancing market which corresponds to offers from suppliers. However for other bidding zones, Countertrading mechanism can either refers to bids only or to bids and offers.

1.2.2.2 Redispatching

(23)

rules. This is the TSO which decides to modify the production dispatch and directly select the most adapted power plants to solve the congestion according to their influence on the congestion and their offered prices. In contrary to countertrading, TSOs play an active role in Redispatching. One can identify three different types of Redispatching according to the concerned power plants and their location across the border:

• Cross-border Redispatching

Considering two bidding zones, A that is exporting and B that is importing, with a congested transmission line located in B. Cross-border Redispatching consists of decreasing the production of a particular power plant in A and increasing a particular power plant in B as illustrated in Figure 1.2. This kind of Redispatching requires coordination between the two concerned bidding zones.

Figure 1.2: Cross-border Redispatching

• Internal Redispatching

Considering the same situation with two bidding zones, the congestion located in B can sometimes be directly solved with Internal Redispatching. In the same way the production of a first power plant is lowered whereas the production of the second one is increased as illustrated in Figure 1.3. However in this situation the concerned power plants are located in the same bidding zone where the congested transmission line is. Therefore, with Internal Redispatching, there is no direct need to be coordinated with the neighboring TSO.

• External Redispatching

(24)

Figure 1.3: Internal Redispatching

solution is to modify the production of two power plants located in another bidding zone, here B. This kind of Remedial Action is called External Redispatching as illustrated in Figure 1.4.

(25)

Table 1.1: Different types of Redispatching

Type Congestion location Upward and Downward power plants

Cross-border B A and B

Internal B B

External A B

To summarize, Counter Trading is a cross-border exchange between TSOs of two bidding zones to resolve congestion by changing their net positions. Counter Trading may be less effective than Redispatching because power plants are selected according to the merit order and not their influence on congestion as it is the case for Redispatching. Counter Trading is thus often activated on a larger volume but for a lower price than Redis-patching. The economic efficiency of these two costly Remedial Actions often depends on the situation: Counter Trading would be more adapted for a simple border with only two bidding zones whereas Redispatching is more adapted to solve one particular congestion on a meshed part of the power grid. For example in France, the congestion concerning the Spanish border is managed with Countertrading whereas the German or Italian border is handled with Redispatching agreements.

1.2.3 Remedial Actions regulation

1.2.3.1 European Regulatory environment

(26)

set up the internal energy market and ensure its optimal operation while tackling the new challenges of the energy sector. [7]

1.2.3.2 Third energy package and network codes

Proposed in 2007 by the European Commission and established in 2009 by the European Parliament, the third energy package gathers the new objectives of the European energy domain that aims at further liberalizing the gas and electricity markets. ENTSOE has the legal responsibility of this new energy package and is committed to bringing the most suitable answers to the new challenges such as integrating a large amount of renewable, improving consecutive flexibility and striving to have a customer centric approach. The objectives set by this package in terms of climate improvement are: reduction of 40% of the greenhouse gas emissions compared to 1990 by 2030, at least 27% share of renewable in the consumption mix and also 27% of energy savings compared to the usual values [8]. In order to achieve these objectives of harmonization, integration and efficiency of European markets, ENTSOE works hands to hands with ACER to elaborate Network Codes that are a set of rules for every European TSOs. They are in total 8 Network Codes classified in three categories: Connection, Operation and Market. The regulation of Remedial Actions and their implementation in accordance with the third energy package refers to only two Network Codes that will be detailed further: System Operation Guidelines (SOGL) and Capacity Allocation and Congestion Management (CACM), respectively included in the Operation and Market categories.

1.2.3.3 Relevant articles of Network Codes for Remedial Actions coordina-tion and coordinacoordina-tion

Among all the articles of SOGL and CACM Network Codes, four of them are directly linked to the implementation of Remedial Actions. These articles are the results of the European Commission guideline and constitute the foundation of this master thesis problem.

(27)

This article imposes to the TSOs of a Capacity Calculation Region (CCR) to de-velop a common methodology for coordinated Redispatching and Countertrading. This article implies to define the notion of cross-border relevant Remedial Ac-tions and cross-border relevant network elements, to be able to identify clearly the available resources and the associated costs.

• Capacity Allocation and Congestion Management – Article 74: Redis-patching and Countertrading cost sharing methodology

This article focuses on the issue of Remedial Actions costs sharing. All TSOs in a Capacity Calculation Region shall develop a common methodology for Redispatch-ing and CountertradRedispatch-ing costs sharRedispatch-ing. Only the cross-border relevant actions are eligible to costs sharing and shall be determined in a transparent and auditable manner.

• System Operation Guideline – Article 75: Methodology for coordinating operational security analysis

This article describes the objectives and the framework for implementing a pan-European coordinated operational security analysis. It focuses on the influence of power exchanges, the identification of critical network element, computation of transmission capacity and the exchange of information between TSOs.

• System Operation Guideline – Article 76: Proposal for regional opera-tional security coordination

In the same way as article 75, this article aims to develop a common methodology for coordinating operational security analyses within a Capacity Calculation Re-gion. The defined methodology should respect and complement the article 75 of SOGL by taking into account all the particularities of the concerned region.

(28)

be directly implemented by the TSOs, it is needed to have the approbation of the Regulators. The approbation process of a methodology is the following:

– the guideline of the methodology is explained in the Network Codes proposed by the European Commission.

– in the view of these Network Codes, the concerned TSOs discuss and develop a proposal of methodology which is presented to the National Regulatory Authorities.

– the National Regulatory Authorities can either approve, request for amend-ment, or refer the methodology to the Agency for the Cooperation of Energy Regulators. If the methodology is approved, it shall be implemented by the TSOs. If there is a request for amendment from the NRAs, the proposed methodology is sent back to the TSOs in order that they can work on the weak points of the proposition. Finally, if the methodology is referred to ACER, the agency will propose its own version and submit it to the vote of the NRAs.

1.2.3.4 Capacity Calculation Regions and cross-border relevance

(29)

exchange capacity. The definitions of the European CCRs are detailed by the Network Code CACM in the article 15. This methodology defining the CCRs has the particularity to have been completely written by ACER.

One commonly used principle for capacity calculation is to define cross-border relevant elements in a CCR. Cross-border relevant elements are transmission lines or transformers that are influenced over a certain threshold by the power exchanges within the CCR. The influence coefficient is known as the Power Transfer Distribution Factor (PTDF) and represents the variation of power flow induced by a change of net position in the bidding zones. The higher the influence coefficient threshold, the less network elements are considered in the capacity calculation. With a high threshold, only elements located close to the border will be concerned by the capacity calculation that will result in a large volume of possible exchanges but also in higher security risks further away from the border. In contrary, with a small threshold, even elements located far away from the border will be taken into account that will certainly guarantee a high security but a very low exchange capacity. The value of the threshold should be decided by the TSOs according to the article 35 of CACM. The chosen values for the experimentation of the methodology are 5% and 10%. The cross-border relevant elements are called Critical Network Elements (CNE) as they are the elements highly influenced by the power exchanges and limiting the capacity between two bidding zones. In order to respect the N-1 criterion, the capacity calculation considers for every CNE contingencies that could create an overload. The couple of Critical Network Element and Contingency is called CNEC and is associated to Remedial Actions (RA) that can solve the constraint. The Network Codes impose operational security coordination within each CCR and also between CCRs. Foremost, the articles from CACM stipulate that each CCR should develop a coordination methodology for the Remedial Actions related to the Critical Network Elements and Contingency (CNEC) and their costs sharing.

1.3 Contribution

(30)

in this region, I participated in the working groups organized between all the TSOs of the CCR. My role in these working groups was to participate in the development of calculation and simulation tools, to make predictive studies allowing to estimate the influence of the options discussed during the European meetings and also to propose alternative methods to solve certain problems encountered.

(31)

(32)

Chapter 2

Study on Core Capacity

Calculation Region

2.1 Presentation of Core CCR and Remedial Actions

co-ordination and costs sharing

2.1.1 Core composition and particularities

Created in 2016, Core CCR gathers 16 TSOs from Central Eastern and Western Eu-rope with a view to harmonize and integrate the electricity market in a large scale. Core is composed of the following countries: Austria, Belgium, Czech Republic, Ger-many, France, Croatia, Hungary, Netherlands, Poland, Romania, Slovenia and Slovakia. Located at the heart of Europe, this region plays a key role in the Pan-European mar-ket integration. Initiated by ACER via the European Commission Network Codes and Guidelines, the aim is to work on common day-ahead, intra-day and also long-term (monthly and annual forward capacity allocation) capacity calculation methodologies. These projects should enhance and optimize the transmission capacity in Central Europe while maintaining a high security level and flexibility.

The coordination of the Remedial Actions is a major subject to manage to relieve, in a most efficient way, the congestion in an interconnected and highly meshed network. Indeed, besides the fact that Core is the biggest CCR in Europe, its central position and interconnections makes it a real crossroad for electricity exchanges. The multiplicity of

(33)

Chapter 2. Study on Core Capacity Calculation Region 19

bidding zones, borders, market models and existing methodologies between neighbors induces a lot of particularities and complicates the harmonization of processes. In every country, the network topology and the exploitation logic are different and every particu-lar situation should be taken into account in order to build an egalitarian methodology. Even if all TSOs will naturally defend their own point of view and personal interests, the final decisions should be based on an overall approach and should improve the European network and optimize its exploitation.[13] If the concerned TSOs don’t manage to come to an agreement, the decision is reported to ACER which will submit it to the vote of the NRAs.

The counter trading and redispatching coordination and cost sharing being a complex and sensitive issue, the methodology needs to be tested during experimentation which aims to try the different solutions imagined by the involved TSOs with real situation. Setting up the scope of this experimentation is thus a decisive part, each TSO is invited to propose its ideas, preferably along with public studies results for a sake of transparency and collaborative work. During my internship I have contributed to this experimentation set up by analyzing the solutions proposed by the Core members and also by putting forward my own ideas on the different parts of the cost sharing process.

2.1.2 Chosen cost sharing principle

The global policies for Remedial Action cost sharing in the different CCR are all built around a sharing principle which define the global approach. Among the existing cost sharing principles, the main approaches are equal splitting, requester pays and polluter pays.

• Equal splitting

(34)

ratio since the exchanges in this electric peninsula has no influence on the rest of the European grid and represent a benefit for both sides.

• Requester payer

The second approach is called requester pays and is a good solution when the activated Remedial Action doesn’t represent a benefit for both of the concerned bidding zones. Here only one side of the border supports the cost. If a TSO can’t manage to solve a cross border relevant congestion on its own and asks a neighbor to activate a Remedial Action, the cost will be attributed to the TSO which has requested this Remedial Action (whether it is Countertrading, cross-border, internal or external Redispatching). For example this kind of approach is used for the Italian border.

• Polluter payer

(35)

2.1.3 Main steps of the cost sharing process

The computation of the Remedial Actions cost sharing among the members of Core CCR can be divided in 6 distinctive parts. For each one of these calculation steps, there are many proposed solutions that needs to be analyzed and assessed before setting the scope of the experimentation. The chosen solutions should first and foremost optimize the European power grid and its electricity market by sending the right incentives to the concerned actors and by creating a social welfare surplus. The results of the costs shar-ing process should result in right transmission investment incentives for the TSOs.[14] [15] However, it is primary to take into consideration the technical issues at stake and to design both a consistent and feasible solution. The principal underlying technical constraints are IT limitations. Indeed processing time, reliability and robustness should be coherent with the imagined solution. For this experimentation, the simplicity of im-plementing the proposed solution is an important quality. Adjustment and optimization of the cost sharing process will be tuned later on, for now is to conceive a consistent and implementable solution.

The 6 steps of the cost sharing process are:

1. Files merging

2. Remedial Actions Optimization 3. Power Flow Decomposition 4. Netting

5. Prioritization 6. Mapping

2.2 Files merging

(36)

in order to create a Common Grid Model (CGM) which is the reference file to analyze the forecasted grid situation. Besides this DACF file, each TSO will join its Contigency list, Remedial Actions and Additional Constraints (CRAC) file which is a file that lists all the congestion forecasts and their associated Remedial Actions. This list contains also details on the upward or downward volumes and the corresponding prices. One day ahead, the Core CCR has a good overview of the upcoming power grid situation, its risks of congestion and the possible Remedial Actions to solve them. The final objective is to manage to integrate this methodology also for intraday and close-to-real-time time-frames.

2.3 Remedial Actions Optimization

From these input files, the optimization of the Remedial Actions of Core CCR can be computed. The development of this tool represents a huge project, dedicated teams at RTE and TSC Net are working on a technical solution to find the most efficient set of Remedial Actions to solve all the constraints of the region. The result of this Remedial Actions Optimizer (RAO) is a list of activated, or potentially activated in case of contingency Remedial Actions. These Remedial Actions are either costly (Counter Trading or Redispatching) or non-costly (topology changes or Phase Shift Transformer tap changes) and the associated costs of the optimal set of Remedial Actions should be minimal. Coreso and RTE have developed a Remedial Actions Optimizer as open source projects which need to be tested during the Core experimentation in order to compare them on concrete examples.

2.4 Power flow decomposition

2.4.1 Flow types definitions

(37)

of the power flows exchanges within the region. By managing to identify the power flows induced by each pair of producer and the consumer, it is possible to “decompose” the physical power flows and to have an idea of the sources and the sinks of each MW that flows in the power grid.

A flow type can be defined by three characteristics: its source, its sink and its trans-mission line. The source of a power flow is the power plant that is at the origin of this electricity power. The sink of a power flow is the consumer that uses this electric energy. The line of the power flow is the transmission element where the flow is observed. Based on this decomposition description, ENTSOE has defined 5 different flow types:

– Internal – Loop

– Import and Export – Transit

– Cycle

The Physical Flow that is transmitted in a line can be decompose in these 5 different flow types: Physical Flow = Internal Flow + Loop Flow + Import/Export Flow + Transit Flow + Cycle Flow The definitions of these flow types are:

– Internal: the source, the sink and the line are located in the same bidding zone.

– Loop: the source and the sink are located in the same bidding zone but the line is located in a different bidding zone.

– Import and Export: the source and the sink are located in different bidding zones and the line is located in one of these two bidding zones.

– Transit: the source and the sink are located in two different bidding zones and the line is located in a third bidding zone.

– Cycle: this particular type of flow represents the power flow induced by the tap of a Phase Shift Transformer (PST)

(38)

Table 2.1: Flow Types

Flow Type Source Sink Line

Internal A A A

Loop A A B or C

Import & Export A B A or B

Transit A B C

as a topological situation of the network. The PST are located at the border of the borders in order to manage the unscheduled power flows and can be used as topological Remedial Actions. The Import, Export and Transit Flows can be grouped in a single category of Commercial Flows as they result from the market coupling process whereas the Internal and Loop Flows don’t originate from this capacity allocation mechanism. The characteristics of Internal, Import and Export, Transit and Loop Flows can be summarized in table 2.1.

The different flow types are illustrated in Figures 2.1, 2.2, 2.3 and 2.4:

(39)

The electric power trades can be divided in two different categories depending if this concerns internal or external schedules. Internal Commercial Trade (ICT) Schedules

Figure 2.2: Loop

Figure 2.3: Import and Export

(40)

represent the commercial exchange within a bidding zone between different market par-ticipants or between nominated electricity operators and market coupling operators. Whereas Aggregated Netted External (ANE) Schedules represent the resulting aggre-gation of all external TSO schedules and External Commercial Trades (ECT) between two bidding zones.[16]

These two kinds of trade creates specific types of flow. Indeed the Internal Commer-cial Trade Schedules induce Internal Flows and Loop Flows according to the electrical distance in the meshed network, whereas Aggregated Netted External (ANE) Schedules induce Import, Export and Transit Flows. In a meshed and interconnected network, the measured physical flows are different from the predicted and traded schedule flows. The unscheduled flows are the difference between the physical flows and the schedule flows. On a tie line at a border between two bidding zones, the unscheduled flows are the composed of Loop Flows and unscheduled Import, Export and Transit Flows. The unscheduled flows across a border represent a security risk for the power grid as they haven’t been taken into account during the preventive security analyzes.

It is a real challenge for the European TSOs to manage to respect the scheduled power exchange traded between bidding zones. By slightly changing the network topology, TSOs can match the scheduled exchanges and avoid unscheduled flows. However, during real time operations, it is not feasible to identify the source and the sink and thus the type of power flows. As unscheduled Import, Export and Transit flows over a tie line is the result of technical limitation and lack of coordination from either sides, the responsibility of this difference is shared between both TSOs. In fact, one can consider that these types of flows originate from market coupling process. However, concerning the Loop Flows, these unscheduled power exchanges are due to the incapacity of one TSO to handle its Internal Commercial Trades with its own power grid. Indeed if a TSO network is not enough developed and flexible, the internal power exchanges will be transmitted by its neighbors’ networks. The Loop Flows are unscheduled power flows which are not taken into account during the capacity allocation mechanism and pollutes the European network.

(41)

transmitted through a line in 5 different flow types doesn’t correspond to a real physical reality but rather to a method that enable to understand the interactions between TSOs and their influence on network congestion in a global European point of view. There are two main methods to decompose power flow through a line:

• Power Flow Coloring (PFC) • Full Line Decomposition (FLD)

2.4.2 Power Flows Coloring (PFC)

2.4.2.1 General presentation

The Power Flow Coloring decomposition method has been developed within a research project called FutureFlow in which four TSOs of Core CCR are involved (ELES, MAVIR, APG and Transelectrica). The aim of this FutureFlow project is to implement, within horizon 2020 an efficient power flow decomposition algorithm that stays consistent with the zonal market model used in Europe. However, even if the European electricity market is based on a zonal model, it is really important to manage to create a model which take into account the nodal production and consumption within a bidding zone. This algorithm is based on the electric superposition principle which enables to decompose the power grid diagram into two different fictional and artificial zonal models:

– A balanced model without cross-zonal exchanges – A complementary model with cross-zonal exchanges

2.4.2.2 Algorithm explanation

(42)

Figure 2.5: Power Flows Coloring calculation process

From the initial power grid model, which results from the merging of all TSO grid model, the algorithm creates two different grid models: a balanced model without cross-zonal exchanges and a complementary model with cross-zonal exchanges. These two models are artificial and aims to identify on the one hand the Internal and Loop Flows with the balanced model and on the other hand the Commercial Flows with the complementary model. The algorithm also uses a Nodal Power Transfer Distribution Factor (PTDF) Matrix to represent the influence of the exchanges between two nodes of the network on the power flows through transmission lines. [17] The calculation of the two models and the Nodal PTDF matrix, as well as their importance for computing the different type of flows will be explained in the following sections.

• Load Flow Calculation

(43)

it is possible to compute the power flows through all the lines of the network. The load flow calculation is essential for every kind of analyzes of the electricity network. Every TSO has its own load flow calculation tool which is often based on the Newton-Raphson solving method.

• Balancing of Zones

The main characteristic of the PFC algorithm is to decompose a power grid model into two different fictional and artificial models according to the net position of the bidding zones: a balanced one without cross-border power exchange and a compensating one with only cross border power exchange. In order to illustrate this part of the algorithm, we will consider the example in Figure 2.6 of a power grid composed of 3 bidding zones A,B and C. Pgen and Pload represent respectively the generation and the load of the

bidding zone expressed in M W .

Figure 2.6: Balancing of zones example

Balanced model:

The balanced model is built from the initial grid model by shifting fictitiously the global production and consumption in order to have a net position equal to zero for every bidding zones.

(44)

If a bidding zone is importing, then its net position is negative. In order to create a balanced model with a net position equal to zero, its total consumption is decreased down to its total production.

The resulting grid model is composed of bidding zones with a null net position and no power exchange between these zones as illustrated in Figure 2.7.

Figure 2.7: Balancing of zones example : Balanced Model

Compensating model:

(45)

Figure 2.8: Balancing of zones example : Compensating Model

Figure 2.9 summarizes the decomposition of the initial grid model.

Figure 2.9: Balancing of zones example : decomposition of the grid model

• Nodal PTDF Matrix Calculation

The Power Transfer Distribution Factor (PTDF), also called Distribution Factor is a matrix which represents how the power flows are distributed through the network if the net injection changes.[1] Let’s consider a power grid with n nodes and l lines. The reference node, or slack node, is the node N which will compensate the net injection changes. The PTDF coefficient for a given node i and a given line j is P T DFij and

(46)

Chapter 2. Study on Core Capacity Calculation Region 32 P T DF =      P T DF11 · · · P T DF1n−1 .. . . .. ... P T DFl1 · · · P T DFln−1      (2.1)

This PTDF matrix depends only on the network topology and is often used to understand the influence of a power plant in the network and to compute the feasible generation dispatches. In fact, in the optimal dispatch computation, the constraint implied by the power lines capacity limits can be expressed with the PTDF matrix of the network. Let’s consider the same power grid with n nodes and l lines. For every node i the power generation at this node is Qi and for every line j the capacity is Kj and the power flow

through this line is Fj. The congestion constraint for this example can be expressed by

the following formula: ∀j ∈ [1, l] Fj ≤ Kj n−1 X i=1 P T DFij× Qi ≤ Kj (2.2)

Usually the slack node chosen to compute the PTDF matrix is chosen as the most interconnected node of the network but by changing the slack node, the compensating withdrawal is changed. By changing the slack node position, it is possible to compute for every pair of source and sink the power flow distribution over every line of the network. Node-to-Node Power Transfer Distribution Matrix is constructed by calculating the PTDF matrix with all the possible slack node position. By doing so, the Node-to-Node PTDF matrix of a power grid models how the exchanges’ variation between two nodes is distributed on a given line of the network. The n × n Node-to-Node PTDF matrix of a line l is P T DFnode−to−node,l and its coefficient can be expressed as:

P T DFnode−to−node,l =      P T DF (l, 1 → 1) · · · P T DF (l, 1 → n) .. . . .. ... P T DF (l, n → 1) · · · P T DF (l, n → n)      (2.3)

(47)

• Power flows identification

From the two power grid models, it is possible to identify different power flow types. In the balanced model, there is no cross-border power exchange and thus Internal and Loop Flows can be identified. Whereas in the compensating model, there are only Import, Export and Transit Flows.

Identification of Internal and Loop Flows:

In the balanced model, the generation and the load at a given node i are respectively G(i)balanced and L(i)balanced. In the balanced model, the power flow through a line l

induced by a bidding zone O is F (l)O

balanced and can be calculated as:

F (l)O_balanced=X

i∈O

P T DF (l, i → ref ) × (G(i)balanced+ L(i)balanced) (2.4)

Here the PTDF coefficient is calculated with an arbitrary slack node and represents how a net injection at a node i located in the bidding zone O influence the power flow through the studied line l.

If the line l is located in the bidding zone O then the identified flow F (l)O_balancedrepresent an Internal Flow. Otherwise, if the line l is not located in the bidding zone O then the identified flow F (l)O_balanced represent a Loop Flow.

Identification of Export, Import and Transit Flows:

In the exchange model, the generation and the load at a given node i are respectively G(i)exchangeand L(i)exchange. In this model, it is necessary to identify first the exchanges

(48)

F (l)i→j_exchange= P T DF (l, i → j) × L(j)exchange PN

n=1L(n)exchange

× G(i)_exchange (2.5)

Here the PTDF coefficient is a node-to-node factor representing the influence of the exchange between node i and j on the line l. The proportional pairing of source and sink is done on all the loads of the power grid, here the considered network has N nodes. As both the source and the sink of an exchange can be considered as responsible of the power flow induced on a line of the network, the power flow created by the source i on the line l, F (l)i_exchange, is equal to the power flow created by the sink j on the line l, F (l)j_exchange, as:

F (l)i_exchange= F (l)j_exchange= 1

2× F (l)

i→j

exchange (2.6)

Finally, the exchange power flows created by a bidding zone O on a line l F (l)O_exchange is: F (l)O exchange = P i∈OF (l)iexchange = 1₂ × P T DF (l, i → j) × L(j)exchange PN n=1L(n)exchange × G(i)_exchange (2.7)

If the line l is located in the bidding zone O then the identified power flow F (l)O_exchange is an Import or Export power flow, depending on the direction. Otherwise, if the line l is located outside the bidding zone O, the identified power flow F (l)O_exchangeis a Transit Flow. For the particular case where the line is located in two bidding zones, F (l)O_exchange is an Import or Export power flow if one end of the transmission line is located in bidding zone O and F (l)O_exchange is a Transit Flow if both ends of the transmission lines are outside the bidding zone O.

(49)

2.4.3 Full Line Decomposition (FLD)

2.4.3.1 General presentation

The Full Line Decomposition (FLD) algorithm is an alternative method developed by Tennet, the Netherland’s TSO. The problem of decomposing power flow in a multi-zonal meshed network is tackled with a totally different approach. Indeed, contrary to PFC method which is based on the net position of the bidding zones, FLD aims to model the power exchange at the nodal level. This method is based on a matrix calculation that involves the following matrices:

– Node-to-node Power Transfer Distribution Factor (PTDF) matrix: represents how the power flows are distributed on the network when the net injection is modified.

– Power Exchange (PEX) matrix: represents the power exchanges between each source and sink of the network for a given situation with a particular produc-tion and consumpproduc-tion dispatch.

– Power Flow Partitioning (PFP) matrix: represent for a given transmission line, the decomposition of the power flow seen as an exchange between every bidding zones.

The computation of these matrices and their role in the decomposition of the power flows will be explained in detail in the algorithm description.

2.4.3.2 Algorithm explanation

The global idea behind this method consists of identifying the sources to sinks flows with the PEX matrix. The Full Line Decomposition algorithm being built on a complex matrix calculation, the description of this method will be explained on a theoretical transmission network composed of L lines and N nodes. The N × 1 matrices PG and

PD represent respectively the vector of nodal generations and nodal demands on the

(50)

If line l starts at node k then:

clk= 1

If line l ends at node k then:

clk= −1

If line l is not connected at node k then:

clk= 0

This connectivity matrix can be decomposed in two different matrices Cd and Cu

rep-resenting respectively the downward and upward direction of the network. Cd is thus

only composed of 1 and 0 whereas Cu is only composed of −1 and 0 coefficients such as

C = Cd+ Cu. The N × N nodal downward flow matrix Fd is defined as follows [18]:

Fd= −CdTdiag(F )Cu (2.8)

The diag() operator creates a N × N diagonal matrix from the values of the N × 1 vector. The coefficient Fdij represents thus the power flow in line from node i towards

node j.

The nodal power of a bus is defined as the sum of the nodal inflows and local generation or as the sum of the nodal outflows and the local consumption. The nodal power of all the buses are gathered in a N × 1 matrix P defined such as:

P = PD+ Fd1 = PG+ FdT1 (2.9)

(51)

Chapter 2. Study on Core Capacity Calculation Region 37 PD = P − Fd1 = P + C_dTdiag(F )Cu1 = [I + C_dT diag(F ) Cudiag(P−1) ]P = AdP (2.10)

Where Ad= I + C_dT diag(F ) Cudiag(P−1) and represents the downstream distribution

matrix. This matrix enables to compute the vector of nodal power with the vector of power demands. In the same way, it is possible to compute the upstream distribution matrix Au. Indeed: PG = P − Fu1 = P + C_uTdiag(F )Cd1 = [I + C_uT diag(F ) Cddiag(P−1) ]P = AuP (2.11)

In order to better understanding the meaning of these distribution matrices, it is possible to express them as:

∀(i, j) ∈ N × N If i = j : Ad ij = 1 If j ∈ αd_i : Ad ij = Pij Pj Otherwise Ad ij = 0

Where αd_i is the set of indexes of all the downstream nodes supplied by node i. Pij is

the power flow on the line between node i and i and Pj is the nodal power at node j.

(52)

Chapter 2. Study on Core Capacity Calculation Region 38 If i = j : Au ij = 1 If j ∈ αu_i : Au ij = Pij Pj Otherwise Au ij= 0

Where αu_i is the set of indexes of all the upstream nodes supplying node i.

This upstream distribution matrix gives the relation between the vector of power gener-ations and the vector of nodal powers. Assuming that matrices Adand Auare invertible,

the important result of these calculations is thus the following relation:

P = A−1_d PD = A−1u PG (2.12)

The proof that the distribution matrices can actually be inverted is reported in paper [19].

The importance of this result is that every element A−1_dij of the inverted downstream distribution matrix represents the share of the nodal power at node j supplied by node i and every element A−1_uij of the inverted upstream distribution matrix represents the share of the nodal power at node j supplying node i.

Inverting a matrix is not an efficient solution to implement in a computer program. By considering the problem of decomposing power flows in a multi-zonal transmission net-work with a graph approach, it is possible to express the inverted distribution matrices. Let’s consider only the downstream point of view and introduce the matrix Ddsuch as:

Dd= CdTdiag(F ) Cudiag(P−1) (2.13)

(53)

Dd= I − Ad (2.14)

The coefficient Dm

d ij represents the share the nodal power of node j indirectly supplied

by node i through m − 1 intermediate nodes of the network. For example, the coefficient Dd ij represents the share the nodal power of node j directly (without passing through

another node) supplied by node i. In the same way, the coefficient D_{d ij}2 represents the share of the nodal power of node j indirectly supplied by node i through one intermediate node of the network.

(54)

This equation shows that the inverted downstream distribution matrix can be expressed as the sum of the shares of the nodal power through all the possible path of the net-work. The convergence of the infinite sum can be demonstrated with further calculation explained in paper [19]. However, just with a graph-oriented approach, it is understand-able that the influence of a node on the nodal power will tend toward zero when the number of intermediary nodes is high.

Finally, the Power Exchange (PEX) matrix represents the power exchange from every source to every sink of the grid. The coefficient P EXij is the power power produced at

node i by the local source that is consumed at node j by the local sink. This matrix can be computed in two different ways, by using either the downstream or the upstream inverted distribution matrix.

In fact, the element A−1_{d ij} represents the share of nodal power at node j that is supplied by node i. By applying the proportional sharing principle, the proportion of the nodal power Pi coming from the local generation PGi is equal to P_PG i_i and thus:

P EXij = PD j×

PG i

Pi

× A−1_{d ij} (2.20)

In the same way, the element A−1_{u ij} represents the share of nodal power at node j that supplies node i. By applying the proportional sharing principle, the proportion of the nodal power Pj consumed by the local demand PDj is equal to P_PD j_j and thus:

P EXij = PG i×

PD j

Pj

× A−1_{u ij} (2.21)

Now that the PEX matrix has been constructed, it is possible to compute for every critical transmission lines its Power Flow Partitioning (PFP) matrix. The P F Pl matrix

of a line l gathers the results of the power flow decomposition for every node of the net-work. This matrix is easily calculated by multiplying the Node-to-Node Power Transfer Distribution Factor with the Power Exchange matrix:

∀l ∈ L

(55)

In the context of costly Remedial Action cost sharing, the nodal results can be aggregated by bidding zones in order to visualize easily the types of flows through the studied line. Given l a transmission line, Z number of bidding zones, for every pair of bidding zones (zA, zB) ∈ Z × Z the zonal Power Flow Partitioning matrix coefficient is[20]:

P F Pzonal l zAzB = X i∈zA X j∈zB P EXij (2.23)

Considering that the studied line l is located in the bidding zone zA, the different

coef-ficient of the zonal PFP matrix can be easily associated to a type of flow.

P F Pzonal l zAzA → Internal F low

P F Pzonal l zBzB → Loop F lowf rom zB to zB

P F Pzonal l zAzB → Export F low f rom zAto zB

P F Pzonal l zBzA → Export F low f rom zB to zA

(56)

Table 2.2: Power Flow identification with PFP on a line located in A Line located in A Load

A B C

Generator

A Internal Export Export B Import Loop Transit C Import Transit Loop

Considering now that the studied line l is a tie-line between bidding zones zA and zB,

the different coefficient of the zonal PFP matrix can be easily associated to a type of flow.

P F Pzonal l zAzA → Loop F low f rom zAto zA

P F Pzonal l zBzB → Loop F low f rom zBto zB

P F Pzonal l zCzC → Loop F low f rom zC to zC

P F Pzonal l zAzB → Export/Import F low f rom zAto zB

P F Pzonal l zBzA → Export/Import F low f rom zBto zA

P F Pzonal l zAzC → T ransit F low f rom zAto zC

P F Pzonal l zBzC → T ransit F low f rom zBto zC

P F Pzonal l zCzA → T ransit F low f rom zC to zA

(57)

Table 2.3: Power Flow identification with PFP on a tie-line between A and B Tie-line between A and B Load

A B C

Generator

A Loop Import/Export Transit B Import/Export Loop Transit C Transit Transit Loop

For each transmission line studied, there is an associated Zonal Power Flow Partitioning matrix that represents the different type of flows created by all the concerned bidding zones.

2.4.4 Netting and scaling of flows

Both power flow decomposition algorithms compute, in a different way, the power flow types which flow through the critical transmission lines. Each algorithm has its own particularities and even if the results may differ in some case, the calculated power flows needs to be understood, transformed and analyzed. It is important to keep in mind that the power flow decomposition doesn’t have any physical reality. Indeed, it is a method to understand the influence of every bidding zone on the network congestions. Thus the values of the power flows manipulated may seem abstract and artificial but it is primary for analyzes to be able to understand the implications.

2.4.4.1 Relieving and burdening flows

(58)

However, there are several ways of handling the reliving flows. This problem, called “Netting of flows”, consists of allocating relieving flows to burdening flows in order to have only burdening flows to consider and to penalize.

2.4.4.2 Netting solutions

The discussed solutions are:

– Vertical shift

– Proportional netting

– Proportional netting per category

– Proportional netting per category with credit – Equal netting per category

– Netting of flows per bidding zones

(59)

Table 2.4: Example of Power Flow Decomposition

A B C

Loop 20% 15% -15% Internal 0% 0% 35% Import/Export 30% -10% 50%

Figure 2.10: Example of Power Flow Decomposition

• Vertical Shift

(60)

Figure 2.11: Vertical Shift Netting

This solution of netting involves choosing a prioritization of the power flows to relieve. The choice of this hierarchy will be further discussed in the section 2.5.

• Proportional Netting

(61)

Figure 2.12: Total Power Flows before and after Proportional Netting

Figure 2.13: Final Power Flows Decomposition after Proportional Netting

(62)

The Proportional Netting per category solution is similar to the Proportional Netting except that each flow category (Internal, Commercial, Loop or Cycle) is treated sepa-rately. Indeed, the power flow decomposition aim is to identify the nature of the power exchanges and to be able to differentiate each category in the netting process. Indeed, each category of power flow can be seen as a different kind of power exchange and thus needs to be treated separately. Here the relieving flows will not be distributed on all the burdening flows but only on the concerned power flow category as illustrated in Figures 2.14 and 2.15.

(63)

Figure 2.15: Final Power Flows Decomposition after Proportional Netting per Cate-gory

• Equal Netting per category

(64)

Figure 2.16: Final Power Flows after Equal Netting per Category

2.4.5 Solutions comparisons

The presented solutions for scaling and netting relieving flows are only the most discussed options for the experimentation. There are many other possible ways to handle this question but the influence on the results is not so important. The solution of Vertical Shift is too dependent on the prioritization of the power flow types, which remain one of the most determinant option. This question will be further explained in the next section. Besides, netting proportionally all the flows without taking into account their category is not compliant with the polluter-payer principle. This solution has been rejected by the NRAs, which has requested to only focus on the last two solutions:

– Proportional per category – Equal per category

Costly Remedial Actions Coordination

Costly Remedial Actions

Coordination

QUENTIN WILLAIME

KTH Royal Institute of Technology

Master Thesis

Costly Remedial Actions Coordination

Declaration of Authorship

Abstrakt

Acknowledgements

Contents

List of Figures

List of Tables

Abbreviations

Chapter 1

Introduction

1.1

Background

1.2

European coordination of Remedial Actions

1.3

Contribution

Chapter 2

Study on Core Capacity

Calculation Region

2.1

Presentation of Core CCR and Remedial Actions

co-ordination and costs sharing

2.2

Files merging

2.3

Remedial Actions Optimization

2.4

Power flow decomposition