A harmonized European electricity balancing market. Incorporation of congestion management into cross-border reserve procurement

I. Table of content

II. List of abbreviations

III. List of figures

IV. List of tables

1. Introduction

2. Current electricity market design in Europe
2.1 Forward and future market
2.2 Day-ahead market
2.3 Intra-day market
2.4 Balancing market

3. Operating reserve market design in continental Europe
3.1 Power-frequency control by the TSO
3.2 Operating reserve types and technical conditions
3.3 Economic conditions
3.4 Operating reserve planning

4. Cross-border coordination in reserve markets
4.1 Benefits of cross-border reserve markets
4.2 Principal options for reserve market coordination
4.3 Current practice in Europe
4.4 Cross-border capacity reservation

5. Design framework for the incorporation of network constraints in cross-border balancing markets
5.1 Design space determination
5.1.1 Operating reserve type
5.1.2 Operating reserve scheduling
5.1.3 Integration model and cross-border transmission capacity reservation approach
5.2 Model selection

6. Methodology
6.1 Model assumptions
6.2 Model structure
6.3 Theoretical model formulation
6.3.1 Stage 1 ะ day-ahead energy market
6.3.2 Stage 2: day-ahead reserve procurement
6.3.3 Stage 3: real-time reserve activation

7. Model calibration
7.1 Case study and data validation
7.2 Results
7.2.1 Generation cost and system reliability
7.2.2 Generation and reserve scheduling
7.2.3 Generation utilization of transmission capacity

8. Critical examination and continuation
8.1 Model evaluation
8.2 Model extension proposition
8.3 Available Transfer Capacity
8.4 Flow-Based Market Coupling implementation
8.4.1 FBMC parameters
8.4.2 Expected outcome and conceivable challenges

9. Conclusion and further research perspectives

V. Bibliography

VI. Appendix

VI.I Nomenclature

VI.II Additional figures

VI.III Additional tables

II. List of abbreviations

illustration not visible in this excerpt

III. List of figures

Figure 1: The current electricity market design in Europe subject to the time dimension

Figure 2: Operational distribution within a synchronous area

Figure 3: Activation of different operating reserve products based on time

Figure 4: Overview deployment and functions of different operating reserves

Figure 5: Trade and delivery of different operating reserve products for one week

Figure 6: Potential benefits of Imbalance Netting with two control areas

Figure 7: Common Merit Order-cooperation

Figure 8: Reserve Sharing example with three control areas

Figure 9: Cross-border capacity reservation approaches for balancing energy

Figure 10 Overvieพ of the model structure and considered scenarios

Figure 11: Trade-off between cost efficiency and loss of load

Figure 12: Average procurement of spinning reserves in the different countries under different scenarios

Figure 13: Duration curve of the reserve export and import margins for Germany

Figure 14: Available Transfer Capacity approach applied to the case study

Figure 15: The Flow-Based Market Coupling Mechanism approach applied to the case study

Figure Al: Basic structure of the balancing market

Figure A2: Example of Imbalance Netting of secondary operating reserves

Figure A3: Overview of pilot projects called by the ENTSO-E in 2012

Figure A4: Historical evolution of the IGCC

Figure A5: Probability density functions of the 2013 imbalances in the different countries

Figure A6: Derivation of АТС and RAM

Figure A7: Flow domains of АТС and FB approach

IV. List of tables

Table 1: Overview of imbalance settlement between TSO and BRP

Table 2: Overview of the technical and economical conditions for balancing reserves

Table 3: Overview of the reasonable scenarios for cross-border coordination in reserve markets

Table 4: Overview of options for cross-zonal cooperation for operating reserves

Table 5: Design space determination of the model

Table 6: Operational generation cost ad system reliability indicators for the different model scenarios

Table 7: Generation output of the day-ahead energy market

Table 8: Schematic overview of all possible export/import options of spinning reserves for a control area

Table Al: ENTSO-E members sorted by countries

Table A2: Participating TSOs and goals of pilot projects in cross-border coordination of balancing energy and balancing capacity

Table A3: Installed conventional generation capacity per control area and generation type

Table A4: Technical characteristics per generation type

Table A5: Fuel prices overview

Table A6: Transfer capacities between the different countries (in MW)

Table A7: Reserve requirement per country (in MW)

1. Introduction

European electricity markets have changed profoundly in the last decades. In national markets increased deregulation and privatization have accompanied the decoupling of transmission, distribution and generation activities. Simultaneously, markets in Europe are becoming more interconnected and harmonized, aiming to achieve one European internal electricity market. Until now, the focus was primarily set on the harmonization and integration of spot and forward markets. Nonetheless, the short-term regulation of balancing energy within the balancing market, especially with regard to increased cost-efficiency and enhanced system reliability, is gaining attention in academia and industry. Here, Transmission System Operators bear the final responsibility to balance the generation and consumption of electric energy on an instantaneous basis in order to maintain the system frequency and to prevent blackouts. The need for flexible short-term operating reserves is accelerated by the increased penetration of intermitted renewables, which leads to more short-term deviation in the power grid and hence to operational uncertainty. In spite of the technical significance of balancing markets, they also have a high economic relevance, as without their existence, occurred outages would lead to high costs. Corresponding to the zonal nature of the European power market, today’s operating reserve markets are mainly organized on a national level. Thereby Transmission System Operators in neighboring European countries manage unscheduled network imbalances by dimensioning, procuring and activating operating reserves within their own control areas, while attempting to maintain day-ahead schedules of international power exchange. The dimensioning and procurement, thus the scheduling of operating reserve capacity happens before real-time, whereas the activation happens in real-time in order to deliver or consume energy. Within a European cross-border balancing market, however, Transmission System Operators could coordinate their balancing activities with other control areas. Such cross-border balancing markets increase social welfare and operational reliability, since the amount of reserves required could be decreased through spatial smoothing of imbalances and the costs of reserve procurement and activation can be reduced.

Nevertheless, cross-border reserve markets are constrained by the available cross-border transmission capacity. Therefore, a challenging issue is to account properly for network constraints, especially within the procurement of reserves, which happens before real-time and under uncertainty as neither the remaining cross-border capacity nor the activation status of the procured reserves is known. Therefore, transmission capacity should be reserved for cross-border procurement based on predictions or certain rules.

The purpose of this thesis is twofold. First, it presents a balancing market framework by analyzing the market design and identifying relevant design variables. These include the different operating reserve types and their technical and economical conditions, the different scheduling phases, possible cross-border integration models and transmission capacity reservation approaches. Second, it introduces and discusses one selected deterministic model from the existing literature, dealing with the cross-border coordination within the operating reserve procurement while taking into account cross-border transmission constraints. In conjunction with the model and in order to quantify the benefits, an obtained case study for the 2013 Central Western European system is presented.

Within a subsequent model evaluation, one potential extension is picked out and discussed into more detail dealing with the locational scheduling of operating reserves by taking into account not only transmission constraints between different control areas but also intra-zonal constraints. Such a locational-based coupling would require a nodal market structure, which is currently infeasible in European electricity markets. One possible alternative, however, is the Flow-Based Market Coupling approach that can be seen as a combination of the zonal approach with the physical transmission constraints from the nodal market clearing. Its possible incorporation within the presented methodology and the expected implications for the obtained case study are examined.

The thesis is structured as follows. Section 2 gives an overview of the current electricity market design and Section 3 integrates the balancing market in the framework and presents the economical and technical conditions. Section 4 describes the benefits from cross-border coordination in reserve markets and introduces possible integration models. Within Section 5, model requirements are formulated and one suitable model is selected. Section 6 presents the selected methodology and Section 7 gives the obtained case study. Section 8 evaluates the model and discusses the proposed expansion involving the flow-based approach. Lastly, Section 9 concludes and gives further research prospects.

2. Current electricity market design in Europe

In order to analyze and afterwards model the balancing market, this chapter should give a general overview of the European electricity market structure, integrate the balancing market within the overall market structure and point out its importance.

Historically, electricity markets were organized nationally, each country focusing on self­sufficiency in terms of electric power supply with vertically integrated companies responsible for all steps of the value chain: generation, transmission, distribution and retailing. In order to realize the Internal Energy Market (IEM), European policymakers focused on the liberalization, independent regulation and supra-natural integration of electricity markets. From 1991 on, the European Commission initiated three directives, also called legislative energy packages aiming to liberalize the energy markets, whereas the first concentrated on the liberalization for industrial customers, the second on the liberalization for small customers and the last on cross-border regulation (Boltz, 2013).

Most of the power industry in Europe nowadays is vertically unbundled, whereas in many member states transmission and distribution of electric power are regulated natural monopolies. Generators compete in the wholesale electricity market to sell electricity to large industrial customers and suppliers. Suppliers themselves compete in the retail market to sell electricity to the final customer. This thesis focuses on the wholesale electricity market.

There are several trading platforms for wholesale market participants. On power exchanges or multilateral trading platforms, traders submit generation and demand bids and the market is cleared once per time period and a market price is determined. On the other hand, market players can directly interact and agree within a bilateral or an organized over-the-counter trade (OTC). In the latter, generators and customers submit generation and demand bids to a market platform, which is cleared continuously, resulting in different prices for each trade.

Furthermore, the European electricity market is an energy-only market meaning that generators are remunerated for the generated electric energy. If generators were remunerated for the generation capacity they have available in the market, it would be a capacity market (KU Leuven Energy Institute, 2015b).

illustration not visible in this excerpt

Figure 1: The current electricity market design in Europe subject to the time dimension

Different types of electricity markets are arranged in a sequential order to guarantee an instantaneous balance between electricity generation and offtake (source: KU Leuven Energy Institute, 2015).

Lastly, there are different types of electricity spot markets divided by the time dimension in order to keep the power market balanced on an instantaneous basis, starting years before the delivery and ending after the actual delivery. As shown in Figure 1, it can be distinguished between four different markets: the forward and future market, the day-ahead market, the intraday-market and lastly the balancing market (Forstbach, 2009).

2.1 Forward and future market

Electricity generators sell electricity on forward and future markets to ensure future sales and reduce their vulnerability to possible electricity price decreases. Large industrial consumers on the other hand might buy electricity on the forward market to secure their future electricity consumption upfront and reduce vulnerability to possible electricity price increases.

The forward and future market starts years in advance to the day before delivery and contains contracts to either deliver or consume a certain amount of electricity. Futures are standardized contracts like monthly, quarterly or annual contracts traded on power exchanges. Forwards on the other hand are non-standardized bilateral contracts that are traded over-the-counter. Thereby the contract parties negotiate their entire conditions, which give them more flexibility.

Furthermore, electricity can be either traded between different market zones or within one. The allocation of the transmission capacity between two market zones happens explicitly, which means that transmission capacity is traded apart from electric energy. In consequence, market players first buy the right to use the transmission capacity via auctioning between two market zones before buying or selling electricity in another zone. Regarding the trading within a market zone, it is assumed that intra-zonal trading is never constrained by the transmission capacities (KU Leuven Energy Institute, 2015b).

2.2 Day-ahead market

On the day-ahead market, electricity is traded one day before the actual delivery. As in the future and forward market, both standardized and non-standardized contracts are traded. During the continuous trading, standardized block-contracts for base and peak load-periods are arranged, whereas in the auction trade, non-standardized single hours or hour-combinations are traded. At the end of the day-ahead market, the market zone has to be balanced which means that scheduled generation in the market zone equals forecasted demand in the market zone plus net export to other market zones.

The final responsibility for the generation-consumption balance lies with the Transmission System Operators (TSO) (ENTSO-E, 2014a). Before the actual delivery, it is passed to a Balance Responsible Party (BRP), which is a legal entity, balancing generation and consumption from generators, suppliers and consumers. Such a BRP may be a power supply company or a power plant operator. The BRP keeps a virtual energy account, which establishes a connection between the financial side of electricity trading and the physical power supply and network stability. Hereby the BRP portfolio may consist of own generation, own consumption or contain electricity traded with other BRPs, whereas the physical process of generating and consuming is still managed by generators and consumers themselves. At the end of the day-ahead market, each BRP submits a balanced portfolio to the TSO, which gives a planned schedule of generation or consumption for every unit of the BRP, based on power plant level.

2.3 Intra-day market

The intra-day market enables market participants to correct for shifts in the day-ahead nominations due to better wind forecasts, unexpected power plant outages and other unpredictable parameters. In order to meet fluctuations, the intra-day market gives the opportunity to trade up to 45 minutes to maturity. Standardized one-hour-bids for the base load covering hours 1 to 24 and peak load covering hours 9 till 20 can be submitted. Non-standardized trade is possible for single quarterly-hour time slices (EPEXSPOT, 2016). After the intra-day market clearing, each BRP can submit intra-day nominations to the TSO on a quarter-hourly basis until the day after delivery. The prolongation aims to give market players additional time to close financial transactions. The intra-day nominations are added to the day-ahead nominations of the BRP. Other than in the day-ahead market, the BRP portfolio can be imbalanced after the intra-day market. The portfolio imbalance is dealt within the balancing market.

2.4 Balancing market

Using the three previously illustrated electricity markets, the individual BRP can generate a quarterly hour-schedule for the power fed into the grid. For deviations from the schedule, meaning imbalances between the total injections and offtakes, there is a separate market: the balancing market^[1] or operating reserve market. Thereby the TSO maintains the system balance by activating reserves. This mechanism balances out short-term fluctuations within seconds up to fifteen minutes preventing the grid frequency to deviate from its target value of 50 Hertz (Hz) and to cause an outage. Furthermore the balancing market can be subdivided in balancing service provision before the time of delivery and balance settlement, where the financial transactions between two parties are being settled. The auctioning of balancing reserve starts one week up to one day before the actual delivery.

This thesis will focus on the balancing provision side rather than on the imbalance settlement. The following chapter intends to take a closer look at the technical and economical conditions of the balancing market.

3. Operating reserve market design in continental Europe

The balancing market serves as a technical instrument on the one hand. On the other hand, due to the fact that without the balancing market occurred outages would lead to high costs, it has as well a high economic relevance. These opportunity costs may be seen as a potential value of the operating reserves (Next Kraftwerke, 2015). In the following, both the technical and economic conditions of the operating reserves for continental Europe are going to be observed.

3.1 Power-frequency control by the TSO

In addition to the provision of network infrastructure for the electricity transport, one of the main tasks of TSOs is to provide ancillary services in order to safeguard the power supply system.^[2] Amongst these system services, the so-called performance-frequency control stands out, not only because of its technical complexity, but also because of the significant cost-relevance and the interactions with electricity generation and electricity distribution.

The stable operation of the power supply system implies that the power balance of feed-ins, withdrawals and losses in the overall system is balanced at all times or is returned to the equilibrium state after deviations within a few seconds. Excessively injected power cannot be stored directly as such, unlike within a gas supply network. In general, indirect storage, for example by pumping water from a basin into pumped storage plants or by other storage techniques, is possible. In the present power supply system, however, such procedures are rarely implemented. Therefore, all participants in the power supply system rely on a permanent supervision of the power balance and suitable regulation in case of a deviation. In order to do so, the control systems must have access to controllable feed-ins or consumption devices.

From a technical viewpoint, the performance balance is guaranteed by maintaining the net frequency within a narrow band around the target value of 50 Hz. Due to several unforeseeable factors such as forecast deviations of renewable energies caused by weather, technical disturbances in the transmission or distribution grid or power plant outages, imbalances cannot be avoided by precise forward planning. Therefore the active continuous regulation of the power balance is absolutely necessary for the supply system stability and resides with the TSOs. Hereby, every TSO manages one control area by continuously deploying and coordinating operating reserves. This concept of centrally organized control responsibility by the TSO has proven to be the only practice-oriented and efficient approach, as it requires far less control engineering and control reserves than the theoretically conceivable concept of regulatory responsibility within individual utilities (Consentec GmbH, 2014).

Within a synchronously interconnected system, which coincides with the control area, in order to enable a source-based allocation of the regulatory costs, power suppliers and electricity traders form so-called balancing groups, in which the feed-ins and the withdrawals of the consumers are bundled.^[3] Each feed-in, withdrawal and electricity trade quantity inside a control zone must at all times be assigned to a balancing group in order to ensure complete balancing. The management of each balancing group lies with the previously mentioned Balance Responsible Party (BRP). The relation and balancing responsibility between the operational entities is visualized in Figure 2.

illustration not visible in this excerpt

Figure 2: Operational distribution within a synchronous area

The synchronous area, assuming that it coincides with national borders, complies with the control area, whose balancing lies with the TSO. The control area comprises balancing groups, which themselves consist of one or more balancing areas (BA). The management of each balancing group lies with a BRP (after Doorman et ah, 2013).

On completion of each delivery month, the TSO responsible for a control area determines a balance equilibrium for each balancing group and each quarterly-hour period. The BRP are obliged to balance the generation and consumption in their balancing group for every quarterly hour. In case of forecast deviations, balancing energy is required. The net sum of all imbalances within a regulation zone yields the total control area imbalance, which is balanced out by the deployment of operating reserves by the TSO. Therefore Balancing Service Providers (BSP), which are for example power plant operators, submit operating reserve bids to the TSO.^[4] The economic conditions and procedure of the provision of operating reserves is discussed in Chapter 3.3.

The amount of reserves activated by the TSO is called the Net Regulation Volume (NRV). A positive NRV corresponds to grid injections and a negative NRV requires an increase in grid off­takes. The operating reserves used within a balancing group are settled by the respective TSO with the BRP on a quarterly-hour basis. The imbalance settlement takes place after the actual delivery and is based on two prices: the marginal incremental price (MIP) and the marginal decrementai price (MDP). The first one is the highest price paid by the TSO for upward activations for a given quarter-hour. The latter is the lowest price received by the TSO for downward activations for a given quarter-hour. In case of a positive NRV in the whole control area, the TSO buys upward regulation and pays the reserve provider for the extra generation. The most expensive unit of upward regulation sets the MIP. When negative NRV is needed for the control area, the TSO receives a payment from the reserve providers who are willing to reduce their output and avoid generation costs. The last unit of downward regulation sets the MDP (KU Leuven Energy Institute, 2015b). Depending on the imbalance of a separate balancing group and the net regulation volume in the control area, four different price categories emerge, visualized in Table 1. The markups are zero for small imbalances and become nonzero for larger ones.

Table 1: Overview of imbalance settlement between TSO and BRP

illustration not visible in this excerpt

The price and the direction of the transaction depend on the combination of the NRV in the whole control area and the BRP imbalance by comparison. In case of a negative NRV the MDP is paid and in case of a positive NRV the MIP is paid respectively. The direction of the payment depends on the imbalance of the BRP (source: KU Leuven Energy Institute, 2015).

On the European level, the regulatory framework for the load-frequency control results from so- called Network Guidelines and Network Codes on the subject of “Load Frequency-Control & Reserves” and “Electricity Balancing”. The regulatory requirements and foundation of operating reserves is further discussed in Chapter 4.

3.2 Operating reserve types and technical conditions

In the power system, small disturbances of the system balance occur continuously, for example due to the stochastic, uncoordinated feed-ins and feed-outs of network users. However, since the generation capacity of power stations as well as the abstraction of electrical loads by consumers can only be adjusted with a delay, the current balancing is made automatically and exclusively by the kinetic energy of oscillating weights in the compound system. Due to the frequency-constant network coupling of synchronous generators, this process is accompanied directly by a drop or an increase in the net frequency, also known as the inertia of frequency-dependent load. This self­regulating effect is just for short-term compensations (Swider, 2006). For deviations above 10 mHz^[5], operating reserves has to be activated by the TSO.

For this purpose, each system-responsible TSO operates its own power-frequency controller, which permanently measures the power balance and the frequency of the control area and compares them with the corresponding reference values. In the event of deviations, reserves are activated in order to reconcile reference and actual values.

Depending on whether there is an upward or downward deviation of the grid frequency, positive or negative operating reserves are considered. Positive operating reserves result in the ramp-up of power plants to balance out a generation deficit. Negative operating reserves on the other hand imply the shutdown of power plants in order to balance a demand deficit by storing or restraining power (Consentec GmbH, 2014).

One distinguishes three different categories of reserve products, based on the temporal sequence of activation: primary reserves, secondary reserves and tertiary reserves (ENTSO-E, 2013a). Figure 3 visualizes them depending on the activation time.

The primary reserves, also known as frequency containment reserves (FCR)^[6], are used to stabilize the frequency within the time frame of seconds by means of automatic controlled and locally activated reserves. The activation of primary reserves starts automatically after a few seconds and is terminated at the latest 30 seconds after the disturbance appears. It is implemented via decentralized speed controllers at all participating power plants in the pan-European interconnected network. The level of regulation is proportional to the deviation from the target frequency, also called proportional control. The provider of primary reserve measures the network frequency independently at the location of the generation or consumption and reacts directly to frequency deviations (Next Kraftwerke, 2015). Thus, temporal losses, for example due to communication, are avoided in order to effect a rapid compensation. The control range of primary reserve is between 49.8 and 50.2 Hz. The activation starts when the frequency falls below 49.99 Hz or exceeds 50.01 Hz respectively and the provider commits to counteract the frequency. The primary reserve can offset the occurring power imbalance and stabilize the system and the net frequency at a new point. Nonetheless quasi-stationary frequency deviations still remain (Consentec GmbH, 2014).

illustration not visible in this excerpt

Figure 3: Activation of different operating reserve products based on time

Small disturbances are automatically balanced out by kinetic energy. For deviations above 10 mHz, primary reserve is activated within seconds. After 30 seconds, secondary reserves are activated. In case of a drastic imbalance, tertiary reserves are activated after 5 minutes. At the latest one hour after the deviation, the balancing mechanism is terminated and the responsibility is passed to the BRPs (after Next Kraftwerke, 2015).

The secondary reserves or automated frequency restoration reserves (FRR)^[7] are used to balance out these deviations and restore the system balance. Frequency restoration reserves must be provided within five minutes. Therefore, all participating vendors are connected to the control room of the respective TSO via a communication link and exchange data in real time. The
previously mentioned power frequency controller of each TSO distributes the secondary reserve demand on the connected units automatically. Consequently the secondary balancing reserve is activated centrally and source-based by the TSO. In comparison to the primary reserves, the secondary reserves are regulated targeting the exact target frequency. For instance, in case of a power plant outage in a control area, positive secondary reserves are activated until the frequency reaches its reference value and the primary reserves are deactivated. Only after the detachment of the primary reserve, it is available again to control the occurrence of further disturbances. Nevertheless, major imbalances lasting longer than 15 minutes cannot be compensated by any of the two reserves.

In such a case, the tertiary reserves, also known as minute reserves or replacement reserves (RR) are used. They allow the secondary reserves to return to their pre-imbalance status preparing them to be ready for the next short-term imbalance intervention (ENTSO-E, 2013a). Tertiary reserves are activated after five minutes and are completed within 15 minutes. They are controlled manually by the TSOs within a 15-minutes-schedule and activated locally by direct communication. Activation depends occasionally on the actual utilization of secondary reserves and their foreseeable development. The goal is, to replace secondary reserves that have been activated over a longer period of time. In some cases, tertiary reserves may be activated preventatively in order to compensate expected larger imbalances. Figure 4 illustrates the functions and interactions of the three different operating reserve categories.

As shown in Figure 3, at the latest one hour after the deviation, the balancing mechanism is terminated and the responsibility is passed to the Balance Responsible Parties.

In order to provide operating reserves, their potential supplier first have to pass a technical prequalification for each type of balancing energy (Verband der Netzwerkbetreiber VDN e.v., 2003). In addition, the proper performance of operating reserves under operational conditions and economic performance of the potential provider must be ensured. The implementation of a prequalification procedure usually requires a period of at least two months. An indispensable part of the prequalification is operating reserve activation trial run. Furthermore, the prequalification also serves to verify control and communication technology connections, for example to the power frequency controller of the TSO. As soon as the prequalified service exceeds the respective minimum bid size, the TSO makes a framework contract with the provider for each type of operating reserve, which is a precondition for the participation in the tender procedure.

Operating reserves have mostly been supplied by easily controllable and fully automatic power stations such as pumped storage power plants or gas turbines. However, recently especially due to an adaption in the prequalification conditions, virtual power plants composed of biogas plants or cogeneration plants also contribute to the operating reserve supply, especially in regard to the tertiary reserves.

illustration not visible in this excerpt

Figure 4: Overview deployment and functions of different operating reserves

Primary reserves are automatically activated though a frequency imbalance above 10 mHz in order to limit deviations from the target frequency. Secondary reserves are activated by the TSO, in order to replace the primary reserve and return the frequency to 50 Hz. In case or large deviations, TSO manually activate tertiary reserves that free up secondary reserves (after Consentec GmbH, 2014).

3.3 Economic conditions

Due to the high relevance of balancing markets and the previously mentioned value resulting from opportunity costs, this subchapter deals with the economic aspect of balancing markets, focusing on the balancing service provision period between TSOs and BSP.

A prequalified power plant owner or BSP places a bid in a pay-as-bid auction. This means, in case a bid is accepted by a TSO, the BSP receives his actual bid as a payment. In contrast to the Multi-Unit-Auction this auction type does not determine a single market-clearing price for all power plant owners. Here, all accepted bids are placed in an ascending merit order and one by one the cheapest are selected until the balancing reserve level is reached. The pay-as-bid auction ensures that the provision and delivery of secondary reserves remain as cheap as possible for the TSOs, as cost-effective installations in case are preferred by virtue of the merit order list. This also means that the plants with the most expensive prices, which are far behind in the merit order list, are rarely called up, and are therefore often not included in the cost of control. Thereby, provision and delivery are remunerated separately within a two stage process (Wulff, 2006). First, the pay-as bid auction for the provision of operating reserves takes place with regard to the capacity price, which remunerates a power plant operator only for the provision of positive and negative balancing reserves. This payment compensates on the one hand the opportunity costs, which result from the absent offer in other power markets and on the other hand the sale revenue loss resulting from the suboptimal operation of the power plant. After the operating reserve providers are selected, in a second step, the accepted providers have to bid their activation prices in another pay-as-bid auction. Here, again, the bids are placed in a merit order list and the most cost-effective are selected. The activation price remunerates the power plant operator for the actual delivery of positive or negative balancing reserves including the fuel input and ramp-up and ramp-down costs. For all three balancing reserve types, the decision whether a bid by a power plant owner is accepted is based on the capacity price.

Within the auction of primary reserves, there is no distinction between positive and negative capacities and is therefore called a “symmetric product”. Every prequalified provider, which has a framework agreement with a TSO can make a bid and participate in the auction. A bid contains the amount of offered service and the capacity price. Due to the fact that on average the ratio between the positive and negative power provided is balanced out and consequently as much electrical power is fed into the grid as additionally obtained, only the capacity price is paid. Furthermore a constant frequency balancing would lead to high transaction costs when billing the activation price, since it is activated frequently. The auction takes place on Tuesdays for the following week and the resulting performance price applies for the whole week without differentiation between the days or hours. The power plant owners whose bids were taken are obliged to provide the capacities.

At a secondary reserve auction, however, power plant owner bid for positive and negative capacity prices and for positive and negative activation prices. Additionally there is a distinction between peak load for the hours 8 am till 7 pm and off-peak time periods for the hours 8 pm till 7 am. Hence, eight different prices result. The auction for peak and off-peak hours takes place on Wednesdays for the following week.

illustration not visible in this excerpt

Figure 5: Trade and delivery of different operating reserve products for one week

The beginning of the black arrows visualizes the tendering phases, while the end of the arrows symbolizes the delivery. The primary reserve auction takes place on Tuesdays for the following week, the secondary reserve auction takes place on Wednesdays for the following week and the tertiary reserve auction takes place every day for the following day and therefore coincides with the day-ahead market (after Eichhorn, 2013).

Likewise at the tertiary reserve auction, four bids for positive and negative reservation and activation prices are set. The tendering takes place daily until 10 am for six different four-hour time slices for the next day. Therefore it is a day-ahead consideration as shown in Figure 5. The greater product differentiation and the shorter call-up periods for tertiary reserves make the product more attractive and manageable for smaller producers.

Due to the similar tendering and delivery time periods, there are interactions between the tertiary reserve market and the day-ahead market. Additionally one may assume an interaction with the intraday-market, since the tendering time period and the time slices are even shorter than in the tertiary reserve auction. But on the basis of the fact that the latter auction closes at 10 am, the day-ahead auction closes at noon and the intra-day market opens at 3 am the day before, one can assume that there is a significant independence between the intra-day and the tertiary reserve market. Thus, the intraday-market is relevant for larger deviations than are known upfront.

Table 2 gives an overvieพ of the most important technical and economical properties of the primary, the secondary and the tertiary reserves.

illustration not visible in this excerpt

Table 2: Overview of the technical and economical conditions for balancing reserves

3.4 Operating reserve planning

Reserve markets in Europe consist of three different phases. The first phase, known as the dimensioning phase (also referred to as the sizing phase), serves to determine the amount of reserves required in each control area by the responsible TSO. It takes place weeks to months before the actual activation. Thereby, a distinction is drawn between primary reserves, which are provided and activated in the whole network and the secondary and tertiary reserves that are coordinated and deployed in the separate control areas. The amount of primary reserves is determined on a European level and per synchronous area and then allocated to the respective TSO in its area. In accordance with the regulations of the ENTSO-E Operation Handbook, 3,000 MW of primary reserves have to be provided for the continental European synchronous network (ENTSO-E, 2013a). This determination results from the goal of being able to master two overlapping reference events with the primary reserves provided, whereas the reference event is the largest expected performance disruption due to a sudden failure of one of the largest power plants operating in the network. Currently those are large nuclear power plants with an output of approximately 1,500 MW yielding the magnitude of the overall primary reserve demand. Each control area has to provide a share thereof, corresponding to their share of the total power generation in the network (Regelleistung.net, 2017). The guidelines provided by the ENTSO-E regarding the sizing of secondary and tertiary reserves, however, are less substantial.

Consequently, the dimensioning varies within the different European market zones. After the determination of the required reserves, each TSO procures reserves in the reserve market from providers such as conventional generators or large consumers in order to meet the reserve requirement for its control area. This stage is called the procurement phase (also sometimes referred to as the allocation, scheduling or reservation phase) (Van den Bergh et ah, 2016). As described in the previous chapter, reserve providers are remunerated for the reserve capacity they sell to the TSO by the reservation payment compensating for their opportunity costs. The procurement phase happens after the dimensioning phase but before the actual delivery and simultaneously with the day-ahead energy market clearing (Van den Bergh et ah, 2017). Lastly, within the activation phase (also referred to as the deployment phase) the TSO activates the reserves in real-time to maintain the system balance. As discussed above, reserve providers of secondary and tertiary reserves are remunerated by means of an activation payment to cover the cost of activating reserves, whereas the activation of primary reserves is not rewarded with an activation payment, as upward and downward activation and therefore back and forth transactions between the TSO and the power plant operator offset each other.

In each of the three phases, separate market zone can coordinated their described activities. The next chapter serves to discuss possible types of coordination in continental Europe.

4. Cross-border coordination in reserve markets

4.1 Benefits of cross-border reserve markets

Due to the concept of source-based and control-area specific use of secondary and tertiary reserves, opposite activation of operating reserves in neighboring control areas may occur. For instance, a positive reserve could be activated in one control area to compensate for a power plant outage, while a negative reserve is activated in an adjacent control area in order to adjust for an overestimated load prediction. Instead, both control zones could agree upon additional power exchange from the area with power surplus into the one with power deficit and save the operating reserve activation. This would lead to grid control cost minimization of the overall system and consequently to an overall welfare increase for the whole network area, respectively Europe. In addition, the coordination of balancing activities requires standardizing the balancing energy and capacity products, to harmonize the main features for imbalance settlement and to facilitate the cross-zonal exchange of balancing energy beforehand (50Hertz, 2014). To what extent this happens, depends on the observed market design of reserve market coordination. Consequently potential options for cross-border cooperation in operating reserve markets, which are also considered by the responsible authorities, will be discussed.

4.2 Principal options for reserve market coordination

In cross-border reserve markets, TSOs can activate and/or procure reserves in other control areas. If activation is coordinated, scheduled reserves in one market zone can be used to cover forecast errors in another market zone. If procurement is coordinated, conventional generation units can be scheduled in one market zone in order to meet reserve requirement in another market zone. Finally, in coordinated reserve dimensioning, multiple market zones determine their aggregate need for the reserves together. Nonetheless, not all possible combinations of coordination across the three phases are reasonable. For instance, when procurement and activation are not coordinated, coordinated reserve dimensioning is not useful (Van den Bergh et ah, 2017). Due to the top-down characteristic of the three consecutive phases, only combination of coordinated markets from bottom to top are meaningful, as shown in Table 3. Scenario A represents a fully uncoordinated network, whereas scenario D represents on the contrary a full coordination between different control areas.

illustration not visible in this excerpt

Table 3: Overview of the reasonable scenarios for cross-border coordination in reserve markets

Only coupled cooperation possibilities from the bottom up are meaningful. Scenario A represents fully uncoordinated control areas, whereas scenario D mirrors complete integrated balancing markets (source: Van den Bergh et ah, 2017).

Reserve markets in Europe consist of the same market phases and reserve categories. Nonetheless, not negligible differences exist concerning the implementation. On the technical side, there are different technical product definitions, procurement procedures and a different overall balancing approach (50Hertz, 2014). On the economic side, the bid selection and activation, as well as the settlement and imbalance pricing vary across different market zones (Verpootten et al. 2016). In order to couple reserve markets, one needs to identify the differences and develop possible approaches to ensure their compatibility, which is in accordance with the regulatory target of an overall harmonization of European electricity markets. In principle cross­zonal cooperation could be distinguished into options that require harmonization by the TSO and options that do not require any harmonization. Furthermore, a differentiation is made between the common usage of two balancing service types: balancing energy and balancing capacity.^[8] Balancing energy relates to the real-time adjustment of balancing resources to maintain the system balance. Balancing capacity on the other hand refers to the contracted option to dispatch balancing energy during the contract period (van der Veenet ak, 2016).

The four most relevant market designs for cross-border coordination, also considered by the ENTSO-E in its network code and discussed below are (ENTSO-E, 2013b, 2014b): (i) Imbalance Netting; (ii) Common Merit Order; (iii) Exchange of Reserves; and (iv) Reserve Sharing. Imbalance Netting aims exclusively to avoid counteracting deployments of balancing energy. This measure is useful whenever the system imbalance, thus also the balancing need in one control area is opposite to the one in another (Hewicker et ah, 2013). Therefore the TSOs exchange opposite imbalances between control areas in real-time, before any demands are detected and reserves are activated. Neither reserve capacity nor cross-border capacity is reserved for this purpose. The total amount of reserves activated decreases, resulting in cost reduction. This type of cross-border coordination entails the lowest technical complexity and is relatively easy to implement (Consentec GmbH, 2014). Imbalance Netting has an impact on the activation phase of secondary reserves. Assuming that, for example, we have two control areas A and в with respectively one responsible TSO in each. If TSO A is regulating down whilst TSO в is regulating up at the same time, TSO A can export its positive imbalance to TSO в and avoid the use of downward regulation whereas TSO в reduces its need for upward regulation. Figure 6 illustrates the potential benefits.^[9]

illustration not visible in this excerpt

Figure 6: Potential benefits of Imbalance Netting with two control areas

On the horizontal line, the time since an imbalance occurred is plotted and the graphs represent the activated secondary reserves. Without Imbalance Netting, in control area A, TSO has to activate negative reserves, while TSO В has to activate positive reserves in order to deal with a generation deficit. The blue line represents the amount of activated secondary reserves in both control areas. Instead, TSO A can export its positive imbalance energy to TSO B, avoiding the use of downward regulation, whereas TSO в reduces its need for positive balancing energy. The yellow line represents this spatial smoothing (source: Hewicker et. al, 2013).

Secondly, TSOs in different control areas can activate reserves in real-time using a Common Merit Order list, as shown in Figure 7. As such, all selected bids of the participating control areas are involved and the most cost-efficient reserves can be activated, regardless where they are located. More attractive bids from one country would thus replace the activation of less attractive bids from another country, which would otherwise be activated in case of separate merit orders. This would result in a reduction in the aggregate costs of balancing energy for both countries. Moreover, for all involved control areas, the total amount of procured operating reserves is available, ensuring a mutual support in case of a reserve shortage and avoiding a counteracting deployment of reserves. The Common Merit Order has an impact on the activation phase and can be applied to secondary and tertiary reserves. Establishing a Common Merit Order list for balancing energy requires a fair amount of harmonization between the participating control areas: the procurement procedures and time frames need to be harmonized as well as the products. In addition, the TSOs have to align on activation and settlement procedures. The implementation of a Common Merit Order list for balancing energy is a precondition for the common usage of balancing capacity and the consideration of additional bids for procurement of balancing capacity (50Hertz, 2014).

[...]

---

^[1] In the draft Network Code on Electricity Balancing, the balancing market is defined as the „entirety of institutional, commercial and operational arrangements that establish market-based management of the function of „balancing“. Thereby „balancing“ is defined as „all actions and processes, on all timelines through which TSOs ensure, in a

^[2] Ancillary services are a range of functions of TSOs to guarantee system security, maintain grid stability and support the continuous flow of electricity so that supply will continually meet demand. These include black start capability (the ability to restart a grid following a blackout), frequency control, the provision of reactive power and various other services.

^[3] A balancing region itself consists of one or more balancing areas.

^[4] Figure Al in the Appendix gives a more detailed overview of the balancing market basic structure and the relations between the different actors.

^[5] 10 Millihertz (mHz) equals 10-2 Hz. Therefore the inertia of frequency dependent load applies to the range of 49.99 to 50.01 Hz.

^[6] Within this thesis for reasons of simplification the terms primary, secondary and tertiary reserves are used. Nonetheless, in order to align with other literature and regulatory guidelines provided by institutions and policymakers, it is important to mention the other terminologies in the context: frequency containment reserves, frequency restoration reserves and replacement reserves.

^[7] There is a further distinction between automated and manual frequency restoration reserves (aFRR and mFRR). Thereby the manual frequency restoration reserves are assigned to the tertiary reserves.

^[8] Capacity is measured in megawatts (MW), whereas energy is measured in megawatt hours (MWh).

^[9] Figure A2 gives an Imbalance Netting example.