Capacity Calculation

Abstract of Chapter 3: Capacity Calculation

Chapters 1 and 2 laid out the framework and guiding principles of the European electricity market. In this chapter, the basics of capacity calculation are discussed, i.e. how to determine how much (cross-zonal) transmission capacity is available for commercial exchanges. The geographical organization of the European Union into capacity calculation regions (CCRs) is discussed, and the succession of timeframes over which capacity is calculated. The capacity calculation process is described in detail: the inputs required, actors involved, and the different methodologies of capacity calculation used/applied in Europe. Then, the day-ahead capacity calculation methodology applied in Core CCR is explained through several important steps and concepts: non-costly remedial action optimization, individual validations, allocation constraints, evolved flow-based, advanced hybrid coupling. Finally, the future evolution of capacity calculation in Europe is explored.

3.1 Basic Principles of Capacity Calculation

As explained in Chap. 1, the amount of power that can be traded between bidding zones depends on the transmission capacity made available for cross-zonal exchanges. Cross-zonal capacity calculation is the process of determining cross-zonal capacities for market time units (MTUs) in the future. As market outcomes are not known, the process of computing cross-zonal capacities is based on forecasts of the market outcome, grid topology, generation, and load.

Capacity calculation is one of several available measures for congestion management. Other measures available to TSOs include:

• Short-term: Application of remedial actions (including redispatch and countertrade—both of which might be used in conjunction with the cross-zonal capacity calculation)
• Long-term: Adaptation of the bidding zone configuration (ideally, to reflect structural congestion)
• Long-term: Development of the network (reinforcing existing network elements and building new ones).

The capacity calculation process takes place in the environment provided by the grid, the bidding zone configuration, and the CCR determination (ACER, 2024).

In the context of this chapter, “capacities” are referred to as “cross-zonal capacities” rather than “cross-border capacities”. This highlights that even in countries(or control areas) with multiple bidding zones, the same set of processes will apply for capacity calculation.

Furthermore, an “oriented bidding zone border” refers to a bidding zone border with a defined direction (e.g. A to B, but not B to A). In contrast, a “bidding zone border” is generally not oriented (i.e. A to B or B to A refer to the same bidding zone border).

“Net position” refers to a bidding zone’s netted import and/or export (commonly assumed to be negative when a bidding zone imports and positive when a bidding zone exports).

This chapter focuses on the provision of cross-zonal capacities by TSOs and RCCs that serve as boundaries for the market coupling algorithm, as illustrated in Fig. 3.1.

3.1.1 Geographic Organization

As noted by European Commission (2015) in the CACM Guideline (see Chap. 1) requires that cross-zonal capacities are computed in a coordinated way based on a common set of assumptions. The groups within which TSOs cooperate to compute cross-zonal capacities in a coordinated manner are called

“Capacity Calculation Regions” (CCRs). These CCRs are defined as sets of bidding zone borders (note:
not as bidding zones or TSO control areas). Hence, one bidding zone or TSO can be a member of several CCRs, while each bidding zone border is associated with only one CCR.

As an example (see Fig. 3.2), Poland is a member of three CCRs—Core, Baltic, Hansa—while each bidding zone border is attributed to only one CCR (PL—DE/ LU, PL-CZ, PL-SK to Core, PL-SE4 to Hansa, PL-LT to Baltic).

Fig. 3.1 Organization of the European electricity markets. Source: Author’s own illustration based
on CACM Guideline

Ideally, CCRs are set up and defined such that there is no significant impact from one CCR on another in terms of commercial exchanges. This, however, proves to be a challenge, especially in the heavily meshed Continental European power system.

Bidding zone borders involving (at least one) non-EU TSOs are not attributed to any CCR (e.g. bidding zone borders of Switzerland and West Balkan countries). The only exceptions are the Norwegian bidding zone borders after ACER decision 2023¹.

Since Brexit, the former CCRs “Channel” and “Ireland and United Kingdom” were dissolved. Consequently, the SEM bidding zone (comprising the TSOs EirGrid in the Republic of Ireland and SONI in Northern Ireland) was not attributed to any CCR. However, the Celtic Interconnector, a HVDC interconnector, which is being constructed between Ireland and France and is expected to be operational by the end of 2026, will create a new bidding zone border between France and Ireland (FR-SEM). In 2023, all TSOs publicly consulted an amendment to the CCR configuration in light of the Celtic Interconnector, assigning FR-SEM to Core CCR. ACER approved this amendment on 19/03/2024 with decision 04/2024.

Additionally, this decision establishes the new Central Europe (CE) CCR as the merger between the CCRs Core and Italy North, with the initial delivery of a day-ahead capacity calculation methodology (DA CCM).
The current CCR configuration, based on the latest ACER Decision on CCRs, is shown in Fig. 3.2.

Each CCR has established governance and working structures for developing, implementing, and operating the cross-zonal capacity calculation methodologies and processes (a few examples can be found in Sect. 3.2). Some CCRs have established special points of contact for engaging with stakeholders. CCR-specific information can be found on the (ENTSO-E) website and the (JAO) website.
ENTSO-E also publishes an interactive Capacity Calculation Regions map.

Fig. 3.2 Capacity Calculation Regions. Each bidding zone border is attributed to exactly one CCR. Each CCR is composed of all the bidding zone borders of the same colour. Source: Authors own illustration based on ACER decision 04/2024 on CCR determination

3.1.2 Timeframes

Cross-zonal capacity calculation is performed for all timeframes for which allocation of cross-zonal capacity is performed: Year-ahead (once per year), month-ahead (once per month), day-ahead (daily), intraday (one to three times per day), and balancing (at least daily).

While day-ahead and intraday cross-zonal capacity calculation are based on common grid models (as also described in Chap. 5) that include forecasts (also referred to as “scenario-based approach”), long-term cross-zonal capacity calculation may be performed on a statistical basis (i.e. without determining cross-zonal capacities using a common grid model as a basis). Day-ahead and intraday cross-zonal capacity calculation processes are relevant in market coupling.

The day-ahead capacity calculation process typically starts in the evening two days before delivery time and ends in the late morning one day before delivery (hence the name day-ahead). The intraday cross-zonal capacity calculation process may run several times one-day-ahead of delivery and/or during the day of delivery.

Regardless of timeframes, cross-zonal capacities are to be provided at Market Time Unit (MTU)-granularity. The MTU granularity for day-ahead and intraday capacity calculation is currently 60 min. Note, however, that the cross-zonal computation process itself (i.e. the computations) may run on a different granularity than the granularity of cross-zonal capacity allocation (e.g. capacity allocation
might run on 15-min granularity, while cross-zonal capacities are computed on 60-min granularity).

The Transparency Regulation (European Commission, 2013) requires the day-ahead cross-zonal capacities to be published at the latest one hour before SDAC (that is, at 11:00 CET D-1, coinciding with the day-ahead firmness deadline²). Intraday cross-zonal capacities must be published 15 min before SIDC gate opening time at the latest.

3.1.3 Inputs to the Capacity Calculation Process

Executing the cross-zonal capacity calculation process requires (Fig. 3.3):

1. a methodology and process to be followed.
2. relevant input data (typically provided by the TSOs of a given CCR).

Fig. 3.3 Inputs to cross-zonal capacity calculation.³ Source: Authors’ own illustration based on
CACM Guideline, Article 21

The CACM Guideline lists several inputs that the capacity calculation methodologies shall address and detail further:

Individual grid models (IGMs):

• Network models are essential in a scenario-based capacity calculation process. TSOs provide IGMs for different timeframes, and for the sake of capacity calculation, these IGMs always contain a forecast of the expected market outcome (not a fully realized market result). The market outcome might be expressed as (a set of) net positions and/or reference exchanges. Some CCRs use a common forecast, while others rely on the forecasts of individual TSOs.
• IGMs reflect a forecast of the grid topology, power generation, and consumption within the TSO’s control area, and flow on HVDC interconnectors. These IGMs are then merged into a common grid models.

Critical Network Elements (CNEs) under N-1 Contingency (CNECs)

• Previously referred to as Critical Branches and Critical Outages (CBCOs), CNEs and CNECs can constitute transmission lines (both overhead and underground), transformers and HVDC equipment, such as HVDC lines or back-to-back converters. They are provided with some basic characteristics, such as rated voltage level(s) and the maximum permissible current (which can be dynamic and vary over time, e.g. in the case of dynamic line rating). Within a flow-based approach for capacity calculation (Sect. 6.4), CNECs are typically defined via a name (one per direction of power flow), thermal current or power flow limit (Imax / Fmax), and flow reliability margin (FRM, see next point).

Reliability margin (RM)

• To cater for forecast errors, due to issues such as the linear and simplified nature of capacity calculation processes and activations in the balancing timeframe, the available (thermal) capacity may be reduced by a reliability margin (RM). In the various capacity calculation methodologies, RMs
• might be referred to as flow reliability margin (FRM), total reliability margin (TRM), or similar.

Generation and Load Shift Keys (GLSKs)

• GLSKs are used to “transpose” a quasi-nodal power system to a single net position of a bidding zone. GLSKs are a list of factors that translate a change in net position (e.g. increased import or export) to the nodal injections or withdrawals. For example, if a given net position changes by 100 MW, the GLSK “maps” these 100 MW to individual nodes in the grid supplying/ withdrawing this additional power (see Fig. 3.4). The sum of the GLSK factors in each bidding zone is equal to 1. It is worth noting that GLSKs themselves do not consider minimum and maximum generation or load limits of generators or loads. Additionally, there is only one GLSK per MTU and bidding zone, regardless of how large the shifts in net positions are. Since GLSKs are linear (their factors remain static), the approximate shifts in net positions are most accurate around the reference point for which they were created, but less accurate with large net position shifts far away from the reference point.

Remedial Actions (RAs)

• This input refers to actions that TSOs can take to redirect power flows in the grid. As their naming suggests, RAs are deployed to alleviate congestion on network elements. In cross-zonal capacity calculation, the (coordinated) use of remedial actions is expected to increase cross-zonal capacities.
• Remedial actions may include topological measures (e.g. opening and closing bus-bar couplers), tapping of phase-shifting transformers, or redispatch (orders to individual power plants or loads to increase or decrease power infeed or withdrawal). Where applicable, the list of remedial actions might be complemented by countertrading. While topological measures and PST (Phase Shifting Transformer) tapping are generally considered non-costly remedial actions, redispatch, and countertrading are regarded as costly remedial actions.
• It should be noted, however, that remedial actions are not actually dispatched or deployed at the cross-zonal capacity calculation stage; this happens at a later stage when market results are available and firm, and consequently, the actual state of the grid (in terms of power flows and potential overloads) can be assessed.

Allocation Constraints

• Allocation constraints are additional constraints used when operational security limits cannot be effectively translated into maximum flows on CNE(C)s. Such constraints are to be respected during cross-zonal capacity allocation to maintain the transmission system within operational security limits.
• The term “allocation constraint” is an umbrella for different designs of such constraints. Two common types of allocation constraints are limitations on net positions and ramping constraints for HVDC interconnectors. Net Position Constraints are used to limit the maximum import and/or export of a certain bidding zone. This limitation can be applicable on either the SDAC level (i.e. a limitation of the overall net position during cross-zonal capacity allocation) or CCR level (i.e. a limitation of the “regional” net position within a certain CCR, e.g. in a CCR applying flow-based capacity calculation and allocation). Ramping constraints provide a maximum gradient for how much power flows on HVDC interconnectors can change from one MTU to the next. As cross-zonal capacity calculation is commonly carried out for each MTU separately without considering preceding or succeeding MTUs, such dedicated ramping constraints are provided as a separate input to the market coupling algorithm, in which consecutive MTUs are coupled.

The above-listed inputs are each to be provided by TSOs. The Coordinated Capacity Calculator (CCC) then combines these inputs and creates “merged” versions containing the inputs of all individual TSOs. The CCC might further receive information on previously allocated capacity (e.g. from long-term allocation) and/or individual grid models from TSOs not part of a given CCR.

Fig. 3.4 A Individual grid models of three bidding zones A, B, and C. B Generation and Load Shift Key, showing the contribution of three nodes to a net position increase of 100 MW in Zone A. Source: Based on author’s own illustration.

3.1.4 Performing Cross-Zonal Capacity Calculation

Generally, when using a scenario-based approach, the process for cross-zonal capacity calculation consists of the following steps (see Fig. 3.5):

1. Input provision: Provide individual grid model and other input data (see
   Sect. 3.1.3) for capacity calculation.
2. Merging of inputs: Merge individual grid models into a common grid model
   and merge individual TSO inputs to CCR-wide inputs.
3. Capacity calculation: Compute cross-zonal capacities based on CCR and time
   frame-specific methodology and perform intermediate steps (e.g. optimization
   of remedial actions, measures to ensure minimum capacities, etc.).
4. Validation of cross-zonal capacities: Each TSO has the right to validate and
   potentially adjust cross-zonal capacities.
5. Re-calculation of capacities: As a result of the validation step, cross-zonal
   capacities might have to be re-computed or adjusted. Furthermore, if applicable, previously allocated capacities (e.g. long-term nominations, cross-zonal
   capacity allocated for exchanging balancing capacity) are considered.
6. Publication of results and provision to market coupling.

Steps 2), 3), 5), and 6) are performed by the Coordinated Capacity Calculator, while steps 1) and 4) are carried out by individual TSOs.

Fig. 3.5 High-level steps in the capacity calculation process. Source Author’s own illustration
based on CACM Guideline, Article 21

3.1.5 (Coordinated) Net Transfer Capacity Calculation

The coordinated net transfer capacity (cNTC) approach yields cross-zonal capacities as NTC (or Available Transfer Capacity (ATC)) values per oriented bidding zone borders. That is one value for one direction and another for the opposite direction (Fig. 3.6).

Fig. 3.6 Example of NTC domain. Source: Based on author’s own illustration

In its simplest (and uncoordinated) form, NTCs are computed by each TSO individually before these NTCs are shared with adjacent TSOs. Then, the lower value stemming from either TSO on a given oriented bidding zone border is provided to market coupling. The disadvantage of this approach is that TSOs do not mutually share information on the assumptions taken when computing NTCs and do not coordinate. As a result, NTC levels can be very conservative (and low).

To overcome this drawback, the cNTC approach is based on a common grid model (which is, in turn, based on each TSO’s individual grid model). This common grid model serves as an aligned starting point, based on which NTCs are computed. The maximum NTC level is reached when (at least) one critical network element is fully loaded. The (coordinated) application of remedial actions may accompany this approach to increase capacities further. The result is the “NTC domain” (i.e. maximum cross-zonal capacities per oriented bidding zone border).

Even with cNTC, TSOs consciously decide how much cross-zonal capacity is offered on one oriented bidding zone border versus another (or several other) oriented bidding zone border(s). This decision or choice is particularly relevant for (highly) meshed grid topology, where cross-zonal exchanges on a number of bidding zone borders have an impact on grid elements. Hence, such an approach is sensible in power systems with a radial grid structure (e.g. Italian peninsula) and less suited for meshed power grids (e.g. Central Europe).

The cNTC approach typically favours one (or a few) expected market direction(s) (generally referred to as “corner(s)”). The capacity calculation process is designed to increase cross-zonal capacities for this one (or few) corner(s) and other market directions, which may not be analysed explicitly.

3.1.6 Flow-Based Capacity Calculation

The flow-based approach allows for a more detailed representation of the grid in both the capacity calculation and allocation stage, as it considers several individual grid elements rather than bidding zone borders. As opposed to the cNTC approach, cross-zonal capacities are not determined per oriented bidding zone borders but are described via the flow-based domain. This flow-based domain is a (large) matrix depicting the “solution space”, as shown in Fig. 3.7, within which the market can clear. Consequently, cross-zonal capacities are not specific to exchanges on oriented bidding zone borders but to a whole “region”.

The flow-based domain is essentially a list of CNEs and CNECs characterized by two main components:

Remaining Available Margin (RAM): The RAM is the amount of transmission capacity (in MW) that is available on a given CNE (or CNEC) and a given direction for the market to allocate. Note that even though the convention “remaining available margin” might imply something that is left over, the RAM is the quantity to be maximized during flow-based capacity calculation. The RAM is determined by shifting the common grid model from its assumed forecast to a “zero-balance” state. In this state, the assumption is that no (commercial) exchanges take place between bidding zones (i.e. the net position of bidding zones is equal to zero). The RAM is then computed based on how much margin is available on CNEs (and CNECs) in the zero-balance state.

Power Transfer Distribution Factors (PTDFs): PTDFs are the sensitivities of individual CNEs (and CNECs) to changes in bidding zone net positions. A single PTDF value describes how much a CNE (or CNEC)’s loading changes when a given bidding zone exports (or imports) one extra MW.

Different kinds of PTDFs can be defined (as shown in Fig. 3.8).

Fig. 3.7 Example of two two-dimensional projection of a flow-based domain. Source: Based on
author’s own illustration.

Fig. 3.8 Power Transfer Distribution Factors. A Zone-to-slack PTDF. B Zone-to-zone PTDF matrix.
The column A→B is obtained by subtracting column B from column A of the z2s PTDF. Source Based
on author’s own illustration

The zone-to-slack⁴ (z2s) PTDF describes how the change in Net Position in a bidding zone results in flows on certain CNECs. Because of the model’s linearity, the effect of a flow change between two bidding zones is the difference of the effects of the flow changes between each of the zones and the slack node. This can be represented as a matrix composed of one column per bidding zone and one
row per CNEC. In flow-based market coupling, zone-to-slack PTDFs are provided to the market coupling algorithm.

If BZ A exports an extra 100 MW to BZ B, on which CNECs will this power flow? Subtracting the column corresponding to BZ B from the column representing BZ A will give us a new column showing the answer to this question. This column is called zone-to-zone (z2z) PTDF, describing which CNEC’s power flows are caused by an exchange between two bidding zones.

How is the zone-to-slack PTDF obtained? The starting point is the Grid Model, composed of its nodes, transmission elements, generators, and loads. The Generation and Load Shift Keys (GLSKs) relate a change at the nodal level to the overall change in Net Position. Therefore, the zone-to-slack PTDF can be obtained from the node-to-slack (n2s) PTDF using the GLSKs (technically, the z2s matrix is obtained from the inner product of the n2s matrix with the GLSK matrix). The node-to-slack PTDF describes how a change in generation/load at a certain node translates to flows on certain CNECs.

The flow-based domain holds one row for each CNE (or CNEC), one PTDF column per bidding zone, and one column for the RAM. The number of dimensions of a flow-based domain is given by the number of bidding zones encompassing a CCR. Note that the number of dimensions can be reduced by one because the sum of all net positions within a given CCR must equal zero.

The flow-based approach is especially beneficial in highly meshed power systems. In a cNTC approach, TSOs must split cross-zonal capacities between highly mutually impacted oriented bidding zone borders. In flow-based allocation, “the market” decides where the capacity is allocated.

Chapter 16 provides more information on the transparency of the relevant methodologies, processes, and input and output data.

3.1.7 Ever-Evolving Capacity Calculation Methodologies

With the first round of approved capacity calculation methodologies (CCMs), things have not come to a halt. During the implementation of the CCMs, inconsistencies, and other issues were identified that constituted hurdles for implementation. In some cases, TSOs requested amendments to the CCMs and
submitted them to NRAs for approval. So far, amendments to CCMs have been approved in several CCRs already, and this is going to continue.

Another aspect many CCMs had to incorporate after their first approval was the requirements of the 70% min MACZT rule (European Commission, 2019) and (ACER, 2023). By mid-2019, when the Clean Energy Package entered into force, many CCMs had already been approved. Catering to this new rule meant
requesting an amendment.

Further changes are already on the horizon. With ACER’s proposal for a “CACM 2.0”, the 70% rule might also apply in the intraday timeframe. Reconfiguration of CCRs or bidding zones and new (HVDC) interconnectors call for further improvements, and the integration of new features into the CCMs means that there will not be a “steady state” situation.

3.2 Illustrative Examples from Day-Ahead Capacity Calculation in Core CCR

In June 2022, the Core Day-ahead flow-based capacity calculation process went live. At the time of writing, this is the only flow-based process in operation, with intraday flow-based capacity in Core CCR and day-ahead flow-based capacity calculation in Nordic CCR expected to go-live soon. See Chap. 11 for a first analysis of the results after Core FBMC go-live and Chap. 12 for specific information on the Nordic flow-based implementation.

Figure 3.9 provides an overview of the steps of the Core Day-ahead flow-based capacity calculation process, which are further described below.

Fig. 3.9 Steps of the flow-based day-ahead capacity calculation process in Core CCR. Source: Author’s own illustration based on ENTSO-E website

1. Input File Delivery
   • Each TSO delivers to the Core Capacity Calculation tool (CCCt) a series of input files:
     • Day-2 individual grid models (one per hour)
     • Generation and Load Shift Keys (GLSKs)
     • List of Critical Network Elements and N-1 Contingencies (CNECs) including FRM
     • Available non-costly remedial actions (RA)
     • External Constraints or allocation constraints on the maximum import and/or export net position
   • Other inputs are:
     • Long-Term Allocations (LTAs) are provided by JAO, resulting from long-term allocated cross-zonal capacities in the yearly and monthly LT auctions.
     • Long-Term Nominations (LTNs) are provided by JAO for BZBs with Physical Transmission Rights (PTRs) (as of the end of 2023, used only on HR-SI Croatia—Slovenia)
     • Additionally, the CCC obtains the Day-1 individual grid models from TSOs, not in Core CCR and the Continental European synchronous area.
2. Merging and Data Quality Check
   • The CCC checks the IGM data for quality. Afterwards, the IGMs are aligned and merged into one CGM (Common Grid Model). Then, CCCt performs quality checks of the other TSO inputs (GLSKs, CNECs, RAs, etc.)
3. Initial FB computation and CNEC selection
   • The CCCt computes the initial flow-based parameters (PTDF and RAM for each CNEC). Then, a CNEC selection is performed (per timestamp), removing any CNEC with zone-to-zone PTDF lower than 5%⁵ on any exchange combination within Core CCR. This means all CNECs not significantly affected by cross-zonal exchanges between any two bidding zones within Core are removed from the computation. Approximately 60,000 CNECs are thus, selected.
4. Non-costly Remedial Action Optimization
   • The CCCt tries to enlarge the flow-based domain by coordinating and optimizing the use of non-costly remedial actions (see 3.1.3 for details).
5. Intermediate FB computation
   • The CCCt computes the intermediate flow-based parameters, now including the selected remedial actions. Additionally, it modifies the previous calculation to (attempt to) satisfy requirements regarding minimum levels of cross-zonal capacities.
   • If needed, the RAM is increased by the Adjustment for Minimum RAM(AMR) to meet the requirement of:
     • 20% of Fmax for exchanges within Core CCR (‘20% minRAM’)
     • 70% of Fmax for cross-zonal exchanges (both within and outside Core CCR) (‘70% minMACZT’)
     • Note: This second requirement effectively implements the consideration of the 70% rule as introduced by Regulation 2019/943 (Regulation (EU) 2019/943 of the European Parliament and of the Council of 5 June 2019 on the internal market for electricity)
6. Validations
   • The TSOs can validate and, if needed, correct calculated cross-zonal capacities to ensure the grid’s operational security. Before reducing cross-zonal capacities, all available remedial actions must be used.
   • At the time of writing, the validation step is performed by TSOs individually or in consortia. A coordinated validation approach is expected to be implemented in the coming years
7. Pre-final FB computation
   • If validation adjustments have been applied during the previous step, the CCCt computes the FB domain again. Otherwise, the result of the intermediate FB computation is retained.
8. Final FB computation
   • Upon receipt of both long-term allocated capacities (LTAs) from year-ahead and month-ahead allocation and long-term nominations (LTNs), the final flow-based parameters are computed. The flow-based and LTA domains are shifted to LTN (i.e. to account for already allocated cross-zonal capacity).
9. Publication & DA Market Coupling (SDAC)
   • The final FB and LTA domains are published on the JAO publication tool
     and provided to SDAC. The SDAC market coupling algorithm uses the Core
     flow-based domain, LTAs and allocation constraints when performing market
     coupling.

Fig. 3.10 Non-costly remedial action optimization. By applying Remedial Actions (RA), the flow-based domain is enlarged by increasing the RAM on certain CNECs and reducing it on others. Source: Based on authors’ own illustration

3.2.1 Non-costly Remedial Action Optimization

After the initial computation of cross-zonal capacities based on forecasts, the Core Capacity Calculation tool tries to enlarge the flow-based domain by attempting to increase the RAM on the CNEC(s) with lowest relative RAM, then repeating the process until the RAM cannot be increased further on any CNEC (Fig. 3.10). This is done by coordinating non-costly remedial actions (such as tapping phaseshifting transformers or applying topological remedial actions). This process is further constrained by monitored network elements that cannot be loaded further and a loop-flow threshold⁶.

Both RCCs (Coreso and TSCNet) perform the non-costly remedial action optimisation (NRAO) separately, using different algorithms. The best result per MTU is selected and used in the subsequent computation steps.

3.2.2 Validations in Core CCR

After the intermediate computation of the flow-based domain, each Core TSO has the right to validate it. If overloads persist despite the use of all available remedial actions, a Core TSO may reduce the RAM on CNECs in its control area/bidding zone to ensure operational security.

At the time of writing, there is no single way of assessing grid security. Each TSO (or consortia of TSOs) has developed its own methodology. These methodologies describe how Core TSOs individually validate cross-zonal capacities.

Reductions of cross-zonal capacities as individual validation adjustments (IVAs) can be applied in case of:

• Exceptional contingency or forced outages.
• When all costly and non-costly remedial actions are not sufficient to ensure that
  operational security limits are respected.
• Mistakes in input data resulting in overestimation of cross-zonal capacities.
• Potential need to cover reactive flows.

The size of the IVAs and the frequency with which they are used differ between TSOs. Figure 3.11 shows applied IVAs relative to Fmax for Core in 2022.

Fig. 3.11 Use of Individual Validation Adjustments by Core TSOs between June and December Source: Author’s own illustration based on ACER Monitoring Report 2023

3.2.3 Evolved Flow-Based in Core CCR

In recent years, high-voltage direct current (HVDC) interconnectors have been used more frequently as subsea connections between regions (e.g. Germany to Norway) and to connect asynchronous AC grids (e.g. Continental Europe and Nordic). An example of an HVDC interconnector between two bidding zones within Core is ALEGrO, connecting Belgium and Germany, commissioned in 2020.

The flow-based approach inherently considers that changing the power flow on one network element in a synchronous AC network results in the modification of flows in the whole grid (progressively smaller changes for lines further away). On the other hand, an HVDC interconnector is not part of an AC network, as it is connected to two AC/DC converter stations. Because of this, the power flow on an HVDC line must be set explicitly.

The Evolved Flow-Based (EFB) approach is one way to include these characteristics into the flow-based process. This approach essentially builds on modelling the two converter stations as two virtual bidding zones, such that they work as Bidding Zones for the capacity calculation and allocation algorithms (i.e. two additional columns in the zone-to-slack PTDF matrix). Note, however, that no market bids can be placed in the virtual bidding zones.

An exchange from zone A to zone B over an HVDC interconnector is expressed as a trade from A to the virtual hub representing the sending end of the HVDC interconnector plus a trade from the virtual hub representing the receiving end of the interconnector to zone B. In the case of ALEGrO, a trade from Belgium to Germany is modelled as BE →AL_BE + AL_BE →AL_DE + AL_DE→DE (with the bidding zones containing “AL” being the virtual bidding zones).

The combined net positions of the two virtual bidding zones must be equal to zero, such that the net positions of the two virtual bidding zones have equal magnitude and opposite sign.

EFB is included in market coupling by sending the virtual bidding zones to the SDAC algorithm and setting the transmission capacity between the virtual bidding zone and the actual bidding zone to the physical capacity of the interconnector. If needed, losses (both at the AC/DC converter stations and on the interconnector itself) are considered during market coupling, while they are ignored during capacity calculation. Similarly, ramping constraints (i.e. the maximum change in power flow between consecutive MTUs) are communicated to SDAC but disregarded during capacity calculation.

3.2.4 Advanced Hybrid Coupling in Core CCR

In cross-zonal capacity calculation, exchanges on bidding zone borders not attributed to the CCR in question but in the same synchronous area (so-called “external bidding zone borders”) need to be modelled to account for the impact of these exchanges on the cross-zonal capacities within the CCR in question. There are two common modelling approaches:

Standard hybrid coupling (SHC) entails including the forecasted exchanges with external bidding zones as fixed inputs (as generators/loads for the forecasted imports/exports). This approach is currently implemented in the Core day-ahead flow-based capacity calculation process.

Advanced hybrid coupling (AHC) goes one step further. This approach treats the external Core bidding zone borders similarly to internal ones. The converter station at the Core-end of an HVDC interconnector is modelled as a virtual bidding zone. The net position of this virtual bidding zone represents the flow over the interconnector. The flow-based capacity calculation and allocation processes include the virtual bidding zone. Note that for external bidding zone borders, two (independently computed) sets of cross-zonal capacities are provided to SDAC: First, the cross-zonal capacities from the “native” CCR of the external bidding zone border (e.g. for NL-DK1 this would be Hansa CCR, providing NTCs), Second, the extended flow-based domain with PTDFs for every CNECs for the virtual bidding zone (e.g. for NL-DK1, the virtual bidding zone in the Netherlands in case of Core flow-based or the virtual bidding zone in DK1 in case of Nordic flow-based). Note that AHC only works for bidding zone borders for which cross-zonal capacity calculation is performed in SDAC. For example, AHC is not an option for HVDC interconnectors between Continental Europe and Great Britain, as Great Britain is no longer coupled with SDAC.

3.2.5 cNTC Capacity Calculation in South Western Europe

The CCR South-Western Europe (SWE), comprising the bidding zone borders Portugal-Spain and Spain-France, applies the coordinated NTC approach to calculate cross-zonal capacities (ENTSO-E, 2022). Because of the intrinsic grid topology (radial, not meshed), using a flow-based approach in this CCR would bring about limited benefits, as the influence of bidding zone border exchanges from one to the other is small.

In addition to the inputs listed in Sect. 3.1.2, for the cNTC process in SWE, a few extra details apply:

• The reliability margin is defined per-border as the maximum between a fixed value (200 MW on FR-ES, 100 MW on PT-ES) and a percentage of the Total Transmission Capacity (7.5% on FR-ES, 10% on PT-ES).
• No Allocation Constraints are used.
• A sensitivity analysis (performed once per year) is used to select CNECs significantly influenced by cross-zonal exchanges. Only CNECs with sensitivityabove 10% are monitored during the CC process (with few exceptions thatmust be justified to the SWE NRAs).

The capacity calculation process in SWE (applied both in the day-ahead and intraday timeframes) consists of an iterative computation starting from the coordinated schedule in the common grid model. The coordinated capacity calculator tests several levels of cross-zonal exchange by using Generation and Load Shift Keys and determines if this level of exchange respects operational security limits on
all the monitored critical network elements after the occurrence of all the monitored contingencies. Available remedial actions are applied when necessary. At each calculation step, the Remedial Action Optimization monitors the maximum flows, voltage levels, and the maximum voltage phase angle differences defined by the TSOs on all the CNEs.

3.2.6 Allocation Constraints

Allocation constraints constitute additional constraints that the market coupling algorithm must respect (see Sect. 3.1.3). Figure 3.12 depicts the application of two types of allocation constraints in the day-ahead market coupling algorithm.

A. Use of Allocation Constraints in SDAC. The HVDC interconnectors over which ramping constraints apply are highlighted in red.
B. Schematics of Net Position allocation constraints. It is reduced if the Net Position calculated by the CC algorithm is larger than the NP allocation constraint.
C. Schematics of Ramping constraints. If the rate of flow change between consecutive MTUs is larger than the maximum flow difference, it is reduced. The per-MTU requested and realized flows are shown in dashed black and solid red, respectively.

Fig. 3.12 Depicting the application of two types of allocation constraints in the day-ahead market coupling algorithm. Source: Author’s own illustration

Italy North and Poland use net position constraints on the SDAC net position, and the Netherlands uses net position constraints on the CCR net position. Belgium phased out their respective net position constraint in November 2023. The Netherlands is in the process of phasing out its respective net position constraints

Ramping constraints are used on the majority of HVDC interconnectors, forwhich cross-zonal capacity allocation is done in SDAC.

3.2.7 History of Coordinated Capacity Calculation

The introduction of the Third Energy Package or the CACM Guideline did not initiate coordination in cross-zonal capacity calculation. Prior to the CACM Guideline’s entry into force, several voluntary initiatives existed.

As one example, day-ahead flow-based capacity calculation and allocation was operational in Central-West Europe (CWE). Having gone live in May 2015, the CWE region, formally not a CCR, pioneered the flow-based capacity calculation and allocation. Initially, the TSOs of Belgium, France, the Netherlands, and Germany (except 50 Hertz) worked on implementing CWE flow-based. Discussions with Swissgrid were held about joining, but the exclusion of Switzerland from day-ahead market coupling (not SDAC at that time) wiped the idea off the table. Later, the Austrian TSO, APG, joined the consortium, then still in the common GermanAustrian bidding zone border. The introduction of cross-zonal capacity allocation
on the German-Austrian bidding zone border in 2018 was implemented in CWE, showing that the system could be extended by additional bidding zone borders. CWE flow-based capacity calculation and allocation was operational until June 2022, when Core day-ahead flow-based capacity calculation and allocation went live and superseded this first operational flow-based initiative in Europe. Compared to CWE, Core incorporated many lessons learned, reduced individual TSO discretion, further centralized the capacity calculation process, and streamlined the overall process.

Further south, the TSOs adjacent to the Italy North bidding zone, namely RTE, Swissgrid, APG, Eles, and Terna, had developed a concept for coordinated NTC computation along the “roof of Italy” (the Italian border, stretching from France to Slovenia). The initiative was driven by the 2006 European blackout and the development of Italy towards a (net) importing country. The NTC capacity calculation concept focused on maximising imports to Italy. After initially focusing on the day-ahead timeframe, the intraday timeframe followed suit. With the introduction of the CACM Guideline, the coordinated NTC capacity calculation process in Italy North was formalised—both through the creation of the CCR (which formally does not include Swissgrid) and the approval of the capacity calculation methodology through regulators.

Other cooperations and coordinations existed in addition to the examples mentioned above. Poland used to compute its NTC on a profile spanning Germany, the Czech Republic, and Slovakia. The Nordic TSOs have long cooperated when computing their NTCs. Initiatives for explicit day-ahead flow-based processes were launched in Central-East Europe (which merged with CWE to form Core) and South-East Europe.

3.3 What Does the Future Hold for Capacity Calculation Methodologies Across Europe and Its CCRs?

A few key points can be highlighted.

There is a trend to have fewer and, consequently, larger CCRs. The most likely future changes are the attribution of the France-SEM bidding zone border to Core CCR and the merger of the Italy North and Core CCRs (provisionally referred to as “Central Europe” CCR).

Implementing a flow-based capacity calculation approach in areas with a highly meshed grid would be highly beneficial. Important examples are Italy North and countries in the Western Balkans (once joining SDAC).

Improving the consideration of non-EU countries in the capacity calculation methodologies would improve the accuracy and overall quality of the capacity calculation processes. It is not unusual for flows computed in a CCR to be realized in countries outside that CCR. Notable examples are Switzerland (involved in flows between France, Austria, Italy, and Germany) and Serbia and Bosnia Herzegovina
(involved in flows between Romania/Bulgaria and Croatia).

Next to geographic extensions, other timeframes than day-ahead gain more traction, namely long-term, intraday, and balancing. Other ongoing implementation processes, such as intraday auctions, long-term flow-based allocation, and the balancing platforms, call for emphasizing other timeframes to deliver cross-zonal capacities to benefit European consumers while maintaining operational security.

About the authors of this chapter

Reinhard Kaisinger is a Principal at Magnus Energy. He co-leads Magnus Energy’s Technical Advisory practice and is heavily involved in several projects related to electricity market integration. With over a decade of experience in the electricity sector in various roles, he has obtained a wide yet deep understanding of power systems and electricity markets. After obtaining his master’s degree in electrical power engineering at KTH, Stockholm, Sweden, he started working on power systems and electricity markets at Vattenfall in Sweden. At E-Control, the Austrian electricity gas market regulator, he focused on network development (TYNDP, PCI) and the implementation of CACM Guideline and System Operations Guideline. Then, he worked on the implementation of Core day-ahead flow-based capacity calculation and allocation at APG (the Austrian TSO) both as an expert and project manager before joining Magnus Energy in 2022. He currently serves as convener of the Core day-ahead capacity calculation project team.

Dr. Iacopo Bertelli is a Senior Consultant at Magnus Energy. His field of expertise is the European electricity market design and integration, both conceptually and analytically. Currently, he is supporting the newly established Central Europe CCR as PMO of the day-ahead capacity calculation project team. He has experience in supporting CCR accessions by TSOs, from assessing the high-level requirements to drafting a detailed local implementation plan. He is a physicist by training, having obtained a PhD in experimental quantum physics from the University of Leiden, Leiden, the Netherlands, and TU Delft, Delft, the Netherlands, during which he published 7 articles on high-impact journals.

About the editors of the book

Dr. André Estermann has more than 20 years of experience in energy sector and currently leads at the German TSO 50Hertz Transmission GmbH a team in system operations responsible for cross-border activities. Furthermore, he is the TSO Chairman of the Market Coupling (MCSC), in the Single Day-Ahead Coupling (SDAC) and Single Intraday Coupling (SIDC) Steering Committees. André holds a Ph.D. in economics at the European University Viadrina Frankfurt, Oder and received an MBA at the University of Southampton, UK. Moreover, he holds a degree in international politics from the University of Potsdam.

Marius Schrade is an Expert in European Market Coupling projects at 50Hertz Transmission GmbH, one of the four transmission system operators in Germany, within the System Operations Concepts and Analytics department since 2019. At 50Hertz, he started his career as project lead for the integration of the German–Czech and German–Polish border into Single Day-Ahead Coupling, which was completed in 2021. Since then, he focuses on cooperation between TSOs and NEMOs in the Market Coupling Steering Committee where he also lead a task force on day-ahead simulation software. Marius holds a Double Degree in European Economics from the Universities of Maastricht in The Netherlands and
Louvain-La-Neuve in Belgium.

Louise Anderson currently works with Hitachi Energy as an Assistant Vice President in the Government and Institutional Relations department. Previously, Louise spent 5 years managing the World Economic Forum’s Electricity Industry Community, while also launching the WEF’s global initiative on Accelerating Industrial Clusters. Louise started her career working in a power plant before moving to work for EirGrid, Ireland’s Transmission System Operator, managing high voltage asset outage planning until 2014. Louise also spent some time in Belgium working for ENTSO-E on different aspects of the implementation of the CACM Guideline. Louise has a bachelor’s degree in Mechatronics and an MBA specializing in International Business.

Bibliography

ACER. (2023). ACER’s Market Monitoring Report 2023—Cross-zonal capacities and the 70% margin available for cross-zonal electricity trade (MACZT). Retrieved November 17, 2024, from https://www.acer.europa.eu/Publications/2023_MMR_MACZT.pdf

ACER. (2024). ACER Decision No 04-2024 on the amendment to the determination of capacity calculation regions in accordance with CACM Guideline. Agency for the Cooperation of Energy Regulators (ACER). Retrieved November 17, 2024, from https://www.acer.europa.eu/document/acer-decision-no-04-2024-amendment-determination-capacity-calculation-regions-ccrs

ENTSO-E. (2015a). All TSOs’ proposal for the day-ahead firmness deadline (DAFD) in accordance with Article 69 of the CACM guideline. Retrieved November 11, 2024, from https://www.acer.europa.eu/sites/default/files/documents/en/Electricity/MARKET-CODES/CAPACITY-ALLOCATION-AND-CONGESTION-MANAGEMENT/8%20DAFD/Action%201a%20-%20D
AFD%20approved.pdf

ENTSO-E. (2015b). Second amendment of the Day-Ahead Capacity Calculation Methodology of the Core Capacity Calculation Region in accordance with Articles 20ff. of the CACM Guideline. Retrieved November 11, 2024, from https://consultations.entsoe.eu/markets/core_2nd_dafb_ccm/supporting_documents/Core%20DA%20CCM%202nd%20amendment.pdf

ENTSO-E. (2022). South West Europe TSOs common capacity calculation methodology for the day-ahead and intraday market timeframe in accordance with Article 21 of CACM Guideline. Retrieved November 11, 2024, from https://eepublicdownloads.entsoe.eu/cleandocuments/nctasks/SWE%20CCM%20Amendment_January2022_clean-v3.pdf

ENTSO-E. (2024a). Core DA CC process steps. Retrieved June 01, 2024, from https://www.entsoe.eu/bites/ccr-core/explained/#fc11

ENTSO-E. (2024b). Public consultation on 3rd Request for Amendment of the Core Day-Ahead Capacity Calculation Methodology. Retrieved November 11, 2024, from https://consultations.entsoe.eu/markets/core-tsos-proposal-3rd-amendment-dafbccm/

ENTSO-E. (n.d.). ENTSO-E Website. Retrieved November 17, 2024, from https://www.entsoe.eu

European Commission. (2013). Commission Regulation (EU) No 543/2013 of 14 June 2013 on submission and publication of data in electricity markets and amending Annex I to Regulation (EC) No 714/2009. Retrieved November 11, 2024, from https://eur-lex.europa.eu/legal-content/ EN/TXT/uri=CELEX%3A32013R0543

European Commission. (2015). Commission Regulation (EU) 2015/1222 of 24 July 2015 establishing a guideline on capacity allocation and congestion management. Retrieved November 11, 2024, from https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A02015R1222-20210315

European Commission. (2019). Regulation (EU) 2019/943 of the European Parliament and of the Council on the internal market for electricity. Retrieved November 11, 2024, from https://eurlex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A02019R0943-20240716

JAO. (n.d.). JAO Website. Retrieved November 11, 2024, from https://www.jao.eu

Notes

1. Note that while Norway is not an EU Member State, it is a member of the European Economic Area (EEA)
2. The approved day-ahead firmness deadline can be found at ENTSO-E (2015)
3. CNECs are Critical Network Elements (CNEs) under N-1 Contingency as explained further
   down
4. The slack node is the reference point for PTDFs relative to a slack (e.g. node-to-slack or zone-to-slack PTDFs). The slack node is a necessary mathematical construct, but the choice of slack does not influence the result. All PTDFs for the slack itself are zero (flow from slack to slack)
5. A CNE(C) satisfying a maximum for zone-to-zone PTDF 5% is considered significantly influenced by cross-zonal exchanges.
6. The loop-flow threshold is defined in Article 16 of the Core DA CCM (ENTSO-E, 2024).
7. See public consultation on 3rd Request for Amendment of the Core Day-Ahead Capacity Calculation Methodology (ENTSO-E, 2024).