Flow-Based Capacity Calculation

A method for calculating available cross-zonal transmission capacity based on actual physical power flows through the grid, as governed by Kirchhoff’s laws. Replaced the simpler NTC (net transfer capacity) method in the Nordic region in October 2024.

How it differs from NTC

NTC (the previous method) treated each border as an independent bilateral exchange: “how much power can flow from zone A to zone B?” This is a simplification — in a meshed AC grid, power doesn’t flow along commercial paths but follows the path of least resistance.

Flow-based models the entire grid simultaneously. It considers how a trade between any two zones affects power flows on all critical network elements (transmission lines). This means:

• A trade between SE1 and SE4 may be limited not by the direct SE1→SE4 path but by a line in Germany that also carries part of the flow
• Available capacity depends on the full pattern of trades, not just one border

Implications

• More accurate: reflects physical reality better than NTC
• Can increase or decrease available capacity: flow-based can reveal that more capacity is available than NTC assumed (when flows are favorable) or less (when parallel flows consume capacity)
• More complex: requires detailed grid modeling, real-time data, and coordination between TSOs
• Enables better market coupling: traders see capacity that actually exists rather than conservative NTC limits
• Temperature-dependent: transmission line thermal ratings (PATL — the Fmax basis since the 2026 CCM amendment) are ambient-temperature sensitive; hot summer weather reduces line ratings and thus Fmax, meaning available cross-zonal capacity is lower during heat waves than in cooler periods (Source - Svk Elsystemet Under Sommaren (2026))

EU legal basis

Flow-based capacity calculation is mandated by CACM Regulation (EU) 2015/1222 Art. 21 as the primary capacity calculation method wherever capacity between bidding zones is highly interdependent. TSOs must apply it first; the coordinated NTC method may only be used where zones are less interdependent and flow-based provides no added value (Art. 21(1)). The CACM further requires all TSOs to develop individual grid models merged into a common grid model (CGM) as the basis for the flow-based calculation (Arts. 16–18). (Source - CACM Regulation (EU) 2015-1222)

Nordic adoption and go-live

The Nordic TSOs (Svenska kraftnät, Statnett, Fingrid, Energinet) implemented flow-based capacity calculation on October 29, 2024 — the first day-ahead auction with flowbased-derived capacities was for delivery October 30, 2024. A rollback period ended November 20, 2024, after which reversion to NTC was no longer possible. The CORE region (Central Western Europe) implemented flow-based for day-ahead in June 2021, per this page’s own cited CACM source (a commonly-cited “CWE since 2015” precedent refers to an earlier, narrower CWE FBMC pilot project, not the CORE-region implementation this page’s source describes — treat the exact year/region pairing with care). The Nordic transition was a major milestone for market integration.

Flow-based calculation is particularly relevant for the Swedish Bidding Areas because the north-south bottleneck (intersections 1, 2, 4) and east-west flows through SE3 create complex flow patterns that NTC could not adequately model.

First weeks — operational observations (November 2024)

Experience from the first three weeks was presented at the November 2024 Elmarknadsrådet meeting. (Source - Elmarknadsrådet Meetings 3 and 4 2024 (Sep-Nov))

Non-intuitive flows: the most discussed phenomenon. Examples from October 30:

• A binding CNEC in SE2 meant that increasing SE1 production (rather than SE2) relieved the constraint while delivering more capacity southward — producing a counterintuitive “power flow from higher-price to lower-price area.” With NTC, both the Hasle (SE3-NO1) and Kontiskan (SE3-DK1) cross-border elements would have needed pre-restricted capacities; flowbased required no pre-restriction.
• Large east-west flows (Finland→SE3 via Fennoskan, SE3→NO1, SE2→SE3) on the first day when all Nordic prices were low but western Norway, Denmark, and the continent had high prices.

Increased throughput: snitt 2 and snitt 4 flows approximately 30% higher (both directions) in the first weeks with flowbased versus the weeks immediately before — reflecting more capacity released to the market by the physics-based model compared to the conservative NTC approach.

Why non-intuitive flows happen — the welfare-economics mechanism: non-intuitive flows were already anticipated and explained theoretically during the pre-go-live parallel-run phase. Standard intuition says power should flow from cheap to expensive zones (arbitrage); a non-intuitive flow occurs when the welfare loss from that “wrong-way” flow is smaller than the welfare gain from relieving a binding congestion elsewhere in the mesh. Because the market-clearing algorithm (Euphemia) maximizes total pan-European welfare rather than per-border arbitrage, it accepts a locally counterintuitive flow whenever the system-wide result is better.

Formally, the flow-based welfare-optimum condition ties price differences to CNE shadow prices via PTDFs (Power Transfer Distribution Factors — the fraction of a zone-to-slack MW injection that flows on a given CNE): Δprice(i,j) = shadow price(CNE) × PTDF(CNE, i→j). When a trade relieves a binding CNE (its zone-to-zone PTDF is negative), the algorithm prices the export zone higher than the import zone to reflect that relief — a direct mathematical consequence of PTDF-based optimization, not an anomaly. NTC’s border-only constraints cannot express “relieve an internal CNE by routing power the ‘wrong’ way,” so non-intuitive flows are structurally more common under flow-based than they were under NTC.

A real parallel-run case study validates this: on 2022-01-18 (MTU 8), NO3 (29.6 EUR/MWh) exported to SE2 (15.3 EUR/MWh) — a 14.3 EUR/MWh non-intuitive spread, exactly explained by decomposing the shadow-price × PTDF contributions across the binding CNEs (aggregated marginal transmission cost via NO3: 258.0; via SE2 instead: 272.4 — the 14.3 gap matches the price spread). The flow-based case also exported 400 MW more than NTC from the four northern low-price zones to southern high-price zones in the same hour, directly illustrating flow-based recovering welfare that NTC’s cruder border constraints left on the table. (Source - Nordic CCM Phenomena Report (2022))

EPAD premiums: prices in area electricity price difference (EPAD) contracts increased after flowbased go-live, most visibly for the December 2024 SE3 contract, consistent with the day-ahead market development.

Intraday capacity: remaining capacity after the day-ahead auction must be converted to ATC for the intraday market (transitional solution); capacity available for intraday was somewhat reduced vs the pre-flowbased period. Intraday trading volumes in the first weeks were comparable to immediately before go-live, but peak volumes seen in the pre-flowbased period had not yet been reached.

Operational situation: go-live coincided with a major switchyard reconstruction at Kilanda (western SE3), creating a strained initial situation. The control room implemented new tools and processes alongside grid hardware changes. The system was assessed as stable by the start of the interruption-free winter period (November 15, 2024).

HVDC protection measures at go-live

The transition to flowbased increased HVDC pole reversal frequency (higher cross-border flows → more hour-to-hour direction changes). MIND-type cables (Kontiskan, Fennoskan, SwePol Link, Baltic Cable) have a maximum of ~1,000 reversals per year, with minimum hours required between reversals. Measures implemented or prepared for go-live: (Source - Elmarknadsrådet Meetings 1 and 2 2024 (Feb-May))

1. Implicit losses on Kontiskan: approved by regulators; Fennoskan not applicable (parallel AC connections make implicit losses counterproductive)
2. Ramp restrictions: SwePol 600→300 MW/h from go-live; Kontiskan 300 MW/h at mFRR EAM go-live; Fennoskan restricted to prevent large shocks (Fennoskan 1 already damaged at the time)
3. Intraday countertrading via portfolio manager: Svk contracted a third-party portfolio manager; routine agreed with Energinet and Fingrid; countertrading volumes announced via NUCS; preference for IDA auctions to increase liquidity
4. Dynamic Profile Selection (DPS) and Quarterly Polarity Reversal Elimination (QPRE) for specific cables

By the November 2024 meeting, implicit losses on HVDC were approved by relevant regulators but Nordic RCC tests were not yet complete. Svk assessed the impact would be limited in the near term but noted the measures are important long-term tools for protecting HVDC equipment from wear.

Flow-based outage information on NUCS (September 2026)

On 24–25 September 2026 the Nordic TSOs (coordinated by the Nordic RCC) began publishing flow-based estimates of how planned outages affect trading capacity on NUCS, the Nordic Unavailability Collection System: month-ahead (M-1, 24 September) and week-ahead (V-1, 25 September), with domain impact per individual outage. It extends the method beyond the day-ahead market, where it has been used since October 2024, to give consistency across time horizons. The TSOs call the first release basic functionality: the figures are approximate indications, not forecasts, and can differ from the day-ahead result; only selected network elements and outage types are covered (busbars, for example, are not represented), simultaneous outages are not yet combined, temporary reconnections such as nights and weekends are not shown (the intact-grid LTCC domains act as the latest flow-based capacity then), and the existing NTC-based outage publications continue in parallel. Svk’s statement that trading capacity has increased since the introduction is unquantified. (Source - Svk Flödesbaserad Kapacitetsberäkning Avbrottsinformation NUCS (2026))

Relevance to flexibility

Flow-based capacity calculation changes the landscape for Flexibility in several ways:

• More transparent congestion signals: market participants can better understand where and when congestion occurs, which improves the efficiency of flexibility markets
• Cross-border flexibility: more accurate capacity allocation enables better use of cross-border flexibility resources (e.g., Norwegian hydropower for Swedish congestion)
• Dynamic capacity: available capacity changes hour by hour based on the full trade pattern, making the system more responsive but also more complex to predict

The Clean Energy Package‘s Electricity Market Regulation Art. 16(8) establishes the 70% minimum capacity rule: at least 70% of critical network element capacity must be available for cross-zonal trade. Under the flow-based approach, this 70% applies to the margin set in the capacity calculation process for flows induced by cross-zonal exchange. The remaining 30% covers reliability margins, loop flows, and internal flows. This rule is a key driver for adopting flow-based methods — it requires accurate modeling of how trades affect physical flows across the entire grid.

(Source - Svk Network Development Plan 2026-2035)

CCM governance and the 2026 amendment

Flow-based capacity calculation is not self-executing: the CACM requires all TSOs within each Capacity Calculation Region (CCR) to develop and submit a Capacity Calculation Methodology (CCM) for NRA approval under CACM Art. 20(2). In the Nordic CCR this means all four TSOs (Svk, Statnett, Fingrid, Energinet) submit jointly to Ei plus the Danish, Finnish, and Norwegian NRAs. The NRAs have 6 months to decide collectively; if they disagree, the case goes to ACER.

The Nordic CCM was first approved in 2018, amended in 2019 and again in October 2020 (the version that governed flow-based go-live). The 2020 version contained mandatory update obligations for certain articles within 18 months of go-live. In May 2026, the Nordic TSOs submitted a Third Amendment to Ei (ärendenummer 2026-103735; proposal filed under e-diarium 2026-103688), replacing the 2020 version entirely. The NRAs must decide by a firm deadline of 19 November 2026. Ei opened a public comment window for market actors, running to 1 October 2026. (Source - Nordic CCM Third Amendment Package (2026), Source - Ei Flödesbaserad Kapacitetsberäkning Synpunkter (2026))

What the CCM governs

The CCM specifies:

• How the common grid model (CGM) is built and merged from individual TSO IGMs
• How power transfer distribution factors (PTDFs) are calculated
• Which critical network elements (CNECs) are included in the flow-based domain
• How remaining available margins (RAMs) are calculated — including reliability margins and the effect of remedial actions
• How the flow-based domain is converted to ATC values for the intraday market (transitional solution)
• How TSOs validate and publish capacity parameters

Technical mechanics: FRM and CDCs

Flow reliability margin (FRM)

Every CNEC’s RAM is reduced by an FRM to account for forecast errors. The 2026 CCM amendment formalizes FRM as two components:

• RM (reliability margin): uncertainty from load/generation forecasts, GSK assumptions, topology changes, and grid model errors — derived statistically at a 95th-percentile risk level from historical deviations between forecast and realized flows
• FCR margin: unintended flow deviations from FCR-N reserve activations within Nordic LFC areas — computed separately from zone-to-slack PTDFs and FCR-N net position data

The RAM formula for each CNEC before validation is: RAM = Fmax + FRA − FRM − F0 − FAAC (this exact formula is not directly cited to a specific source on this page — a structurally different variant, RAM = Fmax − Fref − FRM − FAAC, appears in a different vault source, ENTSO-E BZRR Nordic 2025; worth tracing to its actual origin before treating this specific form as confirmed)

Where Fmax is the physical capacity limit (now always PATL-based, not TATL), FRA is the flow contribution from remedial actions, F0 is the flow in a zero-net-position scenario, and FAAC accounts for already allocated cross-zonal capacities.

Combined dynamic constraints (CDCs)

Not all grid stability limits can be expressed as flow limits on a single line. CDCs are constraints on the combined flow across a set of network elements — used for:

• Voltage stability: collapse affects a transmission corridor, not one line
• Rotor angle stability (transient and oscillations): inter-area oscillations correlate with corridor flows
• Frequency stability: Nordic dimensioning incident limits span multiple elements

CDCs are mathematically equivalent to CNECs in the flow-based optimization but constrain a weighted sum of flows. They were temporary in the 2020 CCM; the 2026 amendment makes them permanent.

The 70% rule: floor or ceiling?

CACM requires the 70% minimum capacity rule (Regulation 2019/943, Art. 16(8)): at least 70% of each CNEC’s Fmax must be available for cross-zonal trade as RAM. The 2020 CCM interpreted this as a floor to be satisfied by adding costly remedial actions (redispatch, countertrading) to increase capacity — including “virtual capacity” above the physical Fmax if economic efficiency could be demonstrated.

The 2026 amendment reverses this. Costly RAs are now limited to: (a) meeting the 70% minimum, (b) managing temporary outages, and (c) enabling TATL in N-1 scenarios. Virtual capacity is explicitly abandoned.

Two reasons are cited:

1. EU General Court ruling (BNetzA and Germany vs ACER, Cases T-600/23 and T-612/23, 1 October 2025): once the 70% minimum has been reached, the court held that an additional economic-efficiency test for remedial action is legally irrelevant and not binding on TSOs (Raw/Clippings/General Court BNetzA and Germany v ACER T-600-23 press release (2025-10-01).md); the Nordic TSOs’ supporting document reads this as: TSOs providing 70% of Fmax comply, with no legal requirement to go further
2. Nordic welfare analysis: with many small bidding zones and scarce balancing reserves, backing up virtual capacity via real-time redispatch is more expensive than the market gains — a net welfare loss, not a gain

Practical consequence: the Nordic TSOs treat 70% compliance as the limit of what costly remedial actions are used to achieve, not a stepping stone toward full Fmax allocation (70% is the compliance level, not a stated cap: the amendment’s own ceiling is 100% of physical Fmax, and Art. 10(2) says each TSO may define costly RAs in the three cases). This is a significant shift from the 2020 interpretation and constrains available cross-zonal capacity in congested periods.

The Article 14 dispute — Nordic TSO disagreement (2026)

Article 14 of the CCM defines which CNECs appear in the final flow-based domain. CNECs whose maximum zone-to-zone PTDF falls at or below 5% are removed (they are not “significantly influenced” by cross-zonal trade). The 2026 amendment package was submitted with one unresolved disagreement on this article.

The FI–NO4 border CNEC: Fingrid argued this interconnection (Finland–Northern Norway) must always remain in the flow-based domain, even when its PTDF falls below 5%, because:

• It is a structurally relevant cross-border security constraint
• Fingrid cannot apply countertrade or redispatch on flows originating in NO4 — so persistent inclusion in capacity calculation is its only security tool

Statnett, Energinet, and Svk disagreed: CACM Art. 29.3(b) requires exclusion of all CNECs not significantly influencing net positions, regardless of cross-zonal vs internal character. A blanket exemption for cross-zonal CNECs has no legal basis. Permanent inclusion of a low-PTDF CNEC could materially limit NO4 export possibilities.

Resolution: Article 14 reverts to the unchanged 2020 wording (standard 5% threshold, no exemption). Fingrid signalled separately that it will seek clarification of ACER Decision 08/2023 on the NO4–FI border as a distinct matter.

This dispute is notable as one of the few publicly documented technical disagreements between Nordic TSOs on capacity calculation methodology, and it connects to the broader question of how CACM’s significance threshold should apply when operational and market considerations conflict.