CoordiNet

A European Horizon 2020 research and demonstration project (grant 824414) focused on how Distribution System Operators (DSOs) and Transmission System Operators (TSOs) can coordinate to procure and activate grid services through demand response, storage, and small-scale generation. The project ran from 2019 to 2022 across three countries (Sweden, Greece, Spain) with eight demonstration activities. The Swedish demonstration is the primary source of documented experience for flexibility markets at distribution level in Sweden.

Swedish demonstration

The Swedish demonstration operated flexibility markets for congestion management across four areas for three consecutive winters (2019/2020, 2020/2021, 2021/2022), plus peer-to-peer markets during maintenance periods.

Consortium members:

• Vattenfall Eldistribution AB (regional DSO, Uppland and Gotland areas)
• E.ON Energidistribution AB (regional DSO, Skåne and Västernorrland/Jämtland)
• Svenska kraftnät (national TSO)
• Uppsala municipality
• Energiforsk AB (research institute)
• Expektra AB (commercial load forecasting)
• RWTH Aachen University (technical partner — confirmed as a full consortium member, alongside Energiforsk, by the project’s own 2019 kickoff-era consortium list in Source - Backe Bjarup Ruwaida Change Management Sweden Capacity Challenge (CIRED 2019))

Origins — two parallel, deliberately complementary tracks. Before CoordiNet began, Vattenfall and E.ON had already been separately testing two different approaches to the same congestion problem: Vattenfall through bilateral treaties for load/production steering (the direct precursor to Villkorade Avtal), E.ON through digital marketplaces (the direct precursor to SWITCH). A 2019 paper by practitioners from both DSOs frames this explicitly as two learning tracks brought together for the first time under CoordiNet, not independent developments — and gives an early practitioner verdict on bilateral treaties that the later regulatory literature (Ei) would independently reach: “bilateral treaties would result in very time-consuming and manual work… more suitable for specific, smaller-scale agreements.” (Source - Backe Bjarup Ruwaida Change Management Sweden Capacity Challenge (CIRED 2019))

Areas and DSO challenges:

Total participation: 39 flexibility service providers; 477 MW/7 GWh registered capacity

The business use case: subscription to the TSO

All Swedish congestion management markets were driven by the subscription level (abonnemang mot överliggande nät): the annually contracted maximum power a regional DSO may draw from the TSO grid without prior notice. When Svenska kraftnät denies a subscription raise — typically because transmission reinforcement is not yet complete — the DSO must manage demand below the limit.

This creates a direct price ceiling for the local Flexibility Market:

• If flexibility is cheaper than the temporary subscription fee (~240–280 SEK/MWh), the DSO buys flexibility routinely (Uppland model: 248 SEK/MWh average, high volumes)
• If the TSO grants temporary subscriptions, the DSO only buys flexibility when subscription is also denied (~2,285 SEK/MWh, low volumes in Skåne)

The 2021 network-charge structure, precisely. The subscription mechanism has four separate charge components: a usage fee (SEK/kWh, covers TSO energy losses), a regular capacity charge (SEK/kW, annual, based on subscribed capacity), a temporary capacity charge (SEK/kW, weekly, revocable by Svk if it assesses transmission-congestion risk), and a temporary subscription usage fee (SEK/kWh, normally the largest component) for energy consumed above the regular capacity but within a granted temporary increase. When a temporary subscription is granted, that usage fee runs 220–250 SEK/MWh depending on the substation. When not granted, the DSO instead pays a tiered penalty: 560 SEK/MWh for the first hour of overrun in a day, 1,400 SEK/MWh for the second hour, and 2,800 SEK/MWh for every further hour with a limit violation that day — a steep within-day escalation designed to make sustained overrun far costlier than a brief one. (Source - Ruwaida et al TSO-DSO-Customer Coordination Swedish Demonstration (2022))

Market timing, precisely. The multi-level model’s gate sequence: the DA local/regional flexibility market clears ~10:30, ahead of the DA wholesale spot market’s 12:00 CET gate closure; the ID flexibility market then opens at 15:00 (once the TSO’s FCR market has closed) and stays open until two hours before delivery; pre-qualified unused bids can then be forwarded into Svk’s mFRR market, which opens 15 minutes before delivery. This sequencing — local markets clearing well ahead of the TSO’s own gates — is what makes bid-forwarding into mFRR possible at all. (Source - Ruwaida et al TSO-DSO-Customer Coordination Swedish Demonstration (2022))

A price discrepancy worth flagging. Ruwaida et al.’s own results table gives Skåne’s average cleared price for winter 2020/21 specifically as 1,503 SEK/MWh — not the ~2,285 SEK/MWh cited above from D4.7.2. The likely explanation is scope: D4.7.2’s figure may average across all three demonstration winters (with more volatile pricing in other seasons) rather than one winter in isolation, but this has not been confirmed by re-reading D4.7.2 itself — both figures are recorded here pending that reconciliation.

Key outcomes and learnings

See Source - CoordiNet D4.7.2 Swedish Demonstration (2022) for full data. Key conclusions:

• Market-based flexibility successfully managed TSO subscription constraints in Uppland
• Availability contracts (capacity fees) are necessary — energy-only payment is insufficient to keep FSPs engaged across warm winters when activation is rare; the Swedish design settled on weekly availability commitments (FSPs guarantee peak-hour availability for the following week only), giving FSPs planning flexibility in both scheduling and pricing rather than locking them into season-long terms
• Liquidity is a persistent challenge: thin markets, weather-dependence, and high energy price spikes all reduce supply reliability
• Dual-market prequalification was a barrier, not just an opportunity: mFRR’s prequalification criteria were stricter than the local flexibility market’s own, discouraging FSPs from qualifying for both — a concrete cause behind the manual value-stacking workaround documented in TSO-DSO Coordination — The Central Design Problem › Value stacking: the opportunity that requires coordination
• Prequalification burden fell hardest on aggregators: legal treatment of metering data, baseline agreements, and locational information of aggregated resources demanded significant DSO+FSP effort, with no consensus reached on which side (DSO or FSP) should be the metering data source (Source - Ruwaida et al TSO-DSO-Customer Coordination Swedish Demonstration (2022))
• The ten abilities framework (product design, customer engagement, DSO operations, coordination, forecasting/baseline, platform) provides a practical checklist for DSO transformation from distribution network operator (DNO) to distribution system operator (DSO) — a more granular, earlier version of this same idea appeared in the project’s 2019 kickoff-era framing as a ten-area “wide approach on transformation” (operating flexibility, TSO-DSO coordination, interfaces/APIs, real-time metering, product development, planning, grid forecasting, marketplace, load forecasting, customer engagement) (Source - Backe Bjarup Ruwaida Change Management Sweden Capacity Challenge (CIRED 2019))
• The project directly inspired sthlmflex (Stockholm) and Effekthandel Väst (Gothenburg) markets; Uppland and parts of Skåne continued post-project

Load forecasting practice — outsourced, not built in-house

The local flexibility market needs a day-ahead load forecast to know when and how much flexibility to call — but Vattenfall Eldistribution did not build that forecast itself. A 2023 Uppsala University thesis written from inside the company (with a company supervisor) evaluated the machine-learning forecasts used for Uppsala across both CoordiNet winters, and records that this project was the first time Vattenfall Eldistribution had created load forecasts for its grid, and that developing and running the forecasts was outsourced to Expektra. The outsourcing left the company without a detailed internal understanding of how its own forecasts worked — including basic facts like how the day-ahead and intraday models actually differ — a gap the thesis’s first phase set out to map, through interviews with Expektra, Vattenfall operations and R&D staff, and E.ON (on the Switchmarket platform), plus an interactive presentation to operations; a separate workshop discussed candidate error factors. Of seven candidate forecast-error drivers examined, significant grid users deviating from their submitted production plan (mainly the electric boiler) and flexibility activations dominated; forecast error unexpectedly decreased at high load (high prices, extreme cold, and activations when measured against load without flex) — the situations where accuracy matters most, though the thesis’s descriptive analysis cannot establish statistical significance. See Load Forecasting › Build vs. buy — the Swedish forecasting vendor landscape for the broader build-vs-buy landscape this sits inside. (Source - Wiss Utvärdering av Lastprognoser CoordiNet Uppsala (2023))

The SWITCH platform

The digital market platform was built in-house by E.ON Energidistribution. Named SWITCH, it consists of a market tool, flex tool (DSO operator interface), FSP interface, and P2P platform. After CoordiNet, it continued as E.ON’s platform for live flexibility markets. The flex tool was also adopted for sthlmflex (with NODES replacing the market tool component). See SWITCH for technical detail.

CoordiNet sits at the start of a wider lineage of EU-funded flexibility projects — OneNet and EUniversal followed it, and BeFlexible (2022–2026) continued the pattern of building on CoordiNet-era platforms across multiple EU demonstrations, including SWITCH’s continued development in Sweden. (Source - BeFlexible Technology Booklet (2026))

Market succession

Post-CoordiNet, the four demonstration areas had different fates (Source - Sweco Kartläggning av lokala flexibilitetsmarknader (Ei, 2025)):

• Uppland (Uppsala): continued as UppFlex (Vattenfall Eldistribution), one season (V2023/24), then closed when the capacity situation changed
• Skåne: E.ON Energidistribution launched permanent commercial flex markets (Södra Skåne, Hässleholm, Nordöstra Skåne); the 400 kV grid expansion in October 2024 (+600 MW) made the markets much less necessary, and by some accounts unnecessary, for the primary bottleneck
• Gotland: no continuation reported
• Västernorrland/Jämtland: no direct continuation; JämtFlex (Jämtkraft, V2023/24) used NODES for a single season

sthlmflex — a separate Stockholm market (Ellevio + Vattenfall, using NODES) — ran 4 seasons (2020/21–2023/24) and closed: warm winters + high energy prices + TSO subscription grants combined to eliminate sufficient congestion need.

Effekthandel Väst (Göteborg/Mölndal, NODES) emerged from the same period and remains active as of 2025 — the only non-E.ON flex market with sustained operation.

FSP and PFSP perspective — why organizations joined (or didn’t)

A 2023 qualitative study (Palm et al., Utilities Policy 82) interviewed 25 organizations (30 informants) invited to participate in CoordiNet Uppland and Skåne for the V2020/21 season — 12 enrolled FSPs and 13 PFSPs (potential FSPs who declined). The study provides the supply-side account of the project that the CoordiNet final report does not. (Source - Palm et al LFM Drivers and Barriers (2023))

Monetary incentives were less important than expected. Most FSPs did not join primarily for money — activation volumes were too low and too infrequent for a purely economic case. The potential for future revenue mattered more than actual payoff. The main drivers were:

• The champion — an individual inside the organization with personal interest in energy/flexibility. Without an internal champion, most organizations would not have engaged. This was the single most frequently mentioned driver across FSPs and PFSPs alike, and had not been identified in prior LFM research.
• Organizational goals and strategy fit — participation aligned with sustainability commitments or energy management strategies
• Social responsibility — a genuine wish to contribute to grid stability and the common good
• Access to an aggregator — FSPs who had an aggregator partner handling bidding, automation, and technical setup were critical; without them, FSPs could not have participated practically

The biggest barriers for PFSPs were not economic:

• Information — difficulty understanding the market concept and, critically, explaining it internally to get organizational buy-in. One quote: “I find it difficult to explain to others.” In large multi-administration organizations (e.g., municipalities), getting all relevant divisions aligned required near-full-time dedicated staff.
• Automation — manual bidding and activation was universally described as unworkable for long-term participation: “It must work via some control system.”
• LFM design — day-ahead (D-1) gate closure meant real-time information between bidding and activation could not be used; short-term contracts prevented investment in enabling technology; pricing one’s own flexibility was difficult (“I have no idea what price to set”)
• TSO market competition — one FSP estimated TSO balancing market revenue at “at least five times higher” than CoordiNet Skåne’s local market rate, creating rational diversion of capacity to national markets

Baseline as a trust problem: FSPs found baseline calculation genuinely confusing. One story circulating among participants: a sports arena turned on all its lighting, used that as its base load, and then earned money on the resulting falsified reduction (Palm et al. report it as a circulating story, not a verified incident). It sits alongside the FSPs’ confusion about how baselines worked as evidence of a perceived integrity gap in the settlement mechanism. See Baseline Methods for the technical dimensions.

~15 PFSPs declined even to be interviewed, citing lack of knowledge — the information barrier was severe enough to prevent research participation itself.

Policy and regulatory legacy

CoordiNet triggered a national standardization project using IEC 62325 CIM for flexibility products — leading to Sweden’s national product catalog (first draft 2022). The project confirmed the CAPEX bias in revenue cap regulation as a real barrier (Ei is addressing this through the RP5 TOTEX reform from 2028). CoordiNet is the principal empirical reference point for Swedish and EU discussions on local flexibility market design and TSO-DSO coordination.

Related pages

• SWITCH — the digital platform built by E.ON for CoordiNet
• Flexibility Market — concept and Swedish landscape
• Congestion Management — the primary use case
• E.ON Energidistribution — platform developer and Skåne/Jämtland DSO
• Svenska kraftnät — TSO partner
• Balancing Markets — mFRR forwarding and FCR-D coordination developed in CoordiNet
• Source - CoordiNet D4.7.2 Swedish Demonstration (2022) — the primary source document
• Source - Ei Flexibility in Distribution Grids (2023) — Ei’s evaluation of CoordiNet alongside sthlmflex and Effekthandel Väst