Energy Storage

Technologies that absorb electricity at one time and release it at another. Energy storage breaks the fundamental constraint that electricity must be consumed the instant it is generated, making it one of the most versatile Flexibility mechanisms — capable of operating across timescales from milliseconds (frequency response) to seasons (hydrogen, thermal mass).

Storage technologies

Flexibility services from storage

Storage can provide flexibility across all timescales and grid levels:

• Frequency regulation (FCR, aFRR) — batteries respond in milliseconds/seconds, faster than any generation asset. The SO GL classifies these as “limited energy reservoirs” and requires FCR activation for 15-30 minutes during alert state in the Nordic synchronous area (Art. 156(9-11)), with energy reservoir recovery within 2 hours (Art. 156(13)(b))
• Peak shaving — discharge during demand peaks to reduce congestion
• Energy arbitrage — charge when cheap (high wind/solar), discharge when expensive. At household scale a practitioner post puts one home battery’s spot timing at about 300 SEK for September 2026 (Source - Hedelin Home Battery September 2026 (LinkedIn))
• Backup/resilience — uninterruptible power supply for critical loads
• Renewable integration — smooth variable generation output
• Voltage support — inverter-based reactive power at distribution level

Behind-the-meter storage (household batteries paired with solar) adds a distributed dimension. When aggregated through aggregators or Virtual Power Plants, thousands of small batteries become a significant system resource.

EU regulatory framework

The Clean Energy Package treats storage as a distinct market category with specific protections:

• Storage must participate in markets on equal footing with generation and Demand Response (Regulation Art. 3(j))
• No double charging — stored electricity should not be subject to network charges twice (once when stored, once when discharged) (Directive Art. 15(5)(b))
• DSO storage ownership ban — DSOs cannot own or operate energy storage unless the market fails to deliver needed services and the regulatory authority grants a derogation (Directive Art. 36). This preserves the DSO’s role as a neutral market facilitator.
• The same ownership restriction applies to TSOs under similar derogation procedures

The Network Code on Demand Response includes storage within its scope — Controllable Units (CUs) explicitly encompass storage assets, giving them standardized market access pathways alongside demand response and distributed generation. (Source - NC DR Proposal (ENTSO-E and EU DSO Entity, 2024))

Tariff treatment — neither double-charged nor fully exempt. The Commission’s January 2026 network-charges guidance sharpens the Art. 15(5)(b) position into a two-sided test for regulators. Tariff methodologies “should most importantly be designed so as not to hinder the development and roll-out of storage”, and in particular must prevent double-charging (paying both generation and consumption charges). But NRAs should equally “carefully consider whether full exemption regimes, whereby the storage operators do not make any contribution to network cost recovery, are justified in line with the cost-reflectiveness principle” — exempt storage shifts cost onto other users. The guidance also warns that charging and discharging which ignores local congestion status would make a growing storage fleet part of the grid problem rather than a solution, so it invites differentiation by location and by charge/discharge timing to reward grid-friendly operation. (Source - C(2026)126 Future Proof Network Charges Guidelines)

Swedish context

Sweden has significant existing “storage” in the form of hydropower reservoirs (primarily in the north), which provide seasonal and daily flexibility. Battery storage is growing rapidly.

Deployed battery capacity (July 2024): ~370 MW installed, >1,400 MW under construction, >8,300 MW planned in a 3–10 year horizon. (Source - Ei R2025-19 Sweden Electricity and Gas Market 2024 (2025))

Grid-connected third-party storage (DSO-reported, EIFS 2022:5): Ei’s SGI data provides a complementary system-wide count of storage directly connected to the grid and not owned by the SO:

3.4× growth in a single year. Largest 2024 reporters: E.ON (216.5 MW), Ellevio (36.6 MW), Vattenfall (27.9 MW). The figure reflects storage at the DSO connection point and is distinct from the FCR-prequalified figure above — it captures all grid-connected third-party BESS regardless of market participation. (Source - Ei SGI Data 2023-2024)

Quantified potentials to 2030 (hourly timescale)

FlexAbility (2025) provides realistic technical maximum potentials at hourly timescale for 2030:

(Source - FlexAbility Delrapport 1 (2025))

V2G impact on LV networks

FlexAbility (2025) used Plexigrid’s Monte Carlo simulation platform to study how increasing V2G penetration — combined with rooftop solar PV — affects residential LV distribution networks when EVs provide FCR-D upregulation service (~58.4 activation hours/year). Two Swedish networks were compared: (Source - FlexAbility Delrapport 3 (2025))

Both Stockholm-suburb constraints occurred on weekends in Q3, named the critical window in both networks (the southern network’s feeder overcurrent instead came on Q2 weekdays at midday) — the combination of low residential load, high solar output, and FCR-D V2G activation. The Stockholm suburb network was substantially more robust, with constraints only at 80–90% penetration vs 30–40% in the southern reference area. The difference is attributed to transformer and cable dimensioning.

Key implication: results are highly network-specific. National-level V2G assessments are unreliable without per-network analysis. A 2.7× range in substation-overload thresholds (30% vs 80%; 2.25× for overvoltage, 40% vs 90%) between two Swedish LV networks means that some areas can accommodate much higher V2G penetration than others before grid reinforcement is needed.

V2G without smart charging coordination risks destabilizing LV grids while stabilizing the transmission network — directly counteracting the resilience that flexibility services are intended to provide.

V2G structural barriers

V2G’s 5,000 MW potential is largely theoretical for 2030 due to non-technical barriers. See Vehicle-to-Grid › Swedish regulatory grey area for the full analysis. Summary:

• Svenska kraftnät physical address rule: Svk requires a fixed physical address for ancillary service participation — mobile EVs cannot easily comply across multiple discharge locations
• Double taxation (dubbelbeskattning): energiskatt + moms charged when electricity enters the EV battery; a second customer is taxed when the electricity is fed back out. Skatteverket’s refund mechanism only applies within the same concessionary network — cross-area V2G loses the refund (Source - KTH Thesis V2G Sweden 2024)
• Classification ambiguity: no official ruling on whether a V2G-enabled EV is treated as a microproduction facility (like solar) or a mobile injection point; different DSOs interpret this differently (Source - Power Circle V2X Synthesis 2024)
• ISO 15118-20 / OCPP 2.1 not universally deployed: the communication standards enabling V2G are defined but not yet universally implemented; AC bidirectional protocols are particularly incomplete
• Business models immature: no established framework for V2G revenue sharing between vehicle owner, aggregator, energy company, and grid operator
• Battery degradation uncertainty: contrasting research findings on V2G’s impact on battery health; OEMs bear warranty costs and have imposed limits (VW caps energy cycled via V2G)
• Sweden-specific — installed base problem: Sweden’s high EV adoption means many owners already have non-V2G-compatible wallboxes (7,000–30,000 SEK investment); upgrading is a barrier absent in lower-penetration markets (Source - KTH Thesis V2G Sweden 2024)

BESS fire safety — no national siting guidelines yet

Energiforsk fire/explosion-risk research (part of the 2021–2025 Risk- och tillförlitlighetsanalys programme) finds BESS fire risk is not eliminable by any single measure — pressure-relief and deflagration protection are needed, with blast zones (kastzon) up to 25 m if underdesigned; toxic smoke spread is practically limited to 10–20 m in real scenarios, well short of the “hundreds of metres” implied by worst-case theoretical models. The explicit finding: no national Swedish guidelines yet exist for BESS siting and protection — a genuine regulatory gap, distinct from the bygglov permitting friction discussed below. (Source - Energiforsk 2025-1148 Syntesrapport Risk Och Tillförlitlighet (2025))

A March 2026 companion report from the same programme (Energiforsk 2026:1163) narrows in specifically on consequence-distance methodology, using established standards (NFPA 68, SS EN 14994 for explosion venting) and real-incident analysis rather than worst-case theoretical modelling: deflagration debris zones up to ~25 m, harmful toxic concentrations typically confined to the immediate vicinity (<30 m; ~10–20 m from the fire source in the incidents studied), with conservative dispersion models tending to overestimate both when conservative assumptions ignore thermal buoyancy in the smoke plume. Its own conclusion: BESS should not be treated as inherently high-risk when properly designed, but documented risk analysis, correctly dimensioned pressure relief, and räddningstjänst coordination remain essential — and the report explicitly calls, again, for harmonized national guidelines with clear acceptance criteria that still do not exist. The report describes the absence of harmonized Swedish regulation as a source of significant uncertainty, and adds a practical design/procurement appendix (Appendix A). Its scope is large stand-alone outdoor systems; indoor and small installations were not studied.

A companion study in the same programme found the fear that batteries cause large voltage variations is overstated: existing rule-of-thumb connection limits (AMP 2020, originally designed for solar PV) understate battery hosting capacity, meaning batteries can be connected faster and in greater volume without network reinforcement than current practice assumes — though full-power discharge can still cause voltage-angle shifts and oscillation risk where a battery is co-located with synchronous generators (e.g. a CHP plant).

BESS deployment barriers — building permits

The primary practical barrier for deploying BESS in urban environments is bygglov (building permits). For stationary batteries above certain size thresholds (related to fire risk and energy content), Swedish planning law requires building permits — adding months to deployment timelines. Urban locations face the most constraints: heritage buildings, multi-tenant ownership structures, and dense planning conditions delay or block projects precisely where grid congestion is most acute (e.g., Gothenburg, Stockholm).

A finding from Energiforsk research (cited in Göteborg Energi’s Elektrifieringsrapport, 2025): the binding constraint for connecting BESS to the grid is transfer capacity (how much power can flow to/from the connection point), not voltage quality (as was previously assumed in some analyses). This means:

• More BESS can be connected to existing grid infrastructure than previously estimated
• DSO investment to unlock BESS scale-up should target transfer capacity improvement, not voltage quality upgrades
• The voltage quality benefits of BESS (reactive power, harmonic filtering) are real but not the bottleneck

(Source - Göteborg Energi Elektrifieringsrapporten nr 1 (2025))

Svk capacity services for Hisingen/Stenungsund

Svenska kraftnät has announced multi-year capacity service contracts for the Hisingen (Gothenburg) and Stenungsund (industrial cluster north of Gothenburg) areas, beginning around 2027. These are bilateral contracts where BESS operators receive payment for making capacity available to Svk during expected scarcity periods.

Significance:

• First structured long-term revenue certainty for large BESS in Sweden outside spot arbitrage and FCR markets
• Multi-year contracts enable investment financing — revenue can be underwritten against contracted income
• Represents Svk acting as a capacity procurer at the TSO level (analogous to capacity markets elsewhere, but via bilateral contracts)
• Stackable with FCR, aFRR, mFRR, and local flex market revenue streams

(Source - Göteborg Energi Elektrifieringsrapporten nr 1 (2025))

Battery acceleration

The stationary battery trajectory is the most credible near-term number:

• ~750 MW prequalified today → 8,000 MW realistic by 2030
• 9,500 MW already in connection queue at E.ON alone
• Battery prices falling ~30% by 2030
• Green tech deduction (ROT/RUT analogue) accelerating hembatterier (home battery) deployment
• FCR-N prequalification grew +30% in the three months to January 2025 — trend already visible
• CheckWatt (Gothenburg-based) operates the largest Swedish home BESS VPP: 15,000+ sites / ~100 MW FCR-D across the Nordics (Source - CheckWatt Website (2025-2026)); locally in Göteborg, ~500 batteries / 5.5 MW participate with a dual-market strategy (local DSO flex + national balancing) (Source - Göteborg Energi Elektrifieringsrapporten nr 1 (2025))
• Flower (Stockholm-based) operates Sweden’s largest confirmed grid-scale BESS portfolio: 4 operational sites (Bredhälla 42.5 MW, Kungälv 15 MW, Grums 10 MW, Sköldinge 4 MW — summing to 71.5 MW, though Flower’s own disclosure separately headlines this as “63 MW”; the two figures don’t reconcile in the source itself) with 70 MW more in development for end 2026 → ~133 MW total Sweden (per Flower’s own stated total). Primary markets: FCR-N, FCR-D, FFR. March 2026 performance: EUR 9,568/MW net revenue; 86% from FCR capacity, 14% energy markets; 0.8 daily cycles. (Source - Flower Website (2024-2026))

The 86%-from-FCR / 0.8-cycles-per-day operating point is the high-water mark of a maturing market: FCR demand is flat to 2030 while prequalified supply already exceeds need ~14×, so the business case must migrate toward energy-paying markets (mFRR, aFRR, arbitrage) that consume the idle cycling headroom. See The Swedish BESS Business Case — Revenue Stacking and the FCR Saturation Problem.

France is already further down this path: BKW’s Jill Huber (operating across Switzerland, Germany, and France) reports that rapid growth in large-scale batteries is already causing saturation in French ancillary services — explicitly cited as an early warning for other European markets (Source - Powernaut Flex Trends Report (2026)). Sweden’s flat-FCR-demand ceiling makes the same squeeze arithmetically predictable.

Modelled revenue benchmark — Clean Horizon Storage Index

Clean Horizon’s free monthly index (modelled gross revenue of a reference battery, k€/MW/year) gives the wiki its only external revenue time series for Nordic batteries. For a 1h battery in Sweden (SE3) the annual mean fell from about 781 in 2023 to 333 (2024), 220 (2025) and 107 (January–August 2026); Finland fell from 390 to 64, while DK1 and DK2 stayed at 250–270 in 2026. The 4h index runs 1.8 times the 1h index in Sweden for 2025, against about 1.0 times in 2023, so duration is increasingly rewarded. The index changed methodology in January 2026 (average instead of marginal MW), so 2026 is not like-for-like with earlier years, and it does not split revenue by source. (Source - Clean Horizon Storage Index (2026); analysis in Nordic Battery Revenue Trends 2023–2026)

External benchmark — Fraunhofer ISE’s German price simulator

Fraunhofer ISE’s Energy-Charts platform launched a free simulator on 1 October 2026 that reruns each German-Luxembourg day-ahead auction from the real bid curves with an added battery fleet, for 2025 onwards. Its published example, in the scenario with block bids and no cross-border trade: 20 GW at 2 hours cuts the average daily price spread from 130 to 53 EUR/MWh (-60 %), hours above 200 EUR/MWh from 169 to 20 and negative-price hours from 575 to 271, and lifts the market value of solar by 35 %. It is a day-ahead-only what-if with fixed assumptions, not a forecast or a project revenue estimate, it covers Germany only, and its development was funded by storage and solar companies and an industry association. It illustrates the same saturation mechanism as the sections above, from the price side. (Source - Fraunhofer ISE Energy-Charts Battery Price Simulator (2026))

Colocation — grid access as the primary driver

In many European countries, obtaining a new standalone grid connection for a BESS is effectively impossible — connection queues are years long and new capacity is unavailable near demand centres. Colocation with an existing solar or wind asset sidesteps this barrier: the connection already exists, permitting timelines are shorter, and in some markets storage-only charging may be restricted but charging from co-located generation is permitted.

Aurora Energy Research (Hanns Koenig) characterises this as a structural shift: colocation is no longer primarily an economic optimisation — it is a grid access strategy. (Source - Powernaut Flex Trends Report (2026))

Five practical colocation drivers (Sonnedix, Mustapha Obalanlege):

1. Single grid access point — reduces permitting time and connection queue exposure
2. Shared maintenance supply chains — operations and asset management can be combined
3. Direct charging from on-site generation — avoids grid charging restrictions and reduces curtailment
4. Easier land acquisition — adjacent to an existing site rather than a greenfield hunt
5. Revenue optionality — the combined asset can optimize across spot arbitrage, FCR/balancing, and local flex; reduces capture price risk and negative-price exposure

Implication for grid planning: the Swedish BESS build-out trajectory (~750 MW → 8,000 MW by 2030) will increasingly be driven by where grid connections are available, not just where economics are best. This reinforces the interaction between flexible connection frameworks and storage deployment.

Portugal live example (Sonnedix, Vijandren Naidoo): one of Portugal’s first utility-scale batteries faces grid charging restrictions during certain hours; no established grid code for batteries exists; TSO technical specifications lag. Flexible connection agreements are becoming standard but impose revenue model adjustments that were not anticipated in original business cases. The grid cannot be treated as a static input to BESS business case modelling.

Tolling agreements

A tolling agreement is a commercial structure where a battery developer grants a counterparty (the “toller”) exclusive access to the battery’s capacity for a fixed term, in exchange for a guaranteed payment. The toller bears all merchant risk and decides when and how to dispatch the asset. The developer receives predictable revenue independent of market outcomes.

Approximately 20 tolling agreements had been signed across Europe by mid-2026. They are emerging as the dominant BESS financing structure as banks have become reluctant to lend against merchant battery risk — making it harder to finance BESS projects on pure merchant terms. Tolling structures fill this gap by providing a creditworthy counterparty willing to absorb merchant exposure. (Source - Powernaut Flex Trends Report (2026), Source - Konsultrapport Balansansvarigas förutsättningar Merlin Metis (2026))

Hybrid PPAs (bundling renewables + storage in a single contract) remain rare despite years of industry discussion. The reason: full battery optimisation is incomprehensible to most corporate offtakers. Battery dispatch involves simultaneous optimisation across FCR, aFRR, mFRR, arbitrage, and local markets — a complexity that cannot easily be encapsulated in a corporate PPA structure. Tolling agreements isolate the storage layer in a separate bilateral structure, leaving the optimisation risk with a specialist operator.

See The Swedish BESS Business Case — Revenue Stacking and the FCR Saturation Problem for the revenue-stacking model that makes the toller’s optimisation both necessary and valuable.

Hydro — the long-timescale backbone

With 16,400 MW installed capacity, hydro dominates Swedish storage at daily and seasonal timescales. At hourly timescale, 6,000 MW is realistic by 2030 (up from 3,000 MW today, adding Sweco-estimated +3,400 MW efficiency potential plus +700 MW assumed additions). Pumped hydro expansion is constrained: the main candidate project (Juktan, 300 MW) is deferred to 2032, meaning pumpkraft will not contribute materially to 2030 storage.

FCR participation — battery-specific technical rules

The Nordic TSOs’ FCR technical standard (2023) specifies exact requirements for batteries as Limited Energy Resources (LER). These rules determine which battery configurations can qualify for each FCR product and how they must behave during operation. (Source - ENTSO-E FCR Technical Requirements Nordic (2023))

Batteries connecting to the grid are classified as kraftparksmodul under Generator Connection Requirements (RFG type A–D depending on installed power). Type A (< 1.5 MW) requires no FSM capability; type C/D resources must have FSM as a prerequisite for FCR participation. The RFG classification determines applicable frequency ranges, ramp rates, and fault ride-through profiles before the Nordic FCR prequalification process begins.

NC RfG 2.0 — regulatory scope expansion: Under the current RfG (EU 2016/631 Art. 3.2(d)), energy storage is explicitly excluded from the regulation’s scope except pumped hydro. Swedish practice applies PPM requirements to BESS in production mode by interpretive convention, but there is no binding EU statutory basis. ACER Recommendation 03-2023 (December 2023) proposes to remove this exclusion: BESS would be explicitly in scope for both “input” (charging) and “off-take” (discharging) modes, with requirements tracked to asset type (A–D). The recommendation is not yet adopted as law; comitology is ongoing with full application expected around 2027–2030. Additionally, Type B/C/D BESS (≥10 MW or ≥110 kV) would face a mandatory grid-forming capability obligation — requiring the inverter to act as a voltage source for autonomous voltage/frequency stabilisation, not just a current-follower. RoCoF ride-through requirements would also apply: ±4.0 Hz/s (0.25 s) / ±2.0 Hz/s (0.5 s) / ±1.5 Hz/s (1 s) / ±1.25 Hz/s (2 s). See Generator Connection Requirements › NC RfG 2.0 — the Commission’s draft revised RfG (July 2026). (Source - ACER Recommendation 03-2023 NC RfG DC (2023))

LER classification and endurance minimums

A battery qualifies as LER if it cannot sustain full FCR activation for 2 hours. In practice, nearly all commercial BESS installations are LER — the 2-hour threshold would require a 1 MW battery to have ≥2 MWh of usable storage.

Minimum endurance requirements:

Sizing implication: A 1 MW FCR-N contract requires at minimum a battery with 1 MWh of usable FCR energy (1 MW × 60 min). A battery providing 1 MW FCR-N + 340 kW opposite-direction reserve must have sufficient SOC headroom for both the main product and the reserve margin.

NEM and AEM — operational energy management

Two automatic functions control how a battery behaves as its state of charge (SOC) approaches operational limits:

NEM (Normal Energy Management) — voluntary recharging mode:

• Activation trigger: remaining endurance <30 min (FCR-N) or <20 min (FCR-D)
• Entity reduces FCR delivery and uses the headroom to recharge
• Maximum recharging rate: baseline reserved capacity is 34% for FCR-N (up to 50% with TSO approval) or 20% for FCR-D (up to 34% with TSO approval)
• Entity remains in the market and continues partial FCR delivery during NEM

AEM (Absolute Energy Management) — unavailability declaration:

• Activation trigger: SOC falls below the 5/60 × C/E ratio threshold (where C = contracted capacity in MW, E = energy capacity in MWh)
• Entity is declared unavailable; exits FCR delivery entirely
• TSO is notified via 5-minute rolling average frequency reference signal
• AEM-declared entity cannot be penalised for non-delivery while in AEM state

Practical meaning for battery sizing: NEM allows a battery to stay in the market longer by recharging slowly. AEM is the hard floor — below it, the battery exits the market. Well-designed battery management systems (BMS) target NEM activation as a soft ceiling well above AEM.

Central control systems — availability requirement

Batteries that deliver FCR via a central control platform (the standard aggregator architecture) must meet:

• 99.95% central system availability — equivalent to <4.4 hours downtime per year
• Redundancy or local fallback — each battery unit must be able to deliver FCR autonomously if central control fails
• Frequency measurement: must use a sensor within the same LFC (Load Frequency Control) area as the contracted FCR; cross-area measurement is not permitted for FCR control

This 99.95% requirement applies to the control infrastructure, not just the battery hardware. Cloud-based aggregation platforms must maintain this uptime or ensure local fallback capability in each unit.

Type qualification for small batteries (≤100 kW)

Batteries ≤100 kW each can qualify for FCR via type qualification — a single test of one representative unit, with the certificate extending to all identical units of the same technology and model. This pathway significantly reduces the per-unit prequalification burden for home batteries and small commercial BESS.

Combined with the dynamic prequalification extension rule (add up to 25% / 1 MW / max 3 MW of new units without re-testing), aggregators can grow a portfolio of small batteries incrementally without triggering full re-prequalification for each addition. See also Aggregation › FCR qualification pathways for aggregated resources.

Grid resilience capabilities

Batteries and inverter-based storage provide grid resilience capabilities beyond standard market services, formally identified in Energimyndigheten ER 2025:35 as part of beredskapsflexibilitet (preparedness flexibility): (Source - Energimyndigheten ER 2025-35 Förbättra Flexibiliteten (2025))

Synthetic inertia (syntetisk svängmassa / syntetisk rotationsenergi)

Batteries equipped with special inverter control settings can continuously emulate the rotational inertia of synchronous generators. Traditional thermal and hydro generators contribute inertia passively through their physical rotating mass, stabilising frequency against rapid disturbances. As these generators are displaced by inverter-based renewables, system inertia declines.

Synthetic inertia from batteries addresses this gap. It is distinct from FCR and FFR:

• FCR/FFR: respond reactively to frequency deviations after they occur
• Synthetic inertia: acts continuously and preventively, resisting the rate of frequency change (ROCOF) before a deviation develops

Svenska kraftnät explicitly identifies a growing need for synthetic inertia capability as inverter-based generation grows. This is not yet a standardized market product in the Nordic Balancing Markets, but it is on Svk’s technical roadmap.

Ö-drift (island operation)

Batteries combined with solar and EVs can enable distribution grid segments to operate as microgrids independently from the transmission system during major outages. See Island Operation for the full taxonomy, case studies, and regulatory framework. This requires coordinated inverter control but not hardware modifications beyond what is already present. ö-drift capability is most relevant during the återuppbyggnad (reconstruction) phase following a major grid failure.

Two operational Swedish examples confirm the technology at different scales:

E.ON Simris (2015–2018, InterFlex H2020): Grid-connected public LES in southern Sweden. 333 kWh / 800 kW BESS as grid-forming unit; 500 kW wind + 442 kWp PV + 480 kW bio-diesel backup; 150 customers on 10 kV. During a 12-hour islanding test (April 2018), power quality (frequency, voltage, THD) was better in island mode than when grid-connected. Operated by E.ON; research in partnership with RWTH Aachen University. Key cost finding: BESS+PCS up to 4× cheaper than grid upgrade for mitigating RES-induced voltage deviations. (Source - InterFlex Simris Microgrid (2018)) The BESS inverter was deliberately oversized to deliver 2× rated current for 2 seconds to allow conventional directional overcurrent protection to function in island mode — without this, the inverter’s inherent 1–1.5 p.u. current limit would prevent reliable fault detection. This illustrates a fundamental BESS island design trade-off: protection capability requires either inverter oversizing or more advanced protection architectures (differential, distance). (Source - Energiforsk 2023-957 Felbortkoppling i Mikronät (2023))

Vattenfall Arholma (commissioned August 2023): Island in northern Stockholm archipelago, ~250 residents. Two 160 kW / 336 kWh BESS units (320 kW / 672 kWh total); nominal capacity ≈ 2.1 hours at full discharge, with a formal reliability design target of 1 hour of island supply at 99% probability. Fault on submarine cable (failure rate 2.1 f/yr, dominant reliability event) detected in milliseconds; automatic ö-drift activated. Real-time software maintains island voltage/frequency. Next phase (2025): extend control to customer heat pumps and heating systems with sequential load reconnection to prevent cold-load pickup. (Source - Vattenfall Arholma Microgrid (2025))

A pre-commissioning Energiforsk reliability study (Ying He, May 2023) using actual Arholma load data (Oct 2015–Jun 2019) quantified the islanding benefit:

The study also found LOLP 1.5% — the local generation is insufficient to meet island demand for approximately 130 hours/year (primarily winter nights when solar output is zero and load peaks exceed battery discharge capacity). This capacity shortfall is the primary technical rationale for the next-phase customer DSR extension: the battery alone cannot guarantee island supply during winter peaks. (Source - Energiforsk 2023-948 Reliability Analysis Microgrid (2023))

Power-as-a-Service at Arholma: The BESS is owned and operated by Vattenfall Elanläggningar (part of Vattenfall Network Solutions), delivered as a capacity service to Vattenfall Eldistribution (the DSO). This structure is consistent with Art. 36 DSO storage ownership restrictions even within the same corporate group — the DSO does not own the asset but procures the service from an internal affiliate.

Protection engineering challenge for island BESS: A 2025 Lund University MSc thesis (Molin & Thunholm, in collaboration with Vattenfall R&D) identifies the core engineering constraint for battery microgrids in ö-drift: inverter-interfaced DERs produce far lower fault currents than synchronous generators, causing conventional protection devices to malfunction. Synchronous generators deliver ~6 p.u. fault current; grid-forming inverters (GFMIs — necessary for islanding without synchronous machines) are limited to ≤2.0 p.u. The Arholma BESS uses GFMIs, which switch from voltage-source to constant-current-source behaviour when current-limited. This creates two anomalies: (1) at low Thévenin impedance (stiff networks), LL faults can produce smaller fault currents than SLG faults — the opposite of conventional protection assumptions; (2) fault current no longer increases predictably with decreasing fault impedance, invalidating standard relay settings. Practical solutions: programmable LVCBs (low-voltage circuit breakers) whose settings update automatically when the microgrid enters island mode, and — for simple cases — MV-only protection where all LV faults trip via the upstream MV relay within a calculable load threshold. These constraints define the network size limits achievable with a given inverter size and are a central planning consideration for future microgrid design. (Source - Lund Arholma Microgrid Fault Detection (2025))

Dödnätsstart (black start)

Batteries combined with renewables can restart the grid from a completely dead state — dödnätsstart (black start). Traditionally this required specific hydro plants with black start capability and was tightly controlled by Svenska kraftnät. Distributed DER-based black start is now technically feasible and adds resilience by reducing dependence on a small number of designated hydro plants.

Cybersecurity exposure

The same large installed battery capacity (~1 GW / 1.6 GWh at end 2024) that provides flexibility and resilience is also a potential cybersecurity attack surface if those batteries are internet-connected and inadequately secured. RISE Nordic32 simulations (ER 2025:35) demonstrate that coordinated activation of compromised internet-connected DERs can destabilize grid frequency at system level. This risk applies to batteries as much as to heat pumps. See Beredskapsflexibilitet › Cybersecurity risk from connected DERs for the full cybersecurity risk analysis.

Open communication protocols (Flexibility Communication Protocols) are recommended to reduce this risk relative to proprietary alternatives. See Security and Resilience of the Digitalized Flexible Grid for the system-level treatment.

Data gaps

• Ei position on DSO storage ownership (Art. 36 derogation approach)
• Synthetic inertia market product design — Svk’s timeline and compensation framework
• Regulatory framework for ö-drift and dödnätsstart — procurement, testing, compensation
• Whether Ei, MSB, or another authority has responded to Energiforsk’s repeated call (2025:1148, and again in 2026:1163) for national BESS siting/fire-safety guidelines with clear acceptance criteria
• Whether Energiföretagen has revised the AMP 2020 battery connection-limit rule of thumb per the Energiforsk recommendation