Island Operation

Island operation (Swedish: ö-drift) is the mode in which a grid section operates as a self-contained electrical “island” — with local generation maintaining balance and frequency, without connection to the wider grid. In Sweden, ö-drift functions both as a normal operational mode for isolated communities and as a preparedness instrument of last resort for sustaining critical societal functions during prolonged transmission failures.

Framing and context

Svenska kraftnät‘s public position is unambiguous: “A national, interconnected electricity supply is the best outcome for Sweden.” Ö-drift at the local/regional level is a fallback — relevant only in “genuinely severe situations” where national restoration plans cannot quickly reconnect areas after transmission failure. It is not a cost-effective operational mode under normal conditions. (Source - Svk Om Ö-drift (2025))

ö-drift as a localization principle: Svk’s February 2025 localization report explicitly cites civil defense and ö-drift capability as one rationale for siting plannable (dispatchable) production near cities and areas with inflexible consumption. The argument is that co-location improves local access to generation during system restoration scenarios, reducing dependence on north-south transmission. Large industrial establishments with limited flexibility also benefit from nearby plannable production for the same reason. This is notably a transmission-level planning argument, not an operational grid argument — ö-drift readiness is embedded in Svk’s geographic guidance for new connection applications. (Source - Svk Planering för ökad elanvändning (2025))

This framing matters because island operation is sometimes discussed as a proactive resilience strategy for remote communities. In the Swedish regulatory context, the primary driver is emergency preparedness (elberedskap), not normal grid economics.

Four levels of island operation in Sweden

Svenska kraftnät distinguishes four scales, with different responsible actors:

The Arholma and Simris microgrids are level 2–3 cases: DSO-led projects serving a defined island community. (Source - Vattenfall Arholma Microgrid (2025), Source - InterFlex Simris Microgrid (2018))

Technical requirements

Black start (dödnätsstart) — the binding constraint

The single most essential prerequisite. The island must have local generation capable of starting from a completely dead grid — without an external voltage reference to synchronize to. Without black start capability, ö-drift is impossible. In practice, most potential ö-drift areas rely on small hydro, diesel, or (increasingly) BESS-based black start.

Frequency and voltage regulation

Once the island is energized, load and generation must be continuously balanced to maintain 50 Hz and acceptable voltage. This requires dispatchable, controllable generation — not just variable renewables (solar and wind can participate but cannot alone sustain stable frequency). BESS with grid-forming inverter control can fulfill this role.

Weak-grid operation

An isolated microgrid is a svagt nät (weak grid) — fundamentally different from normal grid-connected operation:

• Low fault currents: inverter-based sources (batteries, solar, wind) deliver ≤2 p.u. fault current vs 6+ p.u. from synchronous generators. Standard protection relay settings designed for strong-grid operation may fail to detect faults in island mode.
• No voltage stiffness: frequency and voltage are supported only by local resources; any sudden large load change or generation trip creates larger transients than in the mainland-connected case.
• Coordination complexity: DSO must manage protection, switching, and dispatch under abnormal conditions with reduced margins.

The protection engineering challenge for Swedish microgrids is documented in detail in Source - Energiforsk 2023-957 Felbortkoppling i Mikronät (2023) (Simris and Arholma cases) and Source - Lund Arholma Microgrid Fault Detection (2025) (Arholma PowerFactory simulations). The key finding: a low-voltage circuit breaker (LVCB) is a practical enabling component for fault detection and isolation in inverter-dominated island grids.

Cold-load pickup

When reconnecting to the mainland after an islanding event, thermostatically controlled loads (heating, refrigeration) that have been deprived of power all switch on simultaneously — potentially exceeding grid capacity and causing a new fault. Mitigation: sequential reconnection, where loads are re-energized one by one. The Arholma next-phase plan includes remote-controlled switches at customer premises specifically to manage sequential reconnection. (Source - Vattenfall Arholma Microgrid (2025))

Organizational requirements

Technical capability alone is insufficient. Functioning ö-drift requires:

• A written ö-driftsplan covering all foreseeable contingencies
• Trained personnel who can operate a weak, isolated grid — considerably more demanding than normal operations
• Clear ö-driftsledare (island operation leader) with defined authority and communication protocols
• Coordination with emergency services, hospitals, water utilities — actors outside the electricity supply chain that Svk cannot directly mandate

Limitations

Cannot fully test: a genuine live drill replicating complete isolation is not possible without inducing real outages. Testing is confined to simulations, partial tests, and tabletop exercises. The consequence: function cannot be guaranteed with 100% certainty. (Source - Svk Om Ö-drift (2025))

Hard load prioritization: in most ö-drifts, installed generation is sufficient only for the most critical societal functions. The DSO must decide, in advance, which loads receive power and which do not. Non-critical consumers are expected to arrange their own backup (diesel generators etc.).

Reliability analysis (Arholma, 2015–2019 data): Source - Energiforsk 2023-948 Reliability Analysis Microgrid (2023) modelled Arholma’s reliability with the microgrid system. Results: SAIDI improvement −45% in full island mode, −18% in hybrid mode; but residual capacity shortfall LOLP of 1.5% / ~130 hours per year — the 2×160 kW BESS was not sized to guarantee full coverage of demand peaks.

Swedish case studies

Arholma — Vattenfall Eldistribution

Vattenfall Eldistribution‘s Arholma microgrid in the northern Stockholm archipelago is Sweden’s most documented operational example. Key specifications:

• 2 × 160 kW lithium-ion BESS (total 320 kW, ~2-hour island capacity)
• Solar panels on one battery container
• Real-time control system: detects mainland cable fault in milliseconds, automatically opens breakers and activates island mode
• Commissioned August 2023; ~250 permanent and seasonal residents
• Power-as-a-Service model: Vattenfall Elanläggningar owns and operates the BESS hardware; Vattenfall Eldistribution (the DSO) purchases capacity as a service — consistent with Art. 36 DSO storage ownership restrictions even within the same corporate group

Since commissioning, winter load growth on the island has outpaced the original 2019 design assumptions. Vattenfall’s next phase extends control to customer assets (heat pumps, floor heating) via remote switches, with sequential reconnection and tariff-discount compensation. This is implicit demand response rather than a market mechanism. (Source - Vattenfall Arholma Microgrid (2025))

Simris — E.ON

E.ON’s Simris Local Energy System in Skåne was a 2015–2018 H2020 demonstration (InterFlex project). 333 kWh / 800 kW BESS, 500 kW wind, 442 kWp PV, ~150 customers. The project included a 12-hour islanding test, demonstrating sustained island operation with inverter-based resources. Cost comparison: BESS + power conversion system up to 4× cheaper than a conventional grid upgrade to serve the same reliability requirement. This economic finding established the technical and commercial foundation for subsequent Swedish DSO microgrid investments. (Source - InterFlex Simris Microgrid (2018))

Gotland — total defence island operation planning

Gotland presents the most strategically significant island operation planning case in Sweden, both technically and from a total defence perspective. (Source - Nationell Dialog Flexibilitet Nätkapacitet 12 Maj 2026)

Energy transition context

Gotland is currently connected to the mainland via two DC cables in the regionnät. Wind and solar already cover approximately 50% of the island’s electricity needs. Major change: in 2030, Svk will replace the DC cables with a transmission-level AC connection — two 220 kV submarine AC cables providing substantially higher capacity and better power quality than the existing DC link. The AC connection will also facilitate the planned expansion of wind generation on Gotland, including several repowering and new wind parks under GAIST.

GAIST (Gotland Accelererar i Samverkan för Grön Tillväxt) is the island’s regional collaboration platform for green growth. Current development pipeline: Nåsudden generation shift (~50 turbines, 0.8 TWh), Centrala Gotland (~50 turbines, 1 TWh), Ran (~50 turbines, 2–3 TWh), Boge vindpark (~15 turbines, 0.3 TWh), e-fuel hub, reserve and base power (gas turbines), solar parks with batteries.

Three-month island operation — Typsituation 4

Planning for Gotland’s island operation capability has been formalized under MCF’s Utgångspunkter för totalförsvaret, Typsituation 4 — Anfall mot Gotland (Attack against Gotland). The planning basis is a three-month island operation capability. Energy scenarios indicate that electricity use at heightened readiness may be at least as large as in a normal situation — meaning the island must be capable of sustaining full civilian and military electricity demand without the mainland connection for 90 days.

This is distinct from the level-2 ö-drift cases at Arholma or Jönköping: it is not a backup for a local grid fault but a wartime capability for national defence.

FRaM — wind/defence coexistence

The primary obstacle to expanding Gotland’s generation capacity is the density of military interests on the island: radar systems, communication corridors, flight zones, classified facilities, and sea/blast areas. The FRaM project (Försvarsmaktens projekt om möjligheterna till vindkraft på Gotland) is a Försvarsmakten-led analysis of wind power coexistence conditions:

• Scope: analysis only (not implementation); methodology development; condition-neutral assessment
• Strategic importance: Gotland is both complex and strategically critical for total defence
• Prior work: Försvarsmakten has conducted similar analyses in Värmland and Dalarna
• Output: a methodology that can be applied to specific wind project proposals

The broader challenge — many overlapping interests creating conflicts with the energy transition (nature conservation, cultural heritage, recreation, total defence) — is being addressed through a grid group for streamlined permitting, and through GAIST coordination between the island’s stakeholders.

VindStyR (flexibility in production): wind power capacity constraints from the limited mainland cable connection are currently managed partly through production curtailment; the 2030 AC transmission will substantially relieve this constraint.

Data gaps specific to Gotland

• Whether the 3-month island operation target has been formally agreed as a Svk elberedskapsanslag-funded capability, or is still at planning/analysis stage
• Technical specification for the 2030 AC transmission (system operator — Svk or Vattenfall Eldistribution; capacity beyond the 2 × 220 kV mentioned; schedule)
• FRaM project outputs — has the methodology been published; which wind projects have been assessed

Relationship to flexibility and DER

Island operation creates a direct value case for local generation and storage that is separate from the flexibility market arguments:

• A BESS that provides ö-drift capability is earning a reliability service (grid resilience) for the DSO, not just a balancing market revenue
• Heat pumps, EV chargers, and other controllable loads can extend island duration by reducing peak demand during islanding events
• Solar panels provide daytime generation but cannot alone sustain frequency — they require either storage or dispatchable backup

This points toward a value-stacking architecture: local DERs that participate in Balancing Markets and Flexibility Markets during normal operation can simultaneously provide ö-drift capability as a third revenue/service stream. No Swedish market currently prices ö-drift capability explicitly — Svk’s elberedskap funding covers capability costs, not market payments.

Svk’s elberedskap funding mechanism

For level-2 local/regional ö-drift, Svenska kraftnät in its elberedskapsmyndighet role can compensate the costs of maintaining ö-drift capability, planning, and testing. This is a grants/compensation mechanism, not a market payment. Svk’s jurisdiction extends only to actors within the electricity supply chain (DSOs, generators) — not to municipalities, hospitals, or other societal actors that are stakeholders in ö-drift outcomes. (Source - Svk Om Ö-drift (2025))

Unintentional islanding (oavsiktlig ö-drift)

Unintentional islanding is the mirror problem of intentional island operation: when a grid section unexpectedly disconnects from the upstream network while local generators continue to supply the isolated section. This creates risks of personal injury (workers assume lines are dead), fire, equipment damage, and impeded fault clearing. (Source - Energiforsk 2025-1128 Oavsiktlig ö-drift med Distribuerad Generering (2025))

The Non-Detection Zone (NDZ)

The fundamental challenge is detection. The frequency range in which generators must remain connected — 47.5–52 Hz and 0.9–1.1 Un per RfG — defines an unavoidable Non-Detection Zone (NDZ): if the isolated section achieves near-zero power exchange imbalance at the moment of disconnection, frequency and voltage remain within these bounds indefinitely. Passive monitoring cannot detect the island.

Inverter-based resources face a wider NDZ than synchronous generators: they produce only ~1.0–1.2× rated fault current (vs 4–6 p.u. for synchronous generators), so perturbations are smaller and harder to distinguish from normal fluctuations. Grid-forming BESS makes detection harder still — by actively sustaining voltage and frequency, it suppresses the very signals passive detection relies on.

Detection methods

Passive methods: monitor frequency, voltage, ROCOF (df/dt), phase angle jump, impedance change, or voltage THD. Effective only when power imbalance at disconnection is large enough to drive values outside the NDZ. Insufficient as sole detection method.

Active methods — inject controlled perturbations; the grid’s response differs depending on whether it is connected to an upstream source (which stiffens the response) or isolated:

Active methods are more reliable than passive but can be overwhelmed in large islands where perturbations are diluted.

Hybrid methods: passive trigger → active confirmation. Less power quality impact than pure active; reduced NDZ vs pure passive. Tradeoff: longer detection time and higher complexity. Examples: SFS-based hybrid, ROCOFOP/ROCOVOP variants.

Communication-based methods:

• Phase angle measurement at both upstream and island reference points — angle diverges on disconnection; requires communication link
• Intertrip: direct trip signal to generators when specific breakers open. Permitted under RfG Art. 15(5)(b)(iii) as one component of the detection method but cannot be the sole method — must be supplemented by local detection

IEC 62116 is the test procedure standard required in Sweden: 2-second detection time limit; test setup uses DC source + RLC load resonant at 50 Hz + simulated grid that can be disconnected. Must detect disconnect and shut down within specified time under multiple balanced and unbalanced load conditions.

Regulatory requirements

Under RfG Articles 13–14 (EU 2016/631, implemented in Sweden as EIFS 2018:2):

• Type C and D generators: must have an islanding detection method agreed with the TSO/DSO; the method cannot rely solely on switchgear position signals
• EN 50549 (type A/B generator connection to distribution networks): islanding detection required; must not conflict with fault ride-through requirements

Standards comparison:

Compliance with one standard does not imply compliance with another. Swedish market requires IEC 62116.

Protection challenges in islanded distribution networks

Inverter-based resources drastically change fault behavior:

In high-impedance earthed MV networks, the neutral resistance is typically located outside the island. When unintentional islanding occurs, the earth fault detection system becomes non-functional. Solution: zero-sequence voltage protection (öppet-delta VT) at each generation connection point to the MV network.

Real events — E.ON case studies

Case 1: 2022-09-26 unintentional island (Sweden)
130 kV earth fault → distance protection trips → auto-reclose 1 second later → re-energization transients trigger a wrong-zone relay trip at the 40 kV busbar → Station B (40/20/10 kV switchboards) isolated with 1.6 MW run-of-river hydro on 10 kV. Control room sees Station B still energized at elevated voltage. Operator manually opens the 20 kV transformer breaker → voltage/frequency increase → frequency protection trips the hydro. Island duration: ~6 minutes. Lesson: conventional synchronous generators (not only inverter-based DER) create unintentional islands; the hydro’s frequency regulation capability sustained the island. (Source - Energiforsk 2025-1128 Oavsiktlig ö-drift med Distribuerad Generering (2025))

Case 2: Nätvärn (preventive network protection)
A regionnät topology with three wind farms, one BESS, and two distribution stations on a radial. If the transmission connection is lost, power balance between wind, BESS, and consumption could sustain an island. Preventive measure: nätvärn (network protection) monitors specific breaker positions. If any monitored breaker opens, a trip signal is sent to the BESS (the only unit with frequency regulation capability). Without the BESS, wind farms cannot sustain frequency and trip within their passive protection limits. Lesson: complex topologies require engineering analysis of which combinations can sustain power balance, and specific preventive protection for each. (Source - Energiforsk 2025-1128 Oavsiktlig ö-drift med Distribuerad Generering (2025))

Loading strategies (pålastningsstrategier)

During black-start and island operation startup, load must be connected gradually — connecting all loads simultaneously can cause frequency and rotor angle instability. A simulation study of a real anonymous municipality in central Sweden (Uppsala thesis, 2025) compared four strategies using a 30.6 MVA CHP plant and a hydropower plant, supplying 15 prioritized buildings. (Source - Från dödnätsstart till ödrift Uppsala Thesis (2025))

Key results:

• Maximum frequency dip with “Alla samtidigt” (max load scenario): 49.87 Hz — within RfG limits but near the lower bound
• Frequency and rotor angle stabilize ~80 seconds after last load connection
• Voltage within ±5% p.u. in all scenarios; transients recover in ~1 ms
• CHP was more robust than hydro in this model due to higher available capacity relative to the 15 priority buildings’ demand

Practical implication: loading strategy design is as important as generator sizing. Strategies with time-separated connections are significantly more stable than simultaneous connection. Kluster offers a practical middle ground between operational simplicity and stability.

Feasibility planning — Skåne läns förstudie (2023)

A 2023 feasibility study by Energikontor Syd (funded by Region Skåne) mapped Skåne’s potential for crisis island operation. (Source - Förstudie Krisberedskap och ö-drift Skåne (2023))

Key finding: No suitable object could be identified for further detailed investigation among existing CHP plants or waterpower stations. Main barriers:

Financing: Svenska kraftnät’s elberedskapsanslag is the primary funding source for achieving ö-driftförmåga. CHP plants can additionally benefit from revenue from stödtjänst markets (FCR-N and mFRR where ramp capability and budget volume allow) — investment in control capability serves both preparedness and commercial purposes.

June 2023 development: Svk decided to preserve Öresundsverket (Malmö) for island operation by 2025 — its output nearly covers the full power need of the Malmö-Burlöv distribution area.

Future options: gasturbiner at bio- or vätgas production sites; new CHP converted from heat-only värmeverk; decentralized battery + gas turbine combinations for smaller load clusters; V2G at island network access points for supplying remote priority loads.

Swedish examples — additional cases

Ludvika: hydropower (3.5 MW) + 400 kWh battery (300 kWh reserved for crisis) + 218 kW solar. Battery enables dödnätsstart of the hydro plant; hydro takes over voltage/frequency regulation. Investment ~6 MSEK excl. solar (half from Energimyndigheten). Operational since ~2007.

Jönköping: biomass CHP + waterpower backup. Modifications needed: dödnätsstart capability, cooling at bio boiler, communications. Planned completion end-2024. Municipal plan: 7-day capability.

Skälleryd (Mönsterås): run-of-river hydro (1200 kVA, no reservoir). Proposed modifications include dump load (100 kW) and battery storage (400 kWh) for frequency smoothing. Neither implemented as of 2023.

Cybersecurity threats to island-sustaining DERs

RISE (2023) simulated a coordinated cyberattack on connected heat pumps using the Nordic32 grid test model. Sweden has ~300,000 connectable liquid-based heat pumps — approaching a critical mass where a synchronized activation or deactivation attack could create grid frequency disturbances. (Source - RISE Cyberhot mot Elsystemet (2023))

The mechanism is relevant to island operation: a synchronized step change in consumption from a DER botnet (heat pumps, EV chargers, BESS inverters) in an isolated island network — which has no external frequency support — could drive frequency outside acceptable limits and trigger automatic protection disconnection. The same attack vectors that threaten the main grid are amplified in the low-inertia island context.

Attack vectors for DER botnet recruitment: product vulnerabilities (hard-coded passwords, unpatched firmware), cloud service compromise, user credential theft. Devices with long lifespans (heat pumps: 15–20 years) are particularly vulnerable to outdated firmware.

See Source - RISE Cyberhot mot Elsystemet (2023) for recommended mitigations across myndigheter, energibolag, leverantörer, installatörer, and användare. For how islanding (decentralized resilience) trades off against aggregation (concentrated control) in the overall security picture, see Security and Resilience of the Digitalized Flexible Grid.

Regulatory framework: Under the Cybersecurity Act (SFS 2025:1506, in force January 15, 2026, transposing NIS2), Energimyndigheten is the designated supervisory authority for cybersecurity in the energy sector. Energy sector actors (including DSOs with island operation assets) are required to maintain systematic risk management, incident handling, continuity planning, supply chain security, security effectiveness measurement, and threat reporting. Suspected intrusions must be reported to Myndigheten för civilt försvar (MCF), the successor to MSB for civil defence functions; CERT-SE provides active incident response support. (Source - Energimyndigheten Cybersäkerhet Energisektorn (web, 2026))

Related pages

• Svenska kraftnät — elberedskapsmyndighet role; transmission-based restoration
• Vattenfall Eldistribution — Arholma project; implicit DR in island context
• Energy Storage — BESS as black start and frequency regulation resource; grid-forming vs grid-following
• Demand Response — load management during islanding; cold-load pickup mitigation; V2G for remote priority loads
• Electric Power Distribution — distribution grid as ö-drift network; protection challenges
• Generator Connection Requirements — RFG type A/B/C/D; black start capability; islanding detection requirements; EN 50549; IEC 62116

Data gaps

• Updated Swedish implementing provisions for islanding detection under EIFS 2018:2 — does Ei or Svk publish specific guidance on active vs passive method acceptance criteria for type C/D generators?
• How many Swedish distribution networks with significant DER penetration have implemented network protection (nätvärn) analogous to the E.ON Case 2 topology?
• Outcome of Jönköping island operation implementation — was the end-2024 target met?
• Status of Öresundsverket preparedness readiness — was the 2025 island operation target achieved?
• Whether the Gotland 3-month island operation target has been formally agreed as a Svk elberedskapsanslag-funded capability, or is still at planning/analysis stage
• Technical specification for the 2030 Gotland AC transmission (capacity, schedule, system operator)
• FRaM project outputs — has the Försvarsmakten methodology been published; which Gotland wind projects have been assessed