Great Britain: low-wind margin notice and conditional storage support#

Event: 8–9 November 2026, Great Britain. NESO issued an Electricity Margin Notice for the 9 November morning. This is a precautionary operational market signal, not a customer outage warning.

Evidence, provenance and limits of observation#

  • Contemporaneous news: The Guardian, 9 Nov 2026 reports an initial forecast supply reserve-margin shortfall of almost 1,400 MW over a 06:00–11:00 notice window, an approximately 800 MW call for additional generation, and an updated forecast margin deficit of approximately 500 MW after an early review. An additional market call is not itself the measured response, and should not be subtracted from the initial gap to infer physical dispatch. The initial and revised figures are separate forecast snapshots, not interval observations.

  • Operator context: NESO’s 23 Oct Winter Outlook forecasts sufficient winter capacity and explains reserve margins and extensive scenario stress testing. Its seasonal 5.9 GW derated adequacy surplus is not interchangeable with an intra-day margin notice figure.

  • Public warning text context: osir’s notice mirror, accessed 9 Nov displays the 8 Nov and 9 Nov Electricity Margin Notice windows. It is a third-party display; this contribution does not scrape an official NESO time series.

  • Reported vs modeled: The date, approximate MW figures, and notice windows are reported facts. The constant 1,400 MW × 5 h and 500 MW × 3 h profiles are synthetic, constructed for capacity-sensitivity screening. Their integrals are hypothetical margin-support equivalents, not electric energy unserved or electricity that the system actually lacked.

Business question and physical model#

Can an owner procure a sufficiently powerful and charged BESS to make a specified conditional contribution to reserve capacity during a low-wind event? How do inverter MW and precharged usable DC MWh constrain continuous support over several hours?

The reusable method energy_methods.margin_relief.screen_margin_relief models half-hourly intervals and imposes:

output_MW[t] = min(scenario_margin_gap_MW[t], BESS_power_MW,
                   remaining_usable_DC_MWh * discharge_efficiency / interval_hours)
remaining_DC_MWh -= output_MW[t] * interval_hours / discharge_efficiency

The 95% DC-to-AC discharge efficiency, instantaneous access to usable energy, power ratings (0.5, 1.0, 1.5 GW), and starting usable energy (0–10 GWh) are all assumptions. Batteries are initially charged above minimum SOC, never recharge, have no reserve beyond what is encoded in the initial usable energy, and can provide all modeled power instantly. A zero-storage baseline serves 0% of any positive gap (by construction).

For example, a hypothetical constant 1.4 GW deficit lasting 5 h requires at least 7.37 GWh initial usable DC energy at 95% efficiency, and at least 1.4 GW AC discharge rating for full support. This does not mean Britain lacked 7 GWh of energy. The revised 0.5 GW × 3 h scenario analogously requires 1.58 GWh and at least 0.5 GW discharge. The two scenarios must not be added together: they are successive revisions of the same underlying market situation.

Execute and test#

From a fully installed repo checkout with Python 3.12 or 3.13:

uv sync --locked --all-packages --group docs --python 3.13
uv run --no-sync pytest methods/tests/test_margin_relief.py
MPLBACKEND=Agg uv run --no-sync python examples/analysis/plot_uk_low_wind_margin_2026.py
uv run --no-sync ruff check methods/src/energy_methods/margin_relief.py methods/tests/test_margin_relief.py examples/analysis/plot_uk_low_wind_margin_2026.py
uv run --no-sync sphinx-build -W --keep-going -b html docs docs/_build/html

The gallery produces a two-panel BESS power/usable-energy sensitivity plot and prints the assumptions and theoretical full-support boundaries. The source and tests run offline; no API token or non-permissively licensed third-party dataset is bundled.

Scientific and commercial limitations#

The screen is neither an electricity adequacy model nor a capacity-market qualification tool. A forecast reserve-margin gap is not a physical unserved-load MW profile. BESS availability, charging energy, minimum SOC, state of health, network locality, export headroom, dynamic operating reserve, mandatory services, reactive power, duration constraints, grid codes, offer price, start-up constraints, imports, generator response and competing commitments are omitted. Delivered reserve cannot be monetized by multiplying hypothetical support by wholesale prices; no capacity-market contract, revenue or avoided-blackout value is calculated. There is no probability of a shortage in this example and no statistically calibrated uncertainty band. Real-time margin forecasts, wind production, BESS telemetry and market dispatch from NESO/Elexon would be prerequisites for a verified historical dispatch reconstruction.

The next step is to connect time-stamped NESO market notices and Elexon outturn data, align timestamps and revision history, and evaluate eligible delivery with grid/node constraints and actual dispatch/offer prices. Independent baseline: 0 MW / 0 MWh BESS; conservative battery energy boundary as above.