Hams Hall BESS: nominal capacity versus deliverable constant-power support#

News event: On 2 November 2026, BW ESS announced that Hams Hall’s first phase had entered commercial operation in late October 2026, with 350 MW and 1,243 MWh. Its stated future second-phase total is 400 MW and 1,424 MWh. The developer says the completed project could support approximately 1.3 million homes for up to 3.5 hours. This is a developer claim, not an independently measured household dispatch study.

Primary evidence: BW ESS, commercial operation press release, publication date 2026-11-02; Hams Hall project page. This snapshot is not a live site operational feed. Figures should not be treated as metered AC energy or verified battery state of health.

Business question#

For operators, lenders and off-takers evaluating a public battery-completion announcement, which combinations of declared MW/MWh, service duration, usable capacity, state of charge (SOC), operational reserve and losses are internally consistent with a constant-power household-equivalent claim?

Physical model#

The pure-Python method energy_methods.bess_service_envelope.screen_bess_service returns the energy and power limits of a non-recharging, constant-power, single-event screen:

E_available_AC [MWh] = E_declared [MWh] × f_usable × (SOC_start - SOC_reserve) × η_discharge.

P_max [MW] = min(P_rated [MW], E_available_AC [MWh] / service_hours [h]).

All energy is MWh, time is hours and rated/export power is MW. maximum_households divides the sustainable constant-power envelope by an assumed mean household kW, and is an energy equivalence only, not real simultaneous households served, a reliability guarantee, or a topology/network claim. Zero usable capacity and depleted SOC give zero available energy. Every numeric argument must be finite and consistent.

Critical declared-energy ambiguity: BW ESS publicly reports MWh without an independently certified DC/AC energy basis. If those MWh are already AC deliverable, the illustrative discharge efficiency must be 1 to prevent double-counting conversion losses. The gallery shows both ideal and stressed, conditional assumptions. It must not be used to assert an actual discharge-efficiency loss or 90% state of health.

Reproduction and reported/assumed inputs#

Install the locked workspace, then run:

uv sync --locked --all-packages
uv run --no-sync pytest methods/tests/test_bess_service_envelope.py
uv run --no-sync python examples/analysis/plot_hams_hall_bess_service.py

The offline executable gallery makes two reproducible sensitivity plots. Reported: 350 MW / 1,243 MWh operational phase; 400 MW / 1,424 MWh proposed full build; approximately 1.3 million homes at 3.5 hours. Derived arithmetically: phase-one nominal full-power duration ~3.551 hours, planned full-build ~3.56 hours, full-build power ceiling for the household claim ~0.308 kW per home. Assumed (not measured): constant per-home loads 0.25–0.40 kW; initial SOC 50–100%; stress scenario 10% reserved SOC, 90% usable capacity fraction and 95% discharge conversion. The ideal nameplate case assumes full initial SOC, no reserve, full retention and unity conversion.

The example establishes internal consistency only: 400 MW / 1,424 MWh at ideal full availability sustains at most 400 MW for 3.5 h; to equate that to 1.3 million households, the constant mean power per home must not exceed ~0.308 kW. It does not independently verify the marketing claim. The first phase does not carry the announced full-build household figure.

Scientific limitations and next data#

No battery telemetry or validated BMS SOC/temperature, degradation, auxiliary/parasitic loads, inverter curves, firm grid connection rights, reactive power obligations, thermal derating, availability, real UK settlement data, seasonal/correlated household profiles, reserve obligations, cycling/aging economics or service-stack cash flows are available. This is not a capacity-market or investment-grade guarantee. Verify AC nameplate versus usable MWh and collect operator interval discharge/charge and grid-metered readings before calculating actual dispatch adequacy or financial revenues.