Where dynamic ratings are worth most
14 GB transmission nodes, three years of real hourly weather. What the extra capacity is,
when it turns up, and where it is worth least. September 2026.
The short version
- Median 17% extra capacity at winter-evening peak, and
45% in the windiest tenth of hours. Range across nodes
12% to 24%.
- The windy hours are the ones that matter, because that is when the boundary fills and GB pays to
curtail.
- It runs the other way too: about 281 hours a year the honest limit is below
the seasonal static, when the wind is light and blowing along the line. A static rating cannot see those.
- The annual average (19%) is larger than either and should be ignored. It is inflated
by mild, breezy hours when nobody is short of capacity.
- Across 41 GB site-years the uplift was never absent. Worst single case:
9%.
Method
- Zebra ACSR (quad) at 400 kV, maximum conductor temperature 75 °C — the GB 400 kV standard. A bundle is rated per sub-conductor and multiplied by 4, which ignores the slight mutual heating between them.
- Ampacity from the IEEE 738 steady-state heat balance on hourly ambient temperature, wind and solar
irradiance (Open-Meteo ERA5 archive).
- The baseline is a GB seasonal static rating — winter 5120 A · shoulder 4692 A · summer 4047 A — at 0.5 m/s. An uplift figure means
nothing without stating what it is uplift over.
- Hours are selected first, then averaged: winter evenings (Nov–Feb, 16:00–19:00), and the windiest
decile.
- The wind's angle to the conductor is applied. IEEE 738 solves convection for wind blowing across
the line and then scales it by K(φ), which falls to 0.388 when the wind blows along it. We know where each
node is but not which way each circuit leaves it, so every figure here is averaged over 12 orientations
15° apart, and the table gives the band a specific bearing would land in. Omitting this correction
roughly doubles the apparent uplift.
- One point per node. These are locations, not circuits. A real circuit is limited by its worst
span, so once its route is known it rates at or below the figures here — on the corridors we have tested,
4–13% lower depending on length.
- Figures are the operating limit we would publish, not the conductor's raw ampacity — raw ampacity is
the larger number. The ampacity calculation is standard IEEE 738 and can be rebuilt from what is above;
the margin applied on top of it is not published.
Great Britain
Peak demand in GB is a winter evening — cold, often windy, which is when a conductor cools best. And the
expensive constraint is not heat. NESO reports thermal constraint costs of £1.7bn in 2024/25, up 64% in
a year, on 13.5 TWh of curtailment, driven by wind in the north that cannot reach demand in the south.
Those constraints bind when it is windy.
In the windiest tenth of hours the uplift is 45% —
2.4x the annual figure. The extra capacity arrives in the same hours the
system is paying to curtail.
| Node | Owner | Winter peak | Windiest 10% | Peak, by orientation | +MW at peak |
|---|
| Torness | SPT, east coast | 24% | 51% | 15–32% | +837 |
| Harker | NGET, border | 20% | 52% | 10–27% | +704 |
| Pembroke | NGET, S Wales | 20% | 48% | 15–24% | +699 |
| Bramford | NGET, Suffolk | 19% | 42% | 13–25% | +685 |
| Kemsley | NGET, Kent | 18% | 42% | 10–25% | +637 |
| Ratcliffe | NGET, E Midlands | 17% | 46% | 10–24% | +612 |
| Keadby/Humber | NGET, Yorkshire | 17% | 45% | 11–23% | +594 |
| Hinkley | NGET, Somerset | 17% | 43% | 13–20% | +593 |
| Walham | NGET, Gloucestershire | 15% | 42% | 7–22% | +527 |
| Penwortham | NGET, NW England | 14% | 45% | 11–17% | +505 |
| Moffat | SPT, near B6 boundary | 14% | 45% | 8–18% | +482 |
| Denny | SPT, central Scotland | 13% | 45% | 6–19% | +460 |
| Deeside | NGET, N Wales | 12% | 43% | 11–13% | +443 |
| Beauly | SSEN-T, N Scotland | 12% | 39% | 4–18% | +426 |
Against GB seasonal static ratings. The limit sits above the seasonal static for
60% of the year. Where it is worth least: Beauly at
12% at winter peak. Still worth having, but if your circuits sit at the bottom of
this table the case is thinner, and we would rather say so before a study than after one.
When your seasonal rating is the optimistic one
The uplift is not one-directional. At a median GB node, 3.2% of the year — about
281 hours — the honest limit sits below the seasonal static rating, because the wind
is light and blowing along the conductor rather than across it. Add a heatwave and the gap widens: a GB
summer seasonal rating assumes 20 °C ambient, and 2026 delivered 35 °C in the ERA5 grid at Bramford and
Ratcliffe, with station readings higher still.
Those hours are the ones worth knowing about, and they are invisible to a static rating by construction.
The feed publishes the lower number and flags it. This count assumes a seasonal rating derived at
perpendicular flow, which is the most generous reading; if yours assumes an angle, the count falls. Send
your derivation and we will recompute against it.
Does it hold up?
41 site-years across 2022–2024. No node, in any year, showed no uplift. A node moves
a median of 2.0 points between years, and the worst single case anywhere was
9%.
What this does not show
- Uplift is not constraint relief. Whether extra thermal headroom on a circuit relieves a boundary
constraint depends on the network. That needs a power-flow study. This is not one.
- The baseline is yours. Every percentage here is against an assumed seasonal rating.
Yours will differ, and the honest number is the one computed against it — send us yours and we will
recompute. The feed can publish below a static rating, and does so in the hours described above.
- One conductor, one temperature limit. Zebra ACSR (quad) at 75 °C throughout. A different
conductor, a different bundle or a different limit moves these figures — the percentages least, because
they are a ratio, and the megawatts most.
- Orientation is averaged, not known. The band in the table is wide: at Torness a
circuit's bearing alone moves winter-peak uplift between 15% and
32%. Your circuit has one bearing, and we need it to give you a real
number rather than a fleet average.
- Modelled, not measured. These are ratings computed from reanalysis weather, not readings from a
line. Nobody has yet compared our published limit against a network owner's own outturn; that is what a
first study is for.
Sources
- NESO, 2025 Annual Balancing Costs Report (June 2025): total balancing costs £2.7bn in 2024/25,
up 10%; thermal constraint costs £1.7bn, up 64%; thermal constraint volumes 13.5 TWh, up 81%; balancing
costs projected to peak at about £8bn in 2030.
- National Grid's dynamic line rating trial on the 275 kV Penwortham–Kirkby circuit (2022–2024) was
reported as unlocking about 0.6 GW and saving roughly £1.4m of constraint costs a year, and was extended
to nine further circuits in 2025. Penwortham is one of the nodes in the table above.
- Weather: Open-Meteo ERA5 hourly archive. Physics: IEEE Std 738 steady-state heat balance.
We will run this on one of your circuits, against your own seasonal ratings, free —
name a circuit.