How Much of GB's Inertia Spins on the Distribution Network?

The inertia nobody dispatches

When people talk about grid inertia they usually mean the big spinning machines: gas turbines, nuclear units and pumped storage, all connected to the transmission network and dispatched by NESO. But a large fleet of smaller machines is connected at distribution voltages: gas engines and combined heat and power (CHP) units on industrial sites, biomass and energy-from-waste plants, landfill-gas and anaerobic-digestion engines, and small hydro.

These machines are synchronous. Their rotors are locked to the grid frequency, so while they run they store kinetic energy and resist frequency changes just like a power station does. How much does that add up to?

What the register says

Every distribution network operator publishes an Embedded Capacity Register (ECR), and the national version combines all 14 licence areas. As of September 2026 it lists 41 GW of connected generation and storage at sites of 50 kW and above. Split by how each machine connects to the grid:

• Inverter-based: about 26 GW. Solar, wind and batteries connect through power electronics and provide no natural inertia.
• Synchronous machines: about 12.7 GW. Of this, gas engines and turbines are 6.2 GW, biomass and waste 2.7 GW, landfill gas and biogas 1.3 GW, diesel 1.2 GW, hydro 1.0 GW and small pumped storage 0.3 GW.
• Other or unknown: the remaining 2.4 GW. The register gives these sites no energy source we can classify, so we can't tell whether their machines are synchronous. They are left out of the estimate rather than guessed.

The register has gaps. UK Power Networks' three areas (London, the South East and the East of England) carry no connected figures, so they are missing from these numbers.

From capacity to inertia

The stored energy of a synchronous machine is E = H × S, where S is its rating and H, the inertia constant, is typically 4–6 seconds for a large turbine. Embedded machines are smaller and lighter: a reciprocating gas engine is closer to 1 second.

The bigger unknown is how much of the fleet is actually spinning at any moment. The register says what is connected, not what is running. Biomass and energy-from-waste plants run most of the time, but peaking gas engines and diesel gensets sit idle for long stretches. So we multiply each technology's capacity by a range of inertia constants and a range of running shares.

GB system inertia usually sits somewhere between 100 and 250 GVA·s, so distribution-connected machines plausibly provide around 5 to 13 percent of it, more on low-inertia days. Most of that comes from biomass and waste plant, not from the much larger gas-engine fleet, because those plants run so much more of the time.

Why we don't add it to the headline

GridHz's headline inertia figure is calibrated against NESO's published outturn inertia. That outturn already reflects every machine that was spinning, embedded or not. So the distribution estimate is shown as a share of the headline, never as something on top of it.

It is also not a measure of how much embedded generation might trip off in a frequency event. That depends on loss-of-mains protection settings, which the register doesn't record.

Where to see it

In the dashboard, open System inertia: the line "of which distribution-connected" shows the central estimate, its range and its share of the current total. Analyst subscribers can open the breakdown by technology. Switch the dashboard to the Distribution view and the Embedded Gen panel shows connected capacity by grid supply point. The full method and assumptions are in the docs.