Grid Inertia: Why the Grid Runs on Spinning Steel
Grid frequency is the most closely watched number in any power system, and it is a mechanical quantity before it is an electrical one. At 50 or 60 hertz, it is literally the rotation rate of every large generator connected to the network, all turning together. What keeps that number steady in the first seconds after a fault is not a control system. It is momentum.
Frequency Is a Speedometer
A conventional power station converts heat or water pressure into rotation, and the rotation of the generator shaft produces alternating current. The frequency of that current is directly proportional to the rotation rate. Connect a second generator to the same network and physics forces the two into synchronism: they lock together and turn as one machine.
Across a continental grid this means thousands of generators, from the largest nuclear unit to the smallest hydro turbine, all rotating in step. The frequency reported by the operator is the speed of that single vast assembly.
That speed reflects the balance between supply and demand instantaneously. If consumption exceeds generation, the extra energy can only come from somewhere, and it comes out of the rotation: the machines slow fractionally and frequency drops. If generation exceeds demand, the surplus accelerates them and frequency rises. No measurement device is needed to make this happen; it is mechanics.
Operators keep frequency within a narrow band - typically a few tenths of a hertz - because the connected equipment expects it. Motors run at speeds set by frequency, and protection relays on generators disconnect them if frequency strays far enough to risk damage. That last behaviour is what turns a frequency problem into a cascading one.
What Inertia Buys
A large turbine-generator set weighs hundreds of tonnes and spins at three thousand revolutions per minute. The kinetic energy stored in it is substantial, and across a grid the total is enormous. When generation is suddenly lost, that stored energy is released automatically as the machines decelerate, partially covering the shortfall before any control system has reacted.
The measure that matters is the rate of change of frequency, written as RoCoF and given in hertz per second. High inertia means a slow, shallow fall. Low inertia means a steep one. The difference decides whether the system survives an incident.
The sequence after a large plant trips is a race. Inertia acts instantly and buys perhaps a few seconds. Primary reserve - governors opening valves on other plants, or batteries injecting power - arrives within seconds. Secondary reserve follows within minutes and restores frequency to normal. If the initial fall is too steep, protection relays disconnect further generators before the reserves arrive, and each disconnection deepens the deficit.
When even that is not enough, operators shed load: whole districts are disconnected deliberately to rebalance the system. It is the last automatic defence, and it works, which is why complete grid collapse is rare even though frequency incidents are not.
Why Inverters Changed the Problem
Solar panels produce direct current. Modern wind turbines produce alternating current at a variable frequency. Batteries store direct current. All three reach the grid through an inverter, which synthesises alternating current electronically.
An inverter has no rotating mass and no inherent physical link to grid frequency. In its conventional form - grid-following control - it measures the frequency of the network and injects current in step with it. That works well as long as something else is establishing the frequency in the first place. A grid composed entirely of grid-following inverters has nothing setting the reference at all.
So as thermal plant retired and inverter-based generation grew, system operators watched inertia fall and RoCoF limits tighten. Ireland, which has a small grid and a very high wind share, reached this frontier first and for years capped the instantaneous share of non-synchronous generation for stability reasons rather than economic ones. Britain, South Australia and Texas each encountered versions of the same constraint.
The constraint is genuine, and worth stating precisely because it is often either dismissed or exaggerated. It is not that renewable energy destabilises grids. It is that a specific service previously supplied free as a by-product of spinning machinery has to be procured deliberately once that machinery is no longer there.
How the Service Is Supplied Now
The first answer was to keep the mass without the fuel. A synchronous condenser is a large rotating machine connected to the grid that generates no power, spinning freely and contributing inertia and short-circuit strength. Several operators have installed them at retired coal sites, sometimes reusing the existing generator.
The second answer is control software. A grid-forming inverter behaves as a voltage source that sets its own frequency reference, rather than following one, and can be programmed to emulate the response of a rotating machine - releasing stored energy within milliseconds when frequency falls. Because the response is electronic it is faster than mechanical inertia, though it is limited by whatever energy sits behind the inverter.
Batteries are particularly well suited to this because they can absorb as well as inject, and their response time is measured in milliseconds. The Hornsdale installation in South Australia demonstrated the point publicly: it responded to frequency events faster than the thermal plant it was standing alongside, and changed how regulators valued the service.
Several systems have now run for periods at very high inverter shares while holding frequency stable, which established that the question was engineering rather than physics. The practical requirement is that grid-forming capability be specified in connection standards from the start, since retrofitting control behaviour across an installed fleet is far more expensive than requiring it at the point of connection. That is now the direction of grid codes in Europe, Australia and parts of North America.
Frequently asked questions
What is grid inertia?
The kinetic energy stored in the rotating masses of synchronous generators connected to a power system. Because all of them turn in lockstep at the grid frequency, their combined momentum resists sudden changes in that frequency, automatically cushioning the system in the first seconds after generation is lost.
Why does grid frequency matter so much?
Because it is a direct readout of the balance between supply and demand. Connected equipment is designed around it, and generator protection relays disconnect machines if frequency strays far enough to risk damage. Each such disconnection worsens the imbalance, which is how a single fault can cascade.
What is RoCoF?
The rate of change of frequency, in hertz per second. It measures how steeply frequency falls after a loss of generation. High inertia produces a slow fall that gives reserves time to respond; low inertia produces a steep one that can trigger protection systems before help arrives.
Do wind and solar provide inertia?
Not inherently. They connect through inverters, which have no rotating mass coupled to grid frequency. Conventional grid-following inverters measure the frequency and follow it. Grid-forming inverters set their own reference and can be programmed to deliver an equivalent, faster response.
How do operators replace lost inertia?
With synchronous condensers - large freewheeling rotating machines that supply inertia without generating power, sometimes reusing generators at closed coal plants - and with grid-forming inverters, especially on batteries, whose millisecond response can exceed what mechanical inertia provides. Grid codes increasingly require grid-forming capability at connection.
