An AC grid's frequency stays constant only when generation and consumption are exactly balanced. If generation falls short of demand, frequency drops; if it exceeds demand, frequency rises. What determines how much leeway the grid has to react to this imbalance is inertia. In hydro, thermal, and nuclear plants, the synchronous generator's massive rotating body itself acts as a store of kinetic energy, naturally softening the sudden drop or spike in frequency for the first several seconds after supply and demand fall out of balance. Solar, wind, and batteries, however, connect to the grid through power electronics (inverters), so they don't have this natural buffering effect. The basics of synchronous-generator inertia and the swing equation were introduced in the Power Systems Primer. This article digs into what that equation actually means in practice, and how frequency can be protected on a grid where inverter-connected sources take the lead role.
The Bottom Line in 30 Seconds
- Inertia is the property by which the kinetic energy held in a synchronous generator's rotating body (turbine plus rotor) naturally absorbs and releases energy during a sudden supply-demand swing, softening the resulting frequency change. The larger the inertia, the smaller the rate of frequency change (RoCoF: Rate of Change of Frequency) for a given supply-demand gap.
- Solar, wind, and batteries connected via inverters have no mechanical rotating body, so they don't naturally provide inertia. As their share of the mix rises and synchronous generators shrink in relative terms, the grid's overall inertia falls, and RoCoF becomes faster for the same size of generation loss.
- Maintaining frequency breaks down into three stages on different timescales. In Japan's balancing market, regulating power is procured through product categories: primary regulating power (response within 10 seconds), secondary regulating power ①② (response within 5 minutes), tertiary regulating power ① (response within 15 minutes), and tertiary regulating power ② (response within 60 minutes).
- Synthetic inertia (virtual inertia) is a technology in which an inverter detects the rate of frequency change df/dt and instantaneously charges or discharges power from a battery or similar source to mimic a synchronous generator's response, artificially creating an effect similar to inertia. Grid-forming control is being researched as a further development of this idea.
- The 2016 South Australia blackout and the 2019 Great Britain blackout are both real, thoroughly investigated incidents in which low grid inertia and a high share of inverter-connected sources were cited as factors that widened the damage.
- In Japan too, as renewable adoption expands, securing system-wide inertia has been taken up as a study topic in the Agency for Natural Resources and Energy's deliberative councils.
1. What Is Inertia — a Piggy Bank of Energy in a Spinning Mass
A synchronous generator is a massive rotating body in which the turbine and the generator rotor spin on the same shaft. A rotating mass stores kinetic energy of
where J is the rotor's moment of inertia and \omega is the angular velocity. When demand suddenly exceeds generation, a portion of the kinetic energy stored in this rotating body is first supplied to the grid as electrical energy, and the rotor's rotational speed — that is, the frequency — slowly drops by a corresponding amount. Conversely, if generation exceeds demand, some of the surplus power acts to accelerate the rotor, and the frequency slowly rises.
Thanks to this "physical buffer of a rotating body," even when a supply-demand mismatch occurs, frequency doesn't instantly jump to zero — it changes with a few seconds of leeway. That leeway is exactly what buys time until protection relays and regulating power kick in.
Figure 1 — A synchronous generator's rotating body stores kinetic energy, acting as a natural buffer that softens sudden supply-demand swings. Solar, wind, and battery inverter-connected sources have no mechanical rotating body, so playing the same role requires deliberately simulating it through control.
2. The Swing Equation and RoCoF — Seeing Inertia in an Equation
The swing equation introduced in the Power Systems Primer expresses the relationship between the inertia constant H and the frequency deviation \Delta f as
where P_m is mechanical input, P_e is electrical output, and D is a damping term from the load's frequency characteristics. The inertia constant H represents how many seconds' worth of the generator's rated output the kinetic energy stored in the rotating body corresponds to (its unit is seconds); for large thermal and hydro synchronous machines it typically runs roughly 2–8 seconds.
Simplifying by ignoring the D\Delta f term, the rate of frequency change immediately after a supply-demand gap \Delta P=P_m-P_e occurs — that is, RoCoF — can be approximated as roughly inversely proportional to the inertia constant H:
(where f_0 is the rated frequency). The larger a grid's inertia, the gentler its RoCoF for a given size of generator loss. Conversely, as the share of synchronous generators shrinks and the grid's effective inertia falls, RoCoF becomes faster for the same \Delta P, raising the risk that frequency drops to a dangerous level before protection devices or regulating power can react.
3. Why Inverter-Connected Sources "Don't Have" Inertia
Solar panels output DC, and wind turbines spin at variable speed, so in neither case does the output frequency naturally match the grid's. Both create AC matched to grid frequency through power conversion via a power conditioner (an inverter). This inverter is a device that shapes voltage and current waveforms through semiconductor switching, and unlike a synchronous generator, it has no physical rotating body. Even when the grid's frequency changes, the solar panel or battery cells on the inverter's DC side don't naturally charge or discharge kinetic energy.
Many inverters use a control scheme called "grid-following," which simply tracks the grid's voltage and frequency as an externally given reference, without itself creating or supporting frequency. So as the share of inverter-connected sources rises and synchronous generators shrink in relative terms, the total amount of inertia naturally available to the grid decreases, lowering its resilience against generation loss or sudden demand change. This is what's known as "the declining-inertia problem," and in Japan too it's been taken up in the Agency for Natural Resources and Energy's deliberative councils as a challenge accompanying expanded renewable adoption.
4. The Three Stages of Frequency Maintenance — Primary, Secondary, and Tertiary Regulating Power
The three timescales shown in the Power Systems Primer — "instant to a few seconds," "seconds to minutes," and "tens of minutes to a day" — are institutionalized as concrete product categories in Japan's balancing market.
| Product category | Response time | Duration | Role |
|---|---|---|---|
| Primary regulating power | Within 10 seconds | 5 minutes or more | Immediate frequency support via governor-free operation of synchronous generators |
| Secondary regulating power ①/② | Within 5 minutes | 30 minutes | Resolving supply-demand gaps via load frequency control (LFC) |
| Tertiary regulating power ① | Within 15 minutes | 30 minutes | Medium-term supply-demand adjustment via economic dispatch control (EDC) |
| Tertiary regulating power ② | Within 60 minutes | 30 minutes | Longer-timescale supply-demand adjustment, such as for renewable forecast error |
Primary regulating power plays the role of supporting frequency, during the several-to-tens-of-seconds window while inertia's buffering effect is still holding, through synchronous generators' governors autonomously raising or lowering output. If inertia is "the buffer that buys time until a response happens," primary regulating power is "the first actual move made within that time." Secondary and tertiary regulating power settle the supply-demand books over a longer timescale. Which product category a VPP that bundles batteries and EVs can participate in depends on its resources' response speed (see the VPP (Virtual Power Plant) Primer).
5. Synthetic Inertia and Virtual Inertia — Giving Inverters a Pseudo-Inertia
One solution for keeping the grid stable as inverter-connected sources grow is a control scheme called synthetic inertia (or virtual inertia). This is a technology in which an inverter measures the grid frequency's rate of change df/dt in real time, and instantaneously raises or lowers a battery's or solar installation's output in response, artificially creating a power response as if a synchronous generator's rotor were charging and discharging kinetic energy.
This control has two broad aims. First, to keep the initial RoCoF right after a generation loss within a range that the grid's protection devices and other generators can safely track. Second, to stop the point at which frequency bottoms out (the frequency nadir) before it reaches the threshold that triggers load shedding or generator disconnection.
Control algorithms range from methods that are simply an extension of "grid-following" control, adjusting output in response to frequency changes, to "grid-forming" methods, in which the inverter itself behaves as a voltage source and mimics a synchronous generator's dynamics at the level of the governing equations, including VSG (Virtual Synchronous Generator) control, which explicitly simulates the synchronous machine's equation of motion in particular. Grid-forming control is being researched and developed for its potential to maintain stable voltage and frequency even on weak-inertia grids or on remote islands and offshore platforms that have no synchronous generators.
6. Lessons From Real Frequency Incidents
The risk that declining grid inertia brings isn't purely theoretical — there are real, thoroughly investigated large-scale blackouts.
September 28, 2016, South Australia blackout: a series of transmission-line faults triggered by a severe storm caused multiple wind farms to curtail output one after another, due to protective settings responding to voltage disturbances (settings that restrict restart after experiencing multiple ride-throughs within a set time window). This overloaded the flow on the interconnector (the interstate transmission line), which disconnected, blacking out nearly the entire state — around 850,000 customers. The final report by AEMO, Australia's grid operator, concluded that South Australia, with few synchronous generators running and low grid inertia at the time, saw its frequency decline rate from the generation loss outpace the speed at which the under-frequency load shedding (UFLS) settings could arrest the frequency drop above 47 Hz. In response, AEMO recommended that grids with a growing share of asynchronous sources (wind and solar) need designs that don't rely on synchronous generators alone for inertia and frequency regulating power.
August 9, 2019, Great Britain blackout: triggered by a lightning strike, the Hornsea offshore wind farm (about 737 MW) and the Little Barford gas-fired power station (about 244 MW) were disconnected from the grid at nearly the same time. On top of this primary loss, distributed embedded generation that detects rate of change of frequency (RoCoF) and trips as a protective action lost an additional roughly 350 MW, bringing the total generation loss across the system to 1,878 MW. Grid frequency fell to 48.8 Hz, triggering the first stage of low-frequency demand disconnection (LFDD), automatically disconnecting about 973 MW and more than a million customers. Grid frequency itself recovered to the normal range within 4 minutes 42 seconds of the lightning strike, and nearly all disconnected customers were restored within 45 minutes, but the incident — where a relatively common event, a single lightning strike, cascaded through the simultaneous loss of multiple generation facilities and protective actions into a widespread blackout — has been examined in detail by UK regulators and National Grid ESO.
What both incidents share is that the single trigger (a wind-farm protective setting in South Australia, a lightning strike in Great Britain) mattered less than the fact that the grid lacked enough inertia and regulating-power headroom to stop the chain reaction that followed — or that a cascade of protective-device actions widened the damage. Both reports arrive at the same understanding: inertia isn't a mechanism that prevents an incident from happening — it's a mechanism that creates the time margin needed to keep an incident, once it happens, from becoming catastrophic.
7. Current Technology and Research Focus
- Measuring and visualizing inertia: research is underway to estimate a grid's effective, system-wide inertia in real time. Because inertia shifts moment to moment with which generators are running, it can't be treated as a fixed value.
- Standardizing grid-forming inverters: various countries are studying grid codes that would require batteries, offshore wind, and solar inverters to use control that behaves as a voltage source.
- Reassessing synchronous condensers: there's a movement toward installing synchronous condensers, which supply inertia and short-circuit capacity without generating power, at the sites of retired thermal power plants.
- Revisiting balancing-market product design: institutional designs to procure inertia and frequency-support capability themselves as an independent market product are under discussion in various countries, including Japan.
- Redesigning protection coordination: the 2019 Great Britain incident revealed that protective settings on distributed generation that trip on detecting RoCoF can themselves trigger cascading outages, and various countries are now revisiting protection settings.
Summary: Inertia Is an "Invisible Reserve"
Grid inertia isn't visible generation equipment — it's an "invisible reserve" of kinetic energy held in synchronous generators' rotating bodies. How to protect frequency as this reserve shrinks depends on both institutional design tied to timescale, through primary-to-tertiary regulating power, and technical alternatives such as synthetic inertia and grid-forming control.
When reading news about a grid or an incident, it's worth checking the following:
- What was the state of synchronous-generator operation (≈ grid inertia) at that moment?
- How far did RoCoF and the frequency nadir fall relative to the generation loss?
- Which of primary-through-tertiary regulating power, or synthetic inertia/grid-forming, actually responded?
- Did a cascade of protective-device actions become a factor that widened the damage?
Keeping these four points in mind lets you break down the vague claim that "more renewables makes the grid unstable" into concrete elements — inertia, regulating power, and protection coordination.
References
- AEMO — Black System South Australia 28 September 2016 (final report)
- 2016 South Australian blackout — Wikipedia
- National Grid ESO — Technical Report on the events of 9 August 2019
- GB power system disruption on 9 August 2019: Final report (E3C)
- Electric Power Reserve Exchange (EPRX) — What Is the Balancing Market?
- Agency for Natural Resources and Energy — The Current State of the Power System
- NREL — Research Roadmap on Grid-Forming Inverters
- IEEE Power & Energy Society
Comments
Please log in to post a comment
No comments yet.