Grid inertia and solar: what it is and what grid-forming inverters change

Grid inertia is the kinetic energy stored in a power grid's spinning machines. It slows frequency changes when supply and demand diverge. Solar and batteries add none by default: the European Network of Transmission System Operators for Electricity (ENTSO-E) gives solar an inertia constant (H) of 0 seconds and nuclear 5.9 seconds. Grid-forming inverters can emulate some.

What is grid inertia?

Rotating generators, turbines and motors release or absorb energy when the grid frequency moves. A machine spinning at grid speed resists changes in speed, so a sudden gap between generation and demand moves the frequency gradually, not instantly.

ENTSO-E, the body of Europe's transmission system operators (TSOs), defines the inertia constant H as the kinetic energy of rotating masses divided by the rated power of the machines, in seconds. A higher H means the frequency changes more slowly after a disturbance.

The rate of change of frequency (RoCoF) just after an imbalance, before any control acts, follows this formula from the same document (equation 3.16):

RoCoF = (ΔP / P_load) × f0 / (2 × H)

ΔP is the lost or surplus power, P_load the system load and f0 the nominal frequency, 50 Hz in every European synchronous area. H here is the grid's total inertia constant, with system load as its base: stored kinetic energy divided by P_load (equations 3.12 to 3.14). Table 1 applies the formula to illustrative grid-level values.

Table 1. RoCoF for a given imbalance and inertia constant (illustration)

Halving H doubles RoCoF for the same imbalance. The size matters: in ENTSO-E's review of serious events worldwide, every transient above 1 Hz/s ended in fast grid collapse, while defence systems triggered in the events below 0.5 Hz/s.

H also differs by technology. ENTSO-E's estimation method for Continental Europe assigns assumed typical values to generating units, each on the unit's own rated power. They are not measurements, and they are not the grid-level H in the formula. ENTSO-E weights each unit by its rated power and divides by system load to get a grid value (equation 3.12), so do not insert them directly.

Table 2. Plant-level mean inertia constant by technology (ENTSO-E estimation method, December 2020)

What happens in the first seconds after a large plant trips?

The frequency starts to fall at once, at a rate set by inertia. Reserves then act in order: inertia, frequency containment reserve (FCR), frequency restoration reserves (FRR) and replacement reserves (RR). ENTSO-E notes that a few seconds pass before extra reserves can be deployed, and inertia carries the system through that gap.

The EU system operation guideline (SO GL), Commission Regulation (EU) 2017/1485, sets the numbers for Continental Europe (CE). Table 3 gives the sequence and Table 4 the frequency parameters.

Table 3. Response after a large imbalance in Continental Europe

Table 4. Continental Europe frequency parameters

The reference incident sets how much FCR the area keeps ready.

Inertia matters most when a grid splits. ENTSO-E's 2016 criteria study found that, while the area stays interconnected, lower inertia is not critical for the 3,000 MW reference incident. After a system split, an island can face a large imbalance and a high RoCoF.

Why do solar and batteries add no inertia by default?

They connect through power electronic inverters that follow the grid and hold no rotating mass at grid speed. ENTSO-E describes renewables as mainly inverter-based and, today, equipped with a grid-following control scheme, so they add nothing to system inertia. The Iberian blackout report makes the same assumption: batteries, wind and photovoltaic systems do not contribute to inertia, while thermal solar and some other renewables do.

A grid-following inverter works as a controlled current source. It tracks the voltage and frequency that other machines set and injects current to meet a power target. ENTSO-E's grid-forming report notes that such a controller does not respond to a voltage angle change.

What does a grid-forming inverter change?

A grid-forming inverter is controlled as a voltage source, not a current source. Its internal voltage sets its own amplitude, phase and frequency, so current flows naturally when the grid voltage shifts. ENTSO-E's 2021 paper lists seven capabilities that together count as grid-forming: forming system voltage, adding fault current, contributing inertia, supporting survival after a system split, absorbing harmonics, countering voltage unbalance and avoiding adverse control interactions.

The limits are real. ENTSO-E's technical report cites a typical inverter current limit of 1.1 to 1.2 times rated current, and the 2016 study found very low overload capability compared with synchronous machines.

Table 5. Grid-following versus grid-forming inverters

What are synthetic inertia and fast frequency response?

Synthetic inertia is a prescribed electrical response that emulates a synchronous machine's inertia, so active power changes in proportion to RoCoF. Fast frequency response changes active power as a function of the frequency deviation instead. ENTSO-E's grid-forming report treats them as alternative ways to limit a transient frequency deviation.

Both need energy. A solar plant at its maximum available output has no headroom to add power when frequency falls. It can only cut output when frequency rises. ENTSO-E's report reflects this: a plant with no or very limited inherent storage is required to provide only negative power changes.

More support costs more. In the report's example, emulating a starting time of 25 seconds (twice H) at a 2 Hz/s frequency gradient needs a storage module sized at twice the plant's rated power. A 1 second starting time needs the inverter overrated by 4%.

What do European rules require today?

The Requirements for Generators network code (RfG), Regulation (EU) 2016/631, has no grid-forming requirement. It defines synthetic inertia as replacing a synchronous module's inertia to a prescribed performance level, and lets a TSO require it from type C power park modules (Article 21(2)). It also requires generators to withstand a RoCoF value the TSO specifies (Article 13(1)(b)).

A revision is under way. ACER, the EU Agency for the Cooperation of Energy Regulators, sent its Recommendation No 03/2023 with grid-forming proposals on 19 December 2023. On 8 July 2026 the European Commission published a draft delegated regulation, Ares(2026)6819642, often called RfG 2.0, and took feedback until 25 August 2026. Status as of 10 October 2026: the draft is not adopted, and the Commission lists adoption as planned for the fourth quarter of 2026.

As drafted, Article 29 defines grid-forming capability: within its current and energy limits, a power park module behaves as a voltage source behind an internal impedance. Articles 30 to 32 would make it mandatory for type C and D modules and for larger type B modules (110 kV or above, or 10 MW or more on a dedicated feeder to a 110 kV substation). For other type A and B modules, each TSO would propose a roadmap and the regulator would set the start date. The TSO would specify the synthetic inertia level and, for type C and D, could require energy beyond the inherent storage.

The draft also sets RoCoF withstand for power park modules (Article 14): ±2.0 Hz/s over 0.5 seconds and ±4.0 Hz/s over 0.25 seconds, with lower values over 1 and 2 seconds. Under Article 62, Articles 13 to 18 and 30 to 35 would apply to modules connecting from 24 months after entry into force.

ENTSO-E's grid-forming work feeds into this. After its 2021 paper and a May 2024 interim report, it published a Phase II technical report in November 2025, with the European Committee for Electrotechnical Standardization (CENELEC) and the storage, distribution, solar and wind associations. The report is non-binding and will underpin an implementation guidance document once the code is adopted.

Did low inertia cause the Iberian blackout of 28 April 2025?

The official report does not name low inertia as a cause. ENTSO-E's expert panel published its final report on 20 March 2026. Its summary says the blackout resulted from a combination of many interacting factors: oscillations, gaps in voltage and reactive power control, differing voltage regulation practices, rapid output reductions and generator disconnections in Spain, and uneven stabilisation capabilities.

From 12:32:57 Central European Summer Time, a transformer near Granada tripped on overvoltage protection, and overvoltage disconnections of solar, wind and thermo-solar generation cascaded. Continental Spain and Portugal lost supply at 12:33.

The root cause tree names renewable plants operating at a fixed power factor, which does not react to voltage. It also names some conventional units missing their reactive power reference, manually switched shunt reactors, and overvoltage protection settings that diverged from requirements or system needs.

On inertia, the report estimated an H of 2.21 to 2.71 seconds for the Iberian Peninsula at 12:30. Its simulations found that even with significantly higher inertia the loss of synchronism would not have been avoided, because synchronising torque fell as generators tripped, and a 20% increase had a minimal impact on the final moments. RoCoF stayed within 1 Hz/s until 12:33:20.560, when the frequency was already around 49 Hz.

How does this apply to Hungary?

Hungary is part of the Continental Europe synchronous area. ENTSO-E lists Hungary in its Continental Europe regional group, and MAVIR, the Hungarian TSO, is responsible for balancing the national system and the reserves it needs.

The Paks nuclear power plant supplies close to half of Hungarian electricity generation. MVM Paks describes four VVER-440 units totalling 2,000 MW, with eight turbines. Turbine-driven generators of this kind are synchronous machines, and ENTSO-E's method gives nuclear a mean H of 5.9 seconds. Paks II plans two 1,200 MW units, and the operator reported the first concrete pour for Unit 5 on 5 February 2026.

MAVIR publishes actual solar generation and the renewable share of gross system load, with and without household-scale solar. Published capacity figures differ in what they count, as the solar capacity study shows.

Hungary's duck curve: residual load by hour, 2020 to 2026 reports the monthly solar share of load and the highest solar quarter-hours, from a public series that leaves out most rooftop plants. Hungary electricity imports by hour, 2019 to 2026 shows how midday solar changed net imports.

What this means for owners, investors and grid planners

• Solar and storage owners. Grid-forming capability is a proposal, not yet law. The final text, national thresholds and start date decide which new plants need it, and energy for synthetic inertia drives the cost.
• Investors. Two dates matter: adoption of the delegated act, planned for the fourth quarter of 2026 (status as of 10 October 2026), and application 24 months after entry into force as drafted.
• Grid planners and policy readers. ENTSO-E ties inertia risk mainly to system splits and RoCoF. It lists synchronous condensers, storage-backed static synchronous compensators (STATCOMs) and support from high-voltage direct current (HVDC) links as other options, and says grid-forming control could reduce the devices installed only to add inertia. The Iberian report centres on voltage control, protection settings and oscillation damping, not inertia.

Limits of this explanation

• The H values in Table 2 are assumed plant-level values from ENTSO-E's 2020 method, not grid-level H and not measurements of single plants. Real inertia changes every hour.
• The ENTSO-E studies on RoCoF and system splits date from 2016 and 2020, before the 2025 blackout.
• The RfG 2.0 text is a draft (status as of 10 October 2026). Article numbers, thresholds and dates may change before adoption.
• No Hungarian inertia figure is given, because the sources report inertia for the whole synchronous area.
• This article uses no data of its own. The data studies on this site are documented on the methods and data sources page.