Frequency Hz
Frequency is the rate at which the grid's AC voltage completes its cycle — 50 Hz across most of Europe, Asia, Africa and Australia, 60 Hz across North America — and, unlike voltage, it is a single shared quantity: every generator and load in a synchronous area sees the same frequency at the same moment.
That makes it the grid's real-time balance signal — when load exceeds generation the deficit is drawn from the kinetic energy of spinning machines and frequency falls; when generation exceeds load it rises.
The deviation is information, not a defect. For a BESS the number matters three ways: frequency excursions trigger some of the plant's highest-value services, the frequency bands in the grid code are a stay-connected obligation the plant must be settings-proven against, and the speed at which a battery can push real power against a falling frequency is why storage has become the fastest frequency asset on most grids.
Reviewed August 2026 by Sergey Syrvachev
New to BESS? Start free with the 7-email fundamentals course — no cost, no account.
What it is (precise)
Frequency is the number of complete AC cycles per second: 50 Hz means a 20 ms cycle, 60 Hz a 16.67 ms one. The value descends from rotating machines — a synchronous generator's electrical frequency is its shaft speed times its pole-pair count, so a two-pole machine spins at 3,000 rpm on a 50 Hz grid and 3,600 rpm on a 60 Hz one — and in steady state every synchronous machine in an interconnected AC system rotates in electrical lockstep.
That lockstep is what defines a synchronous area: Continental Europe is one, Great Britain another (tied to the continent only through HVDC links, which do not carry frequency across), ERCOT a third, and Japan famously runs 50 Hz in the east and 60 Hz in the west with frequency-decoupling converter stations between them. A BESS brings no rotating mass to this story: its PCS synthesizes the waveform electronically, and in grid-following mode a phase-locked loop measures the grid's frequency and phase and tracks them cycle by cycle.
The property that makes frequency useful is that it is one number for the whole area. Voltage is local — it differs bus by bus with every impedance in between — but a frequency event caused by a generator tripping in one country is measured, within moments, by every relay and every PCS across the synchronous area.
Every plant can therefore act on the system's state with nothing but a local measurement: no telemetry, no dispatch instruction, no communication path to fail. That is why the fastest grid services are built as autonomous frequency-triggered functions in PCS firmware rather than as commands from a control centre, and why the plant controller's job in those services is mostly to configure and supervise a loop that runs locally.
The balance signal — inertia and ROCOF
At every instant, generation must equal load plus losses — the wires store essentially nothing. The buffer that absorbs any mismatch is the kinetic energy in the rotors of the synchronous machines, and because those rotors are locked to electrical frequency, drawing the buffer down shows up directly as falling frequency: excess load slows the machines, excess generation accelerates them.
Frequency is the fuel gauge of that kinetic buffer, which is what makes it the grid's real-time balance signal. Nothing is broken when it moves — it wanders continuously inside a narrow band around nominal as load and generation breathe, and the deviations are the information the entire frequency-control hierarchy runs on.
How fast frequency falls after a disturbance is set by the ratio of the imbalance to the stored rotation: the initial rate of change of frequency (ROCOF) is proportional to the lost power divided by system inertia, conventionally expressed through the inertia constant H — the number of seconds a machine could supply its full rating from stored kinetic energy alone.
Inverter-based resources add no inertia unless explicitly controlled to mimic it, so as synchronous plant retires, the same lost megawatts produce a faster, deeper excursion. For a BESS this cuts both ways: interconnection studies now quote a ROCOF-withstand value the plant's protection and controls must tolerate without tripping, and the shrinking kinetic buffer is exactly the gap the battery's speed is paid to fill — the catalogue of services for doing so belongs to the frequency-response entry.
Load above generation drains rotating kinetic energy and frequency falls; generation above load pushes it up — the deviation is information, not a defect, and every node in the synchronous area measures the same event. HVDC ties do not carry it across, which is why Japan can run 50 and 60 Hz halves and why GB's frequency is its own. How fast an excursion moves is set by inertia: the initial rate of change of frequency is proportional to lost power over system inertia, and the inertia constant H is the seconds of rated power a machine could deliver from stored rotation alone. Inverter-based resources add none unless controlled to mimic it — and yet a BESS is the fastest responder on the system, because power-electronic actuation on a local measurement works in both directions, with a full discharge-to-charge swing of twice nameplate. On 60 Hz systems the numbers shift but the shape holds: nominal is a 16.67 ms cycle and under-frequency load shedding starts near 59.3 Hz.
- Nominal values
- 50 Hz (20 ms cycle) across most of Europe, Asia, Africa and Australia; 60 Hz (16.67 ms) in North America; Japan runs both, split east-west
- One number per area
- Voltage is local, frequency is shared — every node in a synchronous area measures the same event; HVDC ties do not carry it across
- The signal
- Load above generation drains rotating kinetic energy and frequency falls; generation above load pushes it up — deviation is information, not a defect
- ROCOF
- Initial rate of change of frequency ∝ lost power ÷ system inertia — less synchronous plant means faster, deeper excursions
- Inertia constant H
- Seconds of rated power a machine could deliver from stored rotation alone; inverter-based resources add none unless controlled to mimic it
- Why BESS is fastest
- Power-electronic actuation on a local measurement, in both directions — a full discharge-to-charge swing of twice nameplate
- Grid-code bands
- RfG Continental Europe: unlimited operation 49.0-51.0 Hz, time-limited to 47.5/51.5 Hz; UFLS starts near 59.3 Hz (60 Hz systems) and 48.8 Hz (GB)
- Not the same as
- Frequency response (the paid service), ride-through (the stay-connected envelope), droop (the deviation-to-power law behind both)
Why a battery is the fastest frequency asset
A thermal governor answers a frequency dip through a mechanical chain — measure, move a valve, admit more steam or water, accelerate a turbine — and the chain takes seconds to tens of seconds. A BESS answers through semiconductors: the PCS measures frequency at its own terminals and steps real power on power-electronic timescales, so the limiting delay is measurement filtering and control settings, not any physical actuator.
The battery also responds in both directions with its whole nameplate: a generator already at full output can only back off when frequency is high, while a battery can swing from full discharge to full charge — twice its rated power of usable response — which is why symmetric fast-frequency products fit storage better than any other asset class.
The mapping from deviation to power is the droop characteristic — a proportional line that lets thousands of plants share the burden in proportion to their ratings without communicating, and that in grid-forming control emerges from the inverter's own behaviour rather than from a measure-then-command loop. This page stops at the concept: the settings, deadbands, product definitions and market rules are the frequency-response entry's territory, and grid-forming carries the control-mode story.
What belongs here is the constraint that speed does not remove. Frequency duty is power-intensive but energy-light, so the binding design questions are C-rate, PCS headroom and state-of-charge management — a battery parked full is as unable to absorb a high-frequency event as a generator is, however fast its electronics.
Frequency limits in grid codes are ride-through bands
Grid codes treat frequency the way they treat voltage: as an envelope the plant must stay connected through. Around nominal sits a band of unlimited operation — under the ENTSO-E RfG network code (Regulation 2016/631), Continental European plants must run indefinitely anywhere between 49.0 and 51.0 Hz — flanked by time-limited bands stretching the obligation out toward 47.5 and 51.5 Hz, beyond which disconnection is permitted.
In North America, NERC PRC-024 bounds the frequency settings at which bulk-system protection may trip. The bands are deliberately wide because the system's own defences act inside them: under-frequency load shedding typically begins near 59.3 Hz on 60 Hz systems and 48.8 Hz in GB, so a plant that trips early abandons the grid before the load-shedding scheme has even begun to fight for it.
For the BESS the envelope is a settings-and-evidence obligation. The PCS datasheet's frequency operating range must cover the code's envelope with margin; the protection relays at the POI must be set inside the no-trip boundaries, with the settings sheets kept as commissioning evidence; and the interconnection study tests the plant model against frequency excursions alongside the voltage sags the ride-through entry covers.
None of this is paid — it is the precondition for connection — and the distinction between this obligation and the revenue services triggered by the same measured quantity is one of the cleanest lines a specification can draw.
Common pitfalls
The classic confusion is between the obligation and the product. Frequency ride-through is mandatory, unpaid, and written in the grid code and interconnection agreement; frequency response is optional, paid, and written in a market rulebook — same measured quantity, different documents, different tests, different consequences for failure.
The related error is treating frequency as something the plant controls locally: a grid-following PCS does not regulate the frequency at its terminals, it pushes real power against a system-wide value that only the aggregate balance can move, and one plant's influence on the number itself is negligible on any large interconnection. The exception is genuine — a grid-forming inverter running an island, black start being the extreme case, really does set the frequency, which is precisely what distinguishes that control mode.
The second family of traps is portability between the 50 and 60 Hz worlds. Equipment nameplates state frequency because the magnetics are designed to it: transformer core flux scales with volts per hertz, so iron sized for 60 Hz runs about 20% higher flux density at 50 Hz at the same voltage and moves toward saturation — a transformer does not follow a project from Texas to Germany on its nameplate voltage alone.
Cycle-denominated specifications shift the same way: a 15-cycle requirement is 0.25 s at 60 Hz but 0.3 s at 50 Hz, so copying a specification across markets without converting the time base changes every deadline in it by 20%. Read cycles as cycles, then convert for the grid the plant will actually stand on.
The grid holds frequency at exactly 50 or 60 Hz, and each plant's inverter is responsible for keeping its own output at that value — a deviation at the terminals means something on site is wrong.
In reality: Frequency is one system-wide quantity that only the aggregate balance of generation and load can move, and it wanders inside a narrow band all day by design. A grid-following PCS does not set the frequency at its terminals — it measures and tracks it, and the plant's influence on the number itself is negligible on any large interconnection. The deviations are the working signal, not defects: the entire fast-response layer is built on every asset reading the same excursion locally and pushing power against it, with no communication path in the loop. The one context where an inverter genuinely sets frequency is grid-forming operation of an island — which is exactly why that capability is named as a separate control mode rather than assumed of every plant.
- Frequency response Glossary
- Ride-through Glossary
- How Frequency Regulation Pays Batteries — and Why That Paycheck Shrinks Article
- Interactive: Grid Inertia and RoCoF Interactive visual · bess.engineer
Frequency, in context.
The Grid-Scale BESS course covers frequency — and the rest of the system — from the ground up, the way it actually gets deployed.