Dynamic Containment is NESO's fast, post-fault frequency response service, and with Dynamic Moderation and Dynamic Regulation it forms the main family of frequency services that batteries in Great Britain provide. The control requirement is demanding: according to NESO's service page, Dynamic Containment must start responding within 0.5 seconds of a frequency change and reach full delivery within 1 second.
Meeting that on paper is easy. Meeting it every time, across a site of many inverters, while managing state of energy and producing data that proves it, is controls engineering.
The three dynamic services at a glance
| Service | Initiation / full delivery | Delivery duration (energy-limited) |
|---|---|---|
| Dynamic Containment (DC) | 0.5 s / 1 s | 15 minutes |
| Dynamic Moderation (DM) | 0.5 s / 1 s | 30 minutes |
| Dynamic Regulation (DR) | 2 s / 10 s | 60 minutes |
NESO describes DM as a fast-acting pre-fault service, DR as slower pre-fault, and DC as post-fault. Together they help keep frequency within NESO's licence obligation of 50 Hz plus or minus 1%. The services are procured through auctions on the EAC platform, with results published on the NESO Data Portal.
Where the milliseconds go
A full-delivery time of one second sounds generous until you add up the chain: measure frequency, calculate the required power from the service curve, send a setpoint to every inverter, and wait for the inverters to ramp. Each step needs a known, tested latency.
- Frequency measurement: a fast, accurate meter at the connection point, with filtering that does not add too much delay.
- Controller cycle: the power plant controller must run the response calculation fast and deterministically.
- Communications to inverters: protocol, network design and polling affect delay; test it under load.
- Inverter ramp: check ramp-rate settings and that no limits elsewhere slow the response.
State of energy management
A battery can only deliver if it has energy and headroom. The controller must hold the state of energy within limits that allow full delivery for the contracted duration in either direction, while also following any other service or trading position. That means a state-of-energy management strategy that is written down, simulated and agreed with the optimiser, not left to a default setting.
Proving delivery
Payment depends on evidence. NESO states that operational metering applies to both BM and non-BM units, that BM providers submit availability through the Data Concentrator, and that performance monitoring data is sent through the Data Concentrator API using its template. The site's metering, historian and reporting must produce that data at the required resolution, with timestamps you can trust. Build the reporting path at the same time as the control path, and test both together.
What's Really Holding The Grid Together - Fluence
Modo Energy
This video is hosted on YouTube. Playing it loads content from Google, which may set cookies.
Watch on YouTubeCommon causes of under-delivery
- Hidden limits. A ramp-rate, reactive power or thermal limit somewhere in the inverter or plant controller that caps the response.
- Blocks unavailable. One block offline for maintenance while the site's contracted volume was not reduced.
- State of energy drift. Recovery logic too slow, so the battery runs out of headroom during a long excursion.
- Measurement differences. The control meter and the settlement meter see frequency or power differently.
- Communications delay. A busy network or slow polling adds hundreds of milliseconds during an event.
Each of these can be found in testing, and each is cheaper to fix before the site is contracted than after a performance review. Keep a record of every limit and setting that affects response, with the reason it was chosen, so the next engineer does not change it by accident.
Choosing the frequency measurement
The frequency signal drives everything. Use a meter designed for this duty, located where the service requirements say, with known accuracy and update rate. Avoid deriving frequency from a general-purpose power meter that updates slowly or averages heavily. Log the raw frequency alongside the response so that any disputed event can be reconstructed.
Testing response before it counts
- Simulate frequency profiles, including step changes and recorded grid events, at the controller input.
- Measure the time from frequency change to setpoint at each inverter.
- Run the profiles with real inverters at low power where possible, recording delivery at the meter.
- Check state-of-energy behaviour over long profiles, not just single events.
- Confirm the reporting files match what was delivered.
NESO publishes provider guidance and service terms linked from its frequency response services pages. Use the current versions as the test basis, because requirements and data formats are updated.
Why this matters as the fleet grows
Modo Energy reports that Great Britain's battery fleet reached 7.2 GW after Q1 2026 and forecasts around 20 to 22.8 GW by the end of 2029, depending on connection delays (Modo Energy). More batteries means tighter competition in frequency services and more value placed on sites that deliver reliably and can switch between services quickly. Control flexibility is part of the business case.
For the site-level system around these controls, read our guide to BESS SCADA in Great Britain. For connection-stage controls, see G99 and battery storage controls.
Next step
Send us the controller architecture and the services you intend to provide. We will review the response chain and propose controller logic and test plans through our energy systems service, or contact us.
