A large power station can be healthy for years and still trip in seconds.
If one major generating unit suddenly disconnects, the electricity demand of homes, trains, factories and data centres does not disappear with it. The grid instantly has less generation than load.
Singapore therefore keeps operating reserve available before the failure occurs.
EMA states that Spinning Reserve is required to prevent an outage when a single online generating unit fails, and that the required amount is defined to cover the loss of the single largest online generating unit in the power system.
The operating sequence is: generator trips → generation suddenly falls → frequency begins to decline → physical inertia and fast reserve response arrest the fall → spinning/contingency reserve increases supply or approved interruptible load reduces demand → system frequency recovers → replacement generation and market re-dispatch rebuild the reserve margin.
1. Reserve is capacity deliberately not being used fully
A generator producing at its absolute maximum has no upward headroom left.
Reserve requires some capable resources to remain below their full output or otherwise ready to respond quickly.
The system therefore pays for something that looks idle in normal conditions because its value appears only when ordinary supply disappears unexpectedly.
2. Spinning reserve is already online and synchronised
Spinning reserve comes from resources already connected to and synchronised with the grid and capable of increasing output within the required response envelope.
This matters because starting a cold power station can take far longer than the seconds available after a large unit trips.
The first reserve layer therefore has to exist before the contingency.
3. The largest online unit defines the basic contingency size
EMA’s current operating-reserve policy sizes spinning reserve to cover loss of the single largest online generating unit.
If the largest unit changes, the system’s contingency requirement can change with it.
The logic is simple: if the system can survive the biggest credible single generation loss, smaller single-unit trips should fall within the same reserve envelope under otherwise comparable conditions.
4. Frequency reveals the loss immediately
When a generating unit trips, electricity demand remains almost unchanged in the first instant.
The generation-load mismatch causes system frequency to fall below its nominal 50 Hz operating point unless other resources respond.
Frequency therefore becomes the power system’s fastest common signal that a major imbalance has appeared.
5. Inertia buys time but cannot replace the missing megawatts
Rotating generators and other frequency-responsive resources initially slow the rate of frequency decline.
That physical response is valuable because it gives reserve systems time to act.
But inertia does not supply a sustained replacement for the lost unit. The missing power still has to be restored or demand reduced.
6. Fast reserve arrests the frequency fall
Reserve providers respond according to technical requirements defined in Singapore’s market and system rules.
Resources that were holding headroom increase output so the generation shortfall shrinks.
The first objective is not economic optimisation. It is stabilisation: stop frequency moving toward a level that could trigger wider protection actions.
7. Secondary recovery restores the operating position
The first response prevents the disturbance from cascading.
The system then has to restore frequency, replace the lost generation more sustainably and rebuild reserve so it is prepared for another contingency.
Other generating units can be re-dispatched, additional resources can be brought on and the wholesale market schedule can be updated as the event moves from emergency response toward normal operation.
8. Reserve and regulation solve different disturbances
Regulation Reserve continuously corrects smaller second-by-second deviations between generation and demand.
Spinning and contingency reserve exist for larger sudden losses.
One is fine steering. The other is a safety net designed around a discrete failure.
9. Interruptible load can provide reserve from the demand side
A power system can restore balance by increasing generation or by reducing demand.
Approved interruptible loads can be configured under market arrangements to reduce consumption rapidly during specified contingency conditions.
This makes selected consumers part of the reserve architecture rather than treating reliability as a generator-only responsibility.
10. EMA is exploring a broader role for demand-side contingency resources
In 2026, EMA continued work on how demand-side resources could provide contingency reserve for extended periods.
The resources under study include consumers able to reduce electricity use and premises with small modular generation or storage capable of supporting their own load.
The direction is significant: grid resilience can increasingly come from flexible consumers as well as large central generators.
11. Batteries can respond quickly but energy duration still matters
Battery energy storage can deliver power rapidly.
That makes storage technically attractive for fast grid-support services.
But a battery has finite stored energy. A resource suitable for an initial fast response may still need another generator, import, demand reduction or longer-duration source to carry the system through an extended shortage.
12. Reserve is procured together with energy in the wholesale market
Singapore’s Market Clearing Engine co-optimises energy and reserve.
A generating facility’s capacity cannot be sold simultaneously as full energy output and full upward reserve.
The market therefore decides not only which generators should produce, but how much flexible capability should remain available for uncertainty and contingency.
13. Reserve has an economic cost because unused headroom could have produced energy
If a generator holds 50 MW of upward reserve, that 50 MW is not simultaneously being sold as ordinary energy at full output.
Reserve therefore has an opportunity cost.
Reliability is not free; the power system deliberately pays for flexibility that most consumers hope never needs to be used.
14. Long-term reserve margin is a different planning problem
Operating reserve asks whether the grid can survive a sudden event today.
Reserve margin asks whether Singapore will have enough generation capacity years from now after accounting for demand growth, scheduled maintenance and plant retirements.
EMA’s current Centralised Process uses a required reserve margin of 27% as a planning trigger for new generation procurement under the framework.
15. The 2026 generation RFP shows long-term reliability being planned before the shortfall arrives
EMA’s 2026 generation-capacity RFP states that system peak demand is projected to grow materially through 2034.
Under the upper demand scenario, projected reserve margin for 2031 and beyond could fall below the required 27% threshold.
EMA therefore launched an RFP for private-sector generation capacity for 2031 and potentially 2032.
The system is trying to solve a future capacity shortage before it becomes a future operating-reserve crisis.
16. A worked example: largest online unit trips
Imagine the largest online generating unit disconnects unexpectedly.
The lost generation creates an immediate shortfall and frequency begins falling. The grid’s physical response slows the decline. Spinning reserve providers increase output and any approved demand-side contingency resources respond according to their arrangements. The Power System Operator manages the event, restores frequency and re-dispatches the system. Additional generation is brought into the operating plan and reserve is rebuilt so the grid is ready for the next credible event.
The unit trip remains a plant failure instead of becoming a national blackout because the replacement capability existed before the failure was known.
17. Common misconceptions
Misconception: Reserve generation means a power station sitting completely off until an emergency.
No. spinning reserve is already online and capable of increasing output rapidly; other reserve resources can have different technical forms.
Misconception: Reserve is wasted electricity.
No. it is unused capability, not necessarily electricity being generated and thrown away.
Misconception: Regulation Reserve and Spinning Reserve are the same.
No. regulation handles normal continuous imbalance; spinning reserve protects against larger sudden contingencies.
Misconception: A 27% reserve margin means 27% of generation is spinning reserve at every moment.
No. long-term reserve margin is a planning metric for installed/available capacity; operating reserve is a real-time reliability service.
Misconception: Batteries make conventional reserve unnecessary immediately.
No. storage can provide fast response, but system design must also consider duration, energy availability and the size of the contingency.
18. The deeper idea: a reliable grid contains capacity for an event that has not happened
Efficiency pushes equipment toward high utilisation.
Resilience deliberately leaves some capability unused.
Singapore’s reserve system accepts that tension. A generator may hold headroom that could have earned energy revenue. A consumer may contract to interrupt load. Batteries may remain charged. Long-term capacity may be procured years before it is needed.
That unused capability is not inefficiency by accident. It is the space into which the power system can move when something suddenly disappears.