Balance, frequency and system security

When electrical demand exceeds mechanical or controlled electrical supply, stored energy initially contributes to the imbalance and frequency can fall in a connected synchronous system. A supply surplus can drive frequency upward. A frequency trace therefore contains information about a dynamic imbalance, not just the final dispatch schedule.

For an original aggregate teaching model, define E as equivalent synchronous kinetic energy in MW·s at nominal frequency f0. With deficit ΔP in MW defined as positive, the initial approximation is df/dt = −f0 × ΔP/(2E). This comes from linearising stored kinetic energy near f0. It ignores damping, governor and inverter controls, spatial differences and changing topology; it is not an operating-security assessment or a frequency forecast.

At f0 = 50 Hz, ΔP = 200 MW and E = 10,000 MW·s, the initial estimate is −0.5 Hz/s. Halving E doubles the magnitude to −1 Hz/s with the other assumptions unchanged. Do not extend that straight line for seconds or minutes to claim a frequency nadir: controls, protection, load response and nonlinear effects have been excluded.

An initial response that slows frequency change is not the same as enough sustained power to remove an imbalance. A fast power response may help arrest a fall, while additional response and replenishment are needed to recover and continue supplying demand. The timing, duration, location and performance of each service matter alongside its MW rating.

Suppose a synthetic 200 MW deficit lasts for ten minutes. Holding 200 MW throughout requires 200 × 10/60 = 33.33 MWh at the delivery boundary. A 100 MW resource with ample energy still cannot cover the full power deficit alone. A 200 MW resource with only 10 MWh available can sustain that output for three minutes under a constant-output, loss-free assumption.

Do not add inertia, primary frequency response and every ancillary-service label as if they were interchangeable MW products. Check the actual mechanism, obligation and service specification. AEMO's primary frequency response material is a starting point for the applicable requirements; the examples here do not determine qualification, enablement, compliance or a real dispatch instruction.

• AEMO — primary frequency response

Consider three independent teaching resources at a common delivery boundary. Battery A can provide 80 MW and has 20 MWh usable delivered energy. Plant B can add 50 MW after reaching its response level. Demand response C can reduce consumption by 40 MW, conditional on the customer's verified availability. Their eventual combined response is 170 MW, leaving 30 MW of a 200 MW deficit even when all three are available.

The battery's constant-output duration is 20/80 hour = 15 minutes. The combined 170 MW cannot be assumed available at the instant of the event: each resource has its own detection, activation and ramp profile. Plot delivered response against elapsed time, then compare it with the required envelope. Areas represent energy; instantaneous heights represent power.

Now assume C is unavailable and battery A's usable energy has fallen to 8 MWh. Eventual response falls to 130 MW, leaving 70 MW; the battery can sustain 80 MW for six minutes. Report both the power gap and the duration risk. An average annual capacity factor cannot repair either missing capability in this event. No example authorises interrupting a real customer or changing plant controls.

A useful analyst must recognise when an active-power calculation is answering the wrong question. Voltage behaviour, reactive-power support, protection performance and control interactions need their own models and evidence. A battery's MW rating does not by itself specify every capability needed at its connection point; a technically adequate response in one location may not resolve another location's issue.

AEMO's system security planning page distinguishes system strength, inertia and network support services and links their methodologies. It states that the Transition Plan for System Security incorporates the previously separate system-strength, inertia and NSCAS gap reports from December 2025. Use the current publication route, record the version, and distinguish a planning finding from a site-specific connection determination.

Do not transplant NEM service names and obligations into Western Australia's WEM without checking the applicable framework. Likewise, a site UPS supports a defined internal load and duration; that does not establish permission or capability to provide a grid service. Connect this lesson with Topics 16, 37, 41, 42 and 57 to test services, storage, connection design, large loads and recovery.

• AEMO — system security planning and current methodologies

Independent variant: the illustrative deficit is 150 MW and E = 7,500 MW·s at 50 Hz. Calculate the initial frequency slope and energy needed for eight minutes of constant replacement. Compare an 80 MW battery with 12 MWh delivered energy plus a 60 MW flexible plant. State the steady power gap and the battery's ideal duration. Keep activation delays as unresolved inputs rather than silently assuming an instantaneous response.

Check after attempting: slope = −50 × 150/(2 × 7,500) = −0.5 Hz/s; replacement energy = 20 MWh; eventual response = 140 MW, leaving 10 MW. The battery's ideal duration is 12/80 × 60 = nine minutes. Although that exceeds eight minutes, the portfolio still fails the 150 MW power requirement. Sufficient aggregate energy does not cure a power shortfall.

Submit one page with event, boundary, data sources, base calculation, adverse case and missing evidence. Include a response-versus-time sketch and distinguish known values from assumptions. Ask a competent reviewer to check the arithmetic and challenge the proposed next action. The result is an educational screening note, not a compliance certificate, a protection study or authority to operate the system.