Frequency change associated with coal generation events

In the previous article on the topic, The frequency of coal generation events in 2025, we used “Δ300” to identify generation events. To recap, that is:

When the interval-end power output is more than 300 MW lower than the interval-start reading.

With one reason being that it is a large-enough sudden change to effect a change in system frequency.

And just as this article hits the press we observe another event with a corresponding change in frequency: Loy Yang A2 apparent trip from 443MW on Wednesday 25th June 2025

To understand the extent to which these big MW changes drive frequency we’ve calculated the corresponding change in system frequency from the intervals when each Δ300 coal event occurred.

Change in MW associated with change in frequency

The chart shows a reduced impact on frequency since the year 2000.

To orient ourselves in the chart:

• In each tile, with colour, the events from a single year are highlighted.
• Gray dots indicate the positions of the events from all years. This provides context for reviewing each year.   
• Our archive of system frequency supports analysis of the years 2017 onwards.

We observe:

• Points in 2017, 2018, 2019 and even 2020
  • Tended to occupy positions lower down relative to later years; a bigger negative change in frequency for each change in MW.
  • There is also greater variability in the change in frequency in the se years.
  • Changes in frequency of more than 0.4 Hz were not uncommon.
  • In 2020 there’s one point showing a 0Hz change for a 600 MW change. An error arising from bad data quality.
• Since 2020
  • The larger negative frequency changes are less frequent.
  • Points since 2020 tend to occupy the positions higher up meaning a smaller frequency change for each change in MW.
  • It is less likely that a unit trip of, say 400 MW, will send the system frequency outside the lower NOFB of .15 Hz below 50Hz. Note observations in June 2025.
• Most MW changes greater than 600 represent the sum of more than one unit in an interval.
  • E.g. in 2024 , in February, 4 Loy Yang A units tripped.

Frequency control had gradually slipped since changes to the NEM were made in 1999 and 2000

Kate’s article Fast Frequency Service – Treating the symptom not the cause? contains a thorough description of the frequency control issues faced in 2017, and important background. Suffice to say, certain arrangements in the NEM (including introducing FCAS markets and approaches to recover the costs thereof) drove generators away from providing PFR.

With PFR being complimentary and necessary, there was growing evidence that the FCAS markets alone were not able to replace the contribution that PFR provided. The range (distribution) of frequency had widened considerably relative to 2011. One key reason for this is that the FCAS market services start later; regulating services require a central mechanism to instruct generators what to do whereas PFR acts instantaneously and continually (once outside deadbands).

Mandatory PFR was introduced in 2020

We gave it away in the title of the chart, but PFR is the reason for the improved resilience in system frequency to these sudden generation changes.

The AEMC made the rule in March 2020, and it was by late October that many of the large generators in ‘Tranche 1’ of the implementation schedule had made the required changes to provide PFR.

By the following year, 2021, the improvement was clear. This is evident in our chart and also, uncovered at the time, in analysis of system frequency which was tracking much closer to 50Hz.