The frequency of coal generation events in 2025

A long-term trend of sudden reductions in coal generation output is reviewed.

For this article we use “Δ300” to identify the generation events of interest:

When the 5-minute interval start and end readings indicate a drop of more than 300 MW.

We’ve selected this because:

• It’s simple.
• It represents atypical operation (ramp-down rates of coal units are typically less than 10 MW/min, a change of 300MW is 60 MW/min).
• It’s a big enough drop to challenge frequency-regulating services (potentially more on this later). Noting that smaller deviations can also produce a frequency disturbance.

Elsewhere on WattClarity and in Deeper Insights, for a number of years, we’ve been tracking a range of metrics which each reveal trends in the reliability at an aggregate level of coal units:

• A count of ‘Sudden Trip’ events that was inspired by AEMO reporting in a QED back in 2018.
  • That is “when a unit is generating above its minimum generation level and falls to zero within two Dispatch Intervals”.
• The number of ‘Unit Starts’,
• Megawatts unavailable, and
• The number of large (e.g. 300 MW) ‘AggregateRaw Off-Target’ (AggROT) across coal units.

In this article we stick to the “Δ300” metric to identify events.

A decreasing trend indicates fewer events nowadays

Here’s the trend since 2002, a 23-year history. A visible decline in the monthly counts is evident, extending from the 2000s to the current year.

In the above chart some contributing units are no longer operational. Due to our size threshold of 300 MW the units at Wallerawang and Liddell Power Stations make up that component, the ‘closed’ units.

Just the remaining coal units? Also decreasing

When we filter out the units that have closed, we find the trend is still decreasing. The slope of the red line, representing the linear trend, is about 1 event per month fewer every 5 years.

Notable events

The following table presents a list of notable events where a coal unit has dropped by more than 300 MW. It’s not a complete list right now, and gaps in the table don’t mean there were no events – we’ve seen the counts in the charts above. The list aims to capture some events as examples where we have covered it in a past article.

We might add to this as time permits.