With the aim of clarifying uncertainties around what drove the frequency dip, we inspect unit output. We find we gather more questions than answers.
Market data showed NSW exporting over 1850 MW at the start of the 16:25 interval (usually the snapshot would be at 16:20) we seek to understand whether this was a result of an increase in NSW supply.
Taking a look at unit output in NSW the following chart captures the MW output of (non-solar and non-wind units in NSW). The data is at 4-second intervals. The data suggests little change in unit output at 16:20.
However there is a dip in readings across many units earlier than 16:20 – around 16:18.
The PFR response to frequency decline in the 16:25 interval is seen in many units between 16:20 and 16:25. Steady increases in MW are observed across this interval.
Wind and solar unit output in NSW indicates similar steadiness in output at 16:20. Yet it does also indicate a trip or transient outage near 16:18.
If the NSW1-QLD1 interconnector really swung by 1850 MW (approximately), did unit output in QLD drop to accomodate or necessitate this?
Inspecting QLD units we find some units experienced a prolonged “outage” (as per the MW readings) from the 16:18 point in time.
These units include wind and solar varieties such as BLUEGSF1, MOUSF1, EDENVSF1, DULAWF1 to name a few. Possibly not enough to explain the interconnector flow change?
Non-VRE units exhibit a similar pattern to what we found in NSW at 16:18: a brief dip in readings. Not prolonged. Examples of this include Tarong and Stanwell units. WDBESS2 saw a dip of over 100 MW but had resumed by 16:19.
In general, the driver of the radical change in interconnector flows around NSW for the 16:25 interval start is still not clear.
Market unit output appeared to fluctuate – as we see above – but was this to an equivalent extent as the recorded interconnector flows would suggest, or the noted demand change in NSW for the 16:25 interval?
One major difference between commissioning engineers and plant operators is that commissioning engineers do not believe what they read on meters whereas plant operators typically do.
The widespread dips in power in your charts in NSW and QLD would imply a “data glitch” – the commissioning engineers’ explanation (do not believe the meter).
Not so fast – suppose it was real – the operator’s assumption – what could cause such widespread dips? A power system fault would match the observed dips.
To be visible in the 4-second data it would need to have been sustained for a reasonably long period (> 4 s) which is unusual (for transmission faults) but not unknown.
There would be a protection record somewhere which would have recorded the fault – but AEMO initially would not be aware of this and hence it wouldn’t trigger a market notice.
There are two distinct questions out of this episode:
1/ was there a real power system event causing the power dips shown in the charts here and the enormous swings in recorded boundary flows for the NSW AC interconnectors (VNI & QNI)?
2/ why did the change in recorded boundary flows affect the market demand calculation for NSW only? Exactly the same interconnector data should have fed into demand calculations for Queensland and Victoria – offsetting the NSW drop – but it didn’t. Recorded demand for the neighbouring states appeared to reflect very little change in interconnector flows. Only some obscure glitch in AEMO’s dispatch process could explain these different outcomes.