We were a bit distracted by the unfolding events of yesterday (Thu 22nd Feb 2024) … but before then I had been thinking through a few questions and implications about what happened in the ‘Significant Power System Event’ on Tue 13th February 2024.
1) In the AEMO’s preliminary Operating Incident Report they identified 4 discrete events – and (included within these events) were ‘trips’ (or at least fairly rapid off-loadings) of 4 Generation Stations and 1 Load.
2) Using the AEMO’s 4-second SCADA data, we took at look at each of these … and unit performance in general across a 14-hour period.
3) I noted in the AEMO’s preliminary report that what happened at Dundonnell was called a ‘trip’, whereas the words used to describe what happened at Stockyard Hill Wind Farm, at Loy Yang A and also at Yaloak South Wind Farm were different (and notably not a ‘trip’).
So I wondered … how fast is a trip?
Using the AEMO’s 4-second SCADA data (and remembering all the caveats!), I’ve derived an implied ROCDOWN rate for each of these 9 supply-side units on Tuesday 13th February 2024 as follows:
Clearly what happened at Dundonnell Wind Farm is an order of magnitude faster than what happened at the other units … and:
1) That’s not even taking into account there’s a 3 second gap in the data points between ‘ON’ and ‘OFF’ so the ROCDOWN rate is likely to be considerably larger than what is noted here (will need to wait for subsequent AEMO reports using the high-speed data to see).
2) Previous caveats about the rubbery nature of timestamps in the 4-second data not forgotten!
For interest purposes I have compared this demonstrated ramp with the ‘MAX_RAMP_RATE_DOWN’ for these units currently in the DUDETAILSUMMARY table of the EMMS. Understandably what was demonstrated on 13th February 2024 was faster for all units … because what happened on Tuesday 13th February was not normally intended operation from one DI to the next.
With respect to these fast off-loadings on Tue 13th Feb 2024 I’ve compiled this table to provide easier reference:
| DUID | Timings
Noting caveats on the timestamps on the 4-second SCADA data. |
Notes |
|---|---|---|
|
STOCKYD1 |
As Event 1, the unit shows: 11:34:23 at 470MW … so offline in 88 seconds (ROCDOWN = 320.5MW/min), as noted in the graphic above. |
On Tuesday 20th February we published ‘A first (more detailed) look at Stockyard Hill Wind Farm on Tuesday 13th February 2024’. Not sure we’ll have time (or if there is a need) for us to do more here? Incidentally, we also posted ‘Stockyard Hill Wind Farm briefly off again, on Thu 22nd February 2024 (constraints a factor)’ for yesterday’s event. |
|
DUNDWF1, DUNDWF2, DUNDWF3 |
At the start of Event 2, these units were clearly the fastest… All three units were: 13:08:56 online, with the units: DUNDWF1 = 128MW, before ROCDOWN = 2,560MW/min DUNDWF2 = 37MW, before ROCDOWN = 740MW/min DUNDWF3 = 113MW, before ROCDOWN = 2,260MW/min 13:08:59 offline at 0MW. |
On Friday 16th February, Allan wrote ‘Early observations from examining the AEMO’s four-second SCADA data from Tuesday afternoon’ – which showed both Dundonnell and Loy Yang A. We’d like to dig into what happened at Dundonnell Wind Farm in more detail … but that depends on available time (a scarce resource!). |
|
LYA1, LYA2, LYA3, LYA4 |
Three seconds later in Event 2, we have the four units at Loy Yang A: 13:08:59 online. Then varying times offline, with varying ROCDOWN Rates: LYA1 was … LYA2 was … LYA3 was … LYA4 was … |
On Friday 16th February, Allan wrote ‘Early observations from examining the AEMO’s four-second SCADA data from Tuesday afternoon’ – which showed both Dundonnell and Loy Yang A. Likewise will see if we have time to do more… |
|
YSWF |
Taking 59 seconds longer than DUNDWFx to start ramping down in Event 2, the unit shows: 13:09:55 at 23MW … so offline in 16 seconds (ROCDOWN = 86.3MW/min), as noted in the graphic above. |
There’s been nothing, yet, specifically looking at Yaloak South Wind Farm … but interesting to see that (assuming SCADA timestamps are in sync) that this happened 1 minute after Dundonnell and Loy Yang A. |
|
APD01 |
I did not add into the graphic above, but (for completeness) including Event 4, we see: 15:00:27 at 295MW … so down in 4 seconds (implied ROCDOWN = 4,365MW/min). |
Nothing has been written yet about this, other than a note in ‘From the 4-second SCADA data, a view of Production by Fuel Type on Tuesday 13th February 2024’. Not sure we’re going to have time to look further… |
Hope this helps?
Interesting article — I’m a long-retired EE who worked in and around generating plants for 33y. Five of those years were in fossil and nuclear plants with 11 units. It never occurred to me to ask what the time interval was between the unit trip and the generator trip, nor do I recall ever seeing any records that would provide me with that information. Thanks for answering the question I never thought to ask. (USA)