Worth a quick post, whilst it is happening, to note that both Queensland and South Australia have been islanded from (what remains of) the NEM for reasons unknown at this point in time. Here’s four images from NEMwatch v10 today:
First, we have a Market Notice issued at 13:46 to note that the coincident separation events occurred at 13:12 today (i.e. 34 minutes before the notice was issued):
Note the 1,080 MW total loss of load (presumably via automated under-frequency load shedding) – being 800MW in NSW and 280MW in Victoria. This will have happened because both QLD and SA were exporting into NSW+VIC at the time of the links tripping, as a result of which:
1) Frequency drops in VIC+NSW
2) Frequency rises in QLD
3) Frequency rises in SA, probably by more than in QLD because of smaller-sized region.
(remember this explainer here on WattClarity about how frequency changes with changes in the balance between supply and demand, and remember this happens very quickly – well within the 5-minute dispatch intervals).
Second image is a trend of the target flow over the larger (Heywood) AC interconnector from Victoria to South Australia. As flagged on the image, Heywood had been flowing east over the day prior (because of high wind) but had been trailing down quite quickly as the wind faded. This meant lower drop in supply into Victoria (hence lower loss of load in VIC, though this is coincidence).
Third image is a zoomed-in view of flow on the QNI interconnector, showing that the target flow and limits dropped to 0MW in the 13:25 dispatch interval (which was the same as for Heywood above):
We note that Queensland had been exporting roughly 800MW at 13:15, which is the size of the load shed in NSW. We also note that, at the time the image was taken (14:10 dispatch interval) the interconnection had still not been restored. As noted in the image, this is not a simple thing to do, as it requires AEMO to align the frequencies in the two “islands” before reconnecting.
Final image is a snapshot from NEMwatch taken for the 13:40 dispatch interval (i.e. before the Market Notice above issued) showing some price gyrations resulting in both QLD and SA – and also the loss of load in VIC+NSW:
No time for more…
What will be particularly of interest is why the two interconnectors tripped at the same time.
PS, from AEMO it’s worth linking in this note here on the AEMO website, which explains some of what happened on Saturday.
2nd PS worth also linking in further analysis completed on Monday, which really just seems to raise more questions than it answers.
3rd PS also worth also linking in this subsequent analysis completed on Thursday 6th September, with some assistance from a couple other analysts.
4th PS will link in the draft of the AEMO report, when it is available.
Did the two interconnectors actually trip at exactly the same time or did one trip fractionally earlier? Rapidly falling load on the Heywood interconnector could cause it to open to protect the SA network with consequent drop of frequency in VIC & NSW causing the QNI to trip.
So we didn’t lose a single generator on the system but had to trip more than 1GW of load and on a low load Saturday afternoon.
Mick, as noted to Liam (comment below) I don’t know the exact sequence.
All we know is:
1) Two trips occurred within a 5-minute window. We don’t know how close (but suspect within seconds).
2) Which one came first, we don’t know.
3) What caused the first to trip, we don’t know. There were storms in northern NSW so perhaps lightning was a factor, but that is purely speculation (and was not implied in Market Notices).
4) It seems likely that the 2nd trip was a result of the 1st (perhaps due to rate of change of frequency, as Liam suggests in his comment).
Per AEMO’s market notices, simultaneous tripping of 330kV transmission lines in QLD 8L and 8M Dumaresq to Bulli are now considered credible contingency, synchronous export from QLD to NSW have been constrained. The cause is unknown and AEMO not satisfied this event will not reoccur.
Perhaps the chain of causality is opposite to your suggestion, ROCOF in VIC was too fast, overloading Heywood? I will certainly await the report into this one.
Sorry, Liam. Yesterday when I posted, did not intend to suggest either sequence of interconnector trips – i.e.:
Scenario 1 = whether QNI was first and then Heywood, or
Scenario 2 = whether Heywood was first and then QNI.
The Market Notice you refer to was published at 15:44 (more than an hour after this post).
My reading of it still does not confirm which trip came first (i.e. either Scenario could hold). Your hypothesis is one possibility in Scenario 1.
Like you, we’ll be awaiting the report. A sensible thing for any reader to do, to reduce the risk of forming wrong conclusions….
Cheers Paul, my comment was intended to be a reply to Mick above, adding the later information from the market notice.
A small update. One part of the mystery is solved.
“64012 26 Aug 2018 17:52 RECLASSIFY CONTINGENCY
Cancellation of a Non-Credible Contingency Event: Simultaneous trip of 8L and 8M Dumaresq – Bulli 330 kV Transmission lines as a credible contingency lines in NSW due to protection mal-operation.”
Do we have relay flags from Heywood and Qld as to the causes and PRECISE timing of the tripS?