Drop in forecast wind conditions for Thursday 12th May 2022 briefly triggers forecasts of Load Shedding (LOR3) in South Australia

A short note for Monday morning to highlight three Market Notices published yesterday (Sunday 8th May) briefly flagged the possibility of LOR3 (i.e. ‘load shedding’) Low Reserve Condition for the South Australian region this Thursday morning and evening:

I did not have a chance to investigate the SMS Alerts further yesterday, but the (relatively rare) instance of forecast LOR3 (and question about the even-more-rare instances of actual LOR3) prompted me to have a quick look Monday morning.

(A)  Progressive Market Notices

At 04:35 on Sunday morning, Market Notice 96156 first flagged the possibility of load shedding:

[1.] From 0830 hrs 12/05/2022 to 0900 hrs 12/05/2022.
The maximum load (other than interruptible loads) forecast to be interrupted is 54 MW at 0830 hrs.

At 09:37 on Sunday morning, Market Notice 96158 upgraded the forecast for Load Shedding as follows:

______________________________________________________________________________________________

Notice ID            96158
Notice Type ID        LRC/LOR1/LOR2/LOR3
Notice Type Description    MARKET
Issue Date            Sunday, 8 May 2022
External Reference        STPASA – Update of the Forecast Lack Of Reserve Level 3 (LOR3) in the SA Region  on 12/05/2022
________________________________________________________________________________________________

AEMO ELECTRICITY MARKET NOTICE

The Forecast LOR3 condition in the SA region advised in AEMO Electricity Market Notice No. 96156 has been updated at 0930 hrs on 08/05/2022 to the following:

[1.] From 0800 hrs 12/05/2022 to 0930 hrs 12/05/2022.
The maximum load (other than interruptible loads) forecast to be interrupted is 113 MW at 0830 hrs.

[2.] From 1730 hrs 12/05/2022 to 1800 hrs 12/05/2022.
The maximum load (other than interruptible loads) forecast to be interrupted is 7 MW at 1730 hrs.

AEMO is seeking a market response.

AEMO has not yet estimated the latest time it would need to intervene through an AEMO intervention event.

Manager NEM Real Time Operations

Also on Sunday, at 15:32:37 the Market Notice 96164 cancelled this forecast for load shedding.

(B)  Forecast Supply-Demand Balance

I saw all this evolve yesterday via small screen, but did not have a chance to look further, so have had a brief look this morning – using the ‘Forecast Convergence’ widget in ez2view.

(B1) Available Generation

Seeing as it was probably a change in forecast Available Generation in South Australia that cause the cancellation of the forecast LOR3 condition for Thursday 12th May, I started looking that:

As noted on the image, we can see there was an increase made in forecast capacity available in the ST PASA run yesterday for 15:00 … which seems to confirm that this was the reason for the cancellation of the LOR3 warning in Market Notice 96164 at 15:32.

However we also see that there are a number of other red slices in this grid (i.e. low Available Generation) that have not triggered the same level of concern – so we wonder why this might have been the case…

(B2) Forecast capability of Wind and Solar

… so we flip to look at ‘Wind and Solar’ as follows:

Here we can see two things:

1)  Looking up the vertical where I have marked the approximate point of Thursday morning where the LOR3 forecast was for, we can see that the forecast Wind & Solar capability dropped significantly (i.e. to deeper blue) late on Saturday night, and further still on Sunday morning … hence triggering the LOR3 warning

2)  Looking horizontally across the span, we see that this deep blue patch (spanning Thursday into Friday morning) is easily the lowest period forecast for the coming 7 days – hence pointing out why the other periods of low Available Generation have not triggered the warning.

(B3) Forecast Demand and Non-Sched Gen

For completeness we add in the similar view of ‘Demand and Non-Scheduled Generation’ (i.e. which is something similar to ‘Operational Demand’) in the grid:

Again, Thursday morning is marked – whilst a small increase for morning demand, nowhere near the highest (i.e. ‘reddest’) cells in the grid over this 14-day time range shown here.

———–

PS1  method of the quick analysis performed

Thanks to Jack (here), and also to Alex (here), in highlighting two of the factors involved in comments below.

To avoid any reader confusion, my focus in the above was to come from the perspective of saying what changed to trigger the LOR3 warning (and then its resolution).   Hence a number of underlying factors that made the LOR3 warning more likely (but which did not substantially change through ST PASA updates on Sunday) were not really my focus.