High NSW price forecasts not materialising in dispatch, 29th November 2024

A curious phenomenon we’ve been watching with interest today, the 29th of November 2024, has been the somewhat persistent and alarmingly high predispatch price forecasts for NSW.

If you had looked at the predispatch forecasts last night, you might have been shocked to see NSW Energy prices forecast to spend more or less the whole day at or near the MPC of $17,500.

But, as the day progresses, apart from a short burst of actual volatility at 06:55 and 07:05, these predispatch forecasts have not ended up materialising in P5min forecasts or dispatch, disappearing seemingly “just in time”. Actual prices have been very moderate throughout the day.

What’s going on? We’ll use ez2view to investigate. This screenshot is from ez2view at 11:00.

Using the “Constraint Equations” widget, we can apply a filter to select just those constraints affecting the NSW region, then order by “Maximum MV” over a reasonable forecast horizon, in this case, 4 hours. Here, the first result is the N::N_CNSD_2 constraint, forecast to bind with a large maximum MV of -$18,134.52 in 4 hours.

This is an outage constraint for a very small section of 330kV transmission in the ACT, the 3C line, Canberra – Stockdill, some kind of maintenance scheduled in advance, basically. The constraint aims to provide a “stability limit (Snowy-NSW) for fault of various location between Yass-South Morang area”.

Here in “Constraint Dashboard” we can see that looking back to the actual dispatch outcomes, the marginal values of this constraint in dispatch show that it has been binding, but with continually moderate MV.

Looking into the future P30 horizon here however, the predispatch outlook is quite dire, with very high MV!

The final piece of the puzzle, looking at the “Plain English Translation” for this N::N_CNSD_2 constraint from “Constraint Dashboard” widget offers at least a partial explanation — the constraint RHS terms are vastly different in Predispatch vs Dispatch.

Essentially, NEMDE is forecasting possible outcomes based on one equation in predispatch, but a very different form of equation in dispatch, so the actual outcome in dispatch is that the constraint is not binding anywhere near as much, and so prices are remaining moderate.

As to why this particular N::N_CNSD_2 constraint equation is so different between dispatch and predispatch, we’d welcome further insight or comments from knowledgeable readers!