South Australia has a modest amount of wind power, at around 1.5 GW nameplate capacity, but it also has a modest demand for electricity, and there are times when wind power can meet or exceed total demand. The state has also just lost its last coal-fired power station, and its current generator mix  is one towards which many other grids are moving.

The high penetration of variable wind power, and the recent exit of local coal power makes grid stability and control an important issue. To begin to investigate this issue I decided to look at the electricity data for the morning of 28th April 2016, when there was a power outage during rush hour on an electric train system in Adelaide; was this a problem purely with the railway system, or was it triggered by a grid instability?

The following two figures show data for electricity generation and import/export at 5-minute intervals for the morning of 28th April (a Thursday), and for some immediately preceding and subsequent working days. The upper figure shows total consumption in SA, the sum over all local generation, plus net import/export. The lower figure shows data for the separate generation types and for the individual interconnectors.



The upper figure shows an unusual drop in total consumption at 7:30 am, matching press reports of the time of the power outage of the railway. The size of the drop (around 100 MW) suggests to the author that it might have been a lot more than just the loss of the MW being used at the time by the railway, i.e. there might have been a sudden but temporary drop in supply which tripped the railway electrics.

The lower figure shows that wind power was almost totally absent at 7:30 am, thermal generation supply was falling, and the interconnectors were in a period of anti-phase oscillation, with a very substantial drop of over 200 MW in import via the Heywood interconnector between 7:20 and 7:30 am.


There is not enough data available to the public to determine what caused the power outage at the Adelaide railway on 28th April 2016, but the data shown above suggests to the author the distinct possibility that a sudden and large drop in supply was involved.


See the NEM ELECTRICITY DATA page above for links to the archive data used, the interconnector  data shown is METEREDMWFLOW from DispatchIS files, and the generator data shown is the sum of SCADAVALUE figures from DispatchSCADA files.


This entry was posted in Uncategorized. Bookmark the permalink.


  1. Greg Kaan says:

    The noted anti-correlation between the flows in the Heywood and MurrayLink interconnectors could be due to the different nature of the interconnectors and the layout of the eastern portion of the SA grid.

    Heywood is a straightforward AC link with suitable transformers at each end for the local grid voltages. As such, flow control is merely due to the voltage difference (scaled for the transformer windings) between the South Australian and Victorian grids at the terminating substations.

    MurrayLink is an altogether different proposition. It is a HVDC link that is fully controllable via the AC/DC inverters at the terminations so flows can be directed at will (within the capacity limits of the interconnector).

    This introduces the possibility of anti correlation if the eastern portion of the SA grid ties the 2 interconnectors more strongly than to the rest of the grid since an increase in flow from MurrayLink would increase the general voltage of the eastern portion of the SA grid which would then reduce the voltage differential between Heywood and its SA counterpart.

    Of course, if this were the case, any generation in this portion of the grid would also have an effect on the Heywood interconnector flow. I know there wind farms and a gas turbine plant in the Mt Gambier region.



  2. Greg Kaan says:

    There is an issue with your graphing of output from the Torrens plants.

    For the period from 6:15am to 7am, the dispatch figures only show some small drops (around 4MW) in output for Torrens A1 at 6:20, 6:30, a small sag between 6:40 to 6:50 and then 6:55 to 7:00. The total output from all 3 operating Torrens plant (A1, B2 and B4) climbs fairly steadily over this whole period.

    There is a fairly severe change in output from the Pelican Point plant (PPCCGT) at 6:20 where it dips from 45MW down to 11MW and then recovers to 36MW which is maintained from 6:25 onwards.

    I chose these plants and this time period after looking at the graphs for the day at the aneroid energy site. I could not see any evidence of event B when looking at fossil fuel generation and these seemed to be the active plants during the period I looked at (aside from the Northern Plants which had flat output). I did not investigate event C.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s