24th February 2006 – price spikes in Victoria

On Friday 24th February 2006 there was a peak in demand in Victoria.  On the same day (but several hours before the afternoon peak in demand) there was a pronounced price spike that lasted 2.5 hours.

Both these incidents can be seen in the following image:

(please click on the image for a better view)

Also on the same chart is a stacked aggregate generator output for all generators in Victoria (identified by the current trading company).  Finally, we have also included flow over the three operational interconnectors connecting Victoria to elsewhere in the NEM.

From the information provided in this chart (and extracted from NEM-Review for further subsequent analysis), it has been identified that:

1) From 06:00 that morning, aggregate supplies from within Victoria were insufficient to economically supply total Victorian demand, and hence Victoria began importing power from neighbouring regions:

(a) At 06:00, flow on the Snowy-Victoria line had reversed (Victoria normally exports at night) and Victoria began importing from the Snowy region.  These imports reached a maximum at 08:30 before declining again.

(b) At 08:00, Victoria began importing a small amount of power over the Murraylink DC interconnector.

(c) At 09:00, Victoria began importing a greater amount of power from South Australia over the Heywood interconnector.  This allowed the reduction of flow from Snowy into Victoria (which reflected increased transfers north from Snowy into NSW).

2) Also during the morning, production at Victoria’s peaking capacity began ramping up to meet the anticipated afternoon peak in demand

(a) At 06:00 (with the Victorian pool price at only $24.08/MWh) Southern Hydro began supplies of power from its portfolio of plant.

(b) At 09:00 (three hours later and when prices had climbed to $80.15/MWh) the capacity of Ecogen Energy began supplies into the tight market.

(c) At 09:30 (with prices still climbing at $98.69/MWh) the Alinta plant at Bairnsdale began supplying into the market.

(d) Only 30 minutes later (but with prices now at $2,240.93/MWh), the AGL plant at Somerton had begun supplying power into the market.

3) In the following chart, the aggregate available generation capacity in Victoria is compared with the total demand in the region:

(please click on the image for a better view)

(a) It can also be noted (as shown in the image below) that demand in Victoria exceeded the total aggregate available generation capacity in the region between 14:00 and 17:00.  Hence, imports were required during this period no matter what the price offered by domestic generation.  Yet note that prices were modest during this period.

(b) In contrast, during the time of the (earlier) peak in price, it can be seen that there was about 500MW of surplus available generation capacity in Victoria over and above the level of demand at the time.

(c) Indeed, given that (as seen in the first chart) between 500MW and 800MW of this demand was met by imports during the 2.5 hours of high prices, it can be seen that there was 1,000MW to 1,300MW of plant in Victoria that could have operated but chose not to for commercial reasons.

4) The key observation that can be made with reference to the chart above

is that the price spike experienced between 10:00 and 12:00 (i.e. over 21/2 hours) coincided with a reduction in imports from Snowy.

(a) This can be understood when it is known that NEMMCO had placed constraints on the flows on the Snowy-Vic interconnector over the period – these constraints can be seen clearly in the snapshot we took on NEM-Watch at 10:40 on that day:

(please click on the image for a closer view)

At 10:40 market time, the constraints on flow over ALL interconnectors into Victoria can be clearly seen (through the red arrows)

(b) Given that Victoria was economically isolated from the rest of the NEM for that period, local generators in Victoria had market power and hence set prices substantially above the prices in the neighbouring regions.

(c) This point is clearly illustrated when it is seen that, NEMMCO issued Market Notice #14259 to advise participants that the limiting constraint on Snowy-Victoria flow has been revoked from 12:10 due to the return-to-service of a Wagga to Yanco 132kV transmission line.
As seen at the top of the page, this action immediately led to the end of the price spike in Victoria (the 12:30 price was much lower, at $98.51/MWh).

Given that Victorian generators had market power for a period of about 2.5 hours (remembering that the market is dispatched on a 5-minute basis), it is useful to look at Generator behaviour (by trading company) on the day to see how each company responded to the local shortage of capacity in Victoria.

The following charts illustrate the behaviour of each participant over the day.  We’ll start by looking at the behaviour of the large base-loaded stations:

Loy Yang B station



This charts shows that the aggregate output of the Loy Yang B station was constant across the day, and that the availability of the two units did not change across the day.

From the graph it can be seen that output at the station increased marginally at 07:30 (when Victoria began importing) and again at 11:00 (with the price spike in effect.

Based on analysis we have performed of other events of note in Victoria, this appears fairly typical of the Loy Yang B station, now operated by Internationalpower.

Hazelwood station


Hazelwood (also operated by Internationalpower) is the highest-cost of the 4 large baseload plant in Victoria.  Hence, Hazelwood tends to reduce output overnight slightly in order to ensure prices do not fall substantially.

From 04:30, Hazelwood production across all 8 units increased by about 230MW in time for the usual morning increase in prices.

When the price spike hit the market, Hazelwood was able to increase output by another 50MW in order to gain more of the available (and highly profitable) revenues.

Yallourn station


Early that morning, the Yallourn station was producing a full output of 1,465MW.

However, from 08:30 to 11:30, we see that the Yallourn unit 1 reduced its available generation capacity from 350MW to only 300MW.  We are not aware of the reasons for this temporary reduction in output capability (it was reversed by 21:30 the same day).

However, it can be seen that the output of Yallourn station has dropped by double that amount (i.e. by 101MW from 1,465MW to 1,364MW at 12:00, the last half-hour
of the price spike).  This reduction in output will have had the effect of increasing spot prices – especially over the period of the spike.

Loy Yang A station

It is illustrated in the chart that Loy Yang A had a unit offline until early that morning.

Further analysis in NEM-Review reveals that Unit 1 had been taken offline at 17:30 the previous day (after the previous day’s price spike), and that it had been
returned to service beginning 06:00 on the morning of the 24th.

As seen in the chart, the unit was fully available again by 10:00 (unit availability fluctuated between 500MW and 580MW over the period between 07:00 and 10:00).  Hence, at the commencement of the price spike, all units at Loy Yang A were fully available.

We can also see that between 08:30 and 09:30, output at each of the 4 Loy Yang units dropped from 100% utilisation by an amount totalling 207MW (over a period when the spot price was increasing from $34.20/MWh to $98.69/MWh).  Given that declared availability did not change over the hour, it can only be concluded that this change was as a result of changes to bidding behaviour over the period.

Furthermore (despite the fact that prices rose above $2,000/MWh for the 10:00 and 10:30 trading periods), it can be seen that the output at each Loy Yang A station did not change markedly until the 11:00 trading interval, when the average price over the half-hour was $9,134.13/MWh (output rose 129MW). Output also reduced slightly in the next half-hour (by 31MW), when prices dropped to $8833.89/MWh.

This illustrates how a significant portion of the Loy Yang A output was bid opportunistically at or around $9,000/MWh in the knowledge that imports to Victoria was constrained, and that (by virtue of a large peak in demand) supplies would be very tight.


There are also two smaller base-load plants that operate in Victoria – the Angelsea station (traded by VicPower Trading) and the Energy Brix plant in the Latrobe Valley (traded by HRL).  The behaviour of these two companies is also illustrated below:

Angelsea station


This charts shows that the Angelsea plant operated steadily throughout the day on maximum load – with the exception of the period of the price spike, when the available capacity at Angelsea dropped by about 30MW.

It has not been ascertained why there was a drop in available capacity at this critical time.

Energy Brix (a.k.a Morwell) station


The aggregate output of the Energy Brix plant can be seen to be almost at full capacity throughout the day.

It can be seen that Morwell unit 5 increased its output slightly (by about 14MW) in line with the incidence of the price spike that morning.

Finally, for completeness, we have also included a brief analysis of the output of each of those companies in Victoria who were able to offer peaking capacity into the market on this day:

Southern Hydro stations


This charts shows the aggregate output of the four key Southern Hydro stations (there are several other stations that are not dispatched by NEMMCO and hence whose output is not included in this data set provided by NEM-Review).

As is noted above, Southern Hydro is the first of the peakers to begin production (at the time when Victoria begins importing power).  This chart shows this
energy is being produced at Eildon station.

The other 3 stations can be seen to begin contributing from 08:30 and then to continue contributing at maximum capacity through the time of the price spike, and then on through the afternoon peak in demand, after which production was quickly curtailed (presumably to conserve water supplies).

Ecogen Energy stations


Ecogen Energy (a Babcock & Brown infrastructure company) operates the Newport peaking station in Melbourne and the Jeeralang OCGT plant in the Latrobe Valley.

In contrast to the Southern Hydro stations, these gas-fired stations are seen to run a small amount (only 10-20% of available capacity) through the price spike,
before increasing to 50% capacity factor in the lead-in to the afternoon peak in demand.

As can be seen, the Newport thermal station can be seen to be contributing almost all of this energy, whereas the smaller Jeeralang A&B units are not shown to contribute much at all (despite being declared available in the market, and despite the high prices).

Alinta stations


Alinta operates the Bairnsdale station, which was originally constructed by Duke Energy.

As can be seen in this figure, both Bairnsdale units were running flat-out from the time of the morning’s spike in prices, and then through to the evening’s peak in demand.

AGL stations

AGL has been operating the Somerton OCGT plant in Melbourne since its commissioning in April 2002.

Further analysis in NEM-Review can illustrate how monthly output at the station has been increasingly progressively since its commissioning, and is now above 100MW average per month (or above 60% capacity factor).

On this particular day, we can see that the Somerton plant was producing at full output from 10:30 (just after the start of the price spike) through till 18:00 (after the peak in demand).


About the Author

Paul McArdle
One of three founders of Global-Roam back in 2000, Paul has been CEO of the company since that time. As an author on WattClarity, Paul's focus has been to help make the electricity market more understandable.

Leave a comment

Your email address will not be published.


*