Brown coal generators failed the grid in Victoria heat-wave, blackouts

A report into the Victoria “load-shedding” events in the January heat-wave, which sparked yet another political battle over Australia’s energy policy, has highlighted the critical failure of the state’s main brown coal generators, as well as the important contributions of rooftop solar at key times.

The report by the Australian Energy Market Operator confirms observations made at the time that its efforts to keep the lights on in the face of a record heat-wave were undermined by a series of failures and capacity reductions at all three of Victoria’s big brown coal generators.

In contrast, AEMO notes, renewables performed better at the time of the major load-shedding at 11am on January 25 than had been anticipated when modelling for the anticipated hot summer was completed.

AEMO notes the unprecedented heat over the two states at the time, but it already had one hand tied behind its back because one unit at Loy Yang A and another unit at Yallourn were offline, taking away some 855MW of capacity.

Another unit at Yallourn was lost on the morning of February 25, removing another 355MW of capacity, Loy Yang B struggled in the heat and lost a further 85MW of capacity and eventually Loy Yang A lost a second unit – again to a tube leak – taking away another 370MW of capacity.

The outages – couple with higher than expected demand – cruelled hopes of avoiding any load-shedding, and while the ability to summon emergency reserves reduced the extend of load-shedding by half, more than 260MW of load was lost..

The performance of the brown coal generators contrasts sharply with other generation sources. This graph above highlights the actual generation (in red), and shows (through the orange dot).

As AEMO notes, brown coal performed substantially worse than expected when it was putting its plans together a few months ahead of time, gas and hydro performed as expected, but both wind and solar did a lot better than expected.

This is important. Wind and solar are often criticised because they are variable, or intermittent. They don’t claim to be dispatchable – unless they are paired with storage of some sort – but they can claim to be predictable.

That’s what counts for the market operator. Its major difficulties come from those generators – particularly the ageing coal generators – that claim to be dispatchable but often are not, particularly in the heat.

This is just one of the many missing links in the half-baked public debate about energy supplies. The brown coal generators are not only dirty, but they are not flexible, and struggle to deliver anywhere near their rated dispatchable capacity in the heat.

This graph above also highlights that rooftop PV output was nearly double expectations. This is an important contribution, given that it added another nearly 500MW of unexpected resources. Indeed, as these next graphs show, rooftop solar played a significant role in reducing “actual” grid demand at the critical peak times.

Most of of the rooftop solar generation is consumed on-site in homes and businesses, so is not seen by the grid. Some of the excess is exported into the grid.

This graph shows what the grid demand would have looked like without rooftop solar – the peaks wold have been significantly higher and longer. The scale and time of load shedding would have been much greater.

The same thing can be said on South Australia, which has a very high penetration of rooftop solar, where rooftop solar reduced actual demand (the demand the grid sees and needs to match with supply)  on both Thursday and Friday.

The first series of load shedding occurred in Victoria just after 7pm on January 24, when a 75MW shortfall was met by instructing the Portland smelter in Victoria to shed one potline. That was offline was restored in less than one hour, as it must without ill effects, when a second potline was take off line for another 45 minutes.

The bigger and more widespread load shedding occurred from 11am on January  25, with 100MW and then another 150MW shed in Victoria. The Portland smelter was not an option, because it still recovering from the previous day’s efforts, so the shedding was spread across some 200,000 other consumers on a rotating basis over the next 2.5 hours.

AEMO was able to call on emergency reserves through its RERT mechanism to avoid any load shedding in Victoria, and reduce the potential loss of power by half in Victoria.

The cost of the emergency reserves used – a focus of the Murdoch media – were less than they would have been without them – resulting in payments of $10,000/MWh, rather than the $14,500/MWh that would certainly have gone to any gas and diesel plants available.

AEMO put the total cost at $34 million, but noted that this amounted to a total annual cost of $3.20 to each Victorian household, and just 80c for each consumer in South Australia (which include the cost of firing up the back-up state-owned diesel generators).

Still, there is work to be done. AEMO, the Energy Security Board and the Australian Energy Market Commission are looking at more measures to deal with such heatwave events, which are expected to become more frequent, and more challenging as the coal fleet continues to age.

This includes work on a strategic reserve, and an upgrade of the interconnectors. (Once the Murray hydro plant was switched on, there was little capacity to import anything more from NSW).

The three institutions are also looking at how they can better use and manage distributed energy resources (DER), which include rooftop solar, battery storage and demand management, noting the huge increase in capacity expected from state-based schemes such as in Victoria.