2019 was not a great year at the Liddell Power Station

It was September 2014 that AGL Energy acquired Macquarie Generation (comprising Bayswater and Liddell power stations in the Hunter Valley as operational assets) in a process approved in the Australian Competition Tribunal, but which was not popular at the ACCC.

However AGL’s association with the stations extends back further – with the “gentrader” rights having been secured sometime in 2011 (will link if I can find reference).

(A)  Never out of the news?

It seems that the portfolio (and particularly the Liddell station) has almost never been out of the news since that time.  Especially after the closure of the Hazelwood Power Station, an increased focus has turned to the Liddell power station, being a lightning rod for the large numbers of people seeking a quicker energy transition.

(A1)  Closure plans

Who can forget the ‘shirt fronting’ that happened between former CEO, Andy Vesey, and the Federal Government in Canberra that rolled on for months after AGL Energy announced plans to retire.

Can’t recall specifically when the closure announcement was first made, but did find this AGL Media Release on 9th December 2017 speaking about options for replacing the required capacity.  Despite the ongoing pressure placed on AGL (lasting several PMs and two CEOs) the decision to close has not changed  – though there was a bit of a concession with respect to slightly delaying the closure (discussed on WattClarity on 7th August 2019).

(A2)  Latest announcements

Hence it was not really a surprise to hear about Liddell in the news also this week, on the back of comments from the current CEO (Brett Redman) about ongoing operational difficulties at the plant.

These comments were reported here on RenewEconomy as “AGL says it’s getting harder to keep Liddell coal plant online” , and mentioned in passing (with respect to the brighter news story of the AGL-Vena tie-up over the Wandoan BESS here in the Australian, and in longer coverage here in the AFR).


(B)  The year that was in 2019

Coincidentally, just having released our Generator Statistical Digest 2019 only two days ago, I was curious to open it up and have a look at what was shown of the operations of the plant through calendar 2019 on the relevant ‘B’ pages (remembering that there is also a 10-year history included using different stats on the ‘A’ pages).

(B1)  Snippets for each unit from the GSD2019

In developing the GSD2019 we chose to show results on a unit-by-unit basis (which explains why the document amounts to 660 pages after covering 304 units across the NEM, and the addition of a detailed glossary).  Hence here are some snippets from the top of the relevant ‘B’ Page for each unit:

Liddell unit 1

At the top of the ‘B’ page for each unit are two charts that sum up the outcomes through the most recent year for each unit – showing daily range of operation (min to max) and also the daily capacity factor (months shown in text).


For unit 1 we can quickly observe a number of different findings:

Observation #1 = during CAL2019 we can clearly see the plant offline on 6 different occasions through the year, which (noting that it’s possible the unit could have been offline on multiple occasions through a given day, if it tripped in trying to come back online) is not great performance

Observation #2 = perhaps more importantly, we can also see that the volume of capacity offered to the market through the year did not rise above 75% of its Maximum Capacity (as stored by AEMO) through the year:

2a)  It’s my understanding that this restriction has been in place for a number of years and is a result of limiting boiler pressure, and hence reduce risk of tube leaks in the boiler.

2b)  Also worth noting that it’s operational decisions like this that are part of the reason why the 12-month trailing aggregate partial unavailability has been trending upwards for at least 10 years, as outlined here in June 2019 following from the detailed exploration in the Generator Report Card 2018 .

2c)  However lest readers misinterpret these results, our work (also here in June 2019) in the GRC2018 showed that there was NOT a consistently increasing trend in full unavailability across the 48 coal units across the same period.

Liddell unit 2

Here’s the same view for unit 2:


Again, we can see a pretty poor performance through the year – with 6 discrete periods of outage over several days each.

(B1)  Liddell unit 3

The picture for unit 3 shows that this unit had a better year than the other 3 units…


… though still not a great year given the lower maximum capacity available.

(B1)  Liddell unit 4

Finally, with unit 4, we see that the performance stats are dominated by that 3-month outage through March, April, May 2019:


In addition to that, we see three other discrete periods where the plant was offline.

(B2)  What does ambient temperature have to do with coal failure?

In recent times we’ve witnessed that the rise in temperatures that comes in every summer comes in parallel with rising temperature of rhetoric from those propagating a broader #coalfail meme.  To those who’ve been treating all that’s been said under that meme as fact, then the question above might seem superfluous.

However in reality the question is a good one that needs a factual answer.

(B2a)  We’ve seen no evidence above that ambient temperature is a significant factor in full unavailability!

I’d challenge objective readers to review the trends for each unit and ask their own questions about the seemingly ubiquitous (but nevertheless also erroneous*) claims that “coal fails in the heat”.  To me it seems that we might as well make a claim that “coal fails when water is wet” as it would be just as true – but nowhere near as misleading* as the “fails in the heat” meme is trending to become.

* I fully admit that there might be something I have missed in the evidence of output performance of coal units like these – or in my underlying technical understanding of the mechanical processes in the fuel combustion and steam cycle that come together to make these technologies work to produce power. 

* However even after extensive investigation through the GRC2018 (and now the GSD2019) we have not found any evidence.

* Hence, if there are people who’d like to help me understand the factual basis behind how they assert that “coal fails in the heat”, I would be all ears (and happy to share with readers on WattClarity later, as well).

In the GRC2018 we wrote about how the “transitional challenges require greatly enhanced analytical capability” (as Theme 14 within Part 2 of the 180-page analytical component):


Misleading statements linking coal failures to high ambient temperatures seem just one example where we all need to lift our game.

(B2b)  Source of the misunderstanding?

Finally, worth noting two factors that could be combining to lead to this misunderstanding (which, unfortunately, certain parties seem to have latched onto and propagated for their own purposes).

Source #1)  A general confusion between Probability and Consequence

I’ve written about this before, and hopefully don’t need to reiterate.

Source #2)  Confusion between high-temperature limitations on output and plant tripping

I can understand that, to those who don’t have a technical understanding of the combustion-steam cycle, it’s easy to misconstrue a limitation on efficiency at high temperatures with plant tripping (and/or higher forced outage rate) at high temperatures.

For one there is a link, for the other, there is not.

Thankfully in the GSD2019 we also include (for every unit) some stats in a ‘Temperature Energy Harvest’ illustration that show the effect of any high-temperature limitation on the output capability of the plant.  Here’s the view for Liddell unit 4, which shows the most significant impact:


As can be seen, the maximum availability of the plant was considerably lower where local temperature was above 42.5 degC (which occurred for 2.3 hours during calendar 2019).  However the expected effect across a broader temperature range is masked by the fact that the available capacity bid into the market at the plant was never more than 400MW as shown in the first set of charts above.

Across the road at Bayswater unit 4, we see the analogous illustration for Bayswater unit 4 (it’s less pronounced on the other 3 units):


Given the plant is water-cooled, it is not a surprise to see a gradual (and slight) decline in cooling efficiency, and hence maximum capability, at higher ambient temperatures.


High ambient temperatures affect the majority of generators covered in the GSD2019 in different ways.  We’ll look forward to delivering your copy so you can see the details …

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.

1 Comment on "2019 was not a great year at the Liddell Power Station"

  1. Ambient temperature has zero effect on a boiler or a steam turbine. Heat exchangers in auxiliary services such as lube oil coolers and the condenser hotwell is where the effects of high ambient temperature come in.

    High oil temps can trip a unit quickly, especially if a cooling fan fails, but gradual temperature increase can usually be halted by reducing load.

    Gradually losing vacuum in the condenser is easily managed by reducing load.

Leave a comment

Your email address will not be published.