Origin Energy’s announcement of the early closure of Eraring can be seen from a number of different perspectives: environmental, political and economic. The fact is, they have called the market and perhaps New South Wales’ “bluff.”
Can we really build the capacity to replace Eraring? To do so is going to require excellent and accelerated execution of the NSW road map. It’s also going to require, for a while, access to South Australian generation that Project EnergyConnect will bring.
The costs of poor execution of the NSW and Integrated System Plans are high. Every hour of maximum ($15,000/MWh) prices in the spot market adds almost $2/MWh to annual average prices. As Tom Geiser from Neoen points out, that represents – or should be – the true benefit/cost priced into transmission analysis.
All stakeholders, but particularly those responsible for transmission and social license, are going to have to be working hard.
In this note, out of necessity we briefly look at the energy and power impact of Eraring’s accelerated closure but then move on to consider ISP submissions.
It might seem very boring to be reviewing ISP submissions when such market-moving announcements are being made, but really understanding what has to be done to make enough carbon-free energy and power available to replace coal generation is essential. And so the views of stakeholders, their insight and knowledge is, to me, important.
But first, to Eraring.
Eraring’s early closure shouldn’t be a surprise, since everyone has been predicting early coal closures for years. Nevertheless, it is a surprise. Ever since the NSW Roadmap was announced this kind of announcement has been coming.
Around 70% of AEMO stakeholders, that is people who know what they are talking about, suggested that the “Step Change” scenario, or more aggressive, was what to budget for.
The August 2025 Eraring closure is 1-3 years ahead of when the Step Change model saw a big reduction in NSW capacity. There will be more pennies to drop. ITK sees Gladstone in Queensland closing early. Loy Yang A and Loy Yang B will be struggling down the track a bit. Yallourn and Liddell have gone.
Assuming NSW executes its energy infrastructure plan perfectly, there should be roughly enough state-based new wind and solar energy to replace Eraring’s 15-16 terrawatt-hours. Still, perfect execution is unlikely.
In regards to power, it’s more problematic. This is where the disadvantage of pumped hydro comes in. It just takes forever to approve and then get built. On the other hand, a battery can be built in a year, or indeed much less, when push really comes to shove.
So that is why the 700MW, two-hour Waratah battery has been invoked out of thin air. No one had heard of this until yesterday, no one knows how it will be financed and it doesn’t really matter. The important thing is that it can surely be built and delivered in three years.
Angus Taylor is as correct as the next amateur when he says a two-hour big battery can’t do the job of a baseload power station. It’s not designed to do that job. A coal station does two, or even three jobs. They are: (i) provide energy; (ii) provide power, and; (iii) provide system services. If you ask the unions, they do a fourth job, which is provide jobs for life – but I’m not going there.
What a renewable energy system does is to split up the jobs of the coal generator. Wind and solar produce as available, they provide the energy; sometimes it’s a lot and sometimes not so much. But on average over a year or two the total output can be predicted to within a couple of percent.
The batteries are there to provide the dispatchable component and to absorb the excess power when there is too much wind and solar. This is not news, except it seems to Angus Taylor, who despite having a good university degree seems to prefer to appear obtuse.
The batteries also provide the system services, inertia, frequency control, fault current, and voltage support. They do that job much better than coal generators.
How much battery power and energy are required relative to how much wind and solar is a complex matter, but there is a rule of thumb which says about 1MW of battery for every 3-4MW of wind and solar. Still, that’s before considering interstate transmission and so on.
Prices will increase
Every time a coal station closes, prices go up, because the closures generally come before the replacement capacity has been built. There is clearly room for doubt as to whether enough new capacity can be built in front of the Eraring closure, so that prices will be higher than if the closure came in, say, five years.
Equally, though, that price rise most likely won’t be visible, since there are no public traded futures for FY26. Lower capacity from Eraring over the next couple of years, given its lack of firm coal supply, had likely been factored in. But even so, a further premium may be added.
The good news is that renewable developers, already champing at the bit, will get a huge confidence boost from this. They really will be needed. But will they get the transmission they need?
ISP: What stakeholders are saying
ITK has been reviewing public submissions to the draft 2022 ISP available at Draft ISP consultation submissions.
ITK feels well prepared to assess the quality of these submissions, since most of January was spent reviewing the ISP output models and doing some reading.
ITK is interested in the quality of the submissions because they indicate, (i) how seriously the submitter takes the ISP and, (ii) whether the submitter seems to be on top of the topic or not.
In a sense, actions speak louder than words, and Origin Energy’s decision to accelerate the closure of Eraring demonstrates what many of us have known for years. Coal generation is unviable when it has to ramp too hard. Eraring was close to the most flexible plant in the NEM, that is it has the most ramp-ability, but it also is old and is exposed to coal spot prices.
I, personally, thought the AGL submission was a waste of time for both the reader and the writer(s) and that the EnergyAustralia submission was good. The Origin submission missed the point, but Origin management clearly did not. As usual the Neoen submission was on point.
Leaving aside various “good” and “bad” comments the main messages coming through to me were:
1. An over arching sense that some stakeholders, even those making submissions, just hadn’t spent enough time reading the document, or looking at the output spreadsheets. Many things, like earlier than modelled coal closures (therefore accelerate Humelink), social license issues and how they are managed by focusing on REZs, but that there was still a lot to do, and so on were all discussed in the ISP, sometimes at length. The good news is that after four years of work the ISP is now starting to be taken very seriously by not just some stakeholders but the general public. I’d also recommend all stakeholders who haven’t done so – and this most of all means you, Snowy – take the time to read appendix 4, which discusses “Dunkelflautes”, “Curtailment” under the banner of System operability.
2. A (much) greater degree of urgency is needed by policy makers. Its exceedingly obvious that transmission is more likely to be delayed than be completed ahead of time. As Tom Geiser points out in the Neoen submission, if a coal generator blows up, the RIT-T test only considers the change in fuel costs between, say, coal and gas, whereas in reality the electricity price paid by consumers might double. Queensland in the past 12 months is the obvious example. Eraring’s closure announcement brings forward the absolute necessity of getting replacement supply in place ahead of the closure. Where Eraring goes Vales Point B will follow. The ISP models out an enormous increase from the already world-leading distributed energy resource. Rooftop is already, already 30% of supply at lunchtime in Summer. In 10 years’ time it will likely be double that and from a sector that neither knows the spot price or for the most part cares.
The ISP also models distributed storage to be about 60% of the total storage MW right out to 2050, so the fans of “local generation” have been given full credit, but EVEN SO, LOTS OF TRANSMISSION IS REQUIRED AND IT’S REQUIRED SOON. Although much of Snowy’s submission will rightly be seen as self-serving posturing, its call for accelerating the transmission build needs to be seriously considered.
3. A social license modelling “parameter” might be useful. A disadvantage of REZs is potentially their impact on the local community. Too much transmission may be quite unwelcome, as is transmission that goes through National Parks. Perhaps transmission easement compensation could be considered. These are the sort of issues that the federal government could assist with, if it was so minded. The kind of “lets make it happen” attitude that gets things done in war time, as opposed to the “lets put sand in the gears” approach of the current federal government. ITK believes that getting social license to develop and reconfigure the grid is critical. The social license issue is considered in some interesting detail by Walcha Energy and Star of the South. We look at their points raised below.
4. One of the more interesting things in the ISP new resource modelling, ie how much wind to build how much storage and what duration, was that no offshore wind was ever selected in any of the scenarios. That’s because its costs didn’t exceed its benefits according to the costs and benefits inputs adopted for the ISP modelling. Naturally offshore wind developers were unhappy about this. Just as with the Eraring or now Waratah battery, the ISP does not say what developers can or can’t, will or won’t do, the ISP only takes a set of assumptions and models a set of resources that produces lowest cost outcomes that meet the required constraints (policy, reliability, and carbon). Still Star of the South, one of, if not the leading offshore wind would be developer in Australia produced some great graphs.
5.ITK has already noted that Queensland’s plans for how or if it will meet the QRET target are a mystery to most of us. A June quarter energy plan from the Minister is may shed some light on matters. Still Powerlink’s submission points the way to early closure of Gladstone power station. Again this will be no surprise to anyone, but the evidence is mounting. The Powerlink submission talks about 2050 but in my opinion that is code for a lot sooner.
1. Some of Snowy’s comments were over the top
Snowy, one of the largest generators in Australia, was hostile to the modelling outcomes from the ISP. Consider this statement made by Snowy:
“The 620GWh of storage assumed by 2050 presumably includes Snowy 2.0’s 350GWh; therefore the NEM’s storage needs are assumed to be fulfilled by an additional 270GWh of storage, spread over 43GW of generation. This represents an average of 6.3 hours of storage (excluding Snowy 2.0) across the NEM. The probability that this will be enough to last even a single week of low wind or solar output, by 2050, is zero.”
To start with the output spreadsheets released by AEMO clearly break out Snowy are a line item, so there is no “presumably” about it.
AEMO devoted Appendix 4 of the draft ISP “System Operability” to discussing “residual demand” and the like. This appendix provided graphs of how storage will charge and discharge by month and seasonality. Here’s the unserved energy analysis figure taken from that appendix. Yes its only to 2040.
A summary of the ISP modelling process, as presented by AEMO, is shown at the end of this note.
Unfortunately the stridency adopted by Snowy takes away from what ever useful points Snowy might have made. It’s unfortunate that Snowy takes such an approach because it has a strong team well capable of making a useful contribution.
All that said, where I personally would like to understand more the strength of Snowy’s claims around the need to build VNI West and Humelink in order for Project EnergyConnect to reach Sydney and Melbourne:
“ProjectEnergyConnect will be stranded until VNI West and HumeLink are built, because it terminates in Wagga, and will be useless if it cannot connect to Melbourne and Sydney (which it cannot do without VNI West/HumeLink).”
Humelink and VNI West should have certainty and not be burdened with “decisions rules.”
Snowy states:
“AEMO’s approach for these two projects (ie. ‘staged with decision rules’) creates open-ended opportunities for transmission planners to drastically alter the timing of project delivery. This uncertainty significantly dilutes the value of the ISP as a planning instrument and its utility to energy market investors. For HumeLink, AEMO’s ISP 2024 could change, yet again, the date for the project under the decision rules, 2 years out from its supposed completion.”
A clear view about the necessity for Humelink and VNI West is important. Humelink is controversial and expensive but in my view probably necessary. If it is, there should be complete clarity about that necessity.
Social license
Several submissions note the risk that lack of social license poses to ISP execution. Examples include Star of the South, Walcha Energy and Queensland Conservation Coucil, (one of only three Queensland-based stakeholders to submit, so far).
The ISP is clearly written with social license at front of mind. The documentation states that concentrating wind and solar into REZs reduces the overall impact. Except, I might add, on the people that live in the proposed REZs. Social license is a problem. No doubt about it.
Star of the South states that the Optimal Development Path [ODP] builds 10,000 km of new transmission.
Star of the South correctly observes that transmission building costs have dramatically increased over the past couple of years and Social License issues are a big part of this.
Of course the underlying argument of Star of the South is that if social license risk and costs were properly allowed for in the ISP then lots of offshore wind would be built rather than onshore wind. The ISP does not stop other projects being built if developers can get them done.
Looking at the impact of social license, which includes Biodiversity, on costs; for instance, Humelink cost has increased from $1.35 billion to $3.3 billion of which over $1 billion is “Biodiversity” payments, due to the transmission route through national parks.
Project EnergyConnect has had costs increase from $1.5 billion to $2.3 billion partly due to social license issues.
No farmer wants anything on their land unless they are paid for it and even then is reluctant. In general, my experience growing up in the country is that the rural sector is conservative and resistant to change.
What damage other than visual, one might ask, does a couple of transmission towers do to a farm? Does it reduce the livestock carrying capacity? No. Does it reduce the plantable area materially? No. Do the farmers need electricity? Yes. Will they be negatively impacted by coal generation impact on global warming? Yes. Will transmission be resisted by farmers anyway. Yes.
Star of the South’s business case depends partly on a greater social license score and partly on the fact that offshore wind is better correlated with peak demand in Victoria, particularly peak demand on hot days when in general onshore wind performs badly.
If I was to give out a second gold star award it would be to Walcha Energy’s submission, which went through some detailed considerations for transmission in the New England region and raised, as did a series of other submissions, the thorny issue of social license.
Social license goes to the concentration of renewable energy in a particular area, it goes to the impact of having lots of transmission connections in one particular land point, it goes to compensation for land owners for transmission easements, and so on.
In my opinion, it goes to the building of a national consensus that the benefits of this transformation outweigh the costs. Building more transmission, some of which goes through national parks, building more wind and solar, perhaps building more dams for pumped hydro on both private and public land, changing the way electricity is billed, moving away from coal and gas exports; all of this is worth doing to provide consumers big and small with lowest cost electricity and a steep reduction in carbon emissions.
ISP modelling process
The ISP modelling process is fairly standard and consists of about four steps. Last year, Energy Insiders discussed with Paul Denholm, from NREL in the USA, the general modelling process, which typically uses PLEXOS for the time sequential model, as it was employed by NREL in the USA.
But actually, the ISP documentation made it clear to most, if not to Snowy Hydro. Previous EnergyInsiders podcasts with Alex Wonhas have made it clear that AEMO has devoted considerable resources (people, consultant inputs, etc) to the modelling which has now been refined over four years. None of that makes it correct, because it is just a model with all the problems all models have, but it’s not to be dismissed for lack of care or effort.
- The capacity outlook model projects the generation, transmission and dispatch outcomes in each scenario, seeking to optimise capital and operational costs.
- The time-sequential model then optimises electricity dispatch for every hourly or half-hourly interval.
- The engineering assessment tests and validates the capacity outlook and time-sequential outcomes using power system security assessments to ensure that investments are aligned and robust.
- The gas supply model may then validate any assumptions on gas pipeline and field developments.
- Finally, the cost-benefit analyses test each individual scenario and development plan, to determine the ODP and test its robustness (see Part C).
Various submissions call for more examples of what Mark Collette calls “edge case” analysis; ie, how the modelled system would cope with wind and solar droughts. No doubt more analysis could always be done and more results presented, but AEMO did present some detailed discussion of a hypothetical week in 2039:
