At the end of last week, I came across a SolarQuotes article about Snowy Hydro 2.0 being more expensive than batteries. Snowy Hydro 2.0 is a pumped hydro project in Australia (New South Wales and Victoria) and is currently under construction. The brunt of the article is that pumped hydro is too expensive compared to (grid sized) batteries and the plea is made to halt the project in favor of batteries.
That was new to me. As far as I know, pumped hydro is the cheapest way of dispatchable backup in order to counter intermittency and definitely cheaper than batteries. Yet, the author of this article argues that it is the exact opposite.
The name of the author of the article, Ronald Brakels, rings a bell. Not even a year ago, I wrote a post on his claim that South Australia has the second cheapest electricity in Australia, despite it having the most expensive electricity. He did this by applying two neat tricks, sneakingly morphing South Australia from by far the most expensive to the “second cheapest”.
With that in the back of my mind, I expected some trick(s) to be performed in this one too.
The article starts with a general overview of the Snowy Hydro 2.0 project. It is an expansion of the current Snowy Hydro project, connecting two existing dams (Tantangara and Talbingo). This expansion would allow for the closure of two gigawatts of coal capacity. I would expect that, for this reason alone, he would like the project, but that isn’t the case. Brakels not only objects to it from a financial standpoint, but it is clear from the intro of the article that he doesn’t like the project at all.
Brakels starts off by badmouthing the CEO of the project . He calls him “incompetent” because he doesn’t know how long the connecting tunnel between the two dams is. He found on the Snowy Hydro 2.0 website that the two dams are connected with 27 km of tunnels. This is the quote that he used in the article:
The project involves linking two existing dams, Tantangara and Talbingo, through 27km of tunnels and building a new underground power station.
Brakels now claims that this is not true. He makes the claim that in reality it is only 24 km and that the 27 km figure is the total length of all tunnels, including a 2.6 km exploratory tunnel. Even if that would be true, it is no reason to call someone incompetent.
This leaves us with the question how long those tunnels are, 24 or 27 km? He doesn’t give a link to where he found that the length of the tunnels is only 24 km. There is a video following that paragraph, but it doesn’t mention the length of the tunnel. It only mentions that an exploratory tunnel will be excavated, but not how long it will be or how long the tunnels connecting the two dams will be. Unfortunately, I didn’t find the details of the 27 km length of the tunnels on the Snowy Hydro 2.0 either.
I did however find an interview of the CEO of the company that will excavate those tunnels in Engineering News-Record. This is how the length of the tunnel was detailed (my emphasis):
A roughly 17-km-long headrace tunnel from the high level Tantangara reservoir will lead to a 550 m drop into a planned power house, which will be excavated about 1 km below ground. From the powerhouse and transformer halls, both about 190 m long, water will flow though a roughly 10 km tailrace tunnel to Talbingo.
They also have an illustration with a cross-section of the tunnels (my annotations in red font):
Unless the CEO of this company also doesn’t know the length of those tunnels, roughly 17 km plus 0.55 km plus roughly 10 km equals roughly 27 km, so the information provided by the Snowy Hydro 2.0 website is correct.
Later in the article, he objects that the transmission upgrade is not included in the price of the Snowy Hydro 2.0 project. I am not sure why that is even controversial. I have never seen a project where this was the case, whether it is solar, wind or conventional power sources. I even agree with him on that one. I often make the objection that solar and wind projects don’t include the transmission upgrade costs, so I am definitely in favor of including the transmission upgrade cost for both alternative and conventional power projects.
That would however be a disadvantage for solar/wind or other intermittent power sources. Solar and wind energy are low density power sources and therefor need large installations requiring large surfaces and their intermittency results in the need for the strengthening of the grid. The transmission upgrade cost for solar and wind projects will therefor intrinsically be higher than the transmission upgrade of projects at a central location. In this case, the Snowy Hydro 2.0 is just the expansion of an existing project.
Brakels also objects to the stated maximum capacity of the Snowy Hydro 2.0 project. This is the quote he uses from the Snowy Hydro 2.0 website:
Snowy 2.0 will provide an additional 2,000 megawatts of dispatchable, on-demand generating capacity and approximately 350,000 megawatt hours of large-scale storage to the National Electricity Market. To provide context, this is enough energy to power three million homes over the course of a week.
He objects to the 350,000 MWh (= 350 GWh) figure in the first sentence. Brakels claims that this number is not the full story. He accepts that, when the upper reservoir is full, there is enough water to provide that amount of energy, but not during normal operation. Therefor the working amount of storage will be much less. He supports his claim by stating that the upper reservoir is much larger than the lower reservoir, meaning that all the water of a full upper reservoir could not fit in the lower reservoir. Also, the lower reservoir would be mostly full anyway, so under normal operation this 350 GWh could never happen. According to Brakels, 40 GWh would be a better estimate of the storage capacity.
He links to this paper for more details. This paper indeed also accepted that Snowy 2.0 might have an energy reserve of up to 350 GWh, but it contends that it would rarely, if ever, be called upon to deliver that amount of energy at once. It then would also take a long time before the reservoir is refilled and able to generating electricity again. According to their calculations, the recyclable storage capacity could vary between about 40 GWh and 200 GWh, depending on the operation of other reservoirs. The storage capacity could go beyond 200 GWh, but then water would be lost by discharging to a lower reservoir and is therefor not recyclable.
In his calculation, Brakels used the (original) Hornsdale battery as a comparison. He however uses two different standards when it comes to the storage capacity. He stated the storage capacity of the Hornsdale battery as 129 MWh, this is the maximum energy that the original Hornsdale battery could store. If the Hornsdale battery would drain all its charge at its maximum capacity (100 MW), then it would take ±75 minutes before it would be completely empty. However, it would rarely, if ever, be called upon to deliver this amount of energy at once. For example, draining a Lithium-ion battery completely empty is a really bad idea. It will shorten the lifespan of the battery and the battery will need to be recharged before it can discharge electricity to the grid again. If Brakels wanted an apples-to-apples comparison between the two, then he should also adjust the maximum storage capacity of the Hornsdale battery to its, ahem, normal operation, as he did with Snowy Hydro 2.0.
Until that point in the article, it is not clear why Brakels needs the storage capacity of the Snowy Hydro 2.0 lowered. That will become clear in his calculations, but that will be for the next post.