Hacker News Reader

Most industries that deal in refining or other heat-intensive processes are nearly hard to impossible to pause or stop. My understanding is that glassmaking plants, for instance, will literally solidify if they ever stop operation--making it an extremely rare occurrence. I wouldn't be surprised if aluminum refining is the same way--fighting entropy keeping things hot is a losing battle, so not very dispatchable.

I confirm that an aluminium refinery is not something you stop lightly. I remember something about taking two weeks to restart? Maybe more. Source, had family working at a large northern plant.

Being able to restart it at all makes it better than a lot of industries. Many of these liquid-metal affairs depend on convection and inertia to keep the metal molten, and if it solidifies inside a pipe—well, then you don't have a pipe anymore; you have a bimetal pole.

Or ceramic cladding for an iron pole, depending.

Either way it's a rebuild, not a restart.

Totally, like magnetic induction pumps that pump molten iron, once the iron solidifies its a composite tube.

> well, then you don't have a pipe anymore; you have a bimetal pole.

I love that way of describing it.

That is true about float glass plants but for a different, very interesting reason.

Float glass plants work by literally floating a thin sheet of molten glass on top of a giant tank of liquid tin. They have a swimming pool of liquid tin, float molten glass on top, and then push it along the tank length wise.

The glass and tin is then gradually cooled until the glass is solidified. It's then cut into pieces, cooled and stacked.

If the thing suddenly stops, or there is a hiccup of some kind, the thermal expansion of the glass, along with its extreme hardness (lack if strength) it will just shatter the whole mile long sheet of glass. My understanding is that it takes the better part of a year to recover from something like this.

In aluminum, its basically electroplating. They basically electroplate the aluminum out of the ore onto the ingots. My understanding is that takes a few weeks to a month to recover from a similar incident.

If I remember right, you can stop the electricity to an aluminum smelting plant. But only for a while.

Right. The wall has to continue to be cooled -- if it reaches the temperature of the melt, it's ruined -- so there's always some heat loss. And if the electrolyte freezes you're also in trouble.

Glass melting furnace never stops after it's started. If it cools down it's basically destroyed so yeah, they run uninterrupted for years. Any maintenance or fixes need to account for that.

How do they account for power outages? Huge generators? Or is it gas fired?

From what I've read its gas

There are electrically heated glass furnaces. They work by immersing electrodes in the molten glass, which is electrically conductive, if somewhat resistive. I imagine they could use gas for startup though. The use I saw for this was in making fiberglass from recycled glass cullet.

Most aluminum for European countries comes from recycled sources. It's way less energy to just smelt old soda cans.

About 75% of all aluminum ever produced is currently is use, because we're so good at recycling it.

We need to get lithium from 5 to 75%

https://en.m.wikipedia.org/wiki/Aluminium_recycling#:~:text=....

> We need to get lithium from 5 to 75%

As long as you store your old batteries (and other lithium containing gadgets), I shouldn't make too much of a difference if you recycle them now or in ten years?

Where and how do we store them safely?

In landfills?

This brings up an aside I've been wondering about for years, when are we going to start mining our old landfills?

We just need more people to accidentally toss their bitcoin keys in the trash! https://edition.cnn.com/2021/01/15/uk/bitcoin-trash-landfill...

> Where and how do we store them safely?

You could probably just stick them in a warehouse? Or a specialised landfill that only keeps electronics or batteries. (Most just to keep the overall volume down, so it's more economical to make whatever special arrangements you need to keep everything safe enough.)

Well if we're comparing the recycling of old soda cans to lithium batteries... I had to throw a few old phones away recently. I Googled and found a mobile phone store near me that accepted e-waste for free, so I went a bit out of my way and took the phones there. I was not, however, given a bit of cash for every phone I returned..!

Soda can deposits are far from universal an do not result in free money... they simply refund the deposit that was paid up.

But they do make sure more bottles amd cans get recycled. In practice it creates a whole ecosystem of people collecting amd cashing them in. If we didn't have the deposit, the government would have to pay for that cleanup

Batteries in Norway have a very high recycling rate and large batteries like car starter batteries attract a specific tax (miljø gebyr, environment charge) that assists in funding the recycling operations. Everyone who sells batteries has to be a member of an approved recycling group and be willing to accept for recycling batteries of the same general type even if they did not sell them. So you can dump your old phone batteries in a box at any supermarket.

Indeed, aluminum refining is a relatively large industry in Iceland for this very reason. Geothermal energy production in excess of what people need means you can devote a bunch to industry cheaply. If Nordural’s website is to be believed, they use 25% of all electricity generated in Iceland: https://nordural.is/en/

>It is a very useful metal, and if it is electricity intensive to define, I wonder if it would fit the use of excess renewable electricity.

No, absolutely not. Aluminum refining, like all industrial processes at scale, takes a long time to get going and needs to run (hence be fed electricity) continuously, not just whenever the wind feels like blowing.

While it is an excellent business opportunity for electricity that is cheap because it is renewable- and why Canada produces more refined aluminum than any other Western country thanks to its electric generation being overwhelmingly renewable for the entire time it's even had a grid- it's not something you can turn on and off whenever you want and not a good candidate to burn off excess power ("just not using as much naturally-pumped stored power" is not an "excess" of power by definition).

If you have a predictable pricing regime can you vary the power load on a diurnal cycle? Say, drawing more at noon and less at 7pm each day? If so it could still be useful to shape the demand curve (ie use power when solar is cheapest so that meeting the evening peak doesn’t require as much overprovisioning) even though this doesn’t help with inter-day variability of supply.

The problem I’ve always seen with this sort of plan is that the refinery is extremely expensive and so you don’t want it idle; it’s cheaper to overbuild solar than to turn off your refinery for 25% of the day. But interested if this applies here, maybe the energy-intensive bits can be made cheaply?

These kinds of refineries take *weeks* to start up. There is no idle. Period.

How large is the tolerance though? Would it be feasible to heat things up a little more from say 9:00 - 17:00 and using more electricity so that you don't need to heat (as much) during peak demand-supply imbalance? Even a couple of hours can help with the duck curve.

Looks like aluminium smelter demand flexibility is already in use, but it doesn't rely on process temperature variation:

https://enpot.com/

https://aluminiuminsider.com/trimet-aluminium-betting-enpots...

Ah, hacker news, from “ No, absolutely not.” to already being done in 3 comments.

To be fair, first operational in 2019. Variable speed cooling fans, for heat loss regulation, are bleeding edge tech, in the world of aluminum.

I just wish people would be more open minded and curious by default.

They only are if they're incentivized: most ignore the rewards, those you admire who don't ignore self-motivated open-minded-ness and curious-ness, they must have already been rewarded previously for having these traits, and so they continue to express them.

Interesting, thanks!

> The “virtual battery” concept relies on installing adjustable heat exchangers that can maintain the energy balance in each electrolysis cell irrespective of shifting power inputs. Since aluminium production requires a constant energy supply, any fluctuation could have heavy consequences for the molten metal. The technology also ensures that grid power fluctuations do not affect the magnetic fields in the electrolysis cells.

This seems to be solving the even-harder problem of handling on-demand/dynamic fluctuations rather than planning for a diurnal cycle. I wonder if there is scope to use a heat reservoir (molten salt or similar) as a buffer for these sorts of process; basically if you need to dump heat into a process perhaps you can shift the heat production but leave the process itself unchanged.

Amusingly I was reading the wrong Wikipedia last night and it’s mentioned right there too:

> Particularly in Australia these smelters are used to control electrical network demand, and as a result power is supplied to the smelter at a very low price. However power must not be interrupted for more than 4–5 hours, since the pots have to be repaired at significant cost if the liquid metal solidifies.

https://en.m.wikipedia.org/wiki/Aluminium_smelting

Here are some pdfs:

How it works: https://enpot.com/assets/pdfs/enpot-how-it-works.pdf

Energy modulation: https://enpot.com/assets/pdfs/Enpot-Energy-Modulation-of-Alu...

First was operational since 2019. If you push past the “virtual battery” marketing nonsense, a summary of the tech is… they modulate the cooling fans for closed loop thermal regulation, regardless of processing rate (+/- 20% long, 30% short term).

Talk about low hanging fruit.

This is particularly interesting to me because we currently have some unanswered questions about how viable renewable energy is for ~100% of generation.

As we get to higher total renewable contribution, presumably we’d see a more dramatic price difference in energy at different time of day.

This tech suggests to me that, as you say, there is low-hanging fruit that could be harvested, and which perhaps isn’t cost-effective yet with a small diurnal energy cost-delta, but with a higher peak-to-trough cost difference might become viable to extract.

It’s worth noting here that seasonal variations seem harder to deal with, and varying the energy intensity of industrial processes doesn’t seem helpful for that issue due to capex/utilization concerns.

And for diurnal fluctuations, batteries are actually not too expensive these days.

Efficiency for industrial chemistry is massively dependent on the processes being continuous. The processes often don’t work, or work poorly, until you reach a stable equilibrium, which can take many hours. An enormous amount of effort is expended to ensure this in real industrial processes. Otherwise, you are just burning resources for negligible output.

The idea that we’ll do industrial chemistry with intermittent energy surpluses is unrealistic unless we are okay with yield per unit of energy being very poor relative to continuous processes.

Exactly. Chemistry can be done that way, in batches. Chemical engineering cannot, it's not a batch process, it's all rate stuff.

Aluminium smelters already adjust their electricity consumption throughout the day to take advantage of changing power prices.

Even if you ignore the difficulties of start / stop the factories and imagine we have a new manufacturing process that can be restarted immediately, having idle factories and employees sitting around doing nothing for half a day would be a huge money sink

No way, you can't run a factory for 1 hours a day. Aluminum is only "solidified electricity" under normal circumstances. Not when capex is 24x because the normal price of electricity in the UK for most of the day is ~7x what you'd pay in Texas (which isn't even that cheap globally)

But don't the crypto-bros have dibs on all the excess renewable energy so they can Get Rich Quick?

When I was a kid there was an aluminium smelter in Anglesey, North Wales that ran off the base load of the local nuclear power station

You're not thinking big enough.

Have a read about the https://en.wikipedia.org/wiki/Akosombo_Dam

Then watch this documentary on the devastating effects: https://en.wikipedia.org/wiki/Pandora%27s_Box_(British_TV_se...

As Wikipedia says about the dam:

"The Ghana government was compelled, by contract, to pay for over 50% of the cost of Akosombo's construction, but the country was allowed only 20% of the power generated."

Having been to Ghana and to Lake Volta & the Akosombo dam in particular, I'm somewhat torn:

- Yes, it has had a severe impact on the ecosystem, and also on agriculture.

- Yes, Ghana itself only gets about 20% of the power generated by the dam (or apparently a bit more in recent times).

- OTOH, the dam gives Ghana access to at least some power – much more in fact than what's available to any of its neighboring countries. I have heard people say that this is one of the major reasons for Ghana's relative economic & political stability.

> I wonder if it would fit the use of excess renewable electricity.

It's exactly the opposite. The pots[1] in the smelter get their lifetime reduced if they have to be restarted, and if restarted multiple times that can be quite significant[2]. They can survive without power for a few hours, but they freeze over after a day or so which leads to the most damage.

As such they really want stable and cheap electricity, like hydro.

[1]: https://en.wikipedia.org/wiki/Aluminium_smelting#Layout_of_a...

[2]: https://aluminiumtoday.com/content-images/news/Oyeweb.pdf

I recall a big part of Iceland’s economy is processing aluminium. Basically producing as much as, if not more than, the US.

https://www.nytimes.com/2017/07/01/us/politics/american-comp...

Not Aluminum. However iron can be reduced in a low temperature aqueous cell. I can't think of why that couldn't be done intermittently. Takes about 3-4kwh per kg of iron produced. That probably works out to $150-250/ton.

You can also run electric furnaces intermittently because they are inherently a batch process.

You can also make iron by direct reduction with hydrogen. IIUC, this is actually more energy efficient (with green hydrogen) than direct electrolytic reduction. And electrolysers for hydrogen are getting cheap enough that running them intermittently will make sense.

Iceland has incredible hydropower resources, and aluminum is the primary way they export their excess energy: https://en.wikipedia.org/wiki/Economy_of_Iceland#Aluminium

Sibling comment mentions, correctly, that these things need to be running at full utilization for a long time to make sense. So, you'd need batteries to smooth wind or solar, which is still economical (and will become increasingly so).

That article snippet doesn't fully do it justice. They sell electricity to aluminum smelters at 1/4th the price of the EU, and a surprisingly large portion of their total energy consumption is industrial.

Iceland has done it correctly, though, in that in order to tap into their natural resources, you need to be an Icelandic company investing in Iceland. Some want to build a cable to Iceland to help the European electricity market, even though it's never been done at that length before. Some in Iceland are (rightly!) worried that would increase demand for hydropower and further industrialize Iceland, etc.