Hacker News Reader

Comment by Gibbon1 | Hacker News Reader

One thing not touched on in the article is the marginal cost of wind and solar is near zero. Meaning when supply exceeds normal demand the price falls to near zero. This usually presented as a problem, but for some businesses this is potentially a big opportunity. If you consider a electric furnace processing scrap steel can easily use 10-100 MW of power for thirty minutes at a time. Potentially zero margin electricity can be used to decarbonize other industries as well.

View on HN

It's definitely an opportunity, but the price at peak production periods is near zero precisely because nobody has yet figured out how to soak up these kinds of transient production peaks effectively, with sufficiently fast ramp-up and scale. It won't stay at zero if some big users figure out how to use that energy profitably. Though that wouldn't be bad overall. If some large users did figure out how to take advantage of these periods, it'd be a step towards smoothing out both grid pricing and usage.

Cryptocurrency miners have the fast ramp-up and scale required.

But they're profitable at high energy demand times too, so there's no incentive for them to run only during low demand.

Cryptocurrent miners have a strong incentive to locate in places where the $ per kWh rate is consistently low 24x7x365, not just for peaks. Such as near a hydroelectric dam in east Wenatchee WA. Mining also requires very little in the way of network connectivity, maybe 10Mbps at most, so it can be located just about anywhere.

Relevant article from this week regarding miner activity in that area:

https://news.ycombinator.com/item?id=16551848

It appears empirically that they don't. If they did, there wouldn't still be peak-production periods when energy was near zero cost, since they would've arbitraged that out.

Does that run the risk of being damaging to the currency network -- could transient spikes of mining push up the difficulty and thus affect functionality once the energy price goes back up?

You don't even need energy spikes for that, it already happens when hashrate is mobile between different networks.

That's actually what made me think of it (Nicehash pushing up the difficulty on the smaller coins); I just wasn't sure if something more cyclical would have the same effect or be easier to counter.

Cryptocurrency mining isn't enetgy constrained. Right now it it is capital constrained. Your primary costs are fixed upfront purchases of ASICs and GPUs.

Is there really nothing that we could do with this excess energy? Couldn't we purify water or run simulations or something similar?

It's not economically effective to run a supercomputer only during peak times.

Some Nordic countries that have large district-heating systems have started experimenting with dumping the excess energy into those.

Electricity to heat is not that efficient; normally district heating systems run off of the waste heat of electricity generation. Are you talking about an example where the waste heat isn't enough to power the district heat?

Actual energy to heat is only useful in a true excess energy availability situation, but usually you'd start building up heat storage to decouple heat availability from electricity demand, which would otherwise be a problem with cogen plants. And heat storage is so much cheaper and efficient than electrical storage, so you want to run your cogen in sync with electricity demand. Once the heat storage is there, it becomes an absolute no-brainer to also use it when you have excess electricity to dump (as electric heating is 100% efficient, it's the production of electricity in thermal plants that makes electric the worst form of heating).

We use heat pumps for district heating in Drammen, Norway. Run this through your favourite translation engine to read all about it: http://www.energi.no/brakeroya-varmesentral-apnet

Pumped storage hydroelectric plants seem the obvious target to me https://en.m.wikipedia.org/wiki/Pumped-storage_hydroelectric...

Desalination does seem like a good target. I think synthetic hydrocarbon production will be a good place to dump excess too. Pharmaceuticals and plastics aren't going anywhere, and I bet they'd appreciate a cleaner starting point than bitumen.

Electric cars have the obvious potential to absorb peaks. With range on recent models far exceeding typical daily use, it becomes increasingly easy to make charging dependent on supply.

If the "hydrogen economy" people ever figure out a way to make the equipment and fuel cells economical, cracking hydrogen from water at times when power is $0.00 per kWh could be one method. Takes a massive amount of amperage. Store it in tanks and use as needed through giant fuel cells.

Energy losses from this scenario are high enough that there are probably other usage scenarios that would beat it. It's certainly the case that there are already many peak-storage solutions deployed and none of them involve hydrogen.

there are already many peak-storage solutions deployed and none of them involve hydrogen

That seems like an overstatement. There has definitely been considerable research into using hydrogen as "peak-storage solution", and I think some approaches are actually being deployed. For example, here's an NREL story from last year: https://www.nrel.gov/continuum/energy_integration/hydrogen.h...

The article agrees with you that the current inefficiencies of the round-trip to the grid are large, but claims that generating hydrogen as a vehicular fuel using excess solar and wind power does make sense. So if the parent was envisioning those "giant fuel cells" as being mobile (or located somewhere really remote), then he's not that far off.

OK, so what % of peak storage solution is hydrogen?

That’s a demonstration project and to the extent that it’s “deployed”, so are tokamaks.

Hydrogen may be feasible for peak balancing but it seems correct to say that it is not currently used this way.

> Store it in tanks and use as needed through giant fuel cells.

Or just push it into the gas pipeline grid, it provides storage capacity and has the applications already waiting (the amount of hydrogen allowed in the final mix varies between regions, but we are a long shot from displacing natural gas by that fraction, worst case you'd have to add in a slightly lossy methanation step).

It works, a well developed gas grid already has all the storage capacity you could ever need and you wouldn't have to scrap your existing gas peakers even after scaling to 100% renewable. Perfect migration paths tend to be disappointingly unspectacular.

It's kind of entertaining that you mostly said the same thing as the person you replied to, but apparently don't realize it.

Why do you see it as the same? Gibbon1 points out that there is a current opportunity for a business that can take advantage of sporadically available almost free power, and _delirium extends this to say that paradoxically this opportunity will disappear as soon as there is a business that can use this power. Gibbon1 may or may not agree that the opportunity will disappear once it's being utilized, but this wasn't mentioned in his comment. So what is _delirium failing to realize?

Random thought, but I wonder why solar panel manufacturering plants aren't powered by solar panels.

they all are, but indirectly. I challenge you to find any manufacturing plant that's powered by a direct energy source instead of grid power - it's simply not efficient to manage demand peaks and valleys yourself when there's a huge amount of infrastructure already dedicated to doing exactly that.

Most large aluminum plants generate their own power.

This isn't strictly true, but isn't strictly false. Power plants are often built to supply power to aluminum smelters, like in New Zealand [1], Oman [2], and Iceland [3], but they're usually connected to the grid in the end. Iceland has a grid circling the island [4], Oman is upgrading its grid and building out interconnects [5][6], and in New Zealand, Manapouri went from exclusively supplying the Tiwai Point smelter to being a power station of the South Island grid.

[1] https://en.wikipedia.org/wiki/Manapouri_Power_Station [2] http://www.powerengineeringint.com/articles/print/volume-17/... [3] https://en.wikipedia.org/wiki/K%C3%A1rahnj%C3%BAkar_Hydropow... [4] https://nea.is/hydro-power/electric-power/transmission/ [5] https://www.omangrid.com/en/Pages/Existing-Transmission-Syst... [6] https://businessgateways.com/news/2017/07/30/Oman-plans-to-i...

The generators are connected to the grid, both for buying & selling but it is absolutely not true they don’t manage the ebbs & flows. Quit the opposite it’s a major part of their business

Until someone puts in a tie line from Roxburgh to Benmore Manapouri's power is really only available to the lower third of the South Island. Mind you the sooner we shut down the smelter and redeploy that power to locally owned industry the better

Most large aluminum plants are located very near giant hydroelectric dams, but I do not know of a single one that actually owns the dam.

http://www.alcoa.com/global/en/what-we-do/energy/default.asp

Alcoa has a whole division for energy generation. Their energy plants are collocated with aluminum plants, most like the Warrick operation are coal but they do have a hydro plant.

The Lochaber Hydro Electric scheme in Scotland was privately built and is still privately owned and feeds water to power the aluminium smelter in Fort William.

Paper mills tend to. They have lots of waste wood and need heat for their process.

The one near here is one of the largest power plants in the region.

Interestingly, I live in rural Japan and went on a trip with some of my neighbours. One of them brought her son along who is a buyer of natural gas for Fuji film. I asked how they are set up. He told me that they generate all of their electricity onsite from natural gas generators. It seems that they also need the waste heat for some other processes so it makes sense for them. I asked him about solar and wind because I know that both are fairly good bets here (more than 200 days of sun per year at 34 degrees latitude, coastal position with near constant moderate wind). He said that because of the waste heat that they need, it's still much cheaper for them to go with natural gas and figured that it would continue for some time (though... I suppose he might be biased given his job ;-) ). Unfortunately my Japanese isn't quite good enough to understand what they were using the waste heat for. Fuji film makes a lot more than film -- they do a lot of electronics too, so I'm not quite sure what they are doing. I also know a few people who work for automobile parts manufacturers (suppliers for the big car companies here) and most of them also use primarily natural gas, apparently (which surprised me greatly). I did notice that a few years ago quite a few factories in the area have covered their roofs with solar panels, so I think it's moving slightly.

I suspect that in many parts of the US, the national grid is so cheap that it makes sense to use it. But at least here in Japan it seems to be common to generate at least some of the energy yourself.

Note that he's not using a carbon price. It's definitely the case that using waste heat from a thermal plant is a big plus. But using natural gas in Japan, which imports all of its natural gas under high-priced long-term contracts, is not cheap.

Energy wise, nothing is cheap, which is why I was quite surprised that Fuji Film weren't already moving over to renewables. Japan basically has no other energy resources. Natural gas, oil, coal, uranium: basically whatever small supply they had is already gone. I suspect the reason for factories using natural gas is because it is on demand and the capital cost for the generators is relatively small. Probably long term Japan will use geothermal for baseload, but a lot depends on how the recent test plants perform and especially how they are perceived by the public.

Solar energy is land-intensive, and Japan is not exactly known for its wide-open spaces.

But presumably marginal electricity is priced at 0 because there aren't many buyers on the market.

As more people figure out how to use bursts of electricity, they'll become buyers and prices may rise.

They are frequently built using feed-in tariffs, which guarantees that electricity will have a positive fixed price. Otherwise they're basically impossible to finance because as you say the generation marginal cost is zero so they'd just have to accept any price otherwise and lose money for the investors.

https://en.wikipedia.org/wiki/Feed-in_tariff

Which will result in solar and wind having a faster Roi and so even larger investments. Despite the faster adoption for solar and wind that we are having currently all over the world. Energy storage is a still a very large bottlekneck.

As one example of a new buyer who can use bursts of electricity: If you have a large number of electric cars with big batteries, which don't have to charge every day, and they're plugged in both at work and at home, you can do a lot of good by having them charge when electricity is cheapest.

And it's not exactly high-tech, it's just a little bit of schedule prediction.

Sure, but your grid operator won't offer you the low rate unless you can convince them that your fleet will actually not charge during high demand and not just conveniently switch over to the fixed rate socket.

Dumb question: How can an electric furnace get hot enough to melt steel, but not melt itself?

Consumable graphite electrodes.

(edit: also, there are plenty of materials that can deal with sustained molten steel content with acceptable service life.)

The electrodes are made of graphite and the lining of the furnace is lined with a ceramic material. Both have melting points higher than the molten steel.

and if you (OP) continue asking "well how was the ceramic material made?": it is produced by dry dust pressing.

There's an xkcd what-if that says, paraphrased, "little is known about the properties of molten tungsten, because -- and this might sound stupid -- we don't know how to make a container that will hold it".

The question was valid. ;)

Reply on news.ycombinator.com