i didn't know what ahi said (that minimills already depend on cheap off-peak power) but intuitively i would expect an arc furnace to be pretty cheap compared to the power it uses; it's just a water-jacketed chamber lined with castable refractory with a lid with three big carbon electrodes lowered through it, and all of those are cheap materials and low-precision (tight tolerances won't withstand white-hot flaming steel for long). the machinery is large and heavy, but only in proportion to the volume of material it processes. the electrical energy consumption, on the other hand, is comparatively enormous

(admittedly maybe the capex for running the power lines to the facility is significant, but in the same proportion to the cost of the energy used as running transmission lines anywhere else)

there's a nice video illustrating the process at https://www.youtube.com/watch?v=T1CJ5NPW8MU. don't be alarmed, the part that looks like a major industrial accident is just what happens normally when they turn it on. a more detailed documentary with explanations, though unfortunately of an atypically large arc furnace, is in https://www.youtube.com/watch?v=eZRuVEfxIVI

in that particular case they say it runs 24/7

"the part that looks like a major industrial accident is just what happens normally when they turn it on"

That was better than most fireworks displays. I checked the second video but don't understand the German explanation. Is that just the initial effects of the massive amounts of voltage?

it's pretty moderate amounts of voltage actually, only about 300-900 volts, just a little higher than i plug my phone charger into. it's just massive amounts of current, resulting in massive amounts of heat, which sets fire to the impurities and/or boils them off. you need low voltage because the arcs are pretty low resistance, so getting the necessary power requires tens of thousands of amps, about the same as a lightning strike. an hour-long lightning strike

(the power of a lightning strike is much higher than that of a steelmaking arc furnace because it has much higher voltage, possible because the conductive path through the plasma is kilometers long instead of centimeters long)

youtube has automatically translated subtitles which worked pretty well here. the relevant yt-dlp flags are i think --write-sub --write-auto-sub --sub-lang en,es (season to taste) and then the j key in mpv cycles through the languages you chose. this is also possible from youtube's web ui but enormously more awkward

I see they use water to cool the actual furnace encasing. They then lead out the warm water to big basins to let the water cool off. It would be good if they could use that warm water for remote heating of houses or create electricity or something.

so, because this is steel recycling rather than smelting, almost exactly none of the energy used in the process ends up in the final product; 99% or more of it is lost as heat during the manufacturing process. one of the things mentioned in the video is that this furnace uses more electrical power than the entire neighboring town of witten (population 0.1 million), which means that you could heat all the houses. but you'd still need the big basins, because all the houses might not be enough to cool the furnace; certainly in the summer they wouldn't want to

creating electricity is unfortunately somewhat less practical due to the low temperatures of the water, although there are some interesting experiments with deriving power from low-temperature geothermal resources using n-pentane-blown turbines, and you might be able to use that work. getting the furnace's water jacket hot enough to boil water would create the risk of a steam explosion as well as increase plumbing corrosion problems

i mean, it would create a larger risk than the one that already exists due to having a water jacket in a place where a furnace damage could mix water with molten steel—as it turns out, cooling water is already the major cause of fatal accidents in electric arc furnaces, causing several fatal electric arc furnace explosion accidents every year: https://www.aist.org/AIST/aist/AIST/Publications/safety%20fi... and this is stimulating a movement away from pressurized water jackets for electric arc furnaces (at least, if we can believe this marketing white paper from a company that sells non-pressurized electric arc furnace water-cooling systems)

you could conceivably use a different, higher-temperature coolant than water, such as molten salts or sodium, as is done in nuclear reactors, so you could have an unpressurized coolant loop (maybe even using a coolant that wouldn't react violently with molten steel the way water does) but these do have disadvantages; water has a really astonishing ability to transfer heat by nucleate boiling

but then you run into the cost issue: the reason coal plants are not able to produce electricity at a price that is competitive with solar farms is, as i understand it, that the steam-turbine and electromechanical generator part of the plant still cost more than solar power, even in very polar countries like germany, whose utility-scale photovoltaic capacity factor was 10% last i checked. it's not because of the cost of coal; even if coal was as free sunlight, they'd be uncompetitive. and the same cost problem would make it uneconomic to generate electricity from the waste heat from electric arc furnaces, even if you could get the heat out at a more useful temperature. every million dollars you hypothetically spend on recycling eaf waste heat into electricity would have produced more electricity if you had spent it on solar farms and wind turbines instead

district heating is a good idea tho. i wonder if witten is already doing it?

IANAMetallurgist. Arc furnaces for steel making already rely on cheaper off peak power, so utilization is already a factor.

I didn't think they varied their use pattern by time of day.

They might rely on cheaper power to lower the average cost make financial sense, but that is a different than soley utilizing excess power.

If they cant rationalize the opex running 8 hours a day, there is still a problem. My understanding is that many of these plants cant even shut down and be restarted.

Arc furnaces tend to be first in line to voluntarily shutdown to deal with grid instability, or projected power shortages. Network operators will basically pay them equivalent amount of money as they would to buy electricity, if the furnace operator is happy to shutdown at a moment’s notice.

After, to a power grid, removing large loads are functionally equivalent to adding additional generation. So if you’re operating an arc furnace, and can shutdown quickly (which arc furnaces can), then grid operators will pay you for privilege of being able to shut you down at moments notice, and then pay even more for the electricity you’re not consuming, if the grid is forced to call upon that additional “capacity” due other issues on grid.

I’m not sure if arc furnaces today vary their usage in direct response to variable electricity prices through the day, rather than only acting as emergency ballast to be jettisoned in an emergency. But I would be very surprised if they didn’t, they’re a large enough load that they’ll have coordinate the usage with their local grid, and large enough that shifting the usage pattern to avoid high cost peaks would save them a very material amount of money.

there's some footage in the documentary i linked of coordinating the usage with their local grid before turning the arc furnace off for half an hour (because the next load of scrap was delayed)

[deleted due to severely condescending comment below]

it's good that you're starting to learn some of the basics of how real-time energy markets work, but most of the things you have said are unfortunately nonsense

What an exceptionally unconstructive and condescending comment.

Deleting yours removes any hope of a conversation.

Didn't seem like there was much hope of that with the response I got anyway. I'm not here to take condescending comments with no substance (i.e. pointing out why it supposedly was nonsense), so I decided it was better to remove the comment and move on.

Okay but this way it's hard not to treat it as if you did in fact post nonsense.

And whether you said something right or wrong I at least would have liked to see it. Especially when it gets a callout like that.

> Okay but this way it's hard not to treat it as if you did in fact post nonsense.

All the more reason for me to stop engaging, isn't it?

The gist of it was that nuclear power is insufficient in and of itself because it is hard to regulate output to match grid requirements, and therefore that we need both nuclear and renewable energy sources, not just one or the other. Maybe I'm wrong, maybe I'm not, I'm out of here either way.

> All the more reason for me to stop engaging, isn't it?

No, I don't think so. The skepticism all comes from your initial comment, not the followups.

> The gist of it was that nuclear power is insufficient in and of itself because it is hard to regulate output to match grid requirements, and therefore that we need both nuclear and renewable energy sources, not just one or the other. Maybe I'm wrong, maybe I'm not, I'm out of here either way.

Oh well that specific point is wrong. Modern nuclear can adjust its output quickly and within a wide range if the operator wants it to. We could run a 100% nuclear grid. The fundamental issue at hand is price, not capability, because reducing output makes the cost per watt go up.

also modern nuclear is probably too expensive to compete with batteries; steam turbines aren't getting cheaper fast enough, while both batteries and power semiconductors are improving at a staggering rate. there's no reason to expect that nuclear power will constitute any percentage of the power grid at all 20 or 40 years from now. 10 or 100 years from now, yes

if you want to read plausible misconceptions about electric grid operations and stability you can get gpt-4 to generate an unbounded stream of them

oddly enough, that's what i thought about your now-deleted misconception-filled comment, but i thought maybe it was worthwhile to try to engage positively

the normal kind of arc furnaces are shut down many times a day; it's a batch process, not a continuous one

Do they operate 24/7?

Even if it is batch process, going from 24 hours to 8 hours is like tripling plant cost vs productivity.

tripling capex vs. productivity, yes, but maybe if capex is small compared to opex, that's a worthwhile tradeoff. the one particular plant i was able to find the answer for does operate 24/7

don't forget wind. there's lots of wind power and very low power consumption at night. it's probably more like 22 hours to 16

They generally will run the cheapest electric rates. They commonly run only the overnight shift when electric is cheapest - it is worth paying the employees that extra money to work the overnight shift. They commonly shutdown the entire month of December for yearly maintenance - when Christmas lights add additional electric demand. They coordinate with the power company for other shutdowns.

You may be thinking of aluminum plants which can’t be shut down and restarted without significant damage, but they are indeed still throttled up and down. In that case it’s electrochemistry at high temperatures to strip the oxygen off of aluminum. If it gets cold the apparatus gets damaged, but it can be throttled down a significant percentage.

thank you, that is almost certainly what they were thinking of

how much do they get throttled up and down, and how frequently? i'd like to read more about this

This is one of the reasons I believe 'base load' demand is more fungible than people assume.

I think in California 6% of electricity demand is pumping water. I'm almost willing to go on record and say that's the California Aqueduct and the actual number is higher. Okay I'm going to look.

https://www.ppic.org/publication/water-and-energy-in-califor...

> The water system uses approximately 20% of the state’s electricity and 30% of its natural gas for business and home use, according to data from 2001—accounting for more than 5% of California’s greenhouse gas emissions.

There's been other responses here pointing out that municipal water delivery pumping probably isn't a large electricity consumer, but it's also worth noting that even if it was you can't voluntarily shutdown water-delivery in large distribution systems - you have to maintain pressure and flow to prevent back contamination of the system from leaks. It's why you get boil-water advisories when there's a general outage - because dirt and bacteria can get back into the pipes and it takes time to be sure they've been flushed out of the system.

Did you miss the stat directly underneath that?

> Heating and other energy-intensive water uses in homes and businesses make up almost 90% of water-related energy use, while treatment, pumping, and conveyance of water and wastewater account for the rest.

That's 2% for everything which isn't heating water. Pumping is some smaller fraction of that.

Great thing about heating water, is that it stays hot after you’ve heated it.

Water heating is basically the poster child for “demand-response” technologies. You can easily heat your water a few hours earlier than normal with basically no consequences to the user. But you need to get reasonable smart about modelling people water usage, as people don’t tend forget or forgive a cold shower.

What about tankless water heaters? You can eliminate greenhouse gas emissions for heating and be more efficient with energy usage overall.

tankless heaters use slightly less energy, but they use it in a really annoying way. the optimal design for a heavily renewable grid is a heat pump water heater with a tank

The optimal design for a heavily renewable grid is solar water heating a tank in home whenever the sun shines, falling back on a heat pump water heater with a tank.

That's for maximizing energy efficiency. Tankless I think might be a tad bit more easier and cheaper to implement than this.

Regardless the point is eliminating the use of natural gas in heating water is itself a benefit.

That's only optimal if you know you'll need 100% of the heated water (or have a way to store it). While photovoltaics are going to be less efficient than direct water heating, you can direct the energy to other uses if you don't need more hot water.

Large tanks of water are about the cheapest way we know of to store energy. Batteries store more energy, but are extremely expensive.

Yes, but that only matters if we can easily consume that energy. If we need the hot water, then that's easy—no conversion loss. Otherwise, it gets trickier, especially in a small-scale setting like a home.

All homes use at least some hot water for showers. In many climates they can use a lot of energy for heating which means a very large tank could be useful in winter.

I read a study looking at the difference between tankless and tank water heaters.

Summary the more hot water you use the smaller the win is for tankless.

Other thing I've read and seems true is gas and heat pump water heaters cost about the same to run. I installed a heat pump unit five years ago. Works okay for two people. If you had three teenage girls and a wife that likes baths a tankless would be a better choice.

you are right. The thing that is interesting is that pumping loads overall take a lot of the grid's energy, it's just that most pumps are refrigeration loads, not water transport.

The desalination debate is always a little interesting for me - first, I’m curious what they’re pricing in as the lifetime of the plant, because I always hear the capital costs are too high, but that’s a function of total water produced, and second is that the demand for water is pretty friggin’ inelastic, and as a Californian, I’m pretty sure the supply is constrained. I’m convinced that at some point we’re going to be very, very sad that we talked ourselves out of building those plants.

Call Mark Cuban, make the investments now!

California has plenty of water, it's just dedicated far too much of it to farming in what would naturally be arid areas. It would be politically impossible to make farmers pay the real price for desalinated water, so what might happen is retail customers have to pay the high price for desalinated water while the remaining ""free"" water is routed to agriculture.