Hacker News Reader

Comment by shaqbert | Hacker News Reader

As the production costs with windpower decrease due to the technological progress, the question of grid storage and grid interconnection becomes more pressing.

E.g. in Germany, they frequently have days of negative energy prices, where producers need to pay to feed energy into the greed because of overproduction of energy on windy/sunny days [0](in German). Without storage and grid interconnection, there is only so much renewable energy from intermittent sources like wind and solar you can handle.

[0]: https://blog.energybrainpool.com/bereits-103-mal-in-2017-neg...

View on HN

Germany has at least one example of an aluminium producer testing dialing the power use of their melter up/down by 25% in response to input from the grid [1]:

> Trimet says that implementing its technology across Germany’s four aluminum smelters, three of which it owns, could provide a demand response capacity equal to a third of Germany’s 40 gigawatt-hours of pumped hydro storage.

There was a lot of confusion about this one when it was published, but as the article points out, they're not using their molten aluminium pools as a liquid battery - they're merely dynamically adjusting the electricity usage of their hydrolysis cells, which also requires some other adjustments to keep the process going, but it's still a lot of power and more responsive than what they do today (which involves taking the smelters entirely offline for short periods).

It helps that their test plant is near Hamburg in the North, "close" to most of Germany's wind turbine capacity which helps alleviate transmission/interconnection problems to the South.

[1] https://www.greentechmedia.com/articles/read/german-firm-tur...

As turbines get larger, not only do the costs go down, but the reliability of supply goes up. It becomes more economical to place them out at sea, where winds are more reliable, and the size allows them to run more effectively off low speed winds. These turbines are claiming 60%+ capacity factor, compared to 20-30% for smaller, onshore turbines.

The UK is already up to about 15% of its electricity from wind right now, these developments will open that up significantly more. Obviously at high levels of penetration there are going to be issues, we'll just have to see how far we can go, but even 25-40% of electricity would be a huge contribution.

Combine that with the new and existing nuclear, another 15%, some biomass and tidal lagoons, with the rest from gas (which produces half as much CO2 as coal) and it will go a long way towards meeting the UK's 80% CO2 reduction target from 1990 to 2050.

Isn't this where pumped storage and other forms of storage come into play? Here in Scotland we routinely pump water back up into our lochs in order to store it for future use.

There's also various news articles reporting on German mines beginning to be used as pumped storage sites, with a reservoir on the surface, and water dropping into the mines below when there's sudden demand.

This is a good problem to have, though.

A lot of the best sites for pumped storage are already developed as pumped storage. It's kind of the opposite of other storage technologies. As sites get developed the economics of future sites get worse.

According to Wikipedia we have 15 GWh of pumped storage capacity in Scotland. Per annum we use 27,000 GWh. So by my calculations we have a pumped storage battery that lasts about 5 hours.

(NB: I didn't check whether the power output of this battery is enough to actually meet demand, but wikipedia says it can run at full capacity for 22 hours, so no)

It can be reused more than once a year. You could use it every day, charge at night when there's no demand, discharge it during the peak. Really depends on how fast it can charge and discharge.

Depends more on how variable your supply is. If you have to meet a two week window with below average wind it isn't enough. If you have to just meet day to day variation in wind it is probably enough.

Unfortunately two weeks without much wind are common in general weather patterns. Thus my gut reaction is it isn't enough. However we are now talking weather where local patterns are the important thing which means each situation needs to be considered separately.

Sounds about right. It's used for demand-smoothing over fairly short periods.

(I've been to Cruachan, well worth the visit even though you can't walk around the turbine hall any more)

In Ireland there have been studies using sea-facing valleys to store pumped sea-water. There are an order of magnitude more of these (left-over from the last ice-age) than river valleys for pumped hydro storage. The spirit of ireland group looked at them. When we really need the stored energy, this plan will get dug up again. Reference: http://www.sciencemag.org/news/2012/02/massive-energy-storag...

It is important to note, that here in Germany, wind and solar does not quite reach 100% of our energy consumption yet - I think it has done so only once on a public holiday. The more frequent overproduction comes from high wind/solar output combined with the output of power plants which can't be shout down quickly enough, that is coal/nuclear. If those would be replaced by gas plants with faster reaction times, a much higher amount of reneweable energy could be used without requiring storage.

But indeed, as the amount of renewal energy production increases, more storage is needed. Perhaps less so for wind which has a more smooth output over days, but for solar, which has a strong peak over noon and of course shuts down alltogether in the evening. Distributing the solar curve over more time of the day would be very useful.

Coal power plants can ramp up and down pretty quickly in response to demand. Not quite as quickly as gas, but generally quick enough to respond to changes in renewables supply when operated as "spinning reserve". Nuclear plants, of course, can not.

There are plenty of coal plants that absolutely can not ramp up and down in less than many hours, and will lose money as they do so.

There are also many nuclear plants perfectly capable of "load following", responding in the time frame of minutes to 10's of minutes to load changes.

My understanding is that while load following is possible in nuclear plants, there are significant design, efficiency, and cost compromises involved. For these reasons, load-following nuclear is fairly rare.

In the UK, all of our nuclear power plants operate at their design capacity all the time. Their output only changes during scheduled periods when they are shut down for maintenance, refuelling, etc.

In contrast, all of the remaining coal-fired power plants in the UK effectively operate as load-following reserve. They ramp up only on cold winter days when demand is high and when wind generation is low. There are now many "coal free" days in the warmer months where there is no coal-fired power generation in the UK at all (it looks like today will be one of them!)

> frequently have days of negative energy prices, where producers need to pay to feed energy into the greed because of overproduction of energy on windy/sunny days

This is one reason why I'm skeptical for home production of solar and wind for the ability to "feed back" into the grid while getting a credit on your account.

If the power company is also doing this, at scale, I find it really hard to believe that there's no reason the power company won't discredit/discount my rates being returned to the grid.

In which case, I'd rather have a battery installation and not rely on the power company at all.

Most service reps for solar disagree with me. Maybe because battery storage isn't super common/in their inventory?

It's more that batteries are expensive. Especially in places where there is a law saying that the power company has to pay a nice rate for power fed back to the grid.

Understandable, the legislation isn't brought up (but maybe this concern isn't, either) that often. It could certain be a non-issue.

Though I rarely experience power outages, knowing almost all instances would be non-existent is really nice (for a battery backup program), especially as more and more heating solutions are using electricity as the main component (heat pumps; winter power failures are most of the power failures for me).

I think is a totally reasonable point of view and comes down to a classic cost-benefit analysis.

Are you OK with using the grid as a highly-reliable, ~infinite capacity power source to smooth over your uneven solar generation at a cost of selling your excess back at a discounted rate,

do you want to try and capture that discount yourself at the cost of having to buy, install and maintain your own integrated battery backup system?

Maybe the calculus works out in your favor for you, specifically. Maybe not. But it's good to have options.

Can you explain why you think these negative prices are a bad thing?

I think they're an efficient use of market based tools to ensure that the group who can reduce the input to the grid does so, and to incentivize electricity production that can meet the flexible demands of the grid (e.g. slow ramping coal would be penalised over fast ramping gas without any regulation needing to specifically target the coal and the coal plant can do anything it wants to improve this, e.g. fit batteries, since no technology is mandated.

Negative price signals that cause solar or wind production curtailment is essentially throwing clean energy away.

Negative prices are a helpful market signal, but must be combined with battery storage requirements.

Negative price signals encourage people to use that excess energy as they are paid to do so at exactly the times that solar or wind are overproducing.

They are the solution to that problem, not part of the problem.

As for batteries, there needs to be an economic incentive for people to invest in deploying batteries (or any other type of storage). Negative prices are that incentive. Negative prices don't need batteries, batteries need negative pricing.

I think the dream is just to make wind power so ludicrously plentiful that the fact that it's kind of mediocre during daytime is overwhelmed by sheer quantity of generators.

As for storage, shouldn't there be plenty of hydroelectric dams that could run "backwards"? Or is there an interconnection issue you mention?

For a hydroelectric dam to be able to "run backwards", it needs a lower reservoir. Many dams have only an upper reservoir.

Simply turning them off would be running backwards since water would build up behind them.

It’s not about building up capacity alone: you want to take power off the grid at negative energy price times.

You are taking power off the grid: The power otherwise produced by the hydro plant.

Of course with a lower reservoir, you can go into negative territory instead of "just" taking the plant down to 0.

No, you need to consume what other methods, e.g. wind, are producing when there isn't the demand to use it.

There is always demand. The issue is that sometimes demand exceeds supply. Taking away supply has the same net effect as increasing demand.

EDIT: with the one caveat of transmission capacity, which certainly can be an issue.

I don't think you're quite getting this. There are times where there is excess demand, even when removing these stores. What's so hard to understand?

Prices go negative in the US at times and in Germany more often. That is exactly when something like this would be used: to store that capacity.

The point is that you have a given supply and demand. As long as they are in the right positions in the grid that transmission capacity is not the limiting factor, then you can fix that by reducing supply, or by increasing demand - apart from the issue of transmission those options are equivalent from the point of view of bringing the supply under control. They may not be equivalent in terms of cost, but that is a separate issue.

The original point was that you do not need to have a reservoir to pump from to take advantage of the fast start/stop capability of a hydro plant: Closing the sluice gates and/or bypassing the turbines is sufficient to reduce supply by the generating capacity of the plant.

Having a reservoir as the potential to significantly increase the reduction in excess supply you can handle, that is true, but the lack of that option does not change the benefit of shutting off generating capacity.

Hydro plants have been used to adjust grid supply for as long as they have existed in this way. Not least because with grid prices adjusting dynamically, as long as the reservoir is not full, most hydro plants yield better economic result if they respond to falling prices by letting the reservoir water levels build back up.

> Having a reservoir as the potential to significantly increase the reduction in excess supply you can handle, that is true, but the lack of that option does not change the benefit of shutting off generating capacity.

The point is to not turn off generation capacity that we don't control the start of, like wind. Why should we not generate if the wind is blowing at night?

What I commented on, and what you quoted above, was about hydro plants, not wind. With hydro plants, as long as the reservoir has capacity the reservoir acts like a battery, as hydro plants tend to be built with turbine capacity sufficient for a flow large enough to drain the reservoir. As such, unlike with wind, nothing is lost if you close the sluice gates or a while - the reservoir just builds up and you can generate more when needed.

That's the entire point: That hydro plants can act as storage even if there's no lower reservoir to pump from. Which means they're good fast response plants exactly so that types of plants where turning them off wastes potential energy, like wind farms, can keep running.

The original claim I took issue with, was that a hydro plant can't serve this purpose without a second reservoir, which is not true - a second reservoir can be used to increase total effective storage capacity, and to make the plant increase demand rather than just decrease supply, but a hydro plant on itself can improve grid responsiveness, given enough transmission capacity.

This whole discussion has been around the use of pumped hydro plants to store the output of other sources during times of low demand.

Yes and no. The two comments leading up to my first comment was one suggesting that you could achieve a similar effect by simply closing a damn, followed by one implying that did not take power out of the grid. My first comment then stated:

> You are taking power off the grid: The power otherwise produced by the hydro plant. > > Of course with a lower reservoir, you can go into negative territory instead of "just" taking the plant down to 0.

In other words, my point from the beginning was that there is - absent transmissions limitations - no conceptual difference in the outcome of stopping a damn and pumping into a reservoir, as I've gone into at great length in the subsequent comments.

Nope. Adding power to the grid is +X MW. Removing power to the grid is -X MW. Not producing any power is 0 MW.

That's true when looking at an individual plant, but is not directly relevant when you're looking at the grid as a whole (though it may be an issue when you have transmission limitations), where the problem is that at some point you have a supply of +2X MW, and demand is only X MW, and you need to deal with the X MW of oversupply. If one of the suppliers can then turn off X MW of supply rapidly, that has the same effect as if you add X MW of demand.

At a hydro plant you have multiple options: You can bypass the turbines; you can close the dam, as long as it's not near capacity; if you have a lower reservoir you can close the dam and pump water up to the upper reservoir. All of these help reduce the excess supply, but pumping up does so by increasing demand and reducing supply, and the other two options reduce supply.

You can of course do pumped storage without normally producing electricity too - in places where the only convenient location for a reservoir is not "in line" with a water source with sufficient fall height, you can still pump up to a convenient location, in which case the plant is basically "only" storage.

Unfortunately the capacity of these dams in most areas is far from enough to compensate all fluctuation in renewable energy production.

You not only need variable volume reservoirs on both ends, but you also want to have very high drop, because liars from seepage and evaporation scale with volume but not with height. Low-drop pumped storage might be useful for very short term storage, but usually you'd want something that is efficient over a wider range of usages.

This sounds like an opportunity for collaboration between crypto miners and the wind people.

And maybe it can be mediated by fungible electricity transmitted over a grid of conductive wires. Or an electricity market.

What I've meant was mining on the wind mills themselves. So there would be no need of passing through the grid.

I heard that crypto mining hardware has to have decent-ish (80-90%) uptime to be profitable. That would make it infeasible to have it on standby waiting for excess energy on the grid.

After the windmill is paid off, the mining is then practically ( some maintenance will be needed ) for free.

Isn't the number 80-90% based on electricity price?

I'm afraid I have no idea. I vaguely remember reading about it, do not remember where.

Electricity prices matter but lower prices give diminishing returns.

Going from $0.20 per kWh $0.05 to per kWh only increases profits by 50%

http://whattomine.com/coins/151-eth-ethash?utf8=%E2%9C%93&hr...

and going from 0.05 to -0.05 only increases it by 20%

http://whattomine.com/coins/151-eth-ethash?utf8=%E2%9C%93&hr...

By the way, these mining calculators are very "optimistic" but for relative comparisons they are sufficient. In practice you will get significantly less because of increasing difficulty and hidden costs like motherboards, CPUs, etc.

I know. And there's really no difference between colocating mining hardware on wind turbines and just mining using the lower cost electricity which will be on the grid thanks to the turbines. Except the inconvenience of the colocating.

IMHO:

(1) Cheap, large scale storage of electric energy might be a big game changer.

(2) Without such storage, basically a country has to notice what their peak load is and assume that they have installed capacity that is reliable, e.g., independent of the weather, and meets this peak load.

(3) So, net, intermittent sources such as wind and solar don't much lower the capital expense needed. That is, the country still needs the same capacity to meet peak loads without the intermittent sources. So for this capacity they can use nuclear fission, coal, oil, natural gas, hydro, geo-thermal, and that's about it.

So, net, when power from intermittent sources is available, then get to save on the fuel cost of fission, coal, oil, and natural gas and accumulate some water behind hydro. So, e.g., a big fission or coal plant is still there, still with its capital expense and operating expense but is running at reduced power and, thus, saving on direct fuel costs.

At that point, I'm wondering if, really, just for the money part, if the intermittent sources are wanted on the grid even for free.

There is an issue about the intermittent sources: They are not very predictable and, thus, can cause some grid stability problems. E.g., IIRC, can't just change the power level of a fission or coal plant minute by minute to match the minute by minute fluctuation in the intermittent sources; so, can get some grid instability, and that is bad stuff for the whole grid and country.

Worse, the big power plants get much less fuel efficient at lower loads. In some cases you will actually burn more fuel at 50% than 80%!

Reply on news.ycombinator.com