58 comments

[ 2.7 ms ] story [ 133 ms ] thread
In the same thread, there is France with a price of 645€. I don't understand why a country loaded with nuclear reaction can beat Germany in price. Or maybe does it mean that the neighboring country are willing to pay France this much because they are desperate, but in this case why not buy the electricity from Germany at 432€ ?.
I think the answer you may be looking for is: Nuclear reactors are unreliable. Compare them to solar, which is very different.

Solar produces power according to sunshine, with no moving parts, and if a panel breaks the rest of the panels go on working, so the failure is a small deterioration in power output. Nuclear plants are differently unreliable, because they have lots of moving parts and elaborate maintenance procedures, and during maintenance a plant produces 0% power instead of that small deterioration. The unreliability is on a different timescale, both better and worse than solar.

France has lots of reactors, but they're not all up, and you can't expect all to be up, so sometimes they'll have this problem. IIRC they had something similar in 2016, which they were able to solve using imports.

> I think the answer you may be looking for is: Nuclear reactors are unreliable. Compare them to solar, which is very different.

You should come in southern Germany to see how reliable is solar when half of the year there is foggy.

Right. Solar depends on the weather and is unreliable in that respect, but reliable against mechanical problems. Nuclear is the opposite. More and less reliable.
In 2011 the nuclear industry promised rolling blackouts by now:

https://theguardian.com/environment/2011/may/23/germany-nucl...

Pretty sure they didnt envision Germany having cheaper winter electricity than France.

All I can find evidence for is a balloon that short circuited a substation, which is a blackout that is unlikely to have been prevented by a nuclear reactor.

> Pretty sure they didnt envision Germany having cheaper winter electricity than France.

Unfortunately that does not reflect at all in the consumer price. We have been taking top spots in the energy price rankings for years now and there is no real hope for improvements.

Nuke is 'reliable' in a much more important way - the plant outages are scheduled and the generation is scheduled. You can (and do) bet the health of the grid on nuke having the output you plan when you plan for it to be there. This is referred to in generation industry jargon as 'dispatchable' power. This makes a modern zero-sum storageless grid work. Solar is non-dispatchable. You can estimate what power output will be available, you can over-provision generation and store it in expensive batteries, but you simply can not rely on it being there and expect the grid to stay online.

It turns out that regular plant maintenance outages aren't particularly important compared to not being able to properly plan generation.

Of course you can decide that. That's what the French did before the price peak in 2016. The thing is that only some of the decisions involved plannable dates.

They decided that when problems were discovered such that redundant security mechanisms weren't, then the problems should be fixed, with time limits. They decided what kind of problems should require fixes within set time limits, and they decided what those time limits should be.

And they decided that some reactors should be taken down for planned maintenance in specific periods.

When problems were discovered at inoppportune moments, the decisions combined to leave them with too few operational reactors.

Nuclear plants are "dispatchable" in the same way wind farms are "dispatchable" - you can just not feed as much of the power you are creating into the grid.

Gas and hydro are actually dispatchable. You can dial down the power you are using and actually use it later.

False. You can’t make more power out of a wind farm than there is wind.
Of course. But also not much less. If you have a 50-unit wind farm you'll have a lot of problems that reduce your output to 98% of what the wind allows until you can fix them. Not so many that reduce output to 0% of what the wind allows.
Did I say you could?
Dispatchable means you can turn it on and turn it up when you want to.
That was my point.
Colour me confused, a nuclear power plant the operators can choose how much power they make every day and generally choose as much as possible, a wind farm they can set a limit for the maximum power generated but the actual power generated could be anywhere from zero to the set maximum depending on the wind speed.

nuclear is more often considered baseload generation as opposed to dispatchable but wind is not baseload or dispatchable since it is not possible to control the wind speed.

>a nuclear power plant the operators can choose how much power they make every day

No. They cant.

It takes 24-48 hours for them to ramp production up or down and this is both expensive and difficult. This is far too slow to be useful. Natgas takes ~7 seconds-minutes to go from 0-100% and it's cheap and easy. Same for batteries. Hydro is slower but still in the order of minutes/seconds. This IS useful.

When nuclear plants have claimed ramp up/down speeds quicker than that they are always still producing the same amount of energy at the same cost they are just wasting some of it.

The nuclear industry tries to blur these distinctions.

>but wind is not baseload or dispatchable since it is not possible to control the wind speed.

Wind can often pull the same "trick" nuclear does where if the grid only wants 100MW but it's producing 150MW it can just let 50MW go to waste.

This isnt really "dispatchable" either but sometimes wind energy advocates pull the same trick.

Wind farms can’t plan to make 100 MW or 150 MW or 50 MW on Tuesday next week if there is no wind.

Nuclear can.

I agree nuclear isn’t dispatchable, natural gas is. But wind is not dispatchable nor is it baseload so making a comparison to nuclear and pretending there is firm energy out of a wind farm on a given day is false to me.

>Wind farms can’t plan to make 100 MW or 150 MW or 50 MW on Tuesday next week if there is no wind.

Once again, nobody said it could.

You can dispatch nuclear power plants.

Its just that their high costs make that an uneconomical decision. Nuclear fuel is damn near free, so why would you ever want to turn it off unnecessarily? Only if the price of electricity reaches into negative-territory is it economically viable to turn off a nuclear power plant.

Same thing with wind / solar. As long as the price of electricity is positive, there's no point in turning them off.

--------

Now think about coal / natural gas. The price of fuel in these cases dominate. Which means its economically viable to turn off when the price of electricity drops (even if its still above $0).

Nuclear fuel is not free but it is cheap, especially compared to the eye watering capital costs of nuclear power.

It is helpful to turn off power production to stabilize the grid. It's not always about production costs. This is why nuclear plants and wind turbines sometimes just "waste" excess energy.

You clearly don’t work in solar. The panels move on trackers, the panels have to be cleaned of dust and snow, vegetation around the panels has to be cut, inverters fail, etc. there are many moving parts and maintenance required.
That sort of maintenance can easily be done without the need for specialised engineers. A gardener can mow the lawn, a janitor can clean the panels, and most solar plants are not on tracks but static.
But that's not what I had in mind. When an inverter fails, it has to be replaced, and meanwhile a few panels don't produce power. A few, as opposed to all. You get many small failures that cause the plant to produce ninety-some per cent of its possible output instead of the nuclear kind, where the plant is down and produces zero energy, and can't be turned back on even with several days of advance warning.
> Nuclear reactors are unreliable. Compare them to solar, which is very different.

> Solar produces power according to sunshine

So, according to weather. Which isn't exactly renowned for being reliable.

This is about wind power, not solar, but another tweet I saw recently mentioned "Solar availability in mid-winter: 0 %" It's also in Swedish, but the picture speaks for itself. https://mobile.twitter.com/HenrikSundstrom/status/1472654421...

how do European data centers make sense in this scenario?
I'm not sure if they make the most sense for efficiency, but depending on what those data centers are used for, they may be required to be in the EU.
In germany they don't even get subsidized because "they don't produce anything", when you want the most expensive computing power you go to germany.

>Everything you wanted to know about data center design (but did not know who to ask)

https://www.youtube.com/watch?v=s-3YRUFBbrA

(comment deleted)
(comment deleted)
AWS has ~20% higher prices for compute (EC2) in Frankfurt compared to N. Virginia, which doesn't seem that much higher.

Additionally, one of the cheapest dedicated server provider (Hetzner) has two data centers in Germany.

network latency?

Data protection laws?

I wonder how economical it becomes to run their own generation at some point. I know in the US I've seen datacenters that will switch over to backup generators at peak demand times as part of load shedding agreements, but I don't have any idea on the economics (other than running your own diesel generators must be more expensive than grid power).
One thing about load shedding and backup generators... You generally want to load test your generators regularly anyways, to make sure they are a reliable backup.

So you are sorta getting paid to do your load tests. (Though load shedding might be several hours a day for 2 weeks straight, instead of a single 1 hour load test.)

Yep, generally load shedding was a few hours just for peak times. And indeed, someone asked the question if we ever tested our generators and the tech answered that we never need ed to test because they routinely kicked on for one reason or another.
A lot of companies in Germany would not use AWS if the data centers would be not in Europe. By the way tesla in Germany uses gas turbines for energy production as far as I know. So if you are big, I guess you don't pay the 'market price'.
In Sweden we have a surplus of electric power in the northen parts (with lots of hydro). The power grid can't handle more exports.

That said - the surplus will not last forever. We have companies moving there quickly, including facebook but also car battery manufacturers and also all-electric iron production initiatives.

Is hydroelectric generation seasonal with snowmelt?
It is in Sweden yes, but that is why you have magazines (lakes). And those are emptied during winter when the snow does not melt and the general consumption is higher. And at the end of it all you need to spend it before spring because then they star filling up again.

(Edit) Right now in Sweden the magazines are filled to 70% (or slightly lower). In april they were at 30%

https://www.energiforetagen.se/globalassets/energiforetagen/...

At the bottom there you can see how the water accumulates during different months. Area 3&4 is the southern half of Sweden, area 1 and 2 are the northen parts.

By buying wind and solar power capacity instead of paying spot price. That's what Google has done in the north-east of the Netherlands, Microsoft in the north-west and Facebook is now going to build in the middle. All close to wind energy parks on which they buy long term capacity.

Which on planet scale is great, because it puts all those datecenters on renewable energy which is good for everyone. But on country scale it's a problem because it consumes a huge part of the available renewable energy and requires building new wind farms to meet climate change agreements. And people don't like wind farms in their backyards.

For comparison, during last year's freeze in Texas ERCOT prices hit ~€8000 per MWh.

https://www.reuters.com/article/us-electricity-texas-prices-...

Is this a valid comparison? The issue in Texas was due to faulted generation which automatically triggered a cascade of tariffs - all of which were resolved and not billed to customers.
As soon as customers don't pay market rate, because for example a regulator steps in and bails someone out, there is zero incentive to conserve a scarce product.

It's like being given the company credit card and being told "spend whatever you like, bezos will pay the bill".

Again, not a valid comparison. The massive tariff increase was due to a hardware failure, not market pressures. It would like being given a credit card, with unforeseeable terms and conditions occurring outside the scope of your contract which are then resolved later in the month
It's not a exact comparison due to the winter storm, but ERCOT does see wholesale prices $1000+ MWh during the summer peaks and super peaks.
No, it’s not a valid comp at all.

1. Those prices came only from certain resellers, majority Texans were not affected by this.

2. People ended up having to pay that. ERGOT said those prices were valid and people got fucked.

This happened because ERGOT temporarily remove a cap on prices to retain stability during shutdown, allowing certain entities to take advantage of this. In my opinion, while this was not against the law, it was criminal in nature.

Putin is a corrupt criminal and Russians deserve better but he was not wrong when he told the EU that they were dumb.

Closing down coal and nuclear plants without having an alternative that doesn't involve gas (laughs in Gazprom) and signing up to climate change treaties without knowing how to achieve those goals... It looks good on camera we'll deal with the issues later!

But at least Europe shares this misery with China who find themselves in the same predicament.

It's easy to criticise, its not like Putin has been doing a good job with Russian economy
What is your view on the alternative: "Not signing up on climate treaties an considering them optional"?
That's what the US did. And in a pure local view it's the best (but selfish) choice for them.

Climate change is a clear case of the prisoners dilemma. If you decide not to play along, use coal, you are better off from your perspective. But if everyone cooperates and agrees to limit climate impact in the long term everyone is better off.

I guess being part of an attempt to break the stalemate would qualify as "dumb".
There's another alternative: "Push climate treaties that set goals based on current scientific and economic realities".
Putin thinks EU politicians are dumb because HE himself orchestrated nuclear shutdown in Germany https://en.wikipedia.org/wiki/Gerhard_Schr%C3%B6der#Relation...

>In his first term, Schröder's government decided to phase out nuclear power

>As Chancellor, Gerhard Schröder was a strong advocate of the Nord Stream pipeline project, which now supplies Russian gas directly to Germany

>In 2016, Schröder switched to become manager of Nord Stream 2, an expansion of the original pipeline in which Gazprom is sole shareholder.[92]

>In 2017, Russia nominated Schröder to also serve as an independent director of the board of its biggest oil producer Rosneft.

In terms more common to Americans, this is about 49c/kWh. Is this for generation only or generation and delivery?

Here in California I pay about 28c/kWh which is typical (total for generation and delivery), but CA is one of the most expensive states I think.

I pay about 12c/kWh (total, residential rate) in TX