Hacker News Reader

Hacker News

Comment by yazr | Hacker News Reader

Parent Full thread

yazr

Energy is a BIG, HUGE industry (about 5-10% of world GDP). It will always be a mix of technologies.

Cost wise, however, i agree nuclear is late to the party.

In 2018, we had large contracts for solar(+storage!) at 3cent/kw (in the US, Mexico and Saudi Arabia).

European new nuclear came in at 15c/kw.

View on HN

That's a lot closer than you'd think then. While it varies from location to location, the capacity factor of solar seems to be about 20%. That means - sustained - a solar system generates 20% of it's maximum power rating, because the amount of power it produces is lessened during the winter, cloudy days, and of course at night.

The nuclear capacity factor, on the other hand, is rather high. Upwards of 90%, and sometimes even upwards of 100% (due to the powerplants being able to produce more energy than they were designed for).

If you factor that in, a nuclear plant has about a 4.5x capacity compared to a solar plant, meaning you'd have to spend 13.5c/kw for solar with the same capacity.

Of course, this varies greatly based on where you are. In sunny climates, solar's capacity is more like 30%, while in northern climates, it can be 15%. And I'd imagine the cost of new nuclear very much depends on regulations - in the US, it might as well be infinite.

I think you're applying an imaginary capacity factor to a number that already accounts for solar's intermittent yield. Solar is simply way cheaper than nuclear.

It depends if the $/kWh is cost to build or price paid for the energy.

I believe the .03 figure for the solar plants was $/kW and therefore had no bearing on total energy generated.

I think you would need a solar plant that incorporated energy storage to deliver nameplate power overnight for 40 years to be making an apples to apples comparison.

Does that account as well for the longevity of it? The current nuclear reactors lasted around 40 years already without much issues, I wouldn't bet the solar installations to last that long.

When you hear "nuclear reactors last for 40 years" you should instead hear "nuclear technology improves at such a slow pace that it's reasonable to talk about 40 year lifespans".

Almost no one cares if your desktop computer can last 40 years. And PV is improving so fast that making it last 40 years would be nearly as pointless.

Cost of products are averaged for their lifespan, if the solar panels only last 10 years, they need to be 4 times cheaper than nuclear to be at the same price per year.

And nuclear does not improve in a "slow paced", I don't know where you got that impression, there's been a lot of innovations on the last 10 years.

Nuclear technology has improved in cost in the past decade by a factor of 6, like PV has? If not, I will call its rate of improvement "slow".

that's not how percentages are working... There's clearly some math issues here... And the latest generation of plants are already 20% better than the previous ones.

Solar installations are generally rated for a 30 year life, so ¾ of nuclear isn’t bad, especially when you add in decommissioning costs which are going to be a lot lower for solar

UK discounted provision for decommission costs:

£100+ billion in 2013 https://www.theguardian.com/environment/2013/jun/23/britain-...

£161 billion in 2017 https://web.archive.org/web/20170516093449/https://www.gov.u...

£234 billion in 2018 https://www.gov.uk/government/publications/nuclear-provision...

... and counting.

> While it varies from location to location, the capacity factor of solar seems to be about 20%.

Another case for the importance of specifying solar variant.

Solar thermal plants with storage (molten salt, not battery) are above 70% CF already. eg the Torresol in Spain.

Average capacity factor of new utility-scale PV installations in the US southwest is above 30% now, I think. Single axis trackers (rotating on a north-south axis) have become the norm.

Is it still setting birds on fire?

Plus: the value of the power (energy) is variable throughout time: solar produces energy synchronously and tends to drive power prices down due to a very low marginal cost (hence negative power prices when there's a lot of wind and solar in Germany e.g.). When there's no wind and no solar, combined with a high demand (e.g. low temperatures in Europe), prices will be much much higher. So the remaining 70% of your capacity factor is worth more than the first 20%.

I have corrected the the PP

This is 3cent/khw and 15 c/kwh

This means it is cost per final energy produced i.e. after taking into account capacity factors, day/night, etc

EDIT UNIT CORRECTION: this is 3cent/khw and 15 c/kwh

This means it is cost per final energy produced i.e. after taking into account capacity factors, day/night, etc

I understand mostly because it has been regulated to death

I don't know man?

If you're familiar with dispatch stacks, you know that 3c/kw is REALLY attractive. Regulation or no regulation, I don't think you're gonna get down to 2c or 1c with a new nuclear plant.

Maybe I'm wrong, and one day someone will figure out how to build SUPER cheap nuclear plants?

But I doubt it.

Best I would think they could do is some sort of hybrid system that uses nuclear generation, and then storage via a conventional hydro storage system or something maybe? I'd have to do the math on that. But yeah... 2c? That's very hard to do.

Moltex is claiming they can get below $0.04/kWh (for N-th of a kind plants; FOAK would be a bit more expensive.)

They can claim this plausibly for two reasons: first, they can build a MUCH smaller containment building, due to lack of water or other volatiles in the nuclear island. Second, their design isolates the turbines from the nuclear island with a molten salt "thermal battery". This allows some dispatchability in their output, but more importantly it means turbine trips have no effect on reactor stability, which means non-nuclear-grade turbines can be used. This is apparently a very large cost savings.

Might be talking about European regulations. 15c per kw versus 3 in usa - thats 5x magnitude.

Doesnt suprise - gallon of gas in most european countries (sold by liters and funny thing - you first pump then pay!) is about 40% more than same amount in USA. it has been like this forever and doesnt look like will ever change.

I have doubts you’re going to be able to remove enough regulation to get nuclear near the same cost as storage backed renewables (which will continue to decline in cost each year even further).

Without storage, renewables are already close to or below 2 cents/kWh (unsubsidized). Can nuclear startups make generating units as fast as automated fabs can spit out solar panels?

A big argument I have heard in support of nuclear is that it is good for maintaining some amount of base load. As I understand this is particularly important in the current scenario of renewable energy without good storage, since we can't necessarily maintain our energy requirements with solely other clean (relatively speaking) forms of energy.

The race will be to see who can win the day: modular commodity nuclear, or utility scale battery storage (Flow batteries, lithium ion/polymer, etc).

Presumably, both get cheaper as you make more, but anyone can make batteries and ship them around the world with little notice. Tesla’s Hornsdale Power Reserve system was built in 90 days. I don’t know of any nuclear plants that get built faster than 10 years.

> I don’t know of any nuclear plants that get built faster than 10 years.

And the Brits are demonstrating just how woefully optimistic costs, build quality, and schedules can be even with that kind of protracted delivery expectation :

https://jeremyleggett.net/2018/08/22/why-the-uk-nuclear-rena...

I like regulating things that are radioactive or can potentially go boom. Especially both.

[And yes, I know some designs are theoretically not able to melt/blow up]

Reply on news.ycombinator.com