Comment by panosfilianos | Hacker News Reader

Comment by panosfilianos | Hacker News Reader

Parent Full thread

I tried to do some back-of-the-envelope calculations on what this means in regards to energy costs being saved, because I couldn't find a direct source (maybe GPT could help actually).

Anyway, based on ITER [1] to equate the energy production of a 1000MW coal plant you would need 2.7t of coal for that plant or 250kg of deuterium and tritium for the fusion reactor (split equally). Based on [2] deuetrium costs about ~$15k a kg. But tritium is ridiculously expensive, at $30k per gram (!) [3].

This leads to a calculation of ~$700M for the coal plant and ~$3.75B for the fusion plant (of which only ~$1.5M is deuterium)

I have a few questions and I wonder if any can help:

1. Is the above fusion fuel correct?

2. What measures are expected to bring these prices down to price efficiency?

Of course, I am not calculating the cost it would take for the reactor, storage, delivery etc.

Nevertheless, this is an absolutely incredible development and the people working for this progress should be definitely proud of their work. My generation and the ones following will hail this as a breakthrough moment. Thanks!

[1]: https://www.iter.org/sci/FusionFuels

[2]: https://en.wikipedia.org/wiki/Prices_of_chemical_elements

[3]: https://www.science.org/content/article/fusion-power-may-run...

[0] provides economic analysis for this type of power plant. They conclude it would not be unreasonable to get levelized cost of energy could get as low as $25/MWh. For this one really needs a high gain of 1000, although gains of 400 are a bit more reasonable, and a gain of around 100 may be economically competitive in some cases.

While the gain NIF achieved was about 1.5, there is good reason to expect it can be scaled up. Ignition is a runaway process, so small changes in the input can result in large changes in the output. Hydrogen bombs, which also use a burning plasma as was demonstrated here, also suggest that the gain and yield may be scaled up.

[0]https://royalsocietypublishing.org/doi/10.1098/rsta.2020.005...

>> Ignition is a runaway process, so small changes in the input can result in large changes in the output. Hydrogen bombs, which also use a burning plasma as was demonstrated here, also suggest that the gain and yield may be scaled up.

Hydrogen bombs are driven by indirect implosion by a nuclear primary. It isn't a runaway process; the yield of a secondary is limited by the implosion achieved by the primary. Most hypothetical designs for an inertial fusion power plant achieve similar energy gains.

I think capital costs & construction times are going to be extremely important.

I think the biggest reason nuclear energy stopped being built is it takes too long to get political wins.

NuScale and other SMNRs might be able to help with that.

However, the capital costs & construction time for Fusion Energy should theoretically be much lower than nuclear reactors.

But are we too far away from it being a reality that anyone is predicting how much?

It will be quite difficult to make fusion cheaper than present day fission. Fusion reactors are new so significant research work will need to be done. There are difficult engineering challenges related to breeding fuel, which involves moving hot, radioactive, and water reactive molten lithium around. Fusion reactors need to be built to high tolerances and will need to be refurbished occasionally due to radiation/heat and in the case of NIF style fusion, explosion damage.

And one still has to deal with many of the same radiation challenges fission plants must deal with. A large quantity or radioactive tritium must be kept on site and neutrons from the fusion reaction will make the reactor radioactive. In fact, fusion produces more neutrons than fission per unit energy. Even so called 'aneutronic' fusion would have side reactions which would produce quite a lot of neutrons.

Fusion is a lot more complicated than using special rocks to boil water.

The main advantage of fusion is a political one. It politically nigh impossible to build a fission reactor in a suburban industrial park, but Commonwealth Fusion Systems is doing exactly that with a fusion reactor[0]. And there is also the slim possibility this type of reactor could explode. Said reactor uses superconducting magnets which store a lot of energy and if something goes wrong, it would be possible for them to release that energy fast.

But, the NRC hasn't made laws for regulating fusion power yet, so they are able to do this.

[0]https://cfs.energy/news-and-media/commonwealth-fusion-system...

50 years from now, are we going to be using fission or fusion?

Hopefully, both. Even if fusion works extremely well, fission should find niches. Fission power is likely to remain more power dense than fusion for a while, so in situations where one cares about weight and size it is better than fusion.

> However, the capital costs & construction time for Fusion Energy should theoretically be much lower than nuclear reactors.

Maybe then current Gen3+ reactors.

But there is no way they will win economically against GenIV Fission reactors. That are already solving most of the issues with Gen3+ reactors.

They already are much smaller, and therefore much lower CapX. Fuel cost are even cheaper because of better utilization of the fuel. Modern plants operating cost are also less because they are even more automated and need to be refueled less.

There is no practical way fusion can compete in my opinion.

Thanks a ton for that! $25/MWh is great but not the 10x economic impact I expected. Especially since [1] we see PV at $50/MWh currently.

[1]: https://en.wikipedia.org/wiki/Cost_of_electricity_by_source

The economic impact doesn't just come from the fuel to energy ratio. One of the biggest differences is also in the space required. Nuclear fusion reactors could eventually end up being very small - like smaller than an SUV.

Imagine the cost savings in miniaturizing electrical grids.

Yeah but are you beating fission, something like:

https://www.moltexflex.com/flex-reactor/

A small fission reactor can do almost everything a fusion reactor can. Fusion fuel is higher density, but fission density is already not an issue.

Fusion still needs to heat water, that's where much of the cost comes from.

A small fission reactor is also incredibly dangerous and will always require exotic material.

A fusion reactor does not pose environmental risk and could eventually run on highly available basic elements for everything after the initial "startup" once the technology progresses.

What exotic materials? Molten salts and steel is most of it. These reactors have essentially 0% chance to have airborn escape of radiation.

An fusion reactors don't actually run on basic elements. The require a fuel that is specially breed. And when they operate they have radioactive materials that can go airborn.

This seems interesting but not especially impactful.

For me, the question here is: can we get our energy to cost 90% less than it did?

Don't get me wrong, I recognise that this is still a huge win (especially environmentally) and that it can have huge runway effects (eg. much more effective decentralization etc.) but it's quite interesting on how we can get these billions of people out of poverty first (or during).

While not 90%, transmission and distribution are around 40% of the cost of retail electricity, so lowering those substantially would go a long way.

if they can be that small then they might find unexpected use in long range spacecrafts or lunar colonies. you'd still have to build and lift the reactor out of earths gravity well.

Here on earth we could see them used to power container ships as well, which are some of the biggest contributors to greenhouse emissions as far as vehicles go.

1000 time gains is double performance every 2 year, means in 20 year we would be able to achieve fusion with hydrogen from water at home scale min power plants.

I'd imagine commodity tritium breeders would be a part of the infrastructure build out and that it will be a highly competitive space as fusion-to-the-home nears reality.

What is fusion-to-the-home?

Power derived from a fusion power plant delivered through conventional powerlines leading to the consumer.

probably one of those mr fusion units from back to the future.

> tritium is ridiculously expensive, at $30k per gram

D-T fusion almost always breeds tritium in the blanket.

Would that mean that you wouldn't need titium to start with? Or that tritium deposits would replenish?

How would it affect the rough calculations above?

> Would that mean that you wouldn't need titium to start with?

Tritium decays in a decade. To start, you'd need the expensive stuff harvested from the heavy water of spent fuel pools. After that, you'd let your neutrons breed it in lithium (or boron, if you're fancy).

So, the approach here would be that it would just be a much more efficient process.

Your calculations don’t mention energy (J), only power (MW in your post). Fuel supply and cost correlate with energy produced, not power.

Parent post suggests that tritium is a fixed cost, more like a construction cost than a fuel cost. We can’t answer this question without a lot more information. We’d need to know how much tritium is needed to reach the point the reactor breeds more than it consumes, if at all.

I would also wonder if Titium is that expensive... and it generates titium... would that be part of the "break even" equation? Creating something rare to sell?

It is so rare because tritium is not found in nature in any significant quantity. The amount on the market comes from water recovered from water pools used for spent nuclear material storage.

There also is no enormous market for tritium. So in short, fusion reactors exist on both sides of the tritium market, by becoming the primary producers and consumers of it, which should lead to significant drop in price of tritium.

Current demand is a couple hundred grams per year. We just don't need that much of it. The cost per gram would go down a lot if we needed to mass-produce it.

I think what I'm alluding to is, how much can the cost go down. In the case of fossil fuels there value is strongly based on their physical scarcity and cost of extraction/delivery etc.

How much do these things ring true in this case and what are ideas of improvement?

This isn't for power plants, this is to model the efficiency of the second stage of a stored h-bomb in place of periodically exploding them, which we can't do anymore due to the test-ban treaty.

Reply on news.ycombinator.com