Hacker News Reader

Comment by cycomanic | Hacker News Reader

Calling fusion research underfunded is rich(no pun intended). Costs for ITER are estimated to 45 to 65 billion USD it's one of the most ambitious and costly science projects in the world to date. That's not even accounting for all the other fusion projects.

There are many areas in science that receive significantly less funding.

I would also argue that the prevelant perception that fusion would lead to to essentially unlimited free energy is wrong. The extremely high capital costs would likely mean that fusion is unlikely to ever match solar or wind in energy production costs.

I'm not opposed to fusion/plasma research, we have much insights into nonlinear dynamics and chaotic behaviour due to this research, I disagree with the notion that this is a field that achieves incredible outcomes with small budgets, that's far from reality.

View on HN

The new Vogtle reactor in Georgia (the US state) cost about $17B. A single offshore oil platform can cost upwards of $1B and we build hundreds of these.

When weighed against the potential payoff fusion is very underfunded. If we make it work we basically have infinite clean energy until the heat death of the universe, not to mention a source of energy that could jet us around the solar system and beyond at speeds far beyond what anything we currently have can achieve.

The cost of ITER is not crazy for the energy industry in general. Huge energy projects are very expensive. We spend multiples of that on oil drilling every year.

Apple is buying back $110B of stock.

They could build an Apple iFusion Reactor instead and still buy back $50B.

Arguably for much less than that. The current fusion startups are targeting burning plasma and altogether they've raised a tenth as much as ITER's (speculated) 6.5 Bn USD. Granted, very few companies are targeting burning plasma with their current level of funding.

Let's read again what you just wrote and think: "ITER are estimated to 45 to 65 billion USD"

That's not the most costly costly project, not even close (prices in current dollars and total costs)

Apollo project: $257 billion

ISS: $150 billion.

STS: $196 billion.

Fusion energy has been constantly underfunded.

None of those projects was worth the cost.

the ISS was definitely worth the cost.

Really? What result did ISS deliver that justifies its construction? It's an exercise in orbital flagpole sitting.

What's the point of being alive without building monuments?

I find that justification not only not convincing, but utterly repulsive. Governments shouldn't be in the business of praising themselves.

It’s an engineering test platform for long duration space flight. Along with the development of reusable rockets we are now almost to the point of being ready to actually go somewhere and do more than just plant a flagpole.

Apollo was ridiculously ahead of its time and was not sustainable. The entire stack was disposable and it’s pure luck that we never lost a crew. We needed a project like the ISS to do the real work to figure out how to live in space and build long lasting space hardware.

The biggest shortcoming of the project is that we never tried centrifugal gravity. There were plans for a module but it never got there.

That's great, except (1) the thermal environment is all wrong, (2) ditto for the radiation environment, (3) that just pushes back the need for justification to long distance spaceflight. Justifications for space activities that amount to "enables other space activities" are an example of what's called a "self-licking ice cream cone".

There were plenty of non-self-referential justifications for ISS (like growing protein crystals, or space manufacturing) that never really went anywhere.

Ironically, maybe the best justification was it shows large space structures can be assembled from small units, so large launch vehicles aren't needed. NASA has of course totally ignored this lesson with SLS.

It's really hard to put a dollar value on what the value of having something like SpaceX Starship available is, and it's hard to evaluate what the Starship project would look like if we hadn't put the money into the Apollo/ISS/STS/SLS before.

With the current state of things, I would bet that with a fully reusable orbiter like Starship, space-based solar with microwave transmitters will probably be a cheaper option than fusion. But we should work on both because we can't predict it until both are tried. It's easy to say no to a project when we don't know what the outcome will be, but we need to take on a lot of big risky projects to see which ones give outsized returns.

SLS only started in 2011. If SpaceX learned anything from it, it was not much and it was after they'd proven themselves. Starship has been quite from-scratch outside of the engines.

I mean, SLS was definitely a waste of money, and arguably so was STS. But that's easier to say in hindsight, and it's also easy to say in hindsight that SpaceX had an obviously better approach, but that was harder to support in 2011.

Oh, I think plenty of insiders were calling STS a waste in foresight. I recall a story (no source, sorry) that one of von Braun's people lamented about it: "they've reinvented the wheel, and made it square." Minimum Cost Design methodology for launchers dates from the 1960s.

"Estimates of the return on investment in the space program range from $7 for every $1 spent on the Apollo Program to $40 for every $1 spent on space development today."

I guess you take issue with estimates like this?

Yes, because they are basically exercises in bullshit. You might think they went and catalogued in detail what the benefits were, but they're actually just based on macroeconomic assumptions about what the return on R&D is.

I mean, think: for spinoff arguments to work, you need to have some notion of how technology would have progressed without NASA. How could one possibly figure out that contrafactual?

ITER is like electric cars. There’s no point in spending millions to design a new EV until the battery technology exists. That’s the key limiter, not the “rest of the car”.

For fusion, this key technology limiter is the superconducting magnetic tape. Even small improvements in field strength have a dramatic non-linear improvement in fusion power gain. ITER started their design work decades ago and will continue construction for decades more. Meanwhile, SC magnetic tape technology has moved on, making the entire enterprise a dead-end waste of time and money. They got bogged down in the paperwork and bureaucracy of the “rest of the facility” with no hope of making actual working reactor.

There are many instances of this issue in industry, and in my experience large government bureaucracies are pathologically incapable of planning ahead for where the puck will be instead of where the puck was.

I like to imagine what would happen if a rich industrialist like Bezos or Musk approached the problem: First, spend a few billion on superconductor tape research and then start building many small reactors that are minimal test beds while in parallel improving the magnets. Only “go big” once the reactor can produce net power.

See also: NASA spending billions per launch on the SLS with a couple of test launches per decade while SpaceX spends millions per Starship and plans half a dozen test launches this year!

So how's the Lockheed Martin skunkworks project going?

That was the last poster child for "REBCO tape will quickly solve all the problems".

Oh right - incepted in 2010, cancelled in 2021. 11 years - over a decade - and no reactor in sight.

They never mentioned REBCO or HTS. It was obvious bullshit from the beginning. It was more likely a cover for a black project, like the SR-72.

I wonder if some of these whacky military things are actually honeypots for espionage. That's been my assumption for the Navy work on Polywell.

Right but that's the point: the parent post is talking about those things as though they've obviously and clearly obsoleted the ITER design. Except that can't be true, because no reactor based on those technologies exists now, or is likely to complete construction in the near future.

So ITER is the correct bet: take boring technologies and build a functional fusion device with them that gets Q > 1. Because there's a fairly obvious corollary, which is that if you can somehow build better magnets, then the next reactor - DEMO, and actual power plants - can obviously be redesigned to take advantage of those magnets - they will not fundamentally change the internal dynamics of a tokamak, since it's the magnetic field shapes which matter, not the physical generators.

Everyone working with REBCO has always talked about "rapid breakthroughs" in about 5 years, and then 5 years later the story is always the same: it's considerably harder to build with these things then you'd think, they're still working through the engineering etc. It's not that it can't work, but it's not magically faster in the way people keep claiming and nor does it get you any closer to the actual goal, which is to get a Q > 1 fusion device functioning - which will, at all times, fundamentally require a very large vacuum chamber to do so unless magnetic field intensities are being vastly increased (and they're not: low-temperature superconductors can achieve far higher field intensities then high-temperature ones which is why the LHC and other particle accelerators use them[1] - some YBCO tapes might be able to do it, but manufacturing suitable quantities, of suitable quality, in long enough reels, and then winding magnets with them is an unsolved problem).

[1] https://www.researchgate.net/figure/Engineering-critical-cur...

>Everyone working with REBCO has always talked about "rapid breakthroughs" in about 5 years, and then 5 years later the story is always the same: it's considerably harder to build with these things then you'd think, they're still working through the engineering etc.

Please cite the earliest REBCO-based fusion reactor promise date that has lapsed. I don't think reality agrees with your perception.

Importantly you can draw a straight line from REBCO-based fund raising and the amount of REBCO needed to be purchased for them. If "the engineering wasn't there" then that amount of money could not have purchased that amount of tape.

Also, how could ITER be the right bet when it costs so much and takes so long to build that it could never scale to anything more than a multi-generational science experiment (e.g. not a power source)?

ITER is a considerably more complex research device, designed and developed to generate and study burning plasma. It is also the first device of it's kind intended to do this. It is an entirely uncontroversial issue that the first version of literally anything is comparatively expensive compared to the costs of subsequent ones.

This is a problem the US Navy is intimately familiar with when it comes to shipbuilding as it is currently experiencing with it's next generation aircraft carrier design.

It is incredible though to see people saying that the ITER design is "obsolete" when the actual proposed replacement technologies have ongoing papers trying to move towards practical deployment as recently literally this year.

EDIT: It is worth nothing that I could be wrong in the end if the SPARC project succeeds[1], but there's a pretty heavy asterisk on that since ITER is amongst other things, a research project - the plasma physics modelling, supporting fusion research, even manufacturing and magnet winding research and other miscellaneous construction processes are all public knowledge which is shared amongst participating nations. It's been the only game in town for decades, and no one was wrong to try and build a fusion reactor 3 decades ago when it was as desperately needed then as it is even more so now.

[1] https://www.psfc.mit.edu/sparc

No, actually the key technology (for DT fusion) is the blanket and first wall (and possibly limiter/diverter). The problem is maximizing power/area, since if this is too low the reactor is too large and cannot be economical.

The most plausible IMO DT fusion scheme is Zap's, since it allows thick liquid lithium to cover much of the exposed inner surfaces. Zap's scheme doesn't even use magnets, never mind high temperature superconducting ones.

fusion research is underfounded while ITER is overfounded. I worked with it and it is clear that 1) it is too big of a step from the previous nuclear fusion reactor. 2) french regulatory agency is being way too strict 3) the procurement mechanism with in kind contribution has been a total failure.

4) even if it "succeeded" it would be a dead end.

not true. if it works it will demonstrate a huge number of technologies that any type of fusion reactor will face. it is just not efficient in pursuing these

Anything exposed to neutrons is not going to be "demonstrated" in the sense of shown to work in an environment of a working production DT reactor. The integrated neutron flux over the lifetime of ITER, assuming it works at all, is just a few percent what would be needed. This will not be sufficient to show the devices or materials are adequate or to conduct necessary reliability growth.

There is the larger issue of whether anything like ITER, with solid surfaces exposed to DT neutron flux, could ever be successful. ITER itself is very far out of the running as a prototype for a competitive source of heat, with volumetric power density of the reactor 400x lower than existing PWRs. A substantial part of that problem is limits on power/area through the exposed surfaces.

some material in iter are not relevant for future reactors, true. But some yes and will face a fluence that is n order of magnitudes higher than what has been tested in jet, the most powerful tokamak so far. so it's a "half-way" demonstration. in any case materials are not everything, iter will test remote handling, breeding, long pulses, etc etc. we can argue that tokamak are not the right path but at the moment is the most advanced so it is worth pursuing it.

That's a ~12 year construction project. 5.4 Billion USD per year is nothing these days. Only 0.006% of annual GDP for a project that has the potential to transform our civilization.

Reply on news.ycombinator.com