Thread by @enn_nafnlaus

@AukeHoekstra @truth_tesla Huh? Corrosion is one of molten salt's *biggest* problems. And thorium does nothing with respect to corrosion (it's also very immature and has a lot of problems).

But that's just the start of problems.

First off is the fact that you're dealing with toxic materials. And not...

@AukeHoekstra @truth_tesla toxic in the normal sense, but *insanely* toxic, where even vanishingly small amounts of many radioisotopes, quantities you wouldn't give a rat's arse about in any other industry, are too unacceptably toxic for release.

Nuclear produces a broad spectrum of all sorts of these...

@AukeHoekstra @truth_tesla insanely toxic compounds, and each has its own physical and chemical properties. It can be a solid, a liquid, a gas, mobile or immobile in salt, mobile or immoble in water, compatible with steel but not alumium, compatible with alumium but not steel, compatible with this alloy...

@AukeHoekstra @truth_tesla but not that one, etc. You're making these in an environment that by nature has to be hot, and frequently, high pressure.

Now you're generating heat (and byproducts), but unfortunately you can't *stop* generating heat (and byproducts). You can certainly rapidly shut down...

@AukeHoekstra @truth_tesla further fission, but you can't do anything about all of the short-lived daughter products currently in your fuel which will continue giving off many megawatts of heat for quite some time even after you stop all prompt fission.

Nuclear plants have lead times of around a...

@AukeHoekstra @truth_tesla decade or so. New generations of reactors consequently have multi-decadal timeframes. This incredibly slow iteration rate has led to something very unusual in tech, which is a *negative* learning curve. Normally as time goes on, technologies tend to get cheaper, but nuclear...

@AukeHoekstra @truth_tesla has gotten *more* expensive, as we've been learning of new problems faster than we've been evolving the technology and getting the costs down in general.

Nuclear also pairs poorly with renewables. Renewables need *peaking*, but nuclear is fundamentally built around...

@AukeHoekstra @truth_tesla *baseload*, which renewables do not need in any form. Nuclear's already terrible economics (which are not trending in the right direction) are premised on ~90% capacity factors. If you start running at 10% capacity factor as a peaker (even ignoring ramping issues), you're...

@AukeHoekstra @truth_tesla making the power nonsensically expensive.

Conventional nuclear, I see no hope in. I see some potential to improve its economics with small mass-produced modular reactors (economies of scale and faster iteration times offsetting the generally worse economics of smaller...

@AukeHoekstra @truth_tesla reactors), but I'm doubtful it'll be enough to compete with renewables even at their current state, let alone where they're heading. Two avenues where I still think there may be some potential for nuclear:

1) Thermal-only: nuclear tends to get much cheaper when you only care...

@AukeHoekstra @truth_tesla about generating low-grade heat, for industrial / district heating. Might even get up to order-of-magnitude cost reductions. By contrast, generating heat with solar, wind, etc requires (if you want any decent efficiency) an *increase* in capital costs, to pay for heat pumps. ...

@AukeHoekstra @truth_tesla So nuclear *might* carve out a role there.

2) Thermal storage: IF huge quantities of heat could be stored without dramatically raising the capital costs (for example, geothermal storage of heat in bedrock), then you can still run your reactor at >90% capacity factor, but then...

@AukeHoekstra @truth_tesla you can sell power at peaking rates, which would be a massive improvement in nuclear economics, as well as making it suitable to pair with renewables. But there's a big IF in there.

My two cents.

@AukeHoekstra @truth_tesla @threadreaderapp unroll

@AukeHoekstra @truth_tesla ** ED: With *PV* solar