🧵 SMR insanity
Anonymous at Sun, 20 Oct 2024 18:06:41 UTC No. 16441453
Who the hell would think that 80MW TRISO pebble reactors are a good choice to power tens of gigawatts of data center expansion? Is this not the most ideal use case for AP1000s?
Anonymous at Sun, 20 Oct 2024 18:15:52 UTC No. 16441463
>>16441453
In the 1960's everyone read books, we didn't need nuclear reactors to power brainwashing spy boxes in our pockets. We also had real men who smoked, fucked pussy, and went to the Moon when they said they would unlike that no dick Musk.
Maybe progress was a mistake.
Anonymous at Sun, 20 Oct 2024 19:08:21 UTC No. 16441528
>>16441453
TRISO pebble reactors are best used for process heat. That's how the Chinese intend to use theirs and that's how X-energy originally intended theirs to be used. They were supposed to build 4 at a Dow Chemical facility in Texas
Anonymous at Sun, 20 Oct 2024 19:56:39 UTC No. 16441609
>>16441453
Most likely amazon just went with the lowest contractual expense. There's no one company that can comply with the current demand and Amazon was likely turned away from everyone else because of lowball offers.
Anonymous at Mon, 21 Oct 2024 14:23:24 UTC No. 16442575
>>16441609
The Xe-100 has never been built and doesn't even exist as a ready design. The only way it is the "lowest contractual expense" is because it's tiny and Amazon isn't actually serious about nuclear power. It certainly can't be "lowest contractual expense" per watt.
Anonymous at Mon, 21 Oct 2024 17:02:48 UTC No. 16442764
>>16441453
>>16442575
These, assuming Amazon actually wants them, are probably precursors to see if they investing in their own nuclear power infrastructure is worthwhile.
Anonymous at Mon, 21 Oct 2024 19:14:11 UTC No. 16442945
>>16442764
TRISO pebble-bed reactors are a dead end technology for large-scale electrical power generation, because the reactors are hard to scale up. <100MWe reactors are not cost-competitive with >1GW reactors on a per-watt basis. Pebble-bed reactor tech makes sense for microgrids and process heat applications, not for adding multiple gigawatts of electrical power to the local balancing authority area. However, Amazon and Googe have no need for process heat, and data centers don't need to be on a small isolated microgrid.
If Amazon and Google wanted LWRs, they would've invested in LWRs.
Anonymous at Mon, 21 Oct 2024 20:05:45 UTC No. 16443021
>>16442945
TRISO is also very safe, meaning regulatory overhead and physical footprint are both relatively small. This desirable for testing the possibility of using inhouse nuclear energy in the future. Building more economically efficient large PWRs is Herculean undertaking not suitable for a proof of concept.
Anonymous at Mon, 21 Oct 2024 20:35:00 UTC No. 16443044
>>16443021
>TRISO is also very safe, meaning regulatory overhead and physical footprint are both relatively small. This
These reactor designs might be more safe in theory, however for the foreseeable future that is negated by the lack of operating experience. It's not like they lack any conceivable dangerous failure mode whatsoever. For example, one conceivable problem is water ingress from the steam generator into the reactor increasing the moderation.
To economically benefit from the increased safety would require regulatory exceptions that allow it. As far as I know, such exceptions do not yet exist in the US, and it seems to me like it would be politically unlikely that such exceptions would be granted before decades of operating experience have been accumulated.
> This desirable for testing the possibility of using inhouse nuclear energy in the future.
I don't see the purpose in a focus on in-house nuclear reactors that are intended to directly power your data center and nothing else. The data centers are usually built in places like Virginia, not in the middle of Siberian wilderness.
>Building more economically efficient large PWRs is Herculean undertaking not suitable for a proof of concept.
Building large PWRs shouldn't be a Herculean undertaking for multi-trillion dollar market cap companies like Alphabet and Amazon. It wouldn't be a proof of concept because PWRs are well known tech, and mature, already built designs such as the AP1000 already exist
Anonymous at Tue, 22 Oct 2024 17:26:01 UTC No. 16444374
>>16441453
All aboard the SMR/Gen4 hype train!
Anonymous at Tue, 22 Oct 2024 22:34:28 UTC No. 16444851
>>16441453
Bezos is pretty clearly using his public company to fund developments for his space aspirations. These reactors make little sense for data centers, but are a game changer in space.
Anonymous at Tue, 22 Oct 2024 22:44:50 UTC No. 16444863
>>16441528
>use reactor for process heat
>your process happens to be heating steam and turning a turbine which happens to be connected to a generator
nothing personnel
Anonymous at Wed, 23 Oct 2024 14:47:37 UTC No. 16445759
>>16444863
The reason TRISO pebble bed reactors make sense for process heat applications are
(1) Unlike electricity, moving heat long distances means large losses. So you need the reactor to be close to the place of consumption. Since the reactors are hard to share between many plants, it makes sense to build your nuclear reactors "right-sized" for the industrial plant they will power
(2) the industrial process requires temperatures in excess of 320C, which LWRs cannot provide
(3) pebble beds allow online refueling, meaning that you can, at least in theory, operate for long time without turning the reactor off. This is important to avoid having to shut down the factory that relies in the reactor during refueling
Electricity, unlike heat, is easy to transport long distances
🗑️ Anonymous at Wed, 23 Oct 2024 14:49:02 UTC No. 16445760
>>16444851
How are they game changers for space? They rely on Earth gravity to function
Anonymous at Wed, 23 Oct 2024 14:51:19 UTC No. 16445763
>>16444851
How are they game changers for space? They rely on Earth gravity and giant water heat sinks to function
Anonymous at Thu, 24 Oct 2024 16:53:09 UTC No. 16447465
>>16444374
It'll be different this time! Better, safer, cheaper!
Anonymous at Thu, 24 Oct 2024 16:58:27 UTC No. 16447482
>>16445763
Bezos' grand goal is O'Neil cylinders (as retarded as that is, still his goal), so there would theoretically be plenty of cold external surface area for radiators. Also spin gravity. Really it would be good on Mars where you need electricity and water heat, but Bezos is stuck on his outdated fantasy
Anonymous at Fri, 25 Oct 2024 18:24:09 UTC No. 16449273
>>16447482
You have access to continuous or near-continuous sunlight on an O'Neil cylinder. You can use solar panels, like most spacecraft do. If you need more power, then you just add more solar panels.
You can add radiators in space however that's a much less efficient and more inconvenient way to get rid of heat than having a planet-sized radiator that's already there and ready to take heat through conduction and convection