Comment by acidburnNSA | Hacker News Reader

Comment by acidburnNSA | Hacker News Reader

Parent Full thread

I'm a nuclear fission power plant designer. I write a lot about nuclear reactor development history and nuclear economics. One of the most interesting ways to decarbonize with nuclear fission at the scale and pace necessary is to use shipyards to serial construct GW-scale nuclear power plants on offshore floating platforms. This allows us to do Henry Ford serialization while capturing economies of scale. This method was very seriously considered in the 1970s and 1000 people built the production facility in Jacksonville, FL. It was called Offshore Power Systems (a joint venture between Westinghouse and the Newport News Shipyards) and there's a wonderful writeup about it in a 1975 New Yorker [1], plus my summary of it from earlier this year [2].

This form of nuclear production got an actual construction license from the NRC after incredible amounts of environmental study.

It may raise eyebrows, but consider that in deep water tsunamis are a few feet tall, earthquakes are non-existent, and huge ships like the Prelude are designed to handle category-5 cyclones. You also have infinite access to cooling water and can design in a sink-safe worst-case scenario with a engineered recovery operation.

If we were serious about decarbonizing right now and I were in charge, I think I would march over to S. Korea and try to get their world-class nuclear construction people (KEPCO) hooked up with their world-class shipyard people and just start hammering them out.

[1] https://www.newyorker.com/magazine/1975/05/12/the-atlantic-g...

[2] https://whatisnuclear.com/blog/2020-01-26-offshore-power-sys...

Let's not get stuck into one solution. Need to pursue several. Floating power plants are very promising as there is lots of shipbuilding capability and the worldwide commercial cargo fleet numbers in the tens of thousands of ships. This could scale pretty fast. I do dislike the idea of a power plant surrounded by salt water - it just doesn't help with the chemical/corrosion safety even if you can use as coolant for an extreme accident. Obviously, no earthquakes and less population in the area. But once the accident happens, you don't get to do mitigation like collecting the dust or pieces.

Another strong approach gaining traction is production in the 5-50MW range, with 100s/1000s/yr in the same way Boeing and Airbus build ~2000 wide body aircraft each year. Smaller reactors with lower power density can claim inherent safety with respect to various hazards - ie cannot meltdown through system failures.

Must move away from absolutist mind set with single solution and try lots of strategies. Many will be effective.

Yes SMRs look very promising and I believe NuScale has one commercial plant in the early stages of production but it won't even be finished until the end of the decade[1].

The optimal mix using today's technology seems like it would be traditional fission for baseload combined with (solar/wind + battery) / hydro for variable / peak grid load.

[1] https://www.nuscalepower.com/projects/carbon-free-power-proj...

SMRs likely will suffer from diseconomies of scale. If economies of mass production make up for that, it will be awesome if not surprising.

That's why I like combining economies of scale with economies of mass production with large reactors built in shipyards. Decarbonizing requires many new TW of capacity. Exajoules is the unit of import.

I always wondered if the future was something even smaller scale than that.

Something the size of a chest freezer and rated maybe 100kW-- enough to run a few neigbouring homes, or a small office campus. Maybe it's a heat-decay product rather than a steam system, to reduce moving parts and breakdown risk. It's designed as a sealed, maintenance free unit that you dock in someone's shed, it runs for 5 years or whatever, the green "Change Reactor" light comes on, and you call the manufacturer to take it away and drop another one in place. (It could be properly deconstructed and refurbished by the manufacturer, like the old "disposable" cameras). It would probably output something like 240 or 480v natively to integrate in a similar way to rooftop solar.

I figure this can achieve a couple of goals: * By keeping the scale small, we might be able to get the "worst case" risk profile to something people find palatable (they're okay with some medical facilities with X-ray and radioisotope therapy machines in their neighbourhood, but nobody wants to live next to Palo Verde)

* Transmission losses are minimized and putting much of the generation near demand reduces load on the grid itself if the demand estimates are faulty and we still need to pull peak power from elsewhere.

100 kW may make sense but as part of a large facility with maybe a few hundred of them. Distributing nuclear material so widely would be imprudent. The materials themselves are dangerous, there's no getting around that and there's real value in a minimum of centralization and control of the reactors in one place. A turbine generating half a gigawatt is also much more efficient (orders of magnitude) than 5000 turbines generating 100kw.

I wouldn't get hung up on transmission and distribution losses--they're something like 6%--and you're missing out on economies of scale with larger installations.

That's what was always so weird about rooftop solar: if solar is so great, why is there more interest in rooftop than large-scale farms? Especially when you look at the installation cost and sub-par location and orientation, rooftop solar is significantly worse than a farm. Transmission from a farm will be less than inefficiencies because of roof angle.

TBH, I'm astounded. I always went off a discussion we had in a 100-level physics class that basically boiled down to "We're considering deregulating the electrical market here, but any claims of choosing your electrical supplier is nonsense-- if you wanted to get electricity from one state over, you'd need to buy 6 units to get 1 at the point of consumption." I suspect it may be a combination of figures off by a few orders of magnitude and a worst-case of sending power 500km instead of 50.

OTOH, I am still interested in the grid-fragility aspect-- if the main grid is only carrying relatively low peak power needs, instead of the entire base load, won't we be able to better handle failures? Wouldn't extra redundancy be cheaper and easier to build?

> if you wanted to get electricity from one state over, you'd need to buy 6 units to get 1 at the point of consumption.

There's a chance you were talking about transmission of household voltage. What Ohm's Law does for long-distance energy transmission is amazing

Found this:

> For example, a 100 mi (160 km) span at 765 kV carrying 1000 MW of power can have losses of 1.1% to 0.5%.

And also

> As of 1980, the longest cost-effective distance for direct-current transmission was determined to be 7,000 kilometres (4,300 miles). For alternating current it was 4,000 kilometres (2,500 miles)

https://en.wikipedia.org/wiki/Electric_power_transmission#Lo...

> I do dislike the idea of a power plant surrounded by salt water

We've had nuclear powered subs for decades now.

Russian Akademik Lomonosov [0] is already a working example of such technology (well, apart from it being "only" 0.3 GW). While I think this direction has certainly has its niche (somewhat similar to floating storage and regasification units), I don't think we should view it as a general replacement for ground-based plants. It's not only about economics (after all floating plant inevitably will cost more), but providing energy to users deep inside land mass.

[0]: https://en.wikipedia.org/wiki/Akademik_Lomonosov

Reply on news.ycombinator.com