Miniature nuclear reactors could be the answer to sustainable datacenter growth Datacenters use a lot of power and despite our best efforts, a big chunk of that still comes from burning fossil fuels. But what if instead of relying on local utilities for power, these facilities generated their own – maybe using a relatively itty-bitty nuclear reactor? In a recent report, Omdia analysts Alan Howard and …
I don't think this will fly in the way suggested.
If you put (say) 40 SMRs in a single large site connected to the grid, then you only need one set of operators, one set of security, one set of refuelling and waste disposal procedures etc.
If you put 40 individual SMRs in separate locations, then you have a lot more operational cost - not necessarily 40 times, but still substantially more. In return, all you save is the relatively minor cost of grid transmission. (If that was significant, data centres would already be built next to power stations). You also have 40 times as many NIMBYs complaining about a nuke in their neighbourhood.
In the big picture, I don't really see data centres being significant consumers of power here. Why not substitute "city" for "data centre" - wouldn't the same arguments apply? Or perhaps this is a tacit admission that people *really* don't want nuclear power stations on their doorsteps, so having a data centre and a power station colocated in the wilderness is the only acceptable application?
But the biggest flaw here is the huge amount of extra long-term nuclear waste. How does the cost of reprocessing or storing this waste factor into the overall cost? In my opinion, it would be obscene to leave future generations with a waste problem 35 times bigger than it needed to be.
> What happens when the .com that owns the data centre goes bust, who is responsible for the SMR safety and decommissioning ?
To be honest I'm worried about the bit before that ... What happens when the .com that owns the data centre has to train it's staff in operating a SMR. The thought of a fresh-outa-college kid, a 35 MW reactor, and a loose-leaf folder en-titled 'How to handle a reactor scram' does not fill me with a lot of confidence.
I'm *very* pro nuclear power in general, but handing out reactors like laptops is not a recipe for gaining the public's trust. Lets do nuclear, but lets be sensible about it.
-- Chris
These very small ones wouldn't have on site operators, they would be sealed units that would be serviced by a team that comes in or taken away for refurbishing. The design that is showing the most promise is a helium cooled reactor using TRISO fuel.
Electricity generation isn't even the most promising benefit of these micro SMRs, as they would be install on site they could be used purely for their heat output in industrial processes.
Only the larger 400MW SMRs ones like Rolls Royce are planning would have operators.
To some, "No user servicable parts inside" is a challenge.
I remember one time when I walked through the door just as a colleague was about to dig his penknife into the potting compound filling a network switch that "didn't work"(*). I don't think I've ever shouted "NO!" so loudly before or since.
(*) It didn't work because he was testing it with two machines, one with a 10.*.*.* network address and one with a 192.168.*.* address.
All small units running for a local area would be just that a local unit, sealed and computer connected would be very compact etc. and also, these units would be just a local transportable thing. These machines also would be in railway feeders as joint local machines. Of course there is a bonus, nice people like Putin may find a bit of a problem also/ Should someone like Putin damage the machine then the unit could be picked up and dropped into the lap of those that have done the deed. In fact the unit could be dropped from a very great height onto the nefarious thugs and the person or persons that have done the deed will be in a pretty bad way very quickly.
I am afraid that people like The Putin state must be now classified as a criminal terrorist state. In fact, should a terrorist state have a plane downed with a nefarious bomb on it then it must be delivered back to that state. I am getting very fed up with continued aggressive and violent types that really need "help" to stop their aggression!
"Only the larger 400MW SMRs ones like Rolls Royce are planning would have operators."
Oh yeah, they're going to get off-site operated reactors past the regulators. /s
It's near impossible to even get to talk to the regulatory agencies about any type of "unproven" reactors. That is, any reactor that isn't currently in operation. Not having any staff on site of one isn't going to get approved.
they would be sealed units
I'm curious to hear how the first law of thermodynamics will teach you a lesson. (you can get energy out of a thermal exchanger, provided that you have 2 sources of heat. With a sealed unit, the hot source would probably be the reactor, but what would the cold source be ?).
Therefore, no sealed units, it's a physical impossibility
The thought of a fresh-outa-college kid, a 35 MW reactor, and a loose-leaf folder en-titled 'How to handle a reactor scram' does not fill me with a lot of confidence
As Spazturtle has said, that won't be the case.
The big reactors, and the naval ones, are designed on the basis of having multiple safety systems and trained operators etc. The design intent with SMRs is that they will be fully automated and/or remotely operated and will simply shut down if there's a problem. But the key difference is inherent/intrinsic safety ...
In a PWR, you need a number of safety systems and operator procedures as they are inherently unsafe without them - though a water moderated reactor (as used in naval systems) has certain advantages in that it can't continue operating without the water being present. The SMRs are designed to be inherently or intrinsically safe - so if (for example) an "incident" removes all the external systems, it'll sit there, hotter than in normal operation, but will not "melt down" or release noxious stuff into the environment. The difference between inherent and intrinsic safety is (and I can't remember which way round it is) that with one, the reactor will remain safe but could be scrap because it's damaged, while with the other it'll also remain safe but will still be fully functioning once the issues are resolved.
So, for example, one design uses fuel that's in ceramic balls - in the event of losing cooling, it'll heat up (thus increasing heat losses through it's containment etc.), but the ceramic spheres will have a high enough melting point that they'll remain intact and thus contain all the noxious stuff.
The example you give is pebble bed reactors, one of the SMR concept.
In the Molten Salt concept, if the coolant ( which is also the salt containing the molten radioactive elements ) gets too hot hot, two things are occuring. you need first to understand the concept, which involves the molten salt to be both the radioactive element carrier and the coolant.
In one 'chamber' it will heat ( because there's enough volume there ) and and in another chamber it will give out that heat to a secondary circuit coming back cooler. ( yes that's a lot like a PWR reactor... )
If for a reason the molten salt stops moving ( there's way to have it moving without any pumps needed ) the salt witll build heat in the main chamber and the two things I mention will occurs :
- the first is basic science as the salt gets hotter the distance between radioactive atoms become bigger, so less neutrons hit a radioactive atom... lowering the temperature.
- a built in safety plug will melt at a specific temperature and dump the content of the molten salt circuit in a set of tanks small enough that the fission reaction cannot be kept going. ( think of old steam engines that had that kind of feature when fires got too hot )
No active safety system, no power required to activate a safety system, no way to meltdown, ...
"I still think you would need an emergency cooling system of some kind"
There is and it's completely passive. The liquid in the reactor drains into a holding tank that does not have any moderator (graphite) and the fission process comes to a halt. The holding tank is capable of withstanding the heat of the material as it cools down and solidifies.
The current crop of PWR's have lots of daughter products that continue to produce heat after the reactor is shut down. It's enough heat that externally powered cooling is required. There are passive cooling systems, but in the case of Fukushima, the operator was cycling them on/off for some reason and they happened to be off when the site lost all power and couldn't turn them on again even if they realized they were off. As they didn't test these systems periodically, they didn't know how to tell if they were operating or not without power in the control room.
Just for completeness, SMRs will have just the same issue of secondary decay and the need to dissipate in the region of 10% of previous operating power for a while - it decays quite rapidly over the few days after shutdown.
You can deal with this in two ways. In the big reactors, you have big safety systems. In a small reactor, you can design for passive heat removal together with high temperature materials. Basically, you still have heat to get rid of, but you design the system that it will get much hotter than normal without melting, and it can lose heat (enhanced by the high temperature) passively.
"No active safety system, no power required to activate a safety system, no way to meltdown, ..."
There is also no working prototype of this sort of system which means that regulatory agencies aren't going to approve one. It won't be until they're being mass produced and deployed in China before a study is done on allowing them in Europe, UK and US.
after a nuclear reactor shuts down, the initial decay heat can be as high as 10% of whatever the operating power just was. Over an hour or two, it drops to about 1-2% which will slowly drop over several days until it becomes a low enough level that is easily managed.
This initial decay heat at shutdown has to be removed to ensure safety. So a cooling system would have to be present that can reliably remove high levels of decay heat following an emergency shutdown, and then operate without electricity if need be.
I have seen such designs in U.S. Navy vessels, for emergency cooling. All you would need is some 100% reliable way of sinking the heat, even if it is just a standby tank of water.
But ideally it can be managed with a proper design. I like the idea of bringing in a nuclear power plant on a railroad car. You could literally leave it on the train car (inside of a shielded containment building of course) and just "plug it in" like someone suggested. OK, the cables for 50MW of electricity do not just "plug in" _THAT_ easily, though it COULD be just a scale up from shore power cabling on a nuclear ship. Typical shore power hookup was an hour or two as I recall.
As for waste, the tech to reprocess that stuff WILL need to be developed. Otherwise it becomes both wasteful AND dangerous to just store it. No more NIMBY attitudes blocking it, either. Make it SAFE. CAN be done.
And the co-generation idea of using waste heat from the SMRs to "do things" is also pretty cool. [absorption chillers for air conditioning immediately come to mind]. If half of your data center buildings' heating and A/C came from waste heat, imagine the economic benefits of THAT.
Design PRIMARILY for safety IS more expensive but I think in the longer term it will cost less.
[To the best of my knowledge, nobody has yet properly categorized the decomissioning cost of wind farm equipment and spent solar panels, not to mention potential accidents from old embrittled blades and towers that could potentially come crashing down]
This initial decay heat at shutdown has to be removed to ensure safety. So a cooling system would have to be present that can reliably remove high levels of decay heat following an emergency shutdown, and then operate without electricity if need be.
In simple terms ... If you have thermal mass, and a means of passively losing heat (especially when that thermal mass is hotter than normal), then the immediate decay heat goes into heating stuff up and the heat is lost passively. You do the sums, and make sure that the highest temperature reached is below the melting point of the materials involved.
and just "plug it in" like someone suggested
The reactor comes without any ancillaries. What you "plug in" are coolant pipes (plus some controls and stuff) and all you get out is heat. You may use that heat to run turbines (which are part of the site installation), or for heating things (some industries use massive amounts of heat), or some combination.
To the best of my knowledge, nobody has yet properly categorized the decomissioning cost of wind farm equipment and spent solar panels, not to mention potential accidents from old embrittled blades and towers that could potentially come crashing down
That's an SEP - someone else's problem. Easy to do in the non-nuclear world as you "simply" take the things down and sell them for scrap - what happens to it afterwards is not your problem. I do wonder if the windmill enthusiasts realise that some of the windmill towers have been made from recycled "nuclear waste" - stuff that's not been near anything nuclear is still "nuclear waste", but a lot of it is steel and once it's been checked for contamination, gets recycled into the normal steel supply chain.
Pebble Bed reactors have been proven to be problematic,
Sadly. The problem is in making the pebbles. AFAIR the Germans could make the necessary pyrocarbon successfully but everyone else produced pebbles with a risk of breaking up and jamming the machinery that circulates them.
To be honest I'm worried about the bit before that ... What happens when the .com that owns the data centre has to train it's staff in operating a SMR. The thought of a fresh-outa-college kid, a 35 MW reactor, and a loose-leaf folder en-titled 'How to handle a reactor scram' does not fill me with a lot of confidence.
Yes, but also no. There's no shortage of DC operators doing deals with local suppliers or energy co-ops. Just because they're building something to support a DC doesn't mean it will be run or even owned by the DC operator (whether that's a general provider or a specific .com). And nuclear plants are going to be regulated regardless of who owns them - which may be of a higher or lower standard in your jurisdiction but it's not as though Zuckerberg will be able to ring up a provider, have one delivered and get his interns to run it with no supervision.
As an example of on-site, Slough Heat & Power is a 47MW biomass station in the middle of Slough Trading Estate - which is largely populated by several significant bit-barns. Eyeballed on Google Maps, it's on a roughly 4-5acre site. This is comparable with Rolls-Royce's target footprint for their 470MW SMR, which would offer a much better energy density (and less air pollution).
Of course many of the smallest ones will be delivered as sealed units (i.e. the reactor can't be refueled - it's a "black box" with connectors for the working fluids & coolants) which are delivered, guarded and replaced by the provider with relatively few on-site operators who will be predominantly concerned with maintaining the steam turbines. No routine handling of fuel-rods or materials. Once a decade you swap the whole reactor unit.
... seem usually associated with operating companies which did not maintain the nukes as specified in their manuals, and/or took other shortcuts (did stupid stuff) to reduce immediate operating costs. In other words, the corporations who operate the nuke plants haven't been properly incentivised via jail time and major fines, to not fuck it up. Handing out more nukes to the corps seems a recipe for disaster.
"If you put 40 individual SMRs in separate locations, then you have a lot more operational cost - not necessarily 40 times, but still substantially more. In return, all you save is the relatively minor cost of grid transmission. "
First problem is that transmission: The grid needs to be upgraded to carry the extra power that's being demanded by these data centres. It's the current bottleneck and adding 1 or 10 or 100 SMR's won't help if the grid can't carry the extra power.
So having the power generation at the point of demand, with any extra being fed into the grid does seem like a serious option. Certainly it's more viable than trying to generate all the power needed by the site from renewables on site: While it would be nice, that's a lot of turbines and a need for a lot of battery storage.
Now, if the SMR's got small enough and were safe enough to be fitted into lorries, or vans, or cars...
... next would be power suits... like the T45 and T60...or the X01...
on-site placement gives you the opportunity to use waste heat like a co-generation plant.
Hot water can create chilled water with an absorption chiller [LiBr type for example].
Waste heat can also create hot water to heat buildings in winter, and [pre-]heat hot water for various uses [like bathrooms and mop closets]
Waste heat is probably 80% of the thermal output of the reactor, with about 20% becoming electricity. If you can just use the waste heat effectively it can offset a LOT of electricity and/or carbon-based fuel usage, potentially saving TONS of money.
"So having the power generation at the point of demand, with any extra being fed into the grid does seem like a serious option."
I wonder what the security costs of having an SMR on site would be compared to putting multiple ones on a single site and upgrading, where required, the local power lines and having only one security problem and one set of Nimbys to deal with? In an over populated country like the UK, I'd imagine it might be better to simply add them to existing nuclear sites already secured by Civil Nuclear Constabulary
As with the article, you assume Pressurized Water technology being used in SMRs.
If the nuclear industry ever jump into SMR, they will have to start from a clean sheet, and Pressurized Water Reactors are a bad idea...
Their only advantage were that they were easy,( technically ) cheap ( compared to other technologies ) to build and produced all the wastes useful in making nuclear bombs in enough quantities to make it economical.
There's other ( discarded ) technologies like fast breeder or molten salt reactors that produce a low amount of short half life wastes... and that can recycle the long lasting PWR wastes so that it becomes shot half life waste or actual fuel for the fast breeder or molten salt reactor.
Actually the combination of Fast Breeder & Molten Salt is the best tech for SMR. It can't meltdown ( if the temperature gets too high the nuclear reaction slows downs, and at worst the molten salt can be dumped [through a plug that melt at a given temperature, so no high tech control that can be disabled ] into a storage area where it cools down on it's own ), it can recycle wastes of PWR reactors, since the active elements are in the form of molten salt, it's easy to filter out waste and inject new material... and the recycling of the waste is also easier.
Oh and the best : you can't produce bomb material in large quantities and it's contaminated by isotopes that are hard and extremely expensive to remove.
>If the nuclear industry ever jump into SMR, they will have to start from a clean sheet, and Pressurized Water Reactors are a bad idea...
It's not an "if", it's happening. All the SMRs with any kind of commercial funding behind them (RR SMR, NuScale, GEH) are effectively downscaled PWRs.
The Wikipedia article on Nuclear_marine_propulsion indicates a few liquid-Sodium cooled reactors were used on ships, but that most are pressurized water. This may be at least in part a consequence of the rather peculiar limitations of the military marine environment which include stuff like minimizing pump noise (that can be monitored by an enemy for tracking purposes) rather than technology limitations. Still though, the designers went with pressurized water for what seemed to them good reasons. Could be that pressurized water could turn out to be the best of a rum lot even for civilian SMRs. Or not.
There is that, but I just think AGR is fancier and wish it'd been properly developed (thanks yet again for your "the markets will provide!" mantra, Thatch. <_<). But of course it wasn't, and the small number actually built of the many that were envisioned are now ancient. Bitter? Yes!
Is that just theory or are there any working units or prototypes?
It is the very best technology in the sense that it will be many decades before anyone has anything more specific than PowerPoint decks with pictures of shipping containers and some formulae - which means that there are decades to grift money from suckers and the fossil industry.
If SMR's and Molten Salt reactors could be done, they would have been done it in the 1950's when they had both the tech, abundant enthusiasm for everything nuclear and a more relaxed attitude to accidents.
Except that in the 1950s the people with enough money for nuclear reactors ( governments ) wanted reactors that produced the waste materials needed to make bombs... Producing electricity was just a byproduct of plutonium production.
They had zero interest in developing nuclear technologies that produced almost exclusively electricity.
since the active elements are in the form of molten salt, it's easy to filter out waste and inject new material... and the recycling of the waste is also easier.
If that is "easier" than whatever they tried to do at The Thermal Oxide Reprocessing Plant, or THORP, then the state of current, and well known, nuclear tecnology, is even more fucked up than I thought it was!
Instead you are proposing to build a chemical separation plant stuffed with radioactive goo contained by new materials that we don't know how to make yet, with pumps and pipeworks infinitely more reliable and durable than whever is available, because nobody are going in to fix it, not even robots (of course we couod invent some special electronics for them also, since we are in Dreamtime)!?
This is why the grand idea of small nuclear reactors is a monumentally stupid one. "Oh, but they are just like the nuclear reactors on submarines!" they say... but imagine having to decommission more than the handful of nuclear reactors that are currently in use by the current Astute class and the future Dreadnoughts, it's going to be a mess.
Stick to two or three sites with the massive 4-digit megawatt class that make more sense in terms of decommissioning and economy of scale.
If you had next generation molten salt reactors with strong inherent negative temperature coefficients (if the reactor heats up for some reason the rate of reaction slows) you wouldn't need a team of PhD's watching the reactor and holding its reactivity down with control rods. in fact Moltex energy has a 40MW electric output design with "No moving parts". They still move the fuel rods around, but there are no pumps.
getting stuck with waste is A CHOICE as the technology to get rid of the bulk of waste has existed for decades, but the implementation of it is being blocked by the anti-nuclear sentiment and the pression groups who get all their funding from fearmongering desinformation about nuclear. I live in belgium, we could, if we chose to replace our current fleet with GENIV's like the GE PRISM, and power our country for over a century using our current waste as fuel....in the end 98% of the waste will have been converted to energy, the remaining tiny amount will not contain any isotope with a half life over 2 centuries...
the whole energyproblem has been technically solved a long time ago, but politics and ideologies are preventing the implementation...
There really are a fucktonne of reasons to be cautious about allowing large opaque organisations to become energy independent.
Do we really want companies like Amazon and Microsoft and Google to be allowed to secede from a large part of the economic landscape with an armed security force in hand ?
Certainly in the UK such outfits would fall under the aegis of the nuclear police.
Time to (re) watch Edge of Darkness (the BBC version).
Or worse, any company owned by Musk: the idea of that technological fuckwit messing with stuff and/or having a tantrum is terrifying. Though admittedly most $BIG_CORP types aren't really any less terrifying as they don't really do social responsibility; tax-dodging is one thing, irresponsibly handling nuclear materials is another.
Ultimately government is who will pick up the tab, as they always have for any and all long term liability.
Do you seriously think that Google, FB, Amazon, or Tata Nuclear will actually decommission thousands of EOL SMR's in 25 years time, rather than just let the Bahamas/Lichenstein/Curacao based operating compnaies go bust
"Ultimately government is who will pick up the tab, as they always have for any and all long term liability."
In the US, the government is ultimately liable for nuclear waste disposal and storage. It was set up that way to be sure that waste materials were not being diverted. It was also seen as the best way to ensure that waste was properly disposed of. The downside is the US government hasn't kept up its side of the deal and waste is being stored on power plant sites instead of a secure location. Yucca Mountain may be dead at this point. Lots of stuff gets buried at the Nevada Proving Grounds and there doesn't seem to be a bad reason to at least stage material that is in dry cask storage there for now.
“Do we really want companies like Amazon and Microsoft and Google to be allowed to secede “
Do we really want the government controlling our access to our data in the Datacenter? We as individuals dictate how our resources are used, they govern by the laws we agree with. It is a balance, not always good but a balance and it is called freedom.
An example:
During a recent snow storm in an area that does not normally have it or prepared for it, power was always on in the federal buildings, not a sole there but lit up and warm. The power was cut to hundreds of residential homes causing massive damage from freezing pipes. Depending on the person, they may think your need is a luxury.
After COVID, nobody is laughing at Preppers anymore. Their whole concept is controlling the resources they need to survive.
For those of us that live in households where everyone carries the prepper dna mutation, COVID was like christmas. Didn't have to go to the supermarket once in 2 months. (Saved a bomb too as they were shamelessly price gouging).
Only downside was I didn't get to use any of ammo (or arrows), or even sit on the stoop playing banjo to keep the neighbours away.
"For those of us that live in households where everyone carries the prepper dna mutation, COVID was like christmas. Didn't have to go to the supermarket once in 2 months. (Saved a bomb too as they were shamelessly price gouging)."
I've been accused of being a prepper, but I'm just resurrecting what my gran taught me about canning and preserving. I'm buying fresh foods in season when they're cheap and eating like a king all year long. I could last quite a while without going to the grocery store if I needed to. I've got a big fresnel lens that I need to build a mount for that will be awesome to use for cooking if I need it. It will be fun to mess with if I don't. In the event of any type of disaster, it's important to be able to be on your own for at least 3 days and even better if you can go longer.
Yup. They wear camo and bushy beards everywhere, give advice to anyone who will listen, and will stupidly give guided tours of their bunkers detailing their weaponry and defenses. They want to be known as preppers.
True preppers, such as myself, say nothing. We just do things to prepare that are none of your business. We will not speak of prepping, and will make fun of preppers right along with you. You will never know who we are till the nuke hits the city. We may be your closest friend today, but after it goes down our face may be the last one you ever see as the knife goes in because we decided you have something we need. It won't be anything personal, just that someone's gotta live and we will make sure it's us.
Anon, because you don't need to know.
With their recent moves into Healthcare and now buying the company that makes the Roomba... I am sure that it won't be long before they buy into the Power Generation business and then everything else.
IMHO, by the end of the decade, they will be the only place to buy stuff.
All hail our Amazon Overlords. Then Bezos can become POTUS for Life, after all, AMZN will already own the gubbermint.
Stop the world I wanna get off. Unless Musk, Trump and Bezos head off to Mars on a one-way ticket and save us all a lot of grief.
I look forward to the "Who, me?" episode where the datacentre lights go out, and a techie decides to take a screwdriver to the nuclear reactor to fix the problem, instead of replacing the fuse that's just blown.
Oh wait, he wouldn't be around any more to write that, it'd have to be an "On call".
Oh wait, he wouldn't be around any more to write that, it'd have to be an "On call".
Are your sure? When the data centre is near a populated area, which usually is the location of such facility, then we may no longer have that populated area when the techie is done.
So, neither Who, me? or On call will be relevant. There is no "me" or "you" left and certainly no one left to take the call.
icon, what happens with a populated area when combining nuclear facilities, data centres and techies with screwdrivers.
"Who, Me?" for THIS one, for sure.
"Post-accident calculations, as well as examination of scratches on Rod 9, estimate that it had actually been withdrawn approximately twenty inches" (instead of 4 inches).
/me imagines tech yanking upwards on the thing "This... F'ing... thing... is... STU[@#$%^&^]" because with that final YANK, the rod was blown out of the core in a tremendous steam explosion inside the reactor vessel [and rumors have it that the tech was pinned to the ceiling with part of a control rod through his abdomen after being nearly cooked by highly radioactive steam...].
Just another fantasy. Realities of licensing and opex overheads to run a small reactor make little sense, whether attached to a datacentre or the grid.
To say nothing of the not-in-my-back-yard opposition.
The world generates absolutely bucket loads of data, but information is now becoming harder to come by. We would do well to cease cataloguing the pointless...
I think you may be unduly pessimistic - apart from the NIMBYism.
It may not reach the level of "an SMR in the car park for any large business", but it can certainly reach the level of "SMR farm in the field next to a group of datacentres". The SMR manufacturer will be handling all the design stuff to the point where it's an off-the-shelf (almost) commodity - just build the civils, the SMR vendor brings then along, slot them into place, connect up the coolant pipes. There's no fuel handling, no waste handling, in fact really nothing "nuclear" on the site that isn't contained within a big sealed box.
In simple terms, the idea is that the reactor itself will be assembled in a factory and sent by road - already fuelled - to the site. After a few years when it's ready for re-fuelling, the old one will go back to the factory and a new one dropped in place.
The "sent by road - already fuelled" bit may sound scary - as in, what if someone just nicks one and drives off with it. But in reality this isn't something that'll go on a regular sized lorry (truck to our US cousins) - it'll be a very oversize load, needing a large crane to lift it, and even if you could arrange that, the thieves really wouldn't get very far (such a large load can't travel fast as it threads it's way through obstacles) before being confronted by some rather annoyed nuclear police or soldiers with automatic weapons. Hi-jacking or stealing one of these will be no more practical than hi-jacking or stealing something like the 100+ ton fuel transport flasks that are regularly shipped around the country.
Ask the good folks that tend to the Submarine fleet what they think of the idea of a completely sealed and maintenance free unit. It is the stuff of fantasy and lunacy.
And why would you build SMR on proper nuclear sites when you can build properly sized nukes instead?
The idea is potty, the safety case beyond dubious, and frankly the cost alone will kill it.
Seriously, what does the world need exabyte+ level storage for and bottomless compute capacity? Ever more advertising? Let's invest some time in cutting demand; to cut datacentre requirements rather than potty generation schemes.
And why would you build SMR on proper nuclear sites when you can build properly sized nukes instead?
Because building one reactor costs £10-20 billion and takes 10 years to build a 3.2GW plant.
The Rolls Royce SMR generates 0.5GW and is expected to cost £1.8 billion each, and RR hopes to have 10 operational in 2035. That's ~5GW worth of generation for £18 billion which looks reasonably competitive against the overruns in a large plant. The best thing is that if something goes sideways on building a couple of them then the others are still going to be built and delivering power in the meantime.
Ask the good folks that tend to the Submarine fleet what they think of the idea of a completely sealed and maintenance free unit. It is the stuff of fantasy and lunacy.
Being careful how much I say ;-)
Submarine reactors are designed for a different set of optimisations - "sealed and maintenance free" isn't a design goal, though "fuelled for life" is for the next generation.
And why would you build SMR on proper nuclear sites when you can build properly sized nukes instead?
As already mentioned, "properly sized" (who decides what is proper anyway ?) are mind bogglingly expensive and take a long time to build - unless you can simply ignore any objections like the Chinese and throw them up in something like 5 years from cutting turf to online. And they are rather "eggs in one basket" which makes the business case quite tricky.
Each SMR is individually capable of revenue in a much shorter time-scale than a big plant. As already mentioned, if one has a problem, then the others can still generate revenue. And the on-site requirements in terms of operator skills are significantly less. And the biggie - they will be either inherently or intrinsically safe which no design of large plant is, though some come close through passive cooling systems.
"The "sent by road - already fuelled" bit may sound scary - as in, what if someone just nicks one and drives off with it."
I'm not worried about theft, but road collapse. These trucks would be heavy AF. What happens if the road gives way under one side and the truck falls on its side? A whole lot of road closure.
It's also a much better idea to get the installation done and properly tested before loading fuel. You don't want to be chasing down leaks on the hot side that opened up during transport.
And you don't want to the complication of loading fuel on site - it immediately negates several of the advantages of the SMR.
And what "hot side" are you thinking of ? As something of an over-simplification, it'll come as a "box" with two pipe connections - coolant in through one, hot coolant out through the other. Easy to pressure test after you've lowered (I was going to write "dropped" but some might take that the wrong way !) the unit into it's concrete hole.
And these will be designed such that loss of coolant does not result in release of radiative material - things like high-temperature fuel encapsulation (e.g. ceramic pellets), passive cooling sufficient to keep the temperature below the melting point of said encapsulation, and so on. Even simple things like placing both connection at the top, so in teh even of a total loss of the external systems, simple gravity keeps coolant in the box.
And what about the decommisioning costs...?
Mostly the reverse of the installation cost. So unhook your SMR, stick it back on a trailer/barge/railway carriage and transport it to the recycling depot. Where it'll mostly spend time cooling off to a point where it can be cut up for scrap. Or bury it and call it good. But LCOE are generall con tricks anyway, and rarely compare like-for-like, eg-
And on the optimistic end of this equation, this would put NuScale's reactors within spitting distance of natural gas and onshore wind at a LCOE of roughly $37/MWh, according to the US Energy Information Agency [PDF]. However, solar fares a bit better at $33/MWh.
Virtually no onshore wind costs $37/MWh, hence our every increasing energy bills. Solar fares no better, other than perhaps being able to cover the roof of a data centre with solar panels. Both suffer from the usual problem of being unrealiable, intermittent or just weather dependent. Se-
https://gridwatch.co.uk/Solar
minimum: 0.001 GW maximum: 3.853 GW average: 0.337 GW
For this month's solar. The 'renewables' scam relies on selective use of data. So unless you want a datacentre that only operates part-time, you need reliable power. You already have a power density challenge given the land area needed to generate say 100MW via wind or solar. Then you may want to add 1GW of batteries so it can run for 10hrs @ 100MW. Except you'll need more than 1GW because charging/discharging takes energy. Or you may decide to add 100MW of gas turbines, but then depending on the cost (not price) of gas, it's usually cheaper to run the DC from gas in the first place because you don't have to worry about the weather.
See Texas again for more info, or the US in general. Their 'renewables' generation dropped like a rock, just as the cold weather hit. Luckily Texas learned from their last near-miss and their gas network didn't shut down when the wind dropped.
Absolutely nobody is proposing this. You are tilting at windmills. Which is an appropriate irony if nothing else.
I guess you've been sleeping through all the 'Net Zero' nonsense, the constant calls for us to 'invest' more in 'renewables', or just the cost comparison in the article. It isn't a like-for-like comparison because nuclear, either large 500MW units, or SMRs deliver near-constant, base load generation. 'Renewables', such as wind, solar or tidal simply cannot deliver this requirement.
This is the reason why, from today, all home EV chargers must have " the ability to send and receive information and the ability to respond to signals to increase the rate or time at which electricity flows through the charge point.". The UK doesn't generate enough electricity to support datacentre expansion, let alone millions of EVs. Hence 'demand management' so vehicles won't charge (or may discharge) when supply is tight.
I guess it's the fault of our education system that's sold the public on the idea of 'renewables' that they can't see the fundamental connnection between the weather, and our energy supply.
>It isn't a like-for-like comparison because nuclear, either large 500MW units, or SMRs deliver near-constant, base load generation.
Again, absolutely nobody serious is making this comparison. The high and reliable capacity factor of nuclear is exactly why the UK's nuclear sites are awarded such high strike prices versus the apparently massively cheaper pricing of wind on an LCOE basis. This isn't some genius fact only you have noticed while the public is being hoodwinked. We in the industry are smart enough to know the difference between power and energy thank you very much.
This is an integral pillar of both the UK's current and all likely future energy policies. With the exception of "The Green Party" - who are an irrelevance - every major and mainstream climate and political grouping in the UK recognises the current essential role of gas and the future role of nuclear in providing continuous base load.
Again, absolutely nobody serious is making this comparison.
Except the article did.. Or did you miss that part?
The high and reliable capacity factor of nuclear is exactly why the UK's nuclear sites are awarded such high strike prices versus the apparently massively cheaper pricing of wind on an LCOE basis.
No, again it's a deliberate con created by a rigged energy market. Again, it's not a like-for-like comparison, but it's been used for years to promote useless 'renewables'. Best example is probably the crazy CfD system that allowed 'renewables' to lowball bid, giving claims that energy costs are falling and wind is competitive. Most of the <£50/MWh contracts aren't in force, even though some are operating and happily flogging electricity to generate massive windfall profits. This is spectacularly dumb, and CfDs should come into force as soon as the site(s) start exporting to the grid. BEIS is belatedly looking at closing this loophole-
https://notalotofpeopleknowthat.wordpress.com/2022/12/23/beis-look-to-closing-cfd-loophole/
Of the 3,190 hourly periods from the start of the year to 14 May, the IMRP has been higher than £73.71/MWh in 3,072 (96%) and higher than £94.81/MWh in 2,892 (91%) of them, highlighting the opportunity of delaying a CfD.
Suggesting 10 days from exporting to the grid rather than the current up to 3yrs. It still doesn't solve the rigged cost comparison. Assume a month has 730hrs, a 1MW generator should deliver 730MWh. Wind and solar simply cannot deliver this, because they're weather dependent. So the UK has to pay >£73.71 because we also need reliability and electricity when it's dark, or the wind's not blowing. The UK also makes this worse via other subsidies. So a few days ago, it was windy overnight when demand was low. Windmill spinners then got paid constraint payments, and troughed even more profits exporting heavily subsidised energy to EU via the interconnectors.
If the energy market was simply changed to a simple capacity auction, this problem would vanish. Bidders contract to deliver 730MWh @ £73.71/MWh. If they don't deliver 730MWh, they pay a penalty instead of socialising the problem onto our energy bills. Of course that would wipe out most of the 'renewables' scammers unless they were integrated, ie operate both wind and CCGT.
This isn't some genius fact only you have noticed while the public is being hoodwinked. We in the industry are smart enough to know the difference between power and energy thank you very much.
Sorry, but I call BS. This should be obvious by just looking at the marketing claims. 'Renewables' are so cheap, yet as we've poured more and more of our money into these scumbag's pockets, our energy bills have only increased. As has inflation, because energy is an input cost to everything. And as inflation increases, so do 'renewables' scammer's profits because their costs don't increase, and their contracts are indexed to inflation. The more they charge for the energy, the more inflation rises, and the more profit they make.
With the exception of "The Green Party" - who are an irrelevance - every major and mainstream climate and political grouping in the UK recognises the current essential role of gas and the future role of nuclear in providing continuous base load.
Sorry, but that's bollocks. See Goldman Sach's country leader's comment to the world-
https://www.bbc.co.uk/news/uk-63527460
Prime Minister Rishi Sunak is set to urge world leaders at COP27 to move "further and faster" in transitioning to renewable energy.
Sunak doesn't have to worry about heating his homes, because we pay those bills. Or-
https://www.bbc.co.uk/news/av/uk-politics-63025549
Labour leader Sir Keir Starmer had told the BBC's Laura Kuenssberg the UK needs to double onshore wind, triple solar and quadruple offshore wind to achieve the party's ambition of generating 100% renewable electricity.
Sadly, they're all drinking the same Green Kool-Aid, and Starmer's too dumb to realise that having 3x solar generates 3x0MWh for most of the day. Again, he probably doesn't care because we pay his bills, and if not, the 'Renewables' lobby probably will.
The US Navy submarine nuclear reactors have been an amazingly successful design. But Admiral Rickover, speaking from that experience, told members of Congress in 1957: "Any plant you haven't built yet is always more efficient than the one you have built. This is obvious. They are all efficient when you haven't done anything on them, in the talking stage. Then they are all efficient, they are all cheap. They are all easy to build, and none have any problems.
https://wiseinternational.org/nuclear-monitor/872-873/small-modular-reactors-introduction-and-obituary
SMRs, like nuclear fusion and economic nuclear waste disposal, have been on the horizon ever since Rickover said that. It gets harder and harder as the decades pass, to believe that they will not always be just on that same horizon.
I think things have come along somewhat since 1957.
Back then the naval reactor was new - and as suggested, not yet built and all the problems found during development worked out. These days, we've built, operated, and decommissioned, reactors, of numerous types - and our knowledge and technology have come along so much that the next design of reactor for UK subs doesn't even have a prototype facility like PWR1 and PWR2 had in Vulcan NRTE at Dounreay - it's all been modelled in software by boffins with more knowledge in their little finger than I think I'll ever have. Part of that is that PWR3 is designed to not need refuelling during the life of the submarine - making it rather impractical to build a test facility and run it for (say) 25 years when the design needs to be nailed down and units installed in vessels long before that. As a side note, I once got locked in the control room for PWR2 while it was under construction when I was working on a bit of equipment and the keyholder just went on a break without saying anything !
I believe the Rickover quote is more in the eternal truth of
“The best laid schemes o’ mice an’ men. Gang aft a-gley.”
If experience were so useful in foretelling perfect outcomes we would never see budgets overspent on projects like bridges and buildings - for, after all, we've been doing those projects in a modern manner for well over a century.
Admiral Rickover's quote points more to the eternal beliefs that our knowledge, experience and planning will always trump real-world complications. Just ask J. Bruce Ismay and, far more recently and relevantly, Riccardo Morandi (if both were still alive).
Our computer models do NOT account for all possible scenarios, regardless of our blind belief that they actually can. Our human knowledge, and skill, still are not at that level of development. It is always easier to create a design that looks good in a model - hand- computed or computer generated - than it is to create something that lasts the test that our natural world dishes out
It might not be logical but it is often true. [/Spock]
There's also the minor point that nuclear reactors in a submarine replace diesel engines that are very far from problem free ... in a submarine. Either require trained operators. Similar crew count, different training. Plus the diesels compete with the crew and payload for space (no massive fuel tanks for the reactor) and resources including oxygen to breath.
And if the safety systems fail, a diesel can kill the crew in a matter of seconds - simply by out-competing the meatbags for air. What should happen is that a pressure switch should shut down the engines if the pressure inside the boat drops below a safe level. If it doesn't, and the boat is not able to snort fresh air through the snort mast, then the size/speed of the engines means that the pressure drops below that needed for the crew to remain conscious (or even alive) in a remarkably short time.
I think some of the numbers being proposed here are off the mark. The output of the module offered by NuScale - discussed extensively in the report - is 77MWe, twice that detailed in the hypothetical discussion in the mid part of the article. For reasons of economics and operational practice, the smallest form factor offered is four of these modules strapped together for an output of over 300MWe. Likewise the Rolls Royce SMR plant is going to clock in at 470MWe. SMRs are small but they're still grid-scale nuclear reactors with enough output to power a city, not for a single datacentre.
It will still be appealing to operators Virgina or Ireland or whoever, as a way to quickly scale up low-carbon base load capacity, but it won't be strapped to a bit barn. SMRs on the near- or mid-term horizon aren't the self-contained, container-sealed modules of our dreams, they're full scale industrial sites. It's a way to build fusion reactors much more cheaply and predictably, at a larger scale, not a radical evolution of how they work or will be operated.
The U-Battery which is being backed by the UK and Nederland is 4MW electrical / 10MW thermal and is designed.
These micro reactors are exciting because they will be on site which means they can be used for their thermal output as well as their electrical output, so a glass factory or paper mill could directly use the heat from the micro reactor instead of burning gas or needing to use electricity to generate it
U-battery is a neat idea, especially in heavy industry where the "waste" heat that can be reused effectively. The problem is it's never going to come to fruition, because it's derived from experimental reactor designs from the 1970s. It's been something Urenco have brought up from time to time in interviews, on-and-off for over a decade now. It's not real and has no serious funding behind it. Nobody's going to do the leg work to get a novel reactor design approved when it only yields as much power as a single wind turbine.
(And of course in OP when I wrote "fusion" I did in fact mean "fission". Hic.)
"These micro reactors are exciting because they will be on site which means they can be used for their thermal output as well as their electrical output, so a glass factory or paper mill could directly use the heat from the micro reactor instead of burning gas or needing to use electricity to generate it"
It will probably be better to site the reactor first and then locate appropriate businesses nearby that can use the heat and are big power customers. I'm with you on the co-gen aspect. I can easily envision hectares of hot houses near a nuclear power station growing fruit and veg all year long. I'm not sure a glass factory is a good fit, but a paper mill might. Drywall production is another industry that uses lots of process heat. Food packing would work well in conjunction with the hot houses. Heinz could grow the beans and tomatoes, combine all the ingredients and steam cook the tins all in one place. Even if they need to add some energy for the steam cooking process, starting with 80C-90C water would drop the bills a bunch.
To operate does a SMR need an external source of cold (usually water) along with it's source of heat (fully contained nuclear fission of some type) and the temperature difference ultimately drives a supercritical steam/CO2 turbine which in turn spins a three phase electrical generator ? Or is it something else ?
I guess my real question is, does a SMR need to be located near a large body of water as a source of cold (like a conventional coal/oil/peat/gas/nuclear power plant). Or could a SMR be setup in say a desert and use geothermal cooling with a very large ground loop as the cold source. Or even use Air with a large enough surface area as the source of cold for the turbine.
The name says it all. An SMR is a reactor that is small and it built in a factory as modules that are assembled on site. There needs to be some way to provide cooling, but the small size means it won't need as big of a lake, river or be on the sea coast. Ground sinking the waste heat wouldn't work. They would rapidly saturate the ground with heat to the point where there isn't enough delta-T to keep the reactor going efficiently.
Ireland is mentioned in relation to data centres there using more than 14% (and still increasing) of Ireland's total power usage and restrictions being put in place as a result.
Background info:
https://www.theregister.com/2022/05/04/datacenters_in_ireland_energy_draw/
https://www.cru.ie/document_group/data-centre-grid-connection/
https://www.theregister.com/2022/10/24/aws_irish_datacenter_diesel/
Ten years time and some country is having a go at another. Large nuke power sites are well known. SMRs are designed to be delivered by truck thus their locations will not be as well documented. Some future Putin wannabe wants to freeze a country into submission and so takes target practice at civilian infrastructure and hits one of these that it powering a hospital.
The release of nucleotides hardly bears thinking about.
It will be blamed on the locals of course ... but the wind does not care and will blow this everywhere.
The release of nucleotides hardly bears thinking about.
If you hit their radiology department, that may happen anyway. Nuclear medicine already uses and releases a lot of nucleotides, whether for diagnostics or treatments where radioactive materials are fed or injected into patients. The biggest challenge is still really the social/political one in overcoming decades of anti-nuclear propaganda.
Not Chernobyl.
Bhopal. This is exactly what happened in Bhopal.
4000 died 500,000 injured. Far exceeded Chernobyl.
And are so right, this is exactly how a global large scale private nuclear industry ends up: with thousands of aging, borderline uneconomic, end of life plants spread around the globe.
I also immediately thought of Union Carbide when I read this story. And that is just one well known example but there are countless other examples of companies who ruined the environment around them because of their total disregard of safety standards. There was even an incident here in Pretoria where someone illegally stored explosives in his garage and destroyed some of the neighbouring houses. Nuclear is better left to the experts.
"There was even an incident here in Pretoria where someone illegally stored explosives in his garage and destroyed some of the neighbouring houses"
Even the "experts" screw up. Beirut had a bit of a problem when a massive load of Ammonium Nitrate was being held for one reason or another and improperly stored. That went off like a small nuke. It's not the only example of proper care not being taken for AN by a government.
"There was even an incident here in Pretoria where someone illegally stored explosives in his garage and destroyed some of the neighbouring houses."
There was in incident in Los Angeles where the bomb squad overloaded their bomb lorry with illegal fireworks and the bang wrecked a bunch of homes in the neighborhood. The problem was the bomb squad not knowing what they were doing.
I asked a friend of mine (who has a PhD in the physics and in his early career trained the US navy people on operating the submarine reactors) and one thing that they mentioned was that the US Navy reactors ran on far more highly enriched fuel. To the point of being considered weapons-grade. Regular (large) civilian reactors don't need this level of enrichment.
Anyway, the question in my mind is whether governments would be OK with delivering highly enriched uranium to loads of commercial premises around the world.
"one thing that they mentioned was that the US Navy reactors ran on far more highly enriched fuel. "
The military doesn't have to worry about budgets for nuclear fuel and the higher enrichment lets them operate far longer on a single load. They don't do documentaries about submarine reactors, but I've seen how much of an aircraft carrier has to be removed and it's a huge task to burrow down to the reactor spaces to get to the tea kettle for major service/refueling.
...it's a huge task to burrow down to the reactor spaces to get to the tea kettle for major service/refueling.
I think that's a design challenge that's true for pretty much any propulsion on a large ship. I saw a vide where a large ferry needed a diesel engine gearbox replacing, and it was much the same. First cut a large hole through the decks so the gearbox could be lifted out. I guess that's a bigger challenge for warships given you'd also want to protect your propulsion systems well.
Anyway, the question in my mind is whether governments would be OK with delivering highly enriched uranium to loads of commercial premises around the world.
That I think is one of the design challenges. SMRs seem to be derived from existing naval reactors, but would be less constrained by the operating environment. So they use HEU to get the best power density and fuel life into the smallest volume, ie a relatively compact attack submarine like the Astutes. So commercial SMRs may not need to use HEU, or may use alternatives like thorium. But being nuclear reactors, they'd also need to be covered by the usual operator licences and regulation. So the average AlphaGoo DC operator probably wouldn't be allowed to touch them.
Naval reactors like RR's and the French are also designed so they don't need fuelling for the expected life of the submarine. So if an SMR can be delivered fully fueled with a 20yr run time, it wouldn't need to be refuelled at the customer site. That cuts an enormous cost of providing fuel loading/unloading and storage from the site, and can be de-fuelled once the SMR's been removed. There may still be some risk and additional opex for security to prevent nutjobs trying to do stupid things, but we have the Civil Nuclear Constabulary and they're armed.
Personally, I'm not convinced SMRs for 'a' datacenter is a good idea. But consider somewhere like Slough and an RR-sized SMR could probably power their DC's and much of the surrounding area. Which I guess makes the economics more interesting, like who pays for and operates the SMRs. Equinix and their customers would be happy with lower cost/more reliable energy, but probably won't want the cost and regulatory hassle of operating one themselves. So it seems to make more sense for existing energy companies with nuclear experience to do that, eg EDF. The economics just seem to work better as 'muni' energy generators, powering urban areas. Especially if we continue the policy of decarbonisation.
As Threlkeld points out above "Any plant you haven't built is more efficient than the one you actually build." (Rickover-1957).
I'd strongly urge anyone who buys into this SMR fantasy uncritically to read the Wikipedia article on Nuclear_marine_propulsion article from Wkipedia. There's a lot there. Including the fact that commercial demonstration projects using marine SMRs **ALL** failed ultimately due to high operating costs. In at least one case, the reactors were removed and replaced with diesel.
Data point. Way back in the 1960s, the US Navy was concerned about the amount of resource that was required to supply McMurdo station in Antarctica with diesel. The fuel -- copious amounts as one might expect -- had to be hauled in during a rather short period each year when shipping could safely approach the base. So they installed a very small nuclear reactor. After a couple of years of fighting with the thing, they yanked it out and went back to hauling diesel in. Diesel was less aggravation.
=========
Why am I skeptical about SMRs?
1. While shipboard reactors have a reasonable safety record, they are also operated by highly trained crews. Do I believe that the MBAs running AWS, Google, etc are going to pay for appropriately trained personnel to run a bunch of boxes that (most of the time) run themselves? Not if they can avoid the expense, they won't. And I imagine they can find a way.
2. If you think siting and permitting SMRs is going to be one whit easier or faster that siting and permitting a full size power plant, you need to share whatever you are smoking. I'm pretty sure that in either case you'll be able to count on most of a decade supporting ravenous packs of lawyers before the papers that will actually allow you to break ground are in your hands. Is it better to do that once for a 1200MW installation or 20 times for twenty 60MW SMRs?
3. And, BTW, do you seriously think those 20 SMRs are each 20 times safer than the single 1200MW plant? Because they probably need to be. If your reactor containment fails, I doubt it makes much difference what the nominal power output of the plant is. Speaking of containment, do SMRs have any?
4. Who is going to fuel the SMR and haul off the waste? Amazon? I didn't check, but I doubt Amazon is currently peddling enriched Uranium.
5. And then there's waste. A lot of it apparently. As far as I can tell, any realistic hope of rationally dealing with nuclear waste in the US died with the demise of the Yucca Mountain repository. Am I wrong about that? I hope so. If so, feel free to educate me.
6. Who is going to insure these SMRs? My guess, No one.
7 ... And we haven't even got to issues like nuclear proliferation.
Thanks for spelling it out in detail. The issue I have can be summarized as "Civilian". I don't want civilians (including me!) anywhere near a nuclear reactor. Under military or paramilitary control might be justified depending on the risk vs reward. Adding civilians to the equation raises the risk unacceptably.
Be careful here, what do you mean by 'civilian'?
Do you mean, someone not properly trained - in which case I agree with you? Or do you mean someone who is not part of a military or paramilitary organisation, in which case, do you not think said non-civilians might just be susceptible to the 'Yes SIR, no SIR, three bags full SIR' mentality, ie follow orders regardless.....
So why do you think that properly trained and qualified 'civilians' are more problematic or less capable than, what , non-civilians'?
Yes, I really meant military or paramilitary; thanks for the benefit of the doubt. I can see that is an unpopular opinion, but I'm sticking by it. The US Navy has had no nuclear accidents and that is directly attributable to Admiral Rickover. Discipline is a good idea in specific venues. Private enterprise is a wonderful thing, but optimizes for profit. I want nukes secured by soldiers in a separate hardened area, and maintained by people that understand what court-martial means.
No, if you need a miniature nuclear reactor to prop it up then it isn't sustainable.
Most of the stuff that huge MS, Google, or AWS datacentres do could have been done locally on each site, PC, or mobile where any outage affects relatively few people and not everyone. Only artificial centralisation brought about by the push to move to a subscription model has made datacentre outages such a hair-on-fire disaster when they happen because centralisation means everyone is affected.
Centralization of resources started well before the advent of Google, AWS & others, with the beancounters pushing to remove on-site servers to concentrate them in datacenters.
It may be fine for application servers, but just imagine what will happen when you have a secured LAN that relies on communication with your Active Directory servers to allow access, and only 2 AD servers available for 130+ sites all over the world, including offices and factories...
"Hello, is that Mr Alan Quaeda ? Hans Gruber here. You remember you were wanting to build some dirty bombs and set them off all over the decadent western liberal democracies ? But you were having trouble finding the radioactive material to duct tape to your C4 ? Well, have you ever heard of SMRs ?"
"No, tell me more"
"They are small stores of fissile material already in place next to critical infrastructure, at multiple sites. You don't have to do any preparation that is likely to leave you glowing in the dark, you just turn up with a few boxes of semtex, glue to them to the side of a thing like a shipping container, light the blue touch paper and stand well back."
"No, this cannot be, you are an infidel trying to lure me into a trap. Such things will have 24 hour guards with multiple teams, probably armed, with constant active monitoring".
"You'd think so. But you see, if they did that, it would ruin the economics of the project and nobody's stock options would vest".
"Truly, the capitalists will sell us the rope with which we shall hang them".
And, your point is absolutely valid and true.
In reality, so called 'dirty-bombs' have very limited 'strategic value' the destruction they cause is limited to the damage due to the conventional explosive. The radioactive contamination will be restricted to the immediate area, and this can be cleaned up, but at considerable expense, both immediately ie paying specialists to clean it up, and as as a by-product of nobody wanting to be in the immediate area for a while, which may have an economic impact.
I have heard it said that a 'dirty bomb', is the ideal 'terrorists weapon' in as much as it doesn't actually cause too much damage or even casualties - what not does do, is cause considerable economic damage, and well basically , I quite'terrorise', the people!
So realistically stick a load of SEMTEX on one of these things sitting at the middle of a datecentre and set it off - realistically, what will happen is that you blow all the windows out, oh, but DCs tend not to have windows anyway, you have a power failure (presumably mitigated by alternative supplies), and the surrounding area is radioactive.
The data centre carries on as if nothing has happened - the problem is getting staff into the building if necessary, because the obvious response is going to be 'you want me to go where - screw you boss, I quit'
You know, maybe we need to employ teams of greybeards, who a) actually understand this stuff more than then supposed experts, and b) aren't too bothered about radiation exposure on the grounds that they are all in their 70's anyway, so if they are likely to die in 50 years time due to the exposure, well so what?
Note that I am not really advocating the creation of 'suicide squads' of greybeard IT experts, who are paid stupidly excessive sums of money to just be on standby and spend the rest of the time in Tahiti, waiting for the call - although come to think about it....
The radioactive contamination will be restricted to the immediate area, and this can be cleaned up, but at considerable expense,
Though, rather than see it cleaned up, I would expect to be seeing endless government ministers trying to convince us "it's just like winter flu" as they convince us "living with fallout" is just fine and dandy, as they have with covid.
I would have loved to have added a Joke Icon, but ...
The technical arguments are interesting, but aside from those, why would there be a need to site an SMR next to a datacenter rather than the other way around? I recall a story where a bunch of data centers were built in one area and they couldn't build more due to power transmission constraints. Why do all of the data centers need to be in that one place? A couple of decades ago there was a race in the US to install long distance fiber optic cable until engineers became much better at stuffing more and more data down each fiber. That lead to an excess of fiber capacity and several companies going bankrupt as they couldn't attract customers at a profit. There are still corridors with dark cable and lots of inexpensive land. If there is also enough power, Bob's your uncle. I'm not seeing what problem SMR's are solving that couldn't be remedied some other way.
And, you know, that's fine - unpopular options are equally valid.
But, please tell me, what is your alternative solution, and if said solution doesn't work, lights are going out all over the country (whichever country you happen to live in), people are dying in hospital because the power has gone off, and 'the mob' are demanding that 'something be done'- and that includes a massive increase in Nuclear Power stations NOW, and to hell with any and all safely issues - oh and anyone who disagrees is likely to find themselves hanging from the nearest lamppost.....?
Because, I promise you that simply kicking the can down the road indefinitely, and hoping that 'the next lot' in Government will sort it as 'it'll be their problem' - really isn't going to end well.....
The issue of SMRs creating 30 times the waste of a gigawatt scale power station is, perhaps, ameliorated by two future factors:
1) some gen 4 fission reactor proposals can be run from the waste from these SMRs and older plants. Overall the amount of net waste at the end goes down
2) fission may only be a relatively short-lived "bridge" now until the boffins get fusion to work properly and at commercial scale
Public and private investment is pouring into Fission both large scale and small modular on both sides of the Atlantic, but especially in the US. Similar can be said for fusion research.
I don't think we have a flotilla of energy Theranoses about to run aground out there. Investors are a little more scrutinising these days and while many startups may be chasing pipe-dreams, a few will eventually succeed.
Out of curious interest and hoping that someone on here can contribute, I do fully understand how, in theory a 'fission' reactor could be fuelled and 'sealed' and have an operational life of, years, decades!
Now imagine that fusion power becomes commercially possible, be it laser implosion or magnetic confinement, furthermore imagine that this could be miniaturised into a small package.
My question is, how much fuel would be required to sustain a fusion reactor, producing, say 300 MW for a lifetime of five years?
Is there any conceivable system that could make this a self-contained, no maintenance or refuelling system required over this lifetime?
In terms of fusion fuel, sure: less mass than you would need fission fuel for the same output. If you were fusing Deuterium and Tritium and had a closed loop Tritium cycle via a Lithium breeding blanket, you could stockpile enough Deuterium and Lithium and your Tritium seed load.
With a "solid first wall" reactor vessel you are probably going to have to do some maintenance well before 5 years though. There are some liquid first wall ideas out there like First Light and General Fusion which, if they can be made to work, would massively extend your maintenance intervals.
But I am not sure why the need for a sealed, set-and-forget fusion reactor. The materials and radiation levels involved are nowhere near as monster evil bogeyman as in fission reactors, and the nuclear proliferation issue is pretty much non-existent too.
We're straying further off the topic of the article now though.
I've been floating the idea that future charging stations will have a small nuclear plant next to them. Charging levels are starting to hit the 1MW per car in a couple of years and you need at least 20 chargers in an electric charging station to make it economically viable. These SMR (small modular reactors) from Rolls-Royce could come in handy.
"Charging levels are starting to hit the 1MW per car in a couple of years "
Battery technology isn't advancing that rapidly. A couple of things limit the amount of power you can shove into an EV and a biggy is the size of the lead. The cable from the charger can already be rather heavy for some. Even doubling the current will mean cables that a smaller, less strong person my find impossible to use. It will also put more strain on connections so a new standard will have to be put in place. More current will also mean more chances of big problems.
If EV's become able to accept 1MW of input, charging times would drop to a few minutes. That would mean that needing a 20 stall station also becomes moot. If you can go from 10-80% in 6 minutes, you don't even have time to visit the loo and top up your coffee. High power public stations are only really useful for travel corridors that see lots of traffic. Many day to day users now rarely use public charging. EV's are expensive so many that have them also have a home with off-street parking and being able to charge at home is a big advantage of EV's. It will take time, but as EV's are owned by more people, there will be more accommodation for charging in places where people don't have off street parking at their home. More businesses will offer charging for their employees (slow, but you are there for hours). I know the offices for a big ambulance chasing firm near to me have charging for all the pretentious blood sucking lawyers they have on staff.
One of the problems in the past with this was that many of the designs were breeder reactors. Which is great for operational efficiency but essentially means that the waste is close to being weapons-grade, breeder reactors have effectively been kept off the market for that reason.
This is really the nub of the problem: as long as costs, regulation and dependencies are uncertain, investment will be minimal. In the meantime, in many locations renewable solutions + storage are already available and the costs continue to decline. There is even a solution for cooling using solar thermal adsorbtion.
But the nuclear lobby will continue to, er, lobby.
Oh, and El Reg, operationalize, really?
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
