ThorCon – Demonstrated Molten Salt Tech Packaged With Modern Construction Techniques
The dearth of real innovation and focused direction from the established companies in the US nuclear industry in the face of rapidly expanding demand for clean energy solutions has stimulated the formation of a number of start-up companies. The leaders of these companies have backgrounds that have taught them to ask “Why not?” when faced with the standard nuclear industry response of “The NRC will never let you_____” (fill in the blank).
ThorCon is a packaged nuclear power plant concept from Martingale, Inc. that is designed to wring capital costs out of nuclear plant construction. The company visionaries have recognized that the biggest hurdles to building new actinide-fueled reactors are the initial capital investment along with the excessive required construction lead time.
Instead of complaining that “the market” does not reward carefully crafted works of industrial art designed to last for sixty to one hundred years with lucrative paybacks delayed for three or four decades after final investment decisions, the ThorCon design team started with the notion that product designers must create offerings that satisfy market demands.
Today’s energy market rewards financial flexibility, predictable construction schedules, reasonably low investment, affordable operating costs, low or no emissions, and readily implemented upgrade paths. If the offered solution is one that uses actinide fission, customers will also want to clearly understand provisions for handling process leftovers, liabilities, accident prevention, consequence mitigation and regulatory barriers.
The ThorCon conceptual design has several features that will be familiar to regular Atomic Insights readers. They are similar to the choices that Terrestrial Energy has made. Both design teams point back to the Molten Salt Reactor Experiment as a demonstration that molten salt reactor technology works and addresses many of the economic obstacles inherent in conventional light water reactors.
They both have chosen to design reactor systems that are maintained by replacement instead of being designed with expected component lifetimes measured in large fractions of a century.
ThorCon’s reactor modules (referred to be the company as the Can) are low pressure, simple systems manufactured with low cost, commodity materials. Two of the four gas-tight, fission product barriers are included in the Can. Each Can will operate for about four years. After operation they will be allowed to cool for four years and then will be shipped to a centralized facility for decontamination, inspection, and refurbishment.
Both ThorCon and Terrestrial Energy has chosen to locate most of their nuclear systems underground. For ThorCon, the top of the silo will be 29 meters underground, leading to a need to excavate a 50 meter deep hole.
Correction: Terrestrial Energy’s design is “at grade,” not underground. As is the case for any building, there will be a foundation that extends below grade.
Aside: I don’t like the idea of building nuclear plants deep underground. Digging deep holes costs substantially more than is generally estimated by people who have not built major underground projects and it adds unexpected maintenance costs. Building underground also adds site specific requirements that will hamper the ability of suppliers to mass produce identical units.
One additional reason that I am skeptical about underground siting is that the major proponent of the concept was Edward Teller, a theoretical physicist whose memoirs brag about his lack of practical engineering skills and experience. Teller was also the first chairman of the committee that initially saddled the power reactor industry with the almost impossibly expensive requirement of perfect safety. End Aside.
There are significant differences between ThorCon and Terrestrial Energy in other aspects of the design and development plans, with each team building on different core competencies.
Jack Devanney is the principal engineering designer for ThorCon. He has several decades worth of experience in modern large ship manufacturing with a deep understanding of complete project economics. He has led teams that produced high quality products for demanding customers that operate in some of the world’s harshest environments, yet he learned how to satisfy their demands for effective cost and schedule controls.
The block construction techniques that Jack plans to incorporate from the ship building industry are not conceptually different from the modular construction concepts being incorporated in both large (AP1000) and small (NuScale, mPower, Holtec, Westinghouse) light water reactors. The main difference might turn out to be his team’s experience in having successfully implemented blocks on large projects already.
The ThorCon team also includes people with financial savvy, exceptional software skills, education, marketing, and nuclear engineering.
Unlike Terrestrial Energy, ThorCon has not yet decided where it will license and demonstrate its design. It seems unlikely that ThorCon will choose to operate in the US, but there are strengths available here that should be considered.
- There are US shipyards that have experience in modular construction. They might be too costly when operating under Navy rules, but it’s possible that they could apply their skills differently for a different type of customer.
- There are several locations — Idaho, Hanford, and Savannah River — that are ideally suited for demonstration plant siting. There are political challenges associated with each of those sites, but there are also political opportunities.
- One aspect of US basing is the presence of eager, trainable work forces who already have substantial nuclear experience.
- There are political leaders who have recognized that pushing innovative nuclear energy projects off-shore is bad for business here in the US.
- There is still a lot of work to be done to turn this recognition into a supportive development environment, but the US has experienced leaders from other industries that can attest to the fact that this is a pretty good place to do business.
Add one more ‘T’ word to your list of companies to watch.
(Here’s my current list of nuclear plant design companies with a significant US presence worth watching – Areva, Flibe, GE-Hitachi, Holtec, mPower, NuScale, TerraPower, Terrestrial Energy, ThorCon, Transatomic Power, Westinghouse. Some of those names are worth watching for negative announcements. I expect you to let me know who I’ve slighted by failing to include them in the list.)
Update (January 6, 2015 08:20 am):
Martingale issued the following press release:
Martingale reveals a bold approach to solving the global issues of poverty, pollution, energy security, and climate. The ThorCon liquid-fuel nuclear reactor design is detailed at thorconpower.com.
ThorCon is a complete system of power generation modules, interchange maintenance, and liquid fuel service that produces energy cheaper than coal. Principal engineer Jack Devanney led a four-year skunkworks project that has created a new kind of nuclear power plant, integrating proven technologies with breakthrough approaches to manufacturing and licensing. Production can start by 2020. Today Martingale is publishing its design for cheap, reliable, CO2-free electricity at thorconpower.com.
Former MIT professor Devanney’s background in shipbuilding created respect for low-cost, high-precision, block-unit manufacturing at Korean shipyards. He saw how such prefabricated blocks could enable production of enough nuclear power plants to make a global difference, a hundred a year.
Author Robert Hargraves writes that selling so many power plants requires clear, simple economics, cheaper than coal. Coal is today’s energy choice of developing nations, now planning to build over 1400 gigawatt-size coal power plants to enable their economic development.
Lawrence Livermore Lab veteran nuclear scientist Ralph Moir says that today’s nuclear power industry is wedded to expensive solid-fuel nuclear reactors, even though the simplicity of liquid fuels was demonstrated at Oak Ridge National Laboratory. Moir and Devanney modernized that design for mass production. ThorCon uses uranium and thorium fuel dissolved in molten salt to create a power plant that makes electricity cheaper than coal.
Stanford engineering alums Chris Uhlik and Lars Jorgensen contributed to the design of passive safety functions that operate without mechanical or electronic controls, even with no power. The reactor is 30 meters underground. Overheating drains the fuel salt from the reactor. There are four barriers between the fuel salt and the atmosphere. ThorCon is walk-away safe.
Taking another lesson from Oak Ridge, Martingale advocates a return to staged testing of physical prototypes for new nuclear reactor designs. This made the US the world standard for nuclear designs in the 1960s. Martingale supports adoption of the same license-by-test model that has enabled US leadership in aviation and drug discovery.
Martingale is designing ThorCon in the US while targeting its first installations in forward-looking countries that support technology-neutral nuclear regulations and see the benefits of the license-by-test process. ThorCon opens up a practically limitless supply of low-cost, reliable, carbon-free power by 2020.
Media Contact
David Devanney
(772) 285-2210
info@thorconpower.com
End Update.
Thanks for this review, Rod. I would be curious to read a review by you of the following, if time permits. http://www.upowertech.com/
Rod, Here we have an attempt to embody several of the concepts that I have been talking about since 2008 on Nuclear Green. of course I do not claim originality for the concepts, but maybe I was the first to see how they could all fit together in implamenting MSR technology. I disagree with you about underground settings. Abandoned salt mines could offer attractive, safe, and low cost homes for MSRs. I looked at the question of underground reactor settings, some time ago. I had frequently used and underground parking garage, and noted that the bottem level was as deep enough to house a MSR. I also saw the basement being dug to a level that was deep enough to house a MSR. %0 feet excavation is not expensive, certainly less expensive than the massive cement anf steel structures that form the alternative. And just because Edward Teller thought of an idea doesn’t automatically make it a bad idea.
@Charles Barton
I’ve been in a number of underground garages as well. 50 meters is more than 15 stories; the deepest garage I’ve ever been in was five levels below grade, not 15. Admittedly, my experience might be too limited.
Using available underground mines limits the location possibilities considerably.
There is no need for massive concrete and steel structures when all radioactive portions of the system are operating under low pressure and will change from liquid to solid in the case of a leak.
Rod we don’t need 50 meters for safety. This will do. http://www.military.com/video/guided-missiles/launching/missile-silo-for-sale/659487117001/
MSRs are small. The high capacity power lines have to be built, to move electricity from the great planes to market, and wind needs a reliable low carbon, low cost back up. A MSR in a Silo would probably be safe without earth cover, but might be safer with 30 feet of earth on top of the reactor. OIn addition to underground silos, old salt mines would be excellent underground housing.
http://io9.com/5911837/the-vast-abandoned-salt-mines-that-lurk-beneath-detroit
Also see here.
http://www.travelks.com/listings/Strataca-Salt-Mine-Tours/1031/
@Charles Barton
This post is not about “an MSR.” It is about a specific MSR design called the ThorCon. The people who have developed that design have published their initial design concept including numbers and dimensions. Those are the basis for my comments.
Rod,
There is much civil engineering experience with deep pits excavation. Every tall building needs to go down, or it won’t be stable. This includes PWRs with 60 meter tall shield buildings or reinforced concrete. The civil engineering part (not just excavation but also anchoring, piling etc) of the design is always difficult to standardize, as it depends on too many local variables, but the costs for open pit excavation are quite reasonable. When you go totally underground (closed excavations) that is often more tricky and can lead to delays and cost overruns with many surprises of the negative kind.
As for shield buildings of some kind, radiation shields are always needed, so a lot of mass (whatever the material) is always needed. Apart from that protection is needed from external events. While the salt is not volatile and chemically stable, it can still be dispersed and make a big (albeit local) mess if you crash a jet or storm produced missile into the building and your building is merely a sheet of steel. The salt will not turn into a solid if there is no cooling, and cooling systems (even passive ones) require shieldig against external events of various types. We can discuss whether aircraft crash and such should be considered but early on Jack Devanney made it clear that very extreme design bases will be considered as part of the Thorcon philosophy. I think Jack is a firm believer in Murphy’s Law, I’ve always been more into the whiskey, myself.
@Cyril R
A sheet of steel might not be sufficient, but a couple of sheets might be.
There is always some form of cooling available unless you intend to purposely insulate the system like a thermos bottle. Rapid solidification is not required to keep fission products in the salt mixture and to prevent radioactive material releases beyond the plant boundaries.
Its a GWe scale reactor. It has decay heat to manage, a non trivial problem. The salt low volatility helps a lot but losing both main heat sinks is a really bad casualty. At least some heat sink must be available. The heat sink end is problematic because of its interface with the environment. So some stout shield of sorts is needed. Rapid solidification is not required (indeed it is better to avoid it from operational reasons) but the containers, including containment, can fail upon overheating. There are also some parts in Thorcon that do have significant volatility, notably the offgas tanks. These need physical protection.
It would be bad engineering design to have the plant dent its walls after a hurricane or other more or less expected event. Still the outer building is quite lightweight.
There is not much use arguing. Other reactor offerings have physical event shields of one form or another. Thorcon doesn’t offer a lower standard of protection than such other proposals.
I’m also wondering why a Thorcon reactor assembly requires 50m depth. Is it to placate folks who might otherwise see “danger”? Or is there a real danger that requires deep excavation?
Some things to consider with respect to your question.
The ultimate backup cooling system is the membrane wall, a passive natural circulation system. This in turn has a giant cooling pond as its ultimate heat sink. That pond is designed as a hardened, below grade structure. That is one of the things, along with driving head for natural circulation, that sets the height. Then you have equipment like drain tanks, steam cells and the like. Some of this equipment needs to be vertical (drain tank) or it won’t work. It all adds up. Below grade is a natural choice here. Otherwise you end up with the BWR style cooling pools at 50 meters above grade. Which is less attractive for seismic, leakage, refilling etc reasons.
Like Charles, I’ve also talked about some of these concepts 5 and 6 years ago. This was picked up by the nuclear advocacy site coal2nuclear.com later concerning the use of all these abandoned dry-docks and machine spaces at America’s almost dead ship year industry.
By using existing shipyards with these large dry docks and over head gantry cranes both, machine ships, storage and foundry facilities, once upgraded to nuclear standards (and some cases they already have the equivalent of a “N-Stamp” designation for building the nuclear navy) both the reactors AND the means of transportation can be build simultaneously.
By using existing brownfield sites…coal plants in particular have great access often to rail, highway and waterways, sometimes all 3, along with existing cooling water licenses and grid access, these new plants can be produced entirely in a factory/ship yard setting and simply floated into place.
I do share some of your concerns about undergrounding these plants. There are also safety concerns in case of even a standard industrial accidents. but they also have some engineering advantages as well. Structural support can be done horizontally as well as vertically which gives some flexibility in architecture of the plant. It also makes them almost absolutely air-craft accident proof which while not a real issue IMO, goes to satisfying the public’s concern. just say’n…
David Walters
Rod,
You might consider editing your list to provide links to each of the companies you listed. Your list looks pretty solid at a glance. Gen4Energy hasn’t seemed to be very active lately, but they might be worth adding.
http://www.gen4energy.com/
I started a similar list over 2 and a half years ago, at bullet point 3 of the following post.
http://entreprenuclear.blogspot.com/2012/04/2-nuclear-innovation-and.html
I am glad to see more molten salt development, it has a large amount of design space – whether molten salt cooled, molten salt fuelled and cooled, and fuel cycle possibilities. The underground designs bother me for the same reasons mentioned above. The only benefit of being underground is defense against airplane strikes, but I believe above ground structures can be designed to defend against that threat but are far more economical compared to a containment building. The molten salt reactors do not have the H2O high pressures of LWRs and therefore do not need the enormous volume and thick walls of a containment building.
With a lot of luck, perhaps the new Congress can direct some funding that can help molten salt development in particular, and of course fission power plants in general.
The cost of digging into soil would be less than digging into granite or some other hard bedrock, so I suspect these would first be sited where the digging was cheapest. Perhaps that depth specification could be adjusted later for harder sites. I imagine some level of attack resistance is going to be a factor in future nuclear licensing (and in a world where certain enviro-nuts shoot bazookas at nuclear power plants, that isn’t entirely unreasonable) and digging in seems like a cheaper and easier-to-defend option than surface hardening and larger security forces. Even at that, countermeasures against malicious moles tunneling might also be needed.
@Shuttlebug
Until some good conversations with civil engineers, I had the same impression as you about the cost of digging holes and building structures in the resulting hole. Then I learned enough to be curious about soil pressures, reinforcing requirements underground, water intrusion, and a few other factors that escape me right now.
My general impression was that civil engineers can be bemused by assumptions made by mechanical or nuclear engineers in the same way that naval architects can be bemused by estimates made by landlubbers.
Digging holes in the ground is not all that easy. I cite the Alaskan Way Viaduct project in Seattle. The project has had many delays, including the failure of the “Big Bertha” boring machine. The latest problem is with nearby building settling several inches deeper on their foundations.
Of course, many of the problems relate to the tunnel location being constrained to a built-up urban location close to the shoreline.
My intuition is that hardening the modules to withstand soil pressure and groundwater would not be especially more difficult or expensive than hardening them to withstand aircraft impacts and rocket attacks. But I would be interested to see engineering arguments to the contrary.
I’m not sure what the dig costs themselves would be, but I gather the idea here is to compete with coal. Mountain-top coal mining alone has excavated an area about the size of Delaware (some 1.4 million acres) involving the removal of up to 400 ft. of overburden just to reach seams of coal so thin that they aren’t economic to extract by tunnel mining. And after you add the dig cost of removing all the overburden to the thin seams of coal below, coal still comes in as one of our cheapest sources of energy.
Even if all 25 acres at each reactor site was excavated down to 150 ft. this would be a miniscule amount of excavation compared to coal. And the dig cost would not be amortized over the energy output from the thin fuel seams below but over the staggering lifetime output of an entire nuclear reactor facility. You would need a fiscal microscope to find what portion of each kw hour paid for the dig.
Also, burying the reactors constrains access to them. That makes them easier to secure. That could translate into an ongoing savings in security costs which could be deducted from the dig costs.
Burying the reactors to this depth may be security overkill, but if it helps to sell the concept to the general public, seems to me that makes it a dirt cheap bargain, as marketing costs go. And I suspect the average customer would be happy pay the extra .001 cents (or less) per kw hour for the added peace of mind.
Interesting comparison with coal. Of course with coal you just dig out the stuff and be done with it. With construction works the challenge is not the cost to dig a hole, it is the cost to make a structure that will last for 40, 50, perhaps 60 years or even more.
So far the estimates for excavation costs are quite low and innovative methods are envisioned to be used for the retaining/membrane wall and connections. There is no need for horizontal access in this plant topology. The modules are lifted in so the access is vertical. Costs for excavation have been increased recently after consultation. Still low.
In terms of engineering, hurricane force wind loads, that have to be combined (by code) with simulaneous max seismic design basis loading, on a 70 meter tall structure are very serious. In this sense undergrounding helps a lot. The center of gravity is below ground, the ground itself provides both buckling and shearing support for the membrane/retaining wall, and there are obviously no wind loads other than vertical (like tornado type) but those are only compressive so less of a fuss.
One limitation of the Thorcon model is that the modules are very large and must be transported by barge, meaning that sites are restricted to shorelines, as far as I can determine from reading their website. I am also leery of the very deep excavations that are proposed–combined with the shoreline requirement.
Indeed, Nick’s point was my first thought: if you put a powerplant near navigable water for construction reasons, going 50 meters down means a huge overhead in groundwater seepage & flooding protection. Thorcon’s leader Devanny has extensive experience in shipbuilding, so the civil engineering side of things, at least on this point, may have escaped him.
I would have preferred to see a three-pronged approach to the civil engineering side:
1. Fully aboveground, suitable for sites (think Bahamas, Louisiana) with shallow water tables. This would require extra aboveground protection building.
2. Half-depth belowground, with grade level being about the top of the silo. This would also require aboveground protection building (but not so much); and the excavation load would be roughly halved.
3. Fully belowground construction, as in the website proposal. Given likely watertable requirements, I suspect this might be the least popular option when siting is actually contemplated.
I guess I have to rise to a point of personal priviledge.
The design has been reviewed by a civil engineer who has built
large undergroung structures (mainly sewage treatment plants)
in the Middle East. His comment was increase the assumed excavation cost
which we did, but of course this will be very site dependent. In
any event, it is not a large portion of the overall cost.
The building is heavy enough to withstand hydrostatic uplift without
any drain field. That”s one of the reasons for the concrete in the
wall cells and the very thick roof. The spec for the roof which is bascially
a tanker double bottom turned upside down is 1 bar over pressure,
so we can take 10 m above grade flooding. This could be beefed up
if necessary at a particualr site. The buidling itself is basically a ship’s hull
watertight and gastight.
The depth is set by the requirement to have two nearly independent, natural
ciriculation decay heat paths: one from the membrane wall to the pond,
the other via the three normal loops to the TG condensers. When you
work backward from the latter, you can not lift the building higher,
even if you werent worried about airplane strike.
The ThorCon blocks are indeed far larger than can be moved by raiil or truck.
But that is absolutely essential if we are going to achieve shipyard productivity.
in a shipyard, 95% of the man-hours are a the block and earlier stage,
only 5% is expended in the actual erection in the building dock. We
must have similar numbers. What Westinghouse calls modules would
not qualify as sub-sub-assemblies in a good shipyard.
We do have a floating varient, but the water depth requirement
would have limited us to a few fjords or forced us way offshore
which generates a whole new set of problems which we decided
not to face in the first generation.
Let’s let a 1000 flowers bloom and see who wins.
@Jack Devanney
Thank you for taking the time to clarify some of the basis for your design choices.
Perhaps my resistance to underground construction is rooted in my status as a native Floridian where even basements are a challenge. However, you have done the homework I recommend to any of the 1000 flowers that I am trying to fertilize and water to encourage abundant blooming.
Good luck — and I mean that with all sincerity.
@ Jack Devanney
An impressive concept and design.
One question: in light of the desirability of pure Li-7 for the “eventual” salt mix; and the value of Tritium; why did you choose to use boron carbide as a neutron adsorbent instead of lithium orthosilicate?
Tritium may be valuable when already separated and partitioned and packed to sell. Increasing it deliberately in a neutron shield material is not a good way to go.
Silicates are also generally not compatible with fuel salts, whereas boron carbide is (boron is like carbon, inert in fuel salt). Though I haven’t looked into lithium orthosilicate myself.
My thought was that the lithium-containing compound would be encased in 316 steel, so it’s compatibility with fuel salts wouldn’t be to big of an issue.
As for obtaining tritium — it has to come from somewhere (from D2O used in a CANDU or from control rods, etc.) and I’m not sure why neutron shield material would be particularly inappropriate (especially where it will be removed rather frequently).
If lithium orthosilicate would not be the best, then lithia could be looked at (they both have rather high melting points, which is the principal criteria that I was looking at).
@Rick Armknecht
Why does tritium have to come from somewhere?
@ Rod Adams
I could be wrong, but my understanding is that the amount of naturally occurring tritium is negligible — so not only does it “need to come from somewhere”, but that “somewhere” has to be through human activity.
If your question is aimed at the worth of having tritium at all, then I see your point as you are not too keen on fusion. That is a very defensible position, but the fact is that a demand for tritium exists and that selling it can be a profitable “sideline.” Moreover, I’ve read that American reactors have Li-6 in control rods to provide enough tritium for maintaining nuclear weapons — so at least THAT amount of tritium is deemed necessary to produce.
@Rick Armknecht
Tritium has some very useful applications in street signs, emergency exit lights, glowing watches, and beta batteries. However, none of those uses mean that it “has” to come from anywhere.
I have some questions about your final phrase “deemed necessary to produce.” By whom? Why? I reject the assumption that any country needs to maintain “H-bombs.” I accept the need for deterrence. MAD seems to work pretty well to keep leaders from doing things that are really, really stupid. However, long-lived fission weapons that need little to no maintenance would seem a much better investment since all of us hope they will never, ever, be used.
Rick, the neutron shield is inside the reactor vessel, which is sealed. Which is inside the can, which is sealed. The tritium in there is as inaccessible as it is going to be. If you want some tritium then you take a pool type research reactor and put a sealed sample of lithium in the core. Then leave it in for some time and remove it, presto good accessible tritium for sales.
Thorcon is a gigawatt class power reactor concept. Tritium sales would be peanuts and not worth the bother in a design that is inherently sealed and inaccessible.
As well, the high temperature means tritium will simply escape its stainless steel container. Tritium will diffuse through hot metals rather easily.
@ Rod Adams
You say: “long-lived fission weapons that need little to no maintenance would seem a much better investment since all of us hope they will never, ever, be used.”
Not a bad idea as increased accuracy would serve to counter decreased yield.
@ Cyril R.
You emphasize how “sealed” things are, but isn’t the can coming back — to be opened — after 8 years or so? Tritium could be harvested at that point, right?
The diffusion of tritium is also an issue — what if the lithium compound was placed in a slurry (lead, perhaps) that was topped with palladium? The gas would tend to rise to the top where it would be absorbed by the palladium.
Anyway, I regret that this tangential discussion has veered into so much about tritium as the main point I was trying to make was that neutron shield material tends to be selected based upon the sole purpose of getting rid of the neutrons instead of using the neutrons for a valuable purpose. The ideal isotope for such purpose may well be Hg-196: over 3000 barns and neutron capture leads to Au-197.
As you point out, though, such transmutation may not be worth the bother.
@ Jack Devanney
Thanks. That answers my “why?” question above.
In a floating MSR the reactor does not have to be 100 feet below sea level to have a reliable heat sink.
@Bill Hannahan
The driving head for natural circulation requires a certain difference in height to work. 100 feet might be a little more than absolutely necessary.
Rod, why limit yourself to single phase convection? Two phase heat transfer can handle very high energy flow rates in everything from percolating coffee pots to large BWR’s at full power. Didn’t the reactor you worked on run at full power with natural convection using a lot less height? We are just talking about decay heat here.
This would be the ultimate cooling mode, used only after all else has failed. Flood the dump tank room and vent the steam from a pipe above the tank that will vent steam and hot water as in the coffee pot.You will soon have a dump tank filled with solid salt.
@Bill Hannahan
Didn’t the reactor you worked on run at full power with natural convection using a lot less height?
No. Only reactor design I know of that operates at full power with natural convection is NuScale.
“This would be the ultimate cooling mode, used only after all else has failed. Flood the dump tank room and vent the steam from a pipe above the tank that will vent steam and hot water as in the coffee pot.You will soon have a dump tank filled with solid salt.”
This is in fact the emergency response of a membrane wall rupture in the Thorcon design. Such a failure is nearly unreasonable, though, considering the material properties (low stress, all stainless steel) but it makes the membrane wall a fail safe system (even in LOCA).
@Rod Adams
Well… there was Oklo as well, back in the day. But they got shutdown in a licence dispute with the dinosaurs at NRC…
Re single phase vs two phase convection. This is a bit more complicated than it may appear at first glance. While boiling reduces the density in the hot leg (thus increasing the driving force available) it als causes destabilizing (sudden) steam voids which oddly enough tend to increase turbulence pressure drop. The latter negates some, or even all of the driving force increase, depending on geometry. Boiling can also cause undesired (rapid) reactivity changes.
NuScale predecessors considered two phase natural circulation, but the instability issues combined with the increase in pressure drop from turbulence, meant the designers abandoned that option and went for single phase PWR style.
ESBWR operates on natural circulation at full power, but keep in mind the reactor vessel is 27 meters tall and even then they had to shorten the active fuel length to about 3 meters, to cut pressure drop.
Rod, thanks for the correction,
Cyril; I was not proposing two phase flow for normal operation in the reactor, just for the last option of passive cooling using sea water.
Boiling sea water and venting steam only would lead to an accumulation of salt over time. Using the “coffee pot” model to remove both hot water and steam prevents the salt concentration from becoming saturated, so no salt deposition. Using steam bubbles to pump hot water minimizes the height differential required to pump cooling seawater.
Rod – thanks for this post and update. It’s great to see ThorCon making progress, and I’d love it if all of the new, highly maintainable (and therefore adaptable as improvements are made in the ‘cans’ etc.) nuclear heaters got made. Let there be competition!
Oops – I hit the submit button before I was done. I do have a question, Rod. You said
How do the energy markets reward low or no emission energy sources? I’m aware of heavy subsidies for wind and solar technology, but not of a reward for being low carbon emission specifically. IIRC nuclear plants in the US are being hit hard by not being fully credited for their low emissions. I do recall a story about German consumers being offered nuclear electricity as an option for their power, the way I can sign up for wind power here in Alberta. (I told the sales rep I would sign if they offered nuclear electricity, and to take that message back to head office.)
@Nick – my reading of the ThorCon documents says that the cans are small enough to ship by truck or rail, and it’s the nuclear island modules that are the big ones. I would hope that either ThorCon or someone else will see the opportunity to take different approaches to prefab nuclear islands that can be shipped more flexibly, but still use the ‘cans’.
In fact, I’d love to see an industry in which the ‘cans’, potentially from a number of manufacturers, are off-the-shelf modules that can be engineered into industrial facilities of all types. It’ll take years to get to that point, but I think it’s a worthwhile goal. Should we be getting the reactor designers to be thinking interoperability? Sort of like AAA, AA, D, and 9 volt batteries? Consider the ‘cans’ as ‘heat batteries’ rated for a certain number of petajoules, megawatts,and years of service at designated outlet temperatures feeding heat exchanger loops. That image sounds exciting to me, but maybe I’m just naive.
If you consider the nuclear heat island as a battery and the turbines, generator, switch yard, and transmission lines as what we plug into then we are there. The interface is steam. Anything with even a bit of radiation in it has to be part of the licensing and casualty analysis so all of that has to be part of the “battery”. One of the attractive features of the design is that we can reasonably upgrade the can even to already deployed power stations. For example, initially we are using NaBe salt because it is available. FLiBe would be better and when it is available at a decent cost we will switch to that.
@Andrew: “IIRC nuclear plants in the US are being hit hard by not being fully credited for their low emissions.”
Not yet. The US does not yet have carbon emissions rules of any sort. As currently written, nuclear plants would indeed be so penalized under EPA’s proposed Clean Power Plan. But that particular provision has been thoroughly criticized in EPA’s Public Comment process, and such criticism must be addressed in writing when EPA makes its final rule this June. Hopefully, the final rule will simply make all carbon dioxide molecules equal regardless of race, creed, color, or country of origin.
We’ll see. Next up: Methane!!!
THREADWINNER!
In my comment to the EPA, I suggested that so-called renewables should be credited with less than 100% because there does not appear to be any mechanism to attribute to them the fossile fuels burned and CO2 emitted in order to provide spinning backup to compensate for their intermittency.
Andrew,
Just an FYI. 100% of every component or module in a Westinghouse SMR is rail shippable on a conventional car within a standard envelope. This restriction drove our design from day one. We also optimized our excavation depth with cost from day one to set those design requirements as well.
@Cory Stansbury
I like the word “optimized.” There are advantages to building a sufficiently deep foundation for almost any structure. There are also rather enormous costs associated with going deeper than required.
100% agreed.
Thanks all for the replies. You are informative and helpful. I’m also encouraged by the enthusiasm I see!
Glad to see the interest growing for MSR’s. It is easy to make a hole in water, every ship does it.
I believe the successful Model T design concept will be a floating MSR built in an indoor factory consisting of a dry dock where the vessel shell is manufactured and then floated down a canal where internal structures and components are added.
Of the many advantages, the biggest is the fact that site preparation can go on in parallel with plant construction. This could reduce the time from contract signing to full power commercial operation to less than 3 years.
The reactor would be well below sea level, providing a 100% reliable heat sink.
Over 80% of humans live within 500 miles of an ocean.
Correct me if I am mistaken, but the ThorCon writeup suggests it doesn’t need a Hastelloy superalloy for any of its loops including the primary loop. That will fly for the loops other than the primary, but the primary loop is not clean, fission products will build up, after all there are more than just gaseous fission products, and they can be corrosive to non-hastelloy superalloys.
@Joe ChemNuke
My understanding is that the short operating life is designed to account for the corrosion issues that concern you. I think some of the designers have chimed in on this discussion; perhaps one of them can confirm that understanding.
Thanks Rod, that makes sense. I still worry that 4 years at those temperatures may be pushing it. That said this is one of the more complete MSR designs I have seen, well done.
From what I remember from reading the Thorcon Executive Summary last year, the primary loop being replaced with a new “can” every few years is to account not only for the corrosion issues you speak of but also due to damage to the graphite moderator that will occur in the high flux environment.
The previous Thorcon conceptual designs employed Hastelloy N as the primary loop material. I recommended to switch to austenitic stainless steel for a number of reasons. After extensive discussion in the design team, literature review, and contacting a second opinion from a metallurgist, it was decided to switch to stainless.
There are many things to consider in this choice.
First off. While stainless has somewhat higher corrosion rates than Hastelloy N, it is still quite low. In an older test loop that operated for 9.5 years in fuel salt, the uniform corrosion rate of only 22 microns per year was found. Some pitting was observed, but only to a depth of 250 microns. So Thorcon will expect corrosion in the ballpark of 250 microns, possibly much lower due to the improved redox window the reactor will operate at compared to the test loop. These are excellent performance figures, that would be described in most industries as “very low corrosion rates”.
Second, the corrosion type is not detrimental. It is not catastrophic like most non stainless steels can corrode in a runaway reaction. It is just surface leaching of the chromium. The basic structure below the surface is maintained.
Third, the corrosion products themselves are not detrimental; they are typically soluble and stay in the salt or are removed via noble metal removal systems that are already in place to help remove noble metal fission products.
Forth, ORNL did not determine any detrimental fission product corrosion mechanisms, with the exception of sulphur like materials, notably tellurium (and possibly selenium). These materials (they are fission products with reasonably high yields) weaken the alloy structure by forming brittle, low melting point phases. ORNL found this type of corrosion in the post mortem analysis of the MSRE. Stainless steel has a higher chromium content that protects against this type of corrosion. Stainless is expected to be superior in this regard.
Fifth, as mentioned the short operating lifetime. This means that components are replaced regularly by the design.
Sixth the Thorcon uses large corrosion allowances so if corrosion is higher it still isn’t a problem.
Let us imagine I am wrong. Wouldn’t be the first time. Corrosion is far worse than expected. The prototype will prove this soon enough; if it is really bad then Hastelloy N can be used as a drop in replacement for future modules, and I can drop through the ground in shame and disappointment. Ease of module replacement – which must be demonstrated regardless of alloy used – makes this an available fall-back option.
Stainless steel is in many ways superior to Hastelloy N. It is in far bigger production, anyone can supply it and work with it, working with it is easier than Hastelloy N, it has higher neutron radiation resistance than Hastelloy N, it has higher tellurium/selenium resistance than Hastelloy N, and it is widely used in most nuclear plants today – including primary loop material, such as isolation condensers, depressurization valves and other, critical, safety-grade equipment. Cost savings are only a secondary benefit – it is not, surprisingly, a very big difference in a billion dollar powerplant.
Hope this clears a few things up,
Cyril.
The key thing is that these need to be built and tested and then we really know. I think there is plenty of margin, but this reactor will operate at a higher power density than MSRE which will likely change the paradigm on corrosion slightly. The concern is what happens post initial chromium depletion? It’s not a showstopper, just needs to be fleshed out.
Yes, testing will truely confirm this and that is part of Thorcon’s philosophy of license-by-test. Hastelloy N has similar creep and such properties as stainless so it is a drop in replacement should the first module analysis prove that corrosion with stainless is unacceptable. Module 2.0 could then be manufactured of Hastelloy N. This is what the prototype is for – confirming expectations or fallback to the alternative. This would not be an option with a LWR that has to last 60 years, you could never truely reliably extrapolate and it would take a century to do just a few complete design tests at design lifetime.
The corrosion mechanisms are well known, mainly due to the excellent MSRE and related works, which included extensive thermal convection loop testing (including a 9.5 year duration test on stainless steel) apart from the MSRE itself.
Chromium is first attacked as it is available right at the surface, so there is initial leaching. Then the surface chromium depletes so leaching is limited by thermal diffusion of chromium from deeper in the alloy matrix. This is a well known temperature controlled process. The surface layer gets depleted of chromium on the salt side but this isn’t needed on the salt side. The surface is left with some pitting. Well known from tests. It’s perfectly acceptable.
Power density is not important for the structural alloy. It is important for the core and moderator. The reactor vessel is well protected from neutron flux by a reflector followed by a boron carbide neutron shield. It is much lower in dose than what we know the material can take (from LWRs which all use austenitic stainless steels for a variety of demanding applications).
I’ve found a lot of resistance toward stainless steel (and MSRs in general) can be simply rooted down to unfamiliarity (sometimes even obliviousness) to the actual experimental work that has been done decades ago.
Rod, I believe that David LeBlanc plans to build cheap replacable cores, that will be swapped out every few years.
So does ThorCon.
Their knowledge of current industry practice is pretty misguided. Current fleet has excellent NPP maintenance practices, however standardization would help.
Generally agree. There are some serious deadbeats though. Progress Energy decided to cut corners and do the highly specialized tensioned concrete tendon loosening and concrete cutting, to replace steam generators, themselves. They ignored all of their engineering consultants and decided to speed up the tendon loosening sequence by doing it exactly as all basic manuals say you shouldn’t do it. They ended up losing the entire plant, a 2.5 billion dollar rookie mistake that could have been avoided by reading a $5 manual or doing a 10 minute Google research on the subject.
If you leave maintenance to the specialists, its usually performed very well. Somehow though, maintenance anything on nuclear stuff is fiendishly expensive, even when totally succesful.
Crystal River is unique, and a bit more nuanced than this. But they did decide to cut costs by doing it themselves which was clearly a big mistake. Other than this the industry has been very good at upgrades, repairs, and replacements. Just consider the fact that most of the NPPs didn’t anticipate major SSC replacements so they didn’t have appropriate containment hatches, yet look at the record of stream generator replacements. That’s where a design like this will only improve performance.
And these high costs are a function of time and personnel constraints that arise when working in radioactive areas. Low dose reform would reduce this.
It wasn’t a mistake to do it themselves; it was a mistake to ignore their engineering consultants, who were experts on this, whilst Progress Energy knew next to nothing about tendon tensioned concrete. Progress wanted to speed things up by loosening tendons that were too close to each other; that is the first thing any basic manual or handbook on this technology will tell you to avoid!
I don’t see the nuance. Where is it?
In one of his bullet points Rod Adams wrote:
There are political leaders who have recognized that pushing innovative nuclear energy projects off-shore is bad for business here in the US.
It seems as there are just as many (if not more) political leaders that have drunk the Green Kool-Aid who believe things like wind and solar would be even better for business.
Correct me if I am mistaken, but their report says they are using a SUS316 for all piping. That will work with clean salts in the secondary and tertiary loops, but not in the primary loop. That will be quite contaminated with fission products beyond the gaseous ones they bleed off.
So? The evidence we have with fuel salts shows that corrosion rates are low. There is no evidence to support the notion that adding fission products in any concentration would be an issue, except tellurium and possibly selenium. Stainless has better resistance to these sulphur like materials than Hastelloy N. See earlier comment of mine.
Initial salt purification, elimination of galvanic coupling, and adequate online redox control have proven to be extremely important, regardless of alloy of construction. If you get these right, both stainless and hastelloy will work, so stainless is the first choice.
Every time I see an article like this one, I go back and read Hyman Rickover’s discussion of the difference between “academic” (or “paper”) reactors and real ones. The MSRE was not an unqualified success, and it seems likely that significant developmental work remains to be done before the technology can be deployed commercially. Maybe it will ultimately be deployed successfully, but I seriously doubt that it can be done, even as a non-commercial prototype, by 2020.
By the way, Rod, it’s “principal” engineering designer, not “principle.”
Yes! There is a distinct difference between what reactors should look like on paper and what they do look like. Closing this gap is a good goal, but takes work to learn it out. And the MSRE was successful, but it was an experimental reactor, not a full fledged power producing prototype. There is still some work that needs to be done.
I think the wave of new reactor designers is great because the current vendors are losing their innovative edge, but these designers need to remember that real reactors are elegant, beautiful machines, and when they are made of real steel and concrete they are big and can be complex. This is not a bad thing, it is a testament to their functionality.
oldnuke, the ORNL scientists, including my father, who worked on the MSRE, did view it as an unconditional success. The problems that were revealed upon careful inspection were minor, to the extent that the reactor could have operated for a large number of years without showing up. ORNL researchers believed that the materials problems were on the path to fixing them when the DoE shut MSR research down for the last time. As for Rickover, the old fox exagerated the difficulty of the problem. It was something like 6 years from the time that Alvin Weinberg told Rickover about the Light Water Reactor, and the first sailing of a nuclear submarine. Rickover’s problems came from the rush job that lead to the building of the first pressurized water reactor. Rickover had a lot more technological problems to solve.
Apparently ThorCon didn’t get the memo;
The triple core meltouts at Fukushima Daiichi were the last three nails driven into nuclear generated electricity’s coffin.
Every entrepreneur and new technology venture faces steep odds against them. Most such ventures fail, but all of the ones that succeed do so, in part, because they disregard the pessimists and naysayers.
And reactors which melted down will eliminate all interest in reactors which cannot melt down? Really? Sounds more like a selling point to me.
@ Lew,
Yes, if you melt a core it will just keep going to the center of the earth and join with the rest of the Uranium driving the core. No one will ever trust a melted core again! We would never never want to use melted uranium.
These must be zombi reactors. They are afraid of the light, which is why they need to be 30 meters underground. Let’s hope they don’t infect the rest of the power grid! It might all turn ZOMBIE!
The Thorcon design has considered a Zombie plague as a design basis accident. The reactors are walk-away-safe. Although in the case of a Zombie plague, it would be more like shamble-away-safe.
@ Cyril,
With tongue firmly in cheek, I was joking, and truly hoping that the Thorcon reactor would “plague” the world.
They illustrate exactly what I was saying a few days ago, that with reasonable regulation, Nuclear will take over the world quickly!
We have horrible regulation (not the people but the rules and system).
I was also joking, of course. But, hypothetically, a walk away safe plant would be a shamble-away-safe-plant in the event of a apocalyptic pandemic (zombie or otherwise). If there’s still people living to care about that, of course.
It is true the regulatory system is all wrong and Thorcon aims to change that. It is not just a technology offering, it is a way to design, licence (by testing and evidence rather than by computers, paperwork and “quality control”), procure, build and also own and operate nuclear plant systems. This is what sets it apart from other offerings.
@Cyril R
There is a little understood provision in US nuclear regulations that fully supports the idea of demonstration plants and license by testing rather than analysis. I can provide details for interested parties, but it is not the kind of material that makes for a good public blog post.
@Rod Adams,
I for one would be quite interested to know of the alternative regulatory path you mention.
I like the Zombie plague scenario.
A slightly tougher one is that the reactor is taken over by Alan Rickman and a bunch of guys with bad German accents, and Bruce Willis isn’t there in time.
@David, underground may mean cheap and safer.
@ Charles
Underground might mean safer and cheaper, yes but it does limit the site selection greatly. Islands are generally difficult to dig that deep.
But I think the overall design is fantastic. It has application in many places in Africa and Keyna especially comes to mind. They know they need nuclear and are beginning the process of regulation and education to run Nuclear. That is a very progressive country and might be the right place for Thorcon to start.
@Lew Kemia
Thank you for stopping by. If you are actually interested enough in the topic to participate in the conversation, and are not just a “drive-by” commenter, you will find that there is a large body of people that did not get that memo.
There were few new lessons learned at Fukushima. Many BWR operators already knew the mantra “Vent early and vent often” was supposed to be implemented before allowing core damage to occur. Nearly all operators already knew it was a bad idea to have all of your emergency power diesel generators in an area where they could be subjected to a common influence – some should have been elevated, while the basement is a good location for others. We also already knew that distant government officials with no nuclear training should not be included in the operational decision path.
There are no other fuel sources that can provide reliable power without producing vast quantities of gaseous waste products. There are no other proven fuel sources that have a measured abundance that can provide sufficient power to supply an industrialized, developed society with 10 billion or more citizens for thousands of years.
Nuclear energy has a bright future. If people in some parts of the world chose not to allow it to prosper because they prefer to keep burning massive quantities of fossil fuels while enriching the tiny portion of the population that controls access to those fuels, that part of the world will gradually lose its developed prosperity. It might look okay for a while — like Germany — but reality will win in the end.
“Apparently ThorCon didn’t get the memo;
The triple core meltouts at Fukushima Daiichi were the last three nails driven into nuclear generated electricity’s coffin.”
Are you, by chance, also campaigning for the abolishment of sky scrapers and passenger aircraft? The 9/11 events clearly showed both technologies to be very dangerous, with orders of magnitude higher death rate (even per LNT) than Fukushima radiation.
‘Apparently ThorCon didn’t get the memo;
The triple core meltouts at Fukushima Daiichi were the last three nails driven into nuclear generated electricity’s coffin.’
The guys building the 71 reactors under construction must have missed the memo too, you’d better tell them.
This was a great post. Thanks for telling us about Thorcon.
I realize there are a lot of details to be worked out Will the units have instrumentation and control similar to existing plants?.
How are the units to be controlled? I haven’t read everything, but did find a reference to control rods. I assume that there will be some sort of control rod drive assembly and it will not be pneumatic or hydraulic.
In one of the documents, it was stated that these reactors are load followers. I thought the opposite would be true. Won’t the turbine have to follow the constant steam supply from the reactor(s)?
Is there a chance that plant operators do not have to be on site? Could multiple plants (sites) of this type be controlled from a central location?
The silo reminds me of the water walls around conventional steam plants. Tube leaks are very common. Burial may pose some access difficulties.
The salt temperatures and steam supply would not be perfectly constant. They would all run hotter under low load, but presumably within a range accommodated by design specifications.
The turbines function to turn heat into work. Under periods of low load, less work is done, less heat is converted, and more heat is returned to the secondary heat exchanger. That, in turn, returns more heat to the primary heat exchanger, and this raises the temperature of the molten fuel salt returning to the reactor core. Molten salt has a large coefficient of expansion, so entering the reactor core at a higher temp means it is also less dense and thus less reactive as it proceeds through the core–which has the effect of throttling down core heat generation. Control rods throttle down the reaction by absorbing (typically squandering) valuable neutrons, and the neutron budget of these reactors would be tight, so you wouldn’t want to do any more of that than necessary.
I expect each site will need a staff of operators and security at first. Even if scram and reactor output become automated, there would still be a number of auxiliary systems which will need oversight and maintenance, and humans would almost certainly be needed on site for any restart or core swap operation. Control and maintenance would later become more automated, but the speed of transition there would depend on the rate of advances in autonomous machine system technology. Security may always be needed against anti-nuke extremists.
Instrumentation details have not been worked out. Philosophically though the design is intended to be walk-away safe – so that no operator action is required to ensure public safety. So the instrumentation is to optimize performance not to ensure safety. As such, the instrumentation can be more like typical industrial instrumentation rather than nuclear. I expect there will be considerably more performance related instrumentation than existing plants. Continuous process monitoring like what is used in the semiconductor industry should allow much tighter controls of process conditions and earlier detection of any anomalies.
A molten salt reactor will naturally adjust its power output to match the demand. Crudely you can think of an MSR as holding the average temperature of the fuel salt constant. If you supply fuel salt at a faster or slower rate then the power will change proportionately.
For security I think the plant operators will need to be on site but not necessarily inside the reactor building. After the Stuxnet experience I don’t think we want to provide the ability to actually change the reactor remotely. One thought is to provide lots of reactor status information to a central site for analysis and to have expertise available to advise in case of any anomalies. But that actual control of the reactor be local.
While I’m glad of any innovation in nuclear power, I’d caution Thorcon to quit benchmarking their paper reactor against coal power, and instead benchmark against light water reactors, especially the AP1000, which also has many passive heat removal features. Their plant has to be much, much better than the AP1000 for this effort to be worthwhile. It can’t be a little better; it has to be unequivocally superior by a wide margin. Otherwise, there is little point.
I say this because the problems in building a nuclear power plant are regulatory, legal and political, rather than technological, and I see little about Thorcon’s concepts that clears out these hurdles. In the years to come, someone like a Gov. Shumlin or Gov. Coumo, or the German parliament, is not about to say, “Oh, it’s a Thorcon reactor. Ok. Go right ahead. That’s cool.”
Actually, they should benchmark against both coal and the AP-1000.
In benchmarking against AP1000 the figure we would use is capital cost and time to install.
The problem with AP1000 is that its capital costs are much higher than coal so in most of the world you can expect coal will beat the AP1000. In China alone they added 800 GWe of coal power. Worldwide AP1000s are in the process of installing 9GWe. The AP1000 looks to be a winning nuclear design and will build many more plants – but I don’t think even the most optimistic projections for the AP1000s will make a significant dent in the number of coal plants built world-wide. Our target is not the US or Germany – these are replacement markets with serious political opposition. Rather, the target is the rest of the world where cost is going to be the driver. Sure, we’d love to get some sales in the US and Europe as well but the focus is elsewhere.
Can you clarify some aspects of this comment? For example you said: “In benchmarking against AP1000 the figure we would use is capital cost and time to install.”
Does “the figure we would use” mean the capital cost and time to install a Thorcon design, compared to the same for an AP1000? That’s the bottom line and until you have a certified design you have two big unknowns in the US system; your certification cost and the time to get it. I’m not knocking your concept, rather I’m trying to understand your marketing plans.
When I add: “Rather, the target is the rest of the world where cost is going to be the driver” to the discussion, reading between the lines I get you might not go with US certification.
That’s the crux of the whole US problem for new US certified designs. Under the current US regulatory structure that cost and time is literally a black hole; everything in and nothing out.
Do you plan to get US certification for your design?
When digital photography first came out, the image quality was clearly inferior to film, but that clearly was not the end of that story. Whenever a new technology is introduced, it is hard to predict the directions it will take, the possibilities it will create, the branches it may lead to, and how much impact it will ultimately have.
“Better” in this case will depend on one’s needs and priorities. The Thorcon base module is much smaller than an AP1000. That is better for situations needing a smaller power plant. The Thorcon reactor can’t melt down and operates at low pressure. That could easily be a critical difference for people who have a phobia about meltdowns and radioactive pressure blowouts. There would be no on-site spent fuel storage, which would be better to people who care about that. The build times and deployment rate look like they have significant potential advantages over light water. And so long as we remain with light water, we will not be developing real-world experience with molten salt, and its possible advantages will remain a subject of theoretical speculation.
My view is that putting the focus on coal is exactly where it needs to be. Coal is the real enemy, and I tend to agree with Rod Adams that nuclear advocacy in general is ill-served by the various factions of nuclear power kicking the stools out from under each other. (to use very loose paraphrasing) Different kinds of nuclear power will have different advantages, and the more nuclear options there are available covering a greater range of needs, the less chance that will leave any uncovered niche for coal to exploit.
I don’t know if the Thorcon approach (of which their business model may be the most revolutionary part) will ultimately prevail over light water, or find a place alongside it, or whether some other technology will come out of left field and trump them all, but it makes me happy to see another major and significantly different entrant into the field. I’m of the opinion that nuclear power development will be at its most vibrant and healthy with a diverse range of competing projects, where even the experiments which fail can contribute valuable information to the overall effort. And sometimes failed ventures can knock down obstacles and open up pathways for the ventures which follow. I don’t automatically hope that Thorcon succeeds, but I do hope they succeed in demonstrating the worth of their approach. And if it happens to have more worth than other approaches, I hope rewards flow to them accordingly.
The stability of concrete foundations and underground structures is highly dependent on the geological structure, which can vary dramatically even over a small geographical area.
I am familiar with a city that has a geologically stable area underlain by volcanic-origin rock where a number of up-to-15-story buildings are currently under construction in a relatively new and prosperous neighborhood.
The same city has an older section underlain by a deeper layer of soil, with much less rock. Very few buildings over three or four stories are being built in that area, because there is much higher risk of cracking and other structural damage.
Something similar might be said for Manhattan. Something might also be said for seismic siting analysis for any large structure. Seismic is good. Use lots. 🙂
Good and interesting article, Rod. There seem to be many exciting developments in the MSR world — including today’s announcement by Terrestrial Energy of their new arrangement with Oak Ridge Nuclear Laboratory. BTW, in your article, you suggest that Terrestrial Energy’s Integral Molten Salt Reactor is designed to be constructed underground. In fact, this is not the case. TEI recognizes the complications of underground construction, and intends its IMSR to be located at ground level.
Currently, the IMSR’s actual reactor module and most of the containment is designed in a below grade cavity. The rest of the building is at or near grade.
Interesting work. However, in Thorcon’s literature, I have not seen how they will avoid the lithium-6 / tritium issue.
Their website lists the fuel salt as being “a mixture of sodium, beryllium, uranium and thorium fluorides”.
It looks like they avoided tritium production from lithium by avoiding the use of lithium.
Lithium fluoride is replaced with sodium fluoride in the mix. Sodium does not make tritium. Software modelling shows an order of magnitude reduction in tritium production due to swapping lithium fluoride for sodium fluoride.
There is also an extensive defence-in-depth approach to trapping and sequestering tritium, even though tritium is not actually a radiological risk.
I watched a telecast of a speech and question and answer session by Energy Secretary Moniz at the Wilson Center on C-Span last night. Moniz supports research on advanced reactors although he did not specifically mention moltens salt reactors. For the near term (actually the 2020’s) he supports development of small modular reactors and light water reactors. Asked about the impact of low oil prices he said it was good for the consumer and the DOE was developing models to predict the effect on other sources of energy. Our country still has the goal of energy dependence and he expects that we will continue to pump oil. The grid needs to be developed so it can distribute wind and solar energy. Transportation infrastructure needs to be improved to transport some sources.
I meant ENERGYINDEPENDENCE.
Rod,
Thanks for the excellent review. It is wonderful that there are so many innovative nuclear fission reactor designs out ther.
Sadly, nothing will come of any of them in the USA, Germany, France, the UK or Japan. The most likely locations for truly innovative fission reactors will be Canada, Russia, India and China. Even the Czech Republic would be a better bet than the USA.
@gallopingcamel
I refuse to accept that prognosis and will continue working hard to prove you wrong. The USA is a great place full of wonderful, caring, innovative, creative people. There are many challenges to overcome, including the challenge of pushing out elected and appointed officials that are selfishly serving other masters than the ones that elected them, but overcoming challenges is “what we do.”
the company I am really looking forward to seeing the full details on is Moltex Energy. below is a link to a blog article on it which I have included mostly due to the references having the sources of the submissions to parliamentary enquiries in the UK as well as lecture slides on the reactor given by the founder to IchemE (former technical head at Unilever)
http://energyknot.blogspot.com/2014/10/simple-molten-salt-reactor-by-moltex-llp.html
A bit of background the reactor design is also a molten salt but differs completely to the textbook design. there is no pumping of fuel fluid and it never leaves the reactor core. Instead it is contained in fuel tubes (20 to 30mm diameter), convection currents within the tubes allow efficient heat transfer from the centre of the tube to the walls. a coolant salt then transports the heat away from the core to heat exchangers.
The clever bit is the coolant salt contains hafnium, which has a very small cross section to fast neutrons but a large cross section to slow neutrons. The reactor is a fast reactor so the hafnium isn’t an issue from a neutronics perspective however once away from the core the neutrons thermalise sufficiently such that the hafnium then has a large cross section stopping neutron damage to the main structural elements. The fuel tubes are arranged in fuel assemblies conceptually similar to solid fuel reactor fuel rods and can be replaced in a similar fashion.
This reactor is getting a lot of traction in the UK in particular now at a government level. Many of the top nuclear engineers have looked at it with one describing it as the most exciting development in the field of nuclear fission for decades. Atkins consultancy is currently performing cost analysis work with a number of other engineering firms getting interested in the concept.
@Jeremy Owston
Thank you for the tip. If you have any contacts with the company, please send them my way. There is a contact form at the bottom of every page of this site.
Now I have some reading to do.
sent the only contact I have which are on his website http://www.moltexenergy.com/. I have emailed him once in the past and he got back to me straight away and I am just a basic mechanical engineer so I think he should be willing to talk if you contact him
I don’t understand how they can make this work on natural circulation. We’re talking about moving 2000+ MWth with a coolant that has very low nat circ figures of merit compared to water. It must be monstrously tall.
Ow, I see, this is the design with the fuel salt natural circulation but forced coolant over the fuel tubes. We discussed that design some time ago. Didn’t like a lot about it.
http://www.energyfromthorium.com/forum/viewtopic.php?f=5&t=4417
Power density is also crazy high. Looking at multi MPa pressure drop here. These would be monstrous coolant pumps and it wouldn’t be low pressure at all.
Would a variable speed drive reactor coolant pump be needed to vary load? From answer given by Lars Jorgensen on Jan 7th above, this appears to be the case. I have a gut feel that they’d end up with more than a single pump.
@Eino
A basic heat transfer equation for power out of a heat exchanger – like a reactor is that Q (power) = mass flow rate x heat capacity x (Hot Temp – Cold Temp)
Varying power output is generally not accomplished by changing mass flow rate, but by changing the temperature difference between the hot and cold temperatures.
In the zero power condition, no heat is added and both temperatures are the same. In the maximum power condition, there is the largest design difference between Th and Tc, but the average temperature, which is what governs reactivity, is the same as it is in the zero power condition.
Further add to Rod’s comment heat exchanger size is dependent on the following
– volumetric heat capacity of the working fluid,
– membrane resistance to heat (normally wall thickness of your tubes in a shell and tube heat exchanger)
– temp difference.
To transfer a given amount of thermal energy in a small heat exchanger, having high volumetric heat capacity, thin wall tubes (low pressure) and large temp difference is what you want.
Steam generators in a nuclear plant are large because wall thickness is large to withstand the pressure, temp difference is small to achieve acceptable thermodynamic efficiencies. Water is an excellent coolant as it has a high volumetric heat capacity.
Its a design choice to have one or more pumps. Pump power is lower than a PWR so these pumps are relatively small.
One pump is the simplest, no issues with pump communication of any sorts (hydraulic, control/electrical). Multiple pumps has reliability advantage, though single pumps can be made very reliable as well (double or split drives, non-lubricated seals, overengineered thicknesses, etc.).
Variable speed drives would almost certainly be employed, yes. Startup and shutdown is very difficult with only “on” and “off” buttons!
@Rod Adams January 10, 2015 at 1:30 PM
“Varying power output is generally not accomplished by changing mass flow rate, but by changing the temperature difference between the hot and cold temperatures.”
This Ht-X equation is too simplified for advance designs and is not the case at all for natural circ. Virtually all NC applications I am familiar with depend not only on a delta-T increase with Q increase, but also a mass flow rate increase. It’s the basic concept for NuScale’s “pumpless” design, e.g. mass flow rate must increase. In all current “contained fuel” designs, since T-C is ultimately bounded by the heat sink temp, T-H will always be bounded by material properties of the “contained fuel”, and that Ht-X equation does not apply at all to getting the Q out of the fuel containment into the coolant, as there is no mass flow at all inside the fuel container. In those cases it comes down to the thermal conductivity of all the junk the heat must pass through, by conduction, from the source to the sink, and mass flow rate is not even part of that equation.
My understanding of some new advanced designs is that the fuel (fission heat source part) will be moved with a medium. When inside the “magic box” fission will occur and make heat, but when moved outside the magic box fission stops and just medium with fuel and the heat is carried to the next box in the process to extract the heat.
So seems mass flow must be varied. Am I totally all wrong?
@mjd
My comment was not aimed to respond to a natural circulation question, but one asking about load following for the Thorcon MSR.
The flowing mass heat transfer power equation applies to the coolant for contained fuel designs. For fluid fueled designs it applies to the flowing fluid fuel when it is in an area where heat is either generated (the reactor) or transferred (the heat exchanger.)
Practically load following of nuclear plants today involves change in both temperatures and mass flow rates. In case of BWRs, reducing feedwater temperature can increase reactor power, as can increasing the recirc pumps, for example. To get the most load following both can be employed. For natural circulation designs like the ESBWR you only have the feedwater temperature, the mass flow will vary indirectly but can’t be actively fine tuned.
Generally mass flow rate is the easy variable as modern variable speed pumps are common and some way to vary output is necessary for startup reasons (can also be pony motor of course, but this is a bit old fashioned). With temperature you run into all sorts of issues, with molten salt reactors the more because there are limits such as salt freezing temperatures. It is also not clear to me how quickly the thermal fatigue life of the containers and heat exchangers would wear out. So far the focus is on making the Thorcon work in a fail safe and practical manner. Load following is a later question but can be answered by prototype test. Thorcon’s main competition is coal baseload plants and these aren’t good at load following so its not a hard target to be better than coal.
Are you expecting normal operation to have all the rods removed and load variations are accomplished by varying recirc flow? There may be three flows to consider with primary, secondary and tertiary so the relation of this to the reactor must be quite complex. Since it’s all thermal, I could see a slow response time.
Multiple pumps may allow some additional maintenance or repairs to be performed leaving the plant online. A high capacity factor will be desired by this plant.
The control system has not been designed yet for Thorcon. Likely it will operate the pumps in ganged mode. All operating on the same OS/trip logics. Dial down te primary pumps to 50% of nominal RPM, the secondary and tertiary pumps copy that behavior, also cutting down to the same 50% RPM. This avoids “hunting” that an independent loop control would produce, that can result in increased thermal fatigue.
You said: “Practically load following of nuclear plants today involves change in both temperatures and mass flow rates.”
I always find the discussions about load following in a LWR amusing, as in they can’t do it. FWIW, virtually all navy PWRs and the B&W PWR were/are designed to load follow, with a constant T-ave and a constant RCS flow rate. These are not paper reactors, many have been built and are still running. The design basis maneuvering transient of the B&W PWR was a 100% to 15% to 100% transient, at 10%/min, with return to full power at peak Xe. It’s amusing because I licensed on, worked at, and most of our initial training focused on doing exactly that; all while maintaining a constant T-ave and constant RCS mass flow rate.
What was found in commercial plants was there was not really a need to do that because the system load demand did not require it. From there it shifted to economics. The fuel loading needed to provide the ability to load follow does not maximize the fuel burn up, so to maximize the fuel burn up benefit the fuel loads were change to be most efficient at base loading.
So it not a case of they can’t, rather it is a case of they didn’t have to in the ’70s when these decisions were made.
@mjd
Your favorite regulator also played a role in the decision process. As you are aware, B&W reactors were specifically designed to include rapid load following capabilities. One of my former colleagues at usta-company was an operator at Crystal River and involved in a project to demonstrate the value of the load following capability. The project was aimed at showing how a 177 could be a dispatchable generator.
Even though nuclear did not yet have a big market share, it was evident that it would have one if current trends at the time continued. There was therefore incentive to be capable of providing on demand power and not forcing the grid to take unwanted power.
The capability worked as designed, but regulators were concerned about having a dispatcher “controlling reactor power.” Of course, the operators in the control room would actually be the ones making the changes, just as they do in maneuvering on submarines. (The dispatchers were provided the equivalent of an engine order telegraph that sent a signal to the operators to tell them the required power level.)
I am not sure exactly when the decision was made to stop pushing the regulator to accept the concept. My speculation is that the process became fatally “bogged down” sometime in the spring of 1979.
OK, I’m with you. I got lost in the comment thread and thought it had changed to the Moltex design. I see now still on the Thorcon.
btw, the Moltex reference says the reactor runs “barely critical”; that’s something to ponder!
mjd; I saw that too, sort of like the kid who told his parents that his girlfriend was barely pregnant.
I thought it, you said it! But it goes back to some old nuke school arguments about sub-critical multiplication. Is it possible to get to 100% power on SCM and still not be critical? Or if you have an external N source installed is K-eff really 1 at 100%. I’ll stay out of those, don’t have to do requal anymore.
@mjd
In a word – no.
External neutron sources cannot make up for insufficient cycle neutrons to maintain criticality. Some advocate accelerator assisted systems, but I think most of them are just accelerator salesmen.
Criticality (or the effective multiplication factor k-eff) is the ability of a system to multiply neutrons and is independent of the number of neutrons in it (assuming no heating or feedback from energy deposition). A system can be supercritical and have no neutrons in it, or it can be subcritical and driven with a steady state source of a billion neutrons per second.
@Brian Kiedrowski
And how does a steady state source of a billion neutrons per second work?
Its just sales talk. What it referred to is reactivity reserves ie excess reactivity that must be controlled actively. Its not even true. U pu fast cycle has a big reactivity swing over its cycle years. Not poisoning that means the fuel boils.
Hi Rod,
Happy New Year first of all!
I was always wondering why potential customers seems so unaware about the merits of homogenous reactors? As I understand they have the lowest fissile material requirement for criticality. The ThorCon concept is a good example.
What I was further asking myself is: Why does it have to be molten salts for power reactor why can’t you use aqueous reactors as power reactors?
The Argus reactor in Russia e.g. is a highly reliable and economic research reactor producing radioisotopes, mainly Tc-99.
Greetings
Aqueous is possible, but is high pressure and low temperature for a power reactor so that’s bad. It would be less efficient (thermally) than a PWR so that hurts economics, the low temperature means limited process heat applications, and high pressure means a big driving force to push radionuclides into the environment (and with no cladding or non volatile ceramic fuel to hold them up). It doesn’t help that these aqeous reactors make more hydrogen than PWRs, a lot in fact.
Yes this is especially true in transients when a large amount of hydrogen can be produced almost instantly in the core.
However, low power density small aqueous homogenous reactors could be a great idea for research reactors for universities or research labs. They give you a nice neutron flux for NAA and also serves as a production reactor for medical isotopes. The medical isotope side could pay for all operating costs of the reactor with a profit on top. And the more dispersed our supply of medical isotopes, the less likely the world suffers shortages when a single reactor goes down.
Yes, for a medical or research reactor these things are promising.
For a power reactor, the amount of hydrogen is so large that it can be used for superheating – a superheater autocatalytic recombiner, basically, a live steam cooled catalyst. It could probably be made to endure a big transient heatup – its ceramic material, not part of the core. Still, in these latter anti nuclear days such things will be found very scary by the uninitiated.
@Rod, in answer to your why question, in the late 1950’s the AEC discovered that it had three liquid fueled reactor programs. One of the horses involved fuel in heavy water, the second fuel in Molten salt, and the third fuel in liquid metal. It decided to commission a team of scientists, to hold a horse race between these threee steeds. Two of them, the molten salt and the heavy water stallions, were stabled in Oak Ridge, and ny father at various times in his career had fed both of them. The MSR won the race and the other two were sent out to pasture. Later it was discovered that while not good technology for electrical generation, the heavy wwater steed had other uses.
Hey Bill Hannahan, fishing report for Jan 10, 11, 12. Moderate success using a barely pregnant plug on a sub critical multiplication spinner trailing K-eff streamers. And since I’m a “senior” don’t even need a license anymore!
MJD; congratulations, be sure and “gut” those Fukushima CS137 ATOMS, one gamma ray can barely not kill you.
I’m laughing out loud, but just barely.
@ Rod Adams January 15, 2015 at 5:00 AM
You said “(The dispatchers were provided the equivalent of an engine order telegraph that sent a signal to the operators to tell them the required power level.)”
It was actually a bit different than that. At Davis Besse the Integrated Control System (ICS) master station was called the ULD or Unit Load Demand. All B&W ICS control systems were similar at that control station level (everybody had one). The control station had thumb wheel dials to set rollers, like on an odometer read out. The control room operators could set the High Load Limit, the Low Load Limit, and the rate of change in MW/min. The station also had an Up & Down push button. The operator set the target demand with the Up or Down button (which showed on a meter) and then the ULD control station (or called the Integrated Master at some plants) fed that demand forward to the reactor, feed water, and turbine. And off you’d go to the new demand target at the selected rate.
The ULD station also had an Auto and Manual push button. We ran in Manual and controlled the load as described above. But guess what happened if you pushed Auto?
Yup, you transferred control to the ULD station in the Load Dispatcher’s office. Now he had it, control of the Up and Down from the ULD station there. There was a lot of talk “back in the day” about what NRC thought of that idea, but basically it was not needed as there were hard wired communication links between the control room and dispatchers office. In any event Duke Power said it is part of the installed system, we’re going to test that it works. And they put one of the Oconee units in auto, transferred it to load dispatcher control just to verify it worked. So it was not just an engine order telegraph, rather it actually transferred unit load control to the dispatcher.
I don’t know if that capability still exists, as most plants have changed the original analog ULD to digital technology. So they may not have upgraded the equipment in the dispatcher’s office as it was not needed.
@mjd
Thank you for the correction. My misunderstanding is probably due to the fact that I was interpreting, from memory, a conversation that took place in my office over coffee and chocolate several years ago. My source was reminiscing, so there might have been some initial fuzzy areas in his description.
However, I suspect that the interpretation problems were mine; he was a sharp operator who had qualified on and operated several different plants. People like him — and you — tend to have clear memories of operational details.