Wednesday, June 17, 2009

Large vs. Small LFTR Reactors

Large vs. Small Nuclear Reactors

A comradely debate and discussion has been going on for a number of years over on the web site, the pre-eminent web site for all things concerning the Liquid Fluoride Thorium Reactor.

This contribution to that discussion will focus on the LFTR but a few other items of important to note.

This week, Babcock and Wilcox announced their new mPower reactor. It's a light water, pressurized water reactor like almost all currently running reactors but they are very, very small: 125MW modularly build, sited underground, reactors. B&W is suggesting that costs will run between $4,000 and $5,000 a KW installed, or, about $1 billion and change for one of these modules.

B&W joins a host of other companies and proposals for small reactors that include:

Hyperion Power Generations [] Hyperion Power Module, designed for process heat (approx 70MW thermal) and 30 MWs of electricity if hooked up to turbine generator stet.

Toshiba's 4S (Super Safe, Small and Simple) reactor. [] This power plant is designed to provide 10 Megawatts of electrical power.

Even the Russians have developed one off of their advanced lead-cooled submarine reactors: The The SVBR-100 reactor. []

Lastly, but certainly not least, Rod Adams, from Adam's Atomic Engines, also has a very small reactor design using a brayton cycle gas turbine with nitrogen as it's cooling source for the reactor and the propulsion for the turbine.

Modular Construction

All these reactors offer modular construction, transportability of those modules or even whole units to a siting, simplified and passive intrinsically safe designs, and affordability, items that for the most part the larger 1000+ reactors currently being deployed cannot, for the most past, offer.

Before getting into the question of 'large' vs. 'small' reactors, we need to parse out what the various reactors can be used for based on their 'product' and use and how the LFTR fits into all this.


The LFTR is unique from all of the above because it is amazingly scalable...there is no real downward or upward limit to the size or use a LFTR can be employed in. Say, from a small LFTR 'battery' of 20 MWs to a large, base load plant offering 1800 MWs gross base-load power to the grid.

It is my contention that there will be a 'market' for all these sizes. We should first review what these markets are.

On the smaller end, the LFTR, as a high temperature reactor, can provide process heat. A small chemical plant, requiring thousands of tons of steam an hour, can use a LFTR to provide this heat and, to electrically power the plant. A slightly larger version may be able to provide power and vast qualities of heat to an oil refinery or a tar-sands operation thus providing carbon-free process heat to what otherwise would be a huge carbon-spewing operation.

These smaller LFTRs, from 20 to 200 MWs could provide, also, site specific load balancing for a grid that has a lot of load in place but generation many hundreds of miles away. Using a 200 MW LFTR to 'anchor' the grid would be very helpful to any utility. Additionally these smaller LFTRs could be plopped down in various transmission substations to provide quick, peaking power or variable load changing that responds to frequency changes throughout the day.

Additionally, along the process heat scenarios, these small and, larger LFTRs can be used to cheaply crack sea water into various products including and most desirably, hundred of thousands of cubic meters of fresh, potable water for drinking and agriculture, thus allowing a huge increase in agricultural output in drier climates, like California, Egypt or Tunisia. Thus electrical generation would be, or could be, the co-generation product of such plants. Additionally, another such product from process heat could be the cracking of hydrogen from water using high-temperature electrolysis, thus providing yet another alternative fuel for automotive transport.

Larger LFTR units could be would be used for base-load generation. Phasing out gas and coal plants with big 1000+ MW units.

There are many uses for all sizes of LFTRs that are all designed basically the same. has papers and discussion on how all this works, so I won't go into basic LFTR technology here.

One thing that is important for this discussion to note, however, is that LFTRs, from the get go, are cheaper to produce, having a much higher power density than any currently running or under-construction Generation II or III Light Water Reactors. From the reactor core itself to the turbine, size is about 1/2 to 2/3 smaller, thus allowing for a cheaper, and therefore far more efficient, product based on size/cost per MW output. We are looking at, generally a similar ratio in cost reduction. the meat of the issue

It is my contention, or thesis, that there is a use for both the smaller scale, sub-300 MW LFTR units AND the larger, up though 1800MW LFTRs as well.

The argument is poised this way: is it cheaper (safer, easier, efficient, etc etc) to build, for example, 15 100 MW LFTRs vs one 1500 MW LFTR to achieve a 1500 MW requirement for a single location? This is the heart of the debate.

Charles Barton over at and a keen observer of nuclear and 'renewable' energy costs. Many nuclear bloggers have as well. I'm not an economist, financial expert or engineer. My experience is simply one of a interested observer, and commentator, on energy issues. Charles is an advocate of the LFTR deployment being one of small reactors. He has many blog posts on this question and I urge all left-atomics readers to go to his site and look for these posts.

Advantages of Smaller LFTRs...

So...what ARE the advantages of a small, sub-300 MW, or sub-100 MW LFTR? Because the overall size of a LFTR is at least 1/2 the size of the equivalent LWR to start with, and these smaller lifters are 2/3 to 1/10 the out put, small LFTRs may able to be transported, in total, on three tractor-trailers or rail cars (reactor core/housing, turbine, generator). Given that construction costs of large, LWR amount to almost half the costs of a large reactor today, this is a tremendous savings.

Additionally, almost the entire module assemblies may well be assembled in a factory, on a production line similar to that of the building of passenger and cargo aircraft today. Production costs for smaller, but quantitatively greater, components ARE the cheaper form of production based on economy of scale.

The reality, however, is that even for larger reactors of the LFTR variety, almost all their components are also factor built, just like in today's' fleet of reactors. There is something of a 'production myth' that these large reactors are somehow built more primitively. Almost every single component in a modern nuclear reactors is built 'on a line'. But even small pumps, motor, controls, etc are built 'by hand', albeit in a factory. A 100 HP pump is hand assembled for the most part, albeit the parts that compose the whole pump are more likely to be manufactured on a true assembly line process.

I suspect as we parse out the true production costs for larger and small components, even the smaller ones for the small LFTRs that the costs are not that different from the larger units.

The advantage for the smaller LFTR is that many more modules for different aspects of the plant can be cobbled together in a factory as opposed to hand assembled on site. Generally, the rule holds that a system that comes 'ready-to-install' it is cheaper than one assembled by hand on site. A CO2 fire suppression system is an example of this. While the pipe runs to each piece of electrical equipment has to be hand run, the CO2 discharge assembly, composed of dozens of small, say 1" diameter, pipe runs, regulators, pilot and stop valves, on unit that is, say, 10 feet by 10 feet, can easily be built in a factory and shipped to the site. But in a bigger plant, where you need pipe discharge valves and runs that use use, say, 2 and 3" diameter pipe and valves, and is 20 feet by 15 feet, then assembly on site may have to occur. Thus an increase in expense.

But If...

But if...and this is the big "IF", this bigger assembly which can service a 1000 MW plant is more expensive per unit, say by a factor of 3 (for the labor it takes to assemble it on site as opposed to a factory, and the large individual sub-components cost more as well), but you need 10 of the smaller cheaper assemblies for the 10 separate LFTR units, then those cost savings for the 10 smaller units disappear and are actually more expensive by a factor of 3.

So, what all of us need to do, and it would help if actual operations engineers, construction engineers, and manufacturing engineers chimed in here, is actually parse out the true costs of a complete, say, 100MW LFTR and that of a 1800MW LFTR. I argue that it will be *impossible* to determine until all sized units are built in mockups to scale, during the R&D that precedes actual deployment. We simply will not know.

Some more things to think about. A big expense for any reactor and turbine generator set are controls. This including things like monitoring radiation, temperatures throughout the reactor-to-heat exchanger loops to the turbine and, of course, volts/amps/watts on the generator. A generator, regardless of the size, has a voltage relay device that reads the system voltage and the voltage of the generator. A set of leads from the high and low side of the transformer banks "potential transformers" (PTs) runs back to a relay and monitor in the plant's control room. There are *hundreds* of these sorts of systems on any modern power plant, regardless if they are geothermal, conventional steam plants or hydro units.

The PTs generally are all the same size and installation uses the same methods. Their testing, 'stressing' and maintenance are all the same based on conditions at the plant. So, regardless of the size of the LFTR, they will all get the same sort of PTs and all the costs associated with them. So, for 1500 MW plant, the costs for 15 100 MW LFTRs for the PT sets required are *likely* to be 15 times more expensive than for the 1500 MW unit (to be fair, there may be a larger redundancy for the larger units as they often have more than one transformer bank). But even if divided by 2, the costs, or "economy of scale", is now reversed and favor the single larger 1500 MW unit.

And this, fellow nukes, is my point. All these factors have to be taken into account when supplying what the customer wants in terms of gross MWs installed in a single location.

To be fair, this is VERY, very simplified presentation of counter-positions. But this does need to be seriously parsed out, eventually.

I believe the real advantage of the smaller units will be it's flexibility in installations, where they can be located, how, etc. Many times utilities want the bigger units because the are compelled by many reasons to group their "MWs together" in one big unit. The smaller LFTR modules can be "distributed" for, in some/many cases, a more reliable grid. Thus the same 1500 MWs the big public utility may want can now be distributed over a greater, and therefore more reliable area. Because the price difference and environmental impact for doing this with conventional fossil plants was prohibitive, you did not see this occur. Thus the LFTR gives advantage because if it's "infinite scalability" to distribute it's power production.

Secondly, and more importantly, while I will continue to put an "equal sign" between the big and small versions of the LFTRs that will be deployed in the future, for many, perhaps most, parts of the world, the smaller LFTR versions will be highly desirable. Many countries have electric grids that have, because of the legacy of colonialism and imperialism, war and revolution, dilapidated, incomplete or basically retarded development and mass under or non-electrified sections of their countries. This is one of the reasons why India is still interested in nuclear energy in *under* 700 MW units.

A grid that is relying on 200 MWs of diesel powered in 10 locations and subject to blackouts, fuel shortages, can add, for example, as part of a rural and small town electrification program, a set of, say, 3 100 MW LFTRs to boost over all capacity to 500 MWs while load increases from 200 to, say, 300 over a year or two. Multiple non-joined grids could be developed this way and then joined as transmission finally ties the country or region together. Smaller, 50 MW LFTRs can be installed incrementally. Eventually you'd have a larger, regional grid composed of about a dozen or more variable sized LFTRs. As the economy improves because of the addition to the *physical* economy of the nation gets stronger, larger LFTR units could be installed in more centralized locations with the smaller ones previously installed used to balance frequency/voltage and provide peaking power.