gas cooled reactors

=energy =nuclear



Electricity from gas turbines burning natural gas, silicon photovoltaic panels, and big horizontal axis wind turbines is cheap.

Low pressure steam turbines are expensive, because the low gas density means the turbines need to be big. Heat exchangers are also expensive. It makes sense to use them when a heat source is almost free, such as the exhaust gas of gas turbines, so combined cycle gas turbines are common. But if the heat source costs any significant amount of money, or if there's some other cost factor, then the viability of that for power generation is very questionable.

For example, coal is very cheap, selling for what it costs to dig out of the ground and transport it by train, and yet construction of new coal plants isn't considered viable in the USA.

Most nuclear power plants are kind of like coal plants. They have a thing that boils water, steam turbines, and a cooling system that condenses the steam, maybe using cooling towers. This being the case, the actual reactor needs to be very cheap for this to be economically viable, and it's not.



In my opinion, if you want to make an economically competitive nuclear power plant, steam is out.

Historically, steam has been preferred over gas turbines, because compressing gas with turbines decreases efficiency, and the less-efficient older turbine designs made Brayton cycle efficiency lower than using steam. Even now, the temperature has to be significantly higher to get the same efficiency with gas turbines as steam. CO2 can be compressed in its supercritical state which reduces those extra losses, but only a bit; matching the efficiency of a PWR with 315 C water requires maybe 400 C with CO2.

So, temperatures need to be higher. Let's say 600 C to 750 C, which would be 35% to 55% efficiency.

Liquid water also has much higher heat capacity per volume than gases. So, using gas as a coolant requires much higher flow volumes, maybe 20x the flow rate, which potentially requires a bigger reactor, which is expensive. But if using steam is too expensive, and it is, then the only option is to do that affordably.

Water is also a neutron moderator, which is good, because if the reactor overheats, it makes the water expand or boil off, which reduces reactivity. You want reactors to be somewhat undermoderated, with liquid moderator, so that if anything goes wrong the reactivity decreases. Graphite moderator is not safe enough.

You also don't want potential resonance with moderators and reactivity. A big concern about supercritical water reactors has been oscillation of water density and reactivity, which is probably solvable with a proper design, but nuclear reactors haven't been viable anyway, so who cares?

So, if you're using gas as coolant, you still want liquid moderator. One option is heavy water in insulated tubes. (Or, insulated tubes of fuel in heavy water.) About 10% of the heat would still go into moderator, carried by neutrons, so that water would still heat up if the reactor activity increased, expanding slightly and reducing reactivity. And of course, in the case of catastrophic failure, that water would boil off, reducing heat generation to whatever the decay heat is, which is much less than the heat from normal operation.



As I said, heat exchangers are somewhat expensive, compared to the current cost of electricity generation. High temperature ones are more expensive. Ones for corrosive fluids are also more expensive. Heat exchangers for high-temperature corrosive molten salts, such as the fluoride salts proposed for molten salt reactor are, by themselves, too expensive for making electricity. The people who pushed things like liquid fluoride thorium reactors would know this if they weren't blinded by their desire for their pet ideas to work.



Certain principles of nuclear power plant design are apparent from the facts I've listed so far. For an example of failing at all of them, you could look at TerraPower.

TerraPower has a reactor design with nothing that breaks the reaction if it boils off, and potential for increased reactivity if the core geometry changes. The reactivity actually increases as the amount of sodium coolant in the reactor decreases. That's bad. What TerraPower is trying to do is minimize the amount of sodium coolant in the reactor, so that thermal expansion of the fuel can counteract that effect, but that is not a satisfactory solution.

Their whole original concept was a "traveling wave reactor" where fuel would burn in such a way that reactivity is maintained. That was dumb and they've now realized their design wouldn't work, so their new design involves a robot moving fuel rods around inside the reactor. Suppose that robot malfunctions and there's too much reactivity, and the emergency shutdown systems happen to fail at the same time? A Chernobyl-type explosion is then possible. Unlike the Chernobyl plant, they are at least planning a containment structure, but part of their "advantage" is having a much thinner one because the reactor has lower pressure - which isn't exactly an advantage, because pressure differentials increase heat exchanger costs.

Also, TerraPower is doing fast fission, which means a shorter neutron lifetime and thus less time to react to increased reactivity. Overall, their design is unnecessarily unsafe and perhaps unacceptably unsafe.

They're making steam, which as I've said, is too expensive if you have any other expensive parts of a power plant. They have liquid sodium coolant, with a heat exchanger to molten salt, with a heat exchanger to steam, with a heat exchanger to cooling. That's 3 levels of heat exchanger, which is 2 more than a BWR has, and one of them is for molten salt, which is somewhat corrosive. Obviously, that's too expensive.

I suspect the TerraPower people are also big fans of extracting uranium from seawater, despite the fact that the uranium in a cubic meter of seawater produces electricity worth less than even simple processing of a cubic meter of seawater costs.



If you want to use molten salt for energy storage, it can't be very hot because it's corrosive so the heat exchangers are too expensive, and it gotta be hot because steam turbines are too expensive to only run part of the time, so you need to use a different thermodynamic cycle, so the temperature needs to be higher than water boilers use to get decent efficiency. So you see the problem.

This logic also obviously applies to solar thermal power.



The above logic leads us to a CO2 cooled reactor with water as moderator in insulated tubes. This is called a "heavy water moderated gas cooled reactor" (HWGCR) and Wikipedia has pages for 2 such plants: Brennilis and KS 150. Both had some problems, so let's consider the problems they had.

The Brennilis plant was attacked by terrorists using explosives. It worked OK but France decided they wanted to standardize on PWRs because that's what the USA was doing.

KS 150 had several problems, listed on that page:


- two workers were killed due to a leak of carbon dioxide

- ...humidity absorbers covering the rods were not removed, causing local overheating of the fuel (since transmission of heat to the coolant gas was reduced). The active zone was damaged, heavy water came into contact with the coolant and both primary and secondary circuits were contaminated.

- 25% of the fuel elements in a heavy water moderated carbon dioxide cooled 100 MW(e) power reactor were damaged due to operator error. The operators failed to remove silica gel pellets that has fallen into a new fuel element from a damaged pack (there was no procedure available to check the interior of fuel element, therefore only pellets from the top were removed). The silica gel packs were used to keep the unused fuel dry during storage and transport. The silica gel pellets blocked the flow of the coolant resulting in overheating of the fuel and the pressure channel holding it. As a result of overheating the heavy water leaked into the part of the reactor (the gas circuit) where the fuel elements are accommodated, the fuel cladding was subject to corrosion and a considerable amount of radioactivity leaked into the primary cooling circuit (CO2 gas). Through leaks in the steam boilers (similar basic design to a MAGNOX or AGR plant) some parts of the secondary circuit became contaminated.


Yes, if you don't take the covers off the fuel rods, or put a bunch of garbage in the reactor, that can cause problems. And large leaks in power plants and chemical plants can kill people. While KS 150 should've had better diagnostics, and perhaps zirconium cladding would've been preferable, I'd argue that those incidents weren't the fault of the basic design, and if anything demonstrate decent resilience, but the problems at KS 150 were a significant reason why HWGCRs were not developed more. Administrators often see only that a technology was tried and had problems, because people often lie about the reasons for problems.



As I said, CO2 cooling requires greater mass flow through the reactor than water cooling. Is this a serious economic problem? I don't think so. If we compare KS 150 to an AP1000 reactor:

KS 150: 3.56m diameter, 4m height, 773 MW thermal
AP1000: 3.66m diameter, 24.6m height, 3415 MW thermal


KS 150 used metallic uranium with magnesium-based cladding, which made it particularly succeptible to corrosion when water leaked. Metallic uranium has high density and good thermal conductivity, which decreases the relative size of the core, but it also melts at the low temperature of 1132 C, which isn't enough of a margin for a 700 C gas cooled reactor, so it's better to use uranium carbide; its density and thermal conductivity are lower than those of metallic uranium but substantially higher than those of uranium dioxide.

Note that KS 150 used unenriched (0.72%) uranium, while AP1000 uses uranium enriched to 4.25%. Using uranium carbide at 4% enrichment, a hypothetical reactor with similar geometry could be shrunk to as little as perhaps 14 MW thermal with the same power density. Smaller minimum reactor size is good because smaller pressure vessels are generally easier to manufacture and smaller reactors add more design options, but tiny nuclear power plants with 10 MWe output are still dumb.



Anyway, this is my favorite basic concept for nuclear power. If you feel a great need to pursue a new nuclear power plant design, then go for that, I guess. Unless you're making military submarines, in which case you should hire the TerraPower guys or some liquid fluoride thorium reactor fans.



back to index