Different Small Modular Reactor Designs
We look at the various technologic solutions for small water, molten salt and gas cooled reactors and get into the details of why these reactors are potentially much better than existing mega reactors.
Today there are about 50 small modular reactor (SMR) projects being developed around the world. The variations in technologies pursued varies enormously. Here I am going to give you an overview of a few of these reactors that look most promising, and explain the pros and cons of their different technological approaches.
This story was inspired by a recording of conference where Fermi Energia presented their findings about most promising SMRs to pursue for Estonia.
Engineering Choices for a Nuclear Reactor
Before discussing different designs it is worth getting an overview of what are all the important parts of a reactor that we can modify to change what type of reactor we have.
Seen from the most basic level, nuclear reactors are just a fancy way of boiling water. The fundamental operation from a birds-eye view is similar to that of a coal plant or natural gas plant: You heat up water until it boils. The steam drives a steam turbine, which when rotated drives an electric generator producing power.
The engineering choices for different reactor types could be simplified to:
- How do we make nuclear fission happen and sustain it?
- How do we control this fission so it doesn’t run out of control?
- How do we deal with the heath from the fission process. How do we transport this heath to the boiler?
- What sort of fuel can we use? Thorium, Uranium-235, Uranium-238, Plutonium and in what mix and concentration?
To make nuclear fission happen we typically need to concentrate the nuclear material. If e.g. the Uranium atoms are too far apart, then the neutrons from splitting one Uranium atom cannot easily hit another Uranium atom and split that one.
Fuel can be packaged in a lot of ways:
- Solid fuel rods
- Liquid form, such as in a molten salt.
Usually there are multiple layers and coatings. It isn’t like a coal plant or gas plant where you just dump the fuel in raw. For a nuclear power plant, fuel is almost always produced to be in particular way. It can be made into solid pieces of different shapes with different layers and coatings depending on how you intend to use the fuel.
The choice of fuel type and concentration is also strongly aligned with whether you want to split uranium atoms using fast or slow neutrons. Slow neutrons are often referred to as thermal neutrons because they are brought to the same temperature as surrounding coolant (keep in mind temperature is average velocity of particles).
Categories of Reactors
If you start reading about nuclear reactors you will quickly get lost in three-letter abbreviations such as LWR, BWR and PWR. So to help give you an overview I have made hierarchy showing the relation between different reactor types.
Keep in mind there are more ways to categorize reactors. This diagram is categorizing reactors primarily on how they are cooled. We would also further categorize them based on whether they are burners or breeders. But before we go into that rabbit hole, let us try to make sense of this diagram.
The diagram is mainly about the coolant used. What do I mean by coolant? The uranium fuel will generate a lot of heat. You need to get this heat away from the core and somehow move it into a steam turbine to generate power. Thus the job of the coolant is to make sure the fuel doesn’t get so hot that it melts. Secondly it must be able to carry away the heath so it can be made to do useful work, such as generating steam to drive turbines.
The diagram does not represent an extensive list of the nuclear reactor designs that exist. Rather the diagram has been narrowed down to focus primarily on these three reactor types:
- WR — Water Reactor. The coolant used is some kind of water. There are many ways in which it could be used. Most common reactor type today.
- MSR — Molten Salt Reactor. Salt in liquid form is used for cooling. Usually it will be some fluoride salt. Only exists in experimental reactors.
- GCR — Gas Cooled Reactor. Used with early reactors, e.g. the ones use in early weapons programs to produce plutonium. Gas is often CO2 and Helium.
While I personally don’t find water based reactors very exiting they are useful to know about because most nuclear reactors today are water based and the first SMRs to get online will likely be water based such as the one from NuScale.
We could divide water based reactors into:
- LWR — Light Water Reactors. Most reactors in the world is a variant of this.
- HWR — Heavy Water Reactors. In practice these are all of the PHWR type, meaning Pressurized Heavy Water Reactor. Good example of this is the CANDU nuclear reactors in Canada.
Light water in this context means regular water. It is to distinguish it from Heavy water where each hydrogen atom has been replaced with deuterium. This is a heavier isotope of hydrogen consisting of a neutron and proton.
The Gas Cooled Reactors (GCR) can be split into a bunch of different versions. I have only included these two:
- AGR — Advanced Gas Cooled Reactor. This is just a name to distinguish them from earlier GCR reactors which were not efficient at making electricity. They were mainly for making plutonium. AGR is basically a CGR for power production.
- HTGR — High Temperature Gas Cooled Reactor. Operate at very high temperature. Meaning the coolant gets very hot. Also sometimes referred to as very-high-temperature reactor VHTR.
Okay this was just a quick overview without getting into too many detail. That will help keep you oriented when reading about specific SMR reactors. I will focus on a couple of Light Water Reactors (LWR), Molten Salt Reactors (MSR) and one CGR which I believe is a HTGR.
Light Water SMRs
We will look at two different reactors which are close to getting into the market which is the NuScale and the Hitachi BWRX-300. Fermi Energia does a comparison of these two reactors as it considers them good candidates for any country that wants to deploy SMRs quickly.
While both reactors are LWRs, they are quite different, remember LWRs are split into:
- BWR — Boiling Water Reactor. This keeps lower pressure than a PWR and the reactor generates steam directly to drive turbines. Hitachi BWRX-300.
- PWR — Pressurized Water Reactor. Relies on a heath exchanger to heath up water at lower pressure outside of reactor to make it boil and create steam to drive turbine. NuScale.
One useful metric to talk about when comparing reactors is capital costs. Here the cost to preferably beat is $3000/kW. That is roughly what cheap coal power costs. If you get below that you can start outcompeting coal plants in developing countries.
However due to the increased usage of renewable energy it is regarded that storage capacity is valuable. So any system offering storage is seen as competitive at $3500/kW.
NuScale doesn’t beat this cost but has the advantage of being available earlier. It expects capital costs (capex) of 4000–5000/kW. Hitachi claims 3000–4000/kWh in contrast.
However it is worth noting that NuScale offers a more flexible system. Their power stations are made as facilities holding up to 12 individual reactors generating 60MW each, while the Hitachi solution offers a larger single reactor producing 300MW. Thus by going for NuScale one may be able to have smaller initial expenses and can gradually expand later.
The most common nuclear reactor type today is the pressurized water reactor (PWR). The animation below shows how it works. The reactor vessel is where the uranium fuel roads are located which get really hot due to fission occurring.
These rods are inside highly pressurized water, this means that the water doesn’t boil despite reaching temperatures of 315 °C. To achieve this the water must be at 150 to 160 bar. This is one of the key reasons why Nuclear reactors can be dangerous.
This hot water is brought through pipes into the steam generator. The water never interacts with the water in the steam generator. Instead a heat exchanger is used to move heath from this hotter water to the cooler water inside the steam generator. The steam generator water is at much lower pressure which cases the heath increase to cause the water to boil. The steam generated drives a turbine which again drives a generator producing electricity.
What you notice in this schematic overview is that there are pumps everywhere driving the water through each loop. One of the key causes of nuclear accidents is from one of the pumps getting stuck. E.g. if the water doesn’t circulate in the main loop it will quickly get too hot. If it gets too hot, then water gets split into hydrogen and oxygen which gets released. Hydrogen is highly flammable and can cause explosion.
NuScale solves these problems by circulating water without pumps. If you look at the diagram below you can kind of see how this works, although it is not as easy to watch as the earlier animation.
Notice that in the centre, the pressurized hot water moves upwards. But this is not due to a pump but rather due to convection (hot water rises). On both sides you got heath exchangers. This will transfer heath to the steam generator containing lower pressure water, which then boils. This also causes the water in the main loop to cool and thus move downwards, getting back into the reactor core where it is heated up again.
This whole reactor is submerged in water, so that if something goes wrong, the reactor can be cooled down passively by surrounding water. No pumps required. Hence the problem at Fukushima, where the power driving the pumps was lost would not cause problems in this design.
The Hitachi SMR is quite different. It is a Boiling Water Reactor (BWR) which you can see from this animation is based on a simpler principle than the PWR used by NuScale.
In this case we keep lower pressure water in the reactor and generate steam directly in the reactor which drives a turbine. So the main reactor vessel has much lower pressured (70–75 bars) water than a PWR, which is safer. However because water from the reactor goes straight into the turbine, you get radioactive contamination of the turbine, which needs to be well insulate, from any leaks.
The BWRX-300 is based on a larger somewhat more complex BWR reactor type called ESBWR which stands for Economic Simplified BWR. Explaining how this reactor works is a bit more complex as it involved a whole bunch of different water pools and pipes.
However the diagram above of an ESBWR reactor core gives some sense of how it works. You can see that water circulates here without pumps using the fact that water steam naturally rises. In fact this is much easier to achieve than what NuScale relies on which is based on natural convection (hot water being lighter and rising).
I am not expert on this field but based on various criticisms I have read of the NuScale reactor, it may not be able to cause circulation to happen purely without pumps. And submerging the whole reactor in water seems like a potentially complicating factor, which can cause corrosion and other problems. BWRX-300 comes across as a more conservative and proven design as it is an evolution of the older ESBWR reactor.
Molten Salt Reactors (SMR)
The LWR small modular reactors may bring major improvement simply by embracing passive safety systems and potentially allowing economies of scale from mass production of reactors.
However I think the real advantage of SMRs will be realized with Molten Salt Reactors. They are further away in time. We may see NuScale and Hitachi building reactors in 5 years, while Molten Salt Reactors (MSR) will likely take another 10 years and arrive in 2030.
However that wait may well be worth it as these reactors offer a host of advantages not found in LWR reactors:
- Reactor vessel is not pressurized. Removing the dangers inherent in operating with high pressures.
- Passive safety such as cold-plugs, where reactor is shut down immediately if it get too hot or power fails.
- No danger of radioactive fallout spreading far in case of accidents since radioactive elements are bound in salt and not volatile.
- Much higher operating temperatures, allowing a host of new applications: hydrogen generation, thermal storage etc.
There are many ways of making a Molten Salt Reactor (MSR) but most of them will work roughly as shown in the schematic below.
A fluoride salt will contain the fuel itself. Remember a salt is a chemical bond between a gas and a metal. Thus you can bond a Uranium atom with Fluorine gas to form a salt.
Nuclear fission will be sustained as long as there is enough concentration of fissionable material. Basically enough Uranium atoms need to be close to each other for the neutrons that some of them send out to occassionally hit other Uranium atoms and cause a chain reaction.
When pumped into the heath exchanger, the salt will not be concentrated enough to be causing chain reactions. This limits heath generation to the reactor core. Geometry plays a role here. E.g. if you pour radioactive material from a container with low density to one with higher density such as a sphere, you can cause an explosion due to rapid chain reaction occurring.
At the heath exchanger, heath will be transferred to e.g. water to produce steam to drive a generator. Alternatively one can imagine transferring it to another salt for thermal storage. This is a well known technique used with concentrated solar arrays.
With good insulation you could store lots of heath in molten salt for many hours. This allows you to delay when you use the heath bound in the molten salt for turning water to steam and drive turbines.
What makes MSRs much safer than LWR reactor is that in cases where you get some failure of the cooling system or something else which causes the temperature in the core to rise quickly you have much larger buffer to act.
It takes much longer time for salt to reach temperatures which present a problem. At the bottom of the reactor is what is called a Freeze plug which is just salt which is continuously cooled to remain solid. This freeze plug could unfreeze and open up in two cases:
- Salt in reactor gets too hot and thus melt the freeze plug.
- Power fails and thus there is no power to drive the cooling of the freeze plug.
In either case salt with the fuel will pour into emergency storage tanks below. Because the fuel gets separated into multiple smaller tanks, it will no longer be critical. This means the concentration of fuel is too low to sustain a chain reaction. That has the effect of causing a reactor shut down. The beauty of this approach is that it provides entirely passive safety.
Should even this fail and we get leaks of reactor salt into the environment, that is still not a major problem. It sucks for the plant, and it may not be operational anymore. However it will relatively quickly cool and form a hard salt crystal. In here all the radioactive waste is contained. It doesn’t seep into the ground or get dispersed into the air like when an LWR reactor melt down and leaks. There is no volatile material which can be pushed out far as dispersed as fallout over a large area as happened with Chernobyl.
Let us look at some of the interesting companies building these reactors.
Moltex is a UK based company building an MSR they call the Stable Salt Reactor (SSR). The SSR reactor comes in several flavors SSR-W, SSR-U and SSR-Th, which you can read about here. I will however focus on the SSR-W which they intend to put into operation first. As far as Nuclear reactors go, their solution offers several key advantages:
- The SSR-W runs on Nuclear waste. Hence if you are worried about nuclear waste and want to get rid of it, you would actually want to build this reactor type.
- Connects to molten salt thermal storage called GridReserve, to store thermal energy for later use.
The later part requires some further explanation. They pair their nuclear reactor with a steam turbine which can run on 3x the power output of the reactor. Hence if they have a reactor producing 300 MW, their steam turbines can handle 900 MW. Why is that useful?
Because with renewable energy you got excess production of power in periods. In several grids like the German and UK Grid you can have periods of the year and during the day where electric prices go to zero or even negative.
In this scenario it is hard to make a good income with a regular Nuclear power plant as you may have pay people money to take your power.
But with GridReserve Moltex plans to simply store the heath generated from the reactor in molten salt thermal storage tanks. This can be kept for several hours, so that later they can generate 900 MW for up to 8 hours or 600 MW for 12 hours. They can do this in the period of the day where the sun doesn’t shine or there is little wind and where power can be sold at a premium.
This boost the economic proposition of building this nuclear reactor.
With ThorCon the innovation is not really in the reactor itself but in how they plan to deploy and build it. They use well proven Molten Salt Reactor design and don’t try to come up with anything unproven or new.
Instead the ThorCon innovation is in how the reactors are manufactured. One of their key observations is that building reactors on site is really slow. Reactors can be delayed more than a decade. Building reactors in a foreign country with limited expertise in reactor construction can slow down the whole process.
Instead they intend to learn from shipyards. Large shipyards construct enormous ships such as oil tankers very rapidly at low cost.
ThorCon thus designed their whole power plant mostly as you would build an oil tanker. It uses very similar dimensions. Thus the whole power plant is rapidly assembled at a large ship yard and when done it is transported by sea to the shore of the area it will power with electricity.
Because shipyards today have a lot of excess capacity ThorCon plans to utilize this fact to be able to mass produce their power plants quickly.
ThorCon actually uses Thorium rather than Uranium as fuel. However that seems to be primarily for political reasons. Doors get shut in your face when you say you are making a nuclear reactor using Uranium, but politicians and investors have picked up on the Thorium hype, making it easier to get access and acceptance. In principle there is no such thing as a Thorium reactor. There are just different kinds of reactor designs, some of which are able to run on Thorium in addition to Uranium.
Seaborg based on Denmark. Their MSR is called CMSR which stands for Compact MSR. Like ThorCon they can in principle run on either Uranium or Thorium. However they are primarily pushing to have their first version run on Uranium as this is deemed easier and faster to get working. Like Moltex they also plan to be able to fuel it with Nuclear waste. Turning long lived waste into energy and short lived waste.
Their aim is to make small reactors which they can mass produce in a central location and transport to customers. Like ThorCon they want to target developing countries. Both ThorCon and Seaborg remark on attitudes towards Nuclear power is far more positive in this part of the world and they will need to expand power production a lot.
In particular Indonesia is not well suited for Wind and Solar Power which means a lot of coal plants may end up being built. They want to avoid that by being able to offer safe Nuclear reactors instead, which they can transport by ship and install on sight.
Unlike ThorCon however they plan to have the non-reactor part built on land. ThorCon is building the whole power plant with water reservoirs, steam turbines, generators and everything on floating power plant built entirely in a shipyard.
High Temperature Gas Cooled Reactors (HTGR)
HTGR reactors is a type of GCR reactors. The early GCR was generation I and was used for plutonium production for nuclear weapons.
Modern HTGR reactors in contrast are seen as ways of bringing higher safety, efficiency and varied usage cases. Japan has pioneered this kind of reactor type with their high-temperature test reactor (HTTR) in Oarai, Ibaraki.
These reactors have a number of advantages:
- No water used for cooling, so we don’t risk dangerous phase shifts from water to steam.
- Avoid generation of hydrogen gas in case of accident, as there is no water to split hydrogen off from.
- Loss of coolant is not a problem. Graphite which the fuel is embedded in has high thermal conductivity and capacity which means its takes 2–3 days before dangerous temperature levels are reached.
- No danger or radioactive fallout. TRISO fuel particles are used which contain all bad radioactive fission products.
- Very high temperatures allow use for industrial processes, thermal storage of energy, hydrogen production.
- Helium used for cooling can be used to drive turbines directly, increasing efficiency. High temperatures also increase efficiency.
Let us talk a bit about how these reactors work. Below you see a VHTR reactor schematic which is pretty much the same as a HTGR.
This picture is not ideal as it shows a heat exchanger without any turbine for generating power. So this diagram below is a far more accurate depiction. But I included both, as the the other one shows better how the control rods and graphite moderator is placed.
The way this reactor works is that graphite works as a moderator and neutron reflector. Inside the graphite we have tiny pebbles with uranium fuel. These will decay and send out neutrons which when hitting the graphite will slow down (get moderated) and likely reflected, increasing the chance that these slow neutrons will hit another Uranium atom where the will be captured.
A Uranium atom which catch a neutron will typically split shortly after catching the neuron and divide itself into two parts and send out 3 new neutrons. The graphite slows down the neutrons and reflect them so they are more likely to hit some other Uranium atom stored in one of the other small fuel pebbles. This is what continuous the chain reaction. As in most nuclear reactors there are control rods embedded within the graphite which can be moved up and down to reduce or increase the speed of the chain reaction. These rods absorb neutrons thus reducing the number of chain reactions.
The graphite has a dual purpose in that it also conducts heath very well so that heath from the fuel pebbles is quickly spread through the blocks of graphite. Lots of tiny holes going through the graphite allows this heath to be transferred to Helium circulating through these holes. The Helium moves the heath to turbines, hydrogen production facilities and/or district heating.
This is where the high temperature gives a benefit. Each purpose reduce the temperature. E.g. if the temperature gets too low, it may not be capable of driving a turbine anymore, but it could still supply district heating. This way you can with this system get 80% efficiency out of the fuel.
As we have talked about repeatedly, the key problem when a reactor fails is getting the residual heath away from it to avoid meltdown. Dropping the control rods will stop the chain reaction, but there is still residual heath. In the case of Fukushima it was the residual heath which made the water too hot and caused water to split into hydrogen and oxygen. This happened because the pumps failed and the flow of cooling water stopped.
This is not a problem with a HTGR reactor. The graphite will transport the heath quickly away from the fuel and it connects with the reactor walls so that it can transport heath into the surrounds of the reactor. E.g. the MMR reactor is placed under ground so that heath can be spread into the surroundings.
In addition graphite has very high heath capacity, hence it takes a lot of heath to raise its temperature. This means the heath capacity alone buys the operators 2–3 days before they need to act.
This is what gives these reactors walk away safety.
Ultra Safe Nuclear Corporation(USNC) builds fairly small reactors. Keep in mind e.g. that Moltex targets 300 MW, while NuScale intends to have up to 12 reactors of 60 MW each in a facility. The MMR reactors in contrast are a tiny 5 MW of output.
However they have an interesting concept for how they imagine power production can be expanded. Multiple small reactors all heath up the same molten salt reservoir. Thus you could connect this reservoir of molten salt to steam generators to drive turbines, provide industrial process heating or any number of applications.
What you get out of this is similar advantages as Moltex. You could act like a battery.
It is worth nothing however that USNC intends to target a different market than the other companies I have listed. They target remote locations such as mines. These typically run diesel generators for power.
That is an expensive way of generating power. Thus USNC can afford to sell reactors at higher price per MW. That is what allows them to make economic sense with such small reactors. Otherwise that would likely not be economical.
They do however offer a huge advantage to these remote locations such as mining communities. A reactor will last a whole 20 years with no refueling. When the fuel is spent the whole reactor is removed from site and a new one inserted. That means handling of nuclear waste is really simple. The reactors are so small they can be easily put on the back of a truck and driven away.
This makes it easy for maintenance. There is no need to have a crew on site which can refuel the reactor each year. It will just keep humming along for 20 years. A spent reactor can in principle just be trucked straight to long term storage.
SMR reactors come with the promise of cheaper, safer and more flexible reactors. There is however a chicken and egg problem. If reactors are not mass produced they may not end up getting cheap, and if they are not cheap people will not order a lot of them.
However as we have seen, some companies are working around that problem in different innovative ways. Molten Salt and Very High Temperature Reactors will due to the simplicity of their design compared to LWR reactors get major cost reductions.
ThorCon is attempting to drive costs further down by doing power plant construction in shipyards.
USNC in contrast seek to counter the problem by selling reactors to markets where power production costs are usually a lot higher, such as in remote mining locations.
Moltex seeks to be cost competitive by utilizing thermal storage to avoid selling power when prices are low and instead sell when the prices are high.
The non-LWR reactors also seek to gain extra income from ability to produce hydrogen. Usually green hydrogen would be made from electrolysis. But this is a fairly inefficient process.
However electrolysis is not the only way. If you fill water with sulfuric acid, you can create hydrogen by simply heating up the water. One can use the Sulfur–iodine cycle or the Hybrid sulfur cycle. In Both cases you need temperatures above 800 °C. This is hard to get out of a LWR reactor as they cannot be operated at this high temperature.
A HTGR reactor can however. And this is more energy efficient. The MMR reactor e.g. has electric power output of 5 MW, but produce 15 MW of heat, 3 times as much.
Hydrogen has many uses, for energy storage, as a reducing agent when producing metals from ore, it can be used to create synthetic fuels e.g. for use in airplanes which cannot easily be fitted with batteries etc.