The Flexblue nuclear reactor design is a 100 m long, 12 000 t self-contained unit affixed to the sea floor with a capacity of 60 MWe

Flexblue is a novel concept in nuclear engineering: a submarine-derived reactor designed to be installed at a depth of 100 metres below the sea. André Kolmayer, head of French nuclear submarine builder DCNS’s civil nuclear programme, outlines how it works.

André Kolmayer, DCNS, France

DCNS has been making nuclear submarines for the French navy for 40 years, and we have put 18 into service. In a nuclear powered submarine there are several major sections. There are the nuclear reactors, the turbine, the condenser, the weapons system, the living accommodation, the hull and so on.

There are around 1500 suppliers who collaborate in the construction of nuclear powered submarines and DCNS integrates everything in its shipyards in Cherbourg. We build in our own premises the pressure vessels for the reactors, the steam generators, the pressurizer, the accumulators and the heat exchangers for the nuclear steam supply (NSS) system. The design of the NSS of the reactor is done by Areva TA, which is a subsidiary of Areva.

Our first SSBN was launched in 1967 and France pioneered the mass development of civil nuclear power stations in the 1970s. Now DCNS sees a good opportunity to adapt reactor technology from nuclear submarines, carefully demilitarized for subsea civil nuclear reactors. There are about 150 nuclear submarines roaming around the world today, so putting a nuclear plant underwater is not exactly a novelty.

We are currently entering into a two-year design phase with Areva where we are setting up an integrated team with Areva TA, EDF and the CEA to develop Flexblue, a subsea nuclear reactor with a 50 MWe to 250 MWe power output.

Flexblue will not be a totally brand-new reactor. The basis for the Flexblue design is only to use proven technologies in order to reduce technical uncertainties.


The design of the Flexblue reactor will be derived from nuclear powered submarine reactors, although it will be carefully demilitarized owing to some major changes. The reactor fuel will be standard 17×17 civil fuel, because military fuel performance will not be necessary for civil use.

Secondly, the secondary side of the turbine and operating part of the reactor will be adapted. Thirdly, the safety systems will be redesigned in order to use more passive safety systems, as the Flexblue reactor will be immersed and stationary on the seafloor.

The Flexblue reactor is a pressurized water reactor. We will keep the pressure vessel and the steam generators of our operating submarines. For our civil version of this reactor we will reach around 60 MWe.

This is the first time that DCNS attempts to enter the civil nuclear market. Military budgets across Europe, including ours, are generally declining and we are looking, like other companies in this field, to diversify in order to maintain industrial capacity, jobs and competences.

To my knowledge an underwater civil nuclear reactor has never been attempted before. The Russians are developing a surface barge containing two twin reactors of seemingly 38 MWe each, derived from nuclear powered icebreakers.


We will have two different types of reactors within our Flexblue concept, corresponding to two market niches which we have identified; a 50 MWe to 60 MWe reactor for one segment of the market, and a 250 MWe reactor for the other.

The Flexblue units will be 100 m long, and 12-15 m in diameter, weighing 12 000 t and moored on the seabed at a depth of about 100 m. There are a number of different possibilities for how Flexblue will be connected to the grid.

The standard option is utilizing proven technologies for the offshore oil industry like risers and substations, but we have not designed in detail how the connections will be made.

Flexblue will be installed where the depth of the sea is about 100 m. This depth can typically be found from 5-15 km from the shore. This is not too far out to sea, and AC lines will be sufficient.

Before installation, moorings have to be installed on the seabed. The Flexblue modules will then be lowered onto the moorings and locked in place with anchors. Flexblue modules could be ballasted, so they float above the seabed, rendering them totally protected against seismic activity and tsunamis.

Flexblue would be tele-operated from onshore. For the major operations like start-up and shutdown it would be done from a control room in the Flexblue module itself. Exactly like a terrestrial nuclear power plant, Flexblue would be able to load follow with the same kind of pertinence as big plants.


When installed in series the Flexblue reactors will be capable of operating independently of each other. For refuelling, our basic option would be to put the reactor on a ship and sail it back to the shipyards, for example in Brest or Toulon, where we conduct refuelling for nuclear submarines. This is also the standard operation for nuclear powered submarines, which take a little bit longer to refuel than terrestrial nuclear plant reactors because they are integrated reactors.

Flexblue will be installed where the depth of the sea is about 100 m, typically 5-15 km from the shore

We intend to optimize the design of the Flexblue reactor in order to reduce the refuelling period to around a month, which is the same as terrestrial PWRs.

The second option is to develop a special purpose vessel that could perform refuelling on site, i.e. bring the Flexblue unit up to the surface and conduct refuelling on the special vessel.

The third option is to propose a spare Flexblue service to the customer. This spare would replace the one to be refuelled for the entire time the operation takes to complete. This would allow the availability of power to be uninterrupted.

Concerning the cycle length, our goal is to have a refuelling cycle of between two and four years. Operating for four years would still need periodic inspections – we plan to have a mini-submarine allowing people to go for periodic maintenance and visits.


Coal is the least expensive source of power generation in the world, and that’s why 50 per cent of the power generation in the US, China and elsewhere comes from coal. But coal has problems and nuclear energy is the answer.

Gas turbines will be Flexblue’s main competitor, but they also have problems, like the transportation of natural gas, carbon emissions, fluctuating commodity prices and so on. Gas turbines have high operating costs, but the operating costs of Flexblue will be very low.

The market as we evaluate it will support between 100 and 300 modular units depending on the production and commercial costs we can manage with Flexblue.

Our target market is very different from conventional nuclear power plants. It will focused electrically medium-sized countries such as Morocco that may be unsuitable for large nuclear plants. Such plants need a total installed capacity of approximately ten times the size of a large nuclear reactor.

We are also targeting Flexblue for isolated regions of developed countries, like islands all over the world (Malta, Cyprus, but also in the Pacific region or Asia). There are around 60-70 countries around the world who have announced plans to develop nuclear power and most of these countries have opportunities to develop and afford Flexblue units.

The question of the costs per kWh, or the levelized costs, of nuclear power is very controversial because the methods used to compute it are vary widely. To give an example, one OECD report states that the levelized cost of nuclear power is the most competitive of all generation sources, while other reports state exactly the opposite. We are convinced that Flexblue will be competitive (see below).

DCNS expects to build a prototype for testing off the French coast by 2017


At present it is not possible to give an exact figure for the cost per kWh of Flexblue, but there are six good reasons why we think it will be competitive with other sources of power generation in the future. Firstly, we have neither huge site works nor civil works, which amount to 25 per cent of conventional nuclear power plants. Needless to say this hugely reduces the investment costs of Flexblue.

The second reason is that Flexblue reactors will have passive systems with fewer safety-grade pumps, valves, piping and cables together with total modular standardization.

All terrestrial nuclear power plants are unique, at least for their ultimate heat sink, due to site conditions, which require a huge amount of engineering and redesigning work. On the contrary, the site conditions at 100 m deep are quite constant and there are few variations with latitude, allowing a true standardization of our Flexblue modules.

Thirdly, our modules will be fully manufactured in our shipyards and workshop. This makes a huge difference with terrestrial plants. Ina typical plant construction, like Finland or Flamanville for example, there may be more than 1000 people erecting the plant, whatever the weather conditions.

We anticipate that the Flexblue reactors will be much more efficient and low cost to build. Moreover, the fact that Flexblue is transportable could permit authorization to “tile” the different phases of a project, such as the administrative phases (early site permit, combined operating licence) and construction of the module in a shipyard. This could allow a huge reduction in project duration risks and financing costs.

The fourth reason is that Flexblue facilitates progressive investment, i.e. additional modules can be added later to the same site. This is a very important financial advantage.

Fifth, Flexblue can be very easily dismantled. Decommissioning terrestrial reactors can be very costly due to the presence of concrete (the dust can be radioactive).

In our case, we only use steel, which is much easier to clean (DCNS has the experience of deconstructing SSBNs). Moreover, we have to remove only the Flexblue upon decommissioning and the site returns to its initial state.

Finally, as the modules are standardized there will be an important volume effect for the production costs, i.e. economies of scale. We think we have very good arguments to believe that we have a concept that will be truly competitive. Flexblue is the only small modular reactor that can claim all of these benefits.

How many Flexblue modules will be operational in ten years’ time? It is difficult to say as it is a brand-new concept. We have chosen proven technologies, but is a new concept in terms of safety regulations, public acceptance, and it could be a new challenge in terms of licensing.


Being immersed under the sea, Flexblue is naturally protected against external events like aircraft crashes, earthquakes, tsunamis and extreme weather, as well as other natural phenomena that must be taken into account when designing a terrestrial power plant.

In addition, local populations are naturally protected against radiation by the remote location on the sea floor and the expanse of seawater between them and the nuclear modules. At the moment, however, there are not any existing safety regulations for the installation of nuclear power plants on the seabed.

This is a more general question for what the International Atomic Energy Agency call Transportable Nuclear Power Plants, like for instance the Russian barges. This problem is generic and must be solved on a generic basis by international bodies.

Once this is done, in terms of licensing by the safety authorities we do not foresee any big technical problems. Once a Flexblue is in operation under sea it is very easy to organize onboard visits for the regulator.

The design life of DCNS’ nuclear powered submarines is 30 years, but like terrestrial nuclear plants it is possible to extend their life cycle. Like terrestrial nuclear power plants, a key factor limiting the life cycle of Flexblue modules is obsolescence of subcomponents, specifically the electronic instrumentation and control components.

After this two-year design phase is completed and we have safety certification we intend to have an order for prototype and after that we will market the product commercially. We will not fund the prototype ourselves, we will find the customer for it. If the results of validation studies are positive, a prototype could be tested off the French coast in 2016 or 2017.

More Power Engineering International Issue Articles
Power Engineering International Archives
View Power Generation Articles on