Chapter 3
Towards Fourth-generation
Nuclear Reactors 1
,
3.1. Context
In the context of increasing energy needs, an increasing number of countries want to integrate nuclear energy into their energy mix (which will result in a decrease in green house gases emission and avoid the reliance on fossil resources that are progressively being exhausted).
This “nuclear rebirth” should lead to a greater pressure on uranium ore. Thus it must be accompanied by a policy of resource preservation. This is one of the big stakes in fourth-generation nuclear reactors, which appears to be the keystone on which tomorrow's nuclear industry will be able to be built.
This fourth generation will be fast neutron reactors that are able to transform a large portion of uranium 238 into plutonium 239, and can thus be used for electricity generation. Through this transformation it becomes possible to use not just 0.5–1.0% of natural uranium, but >90% for the purposes of electricity generation. It also enables us to consume important stocks of depleted and reprocessed uranium that alone could supply the current level of electricity production in France for a few thousand years. The world “availability” in primary fissile resources can thus be multiplied by 100. These reactors also enable us to recycle plutonium by taking advantage of its energy potential.
The other big stake in fourth-generation reactors is the facilitation of the management of nuclear waste by decreasing the volume and intrinsic long-term radio-toxicity of ultimate waste. These reactors would, in fact, be able to burn a significant part of the long-life radioactive elements that make up the waste — the minor actinides (americium, neptunium, curium, etc.).
The Generation IV International Forum, launched in 2000 by the American Department Of Energy, was born from this desire to create an international research and development framework able to catalyze the research efforts led by different countries to enable emergence of the highest performing technologies at a greater speed.
Four main objectives have been defined to characterize the systems of the future, which should be: sustainable (economical in terms of natural resources with the minimal production of waste); economical (in terms of investment cost and production cost by kWh, which must be competitive with respect to that of other energy sources); safe and reliable; and resistant towards proliferation and easy to protect against external attacks.
In France, program law n° 2006–739 of June 28, 2006 relating to the sustainable management of nuclear materials and radioactive wastes demands that a fourth-generation nuclear reactor be in operation by 2020. The choice has been made to use ASTRID (the Advanced Sodium Technological Reactor for Industrial Demonstration), a sodium fast reactor (SFR)-only path that is able to fulfill this requirement by the imposed deadline because it benefits from maturity and sizeable feedback in France and abroad.
To fulfill the aforementioned goals, the fourth-generation SFR should make significant progress in the area that is of particular interest to us by decreasing the probability of an accident leading to a core meltdown. It will achieve this through increased prevention, improving the inspectability of sodium structures, particularly those that have a safety function, and decreasing the risks linked to the affinity of sodium for oxygen: sodium fire and sodium water reaction.
ASTRID therefore wants to show significant advances in the qualification of innovative options in the areas identified above. Its features should be able to be extrapolated to future high-power industrial SFR, particularly with regards to safety (which should be of a level at least equal to that of third generation power plants) and operability. Its features will allow it to be representative of the core and main components and it will offer experimental capabilities. Its design should therefore have some flexibility so that more innovative options can be tested in a second step that would not have been chosen in the initial design. Novel inspection and instrumentation technologies should be tested, as should new fuels and even new components on the circuits.
3.2. Surveillance and acoustic detection
In this context, the first project, relative to acoustic detection and what it could bring to the early detection of some evolutions intervening in steam generators, has been suggested within the framework of GIS-3SGS.
The second phase consisted of competent laboratories processing acoustic signals recorded in 1994 in the steam generators of the Scottish PFR reactor, during the voluntary injections of water vapor and argon gas in the liquid sodium. It also included processing of signals that were collected during the “end of life” scientific program carried out by Phenix, a French SFR prototype of 250 MWe. This phase was completed in 2010. We want to characterize and classify the different sequences recorded, which will enable us to update the analyses and treatment already carried out in 1996 and 1997 by CEA (the French Atomic Commission) on the recorded PFR signals. The different tools used by the universities involved can then be compared.
The results were conclusive (providing an interesting enlightenment on the applicability of the detection technique of loss-of-coolant in the steam generators by passive acoustic detection). A wider program was then established in order to define the techniques to be used to obtain a gain considered useful by the reactor designers from the point of view of safety.
Acoustic detection is also a promising method under development in the area of instrumentation and surveillance. It could enable us to detect the boiling of sodium within the core and has particularly interesting potential in the earlier detection of accidents linked to this phenomenon, thus constituting a new line of defense. The first line of defense is the detection of delayed neutrons.
It is very useful for the future of sodium reactors for us to ensure that among the stakes linked to this issue that the knowledge gained by the current operators of the Phenix reactor in Marcoule (stopped in 2009) has capitalized on, and that the associated simulator, in its current version, has been properly taken into account. Generally, it would be useful if we could be sure that there is no bias between the physical reality and the hypotheses of the software of the simulator. An ergonomic study would also be interesting.
The decade that will pass between the stopping of Phenix and the starting of Astrid leads us to favor transmission feedback from the operators of Phenix to that from the future operators of Astrid. This action would involve interviewing operators and appraising the Phenix simulator in order to better define the specifications of the future Astrid simulator. For control and surveillance sensors, we could study how the physical parameters measured can/must be completed by parameters from models and simulations (for example, the material balance or thermal energy balance), as well as the possible correlations between different sensors (the data fusion method).
The crosscheck between measurements and models also has a stake. The coherence between the different information layers must be evaluated, for instance, the sensors that enable us to follow the load operation of the reactor and the automatic regulation loops. There are a number of questions that must be answered here. How this information can be taken into account at the “surveillance” level? Is there communication between the diverse layers? Are there methodological locks that must be investigated?
The current reflections on the surveillance conditions of sodium, the coolant for heat transfer in SFR reactors, aim to enable the emergence of new processes to complete, improve and diversify the current systems (towards the rich feedback) and those whose study is ongoing. These new techniques should be adapted to the environment of the coolant sodium by taking into account its specific properties, for instance the risk of inflammability at temperatures close to the fusion point and its exothermal reaction with water generating hydrogen and sodium hydroxide. It is also necessary to define the relationships between the surveillance level required, the potential level of damage and the safety requirements.
The surveillance of SFR (sodium-cooled fast reactor) and the associated controls lies initially on the fulfillment of three safety functions:
— control of the nuclear reactivity:
– insertion of rods absorbing the neutrons at the level of the core (rods referred to as command or control); and
– neutron measurements (in core or deported);
— evacuation of the thermal power by the coolant fluid:
– thermal measurements; and
– flow measurements (or rotational speed of mechanical pumps);
— confinement of products that are dangerous to the environment (radioactive products and sodium):
– detection of clad failure (if any);
– detection of sodium leakages (within the energy conversion system exchangers, and in air); and
– pressure measurements.
Furthermore, it is necessary to have continuous surveillance for inaccessible zones at our disposal. For instance, zones in the main vessel situated under the core are far from the slab through which the sodium instrumentation can enter.
3.3. Inspection during operation
The opacity of sodium forbids the use of optical processes for inspection when a reactor is in operation. This is why ultrasonic techniques are used: it enables us to test the sodium zones (telemetry, metrology of immersed surfaces, defectometry and imaging).
The associated accuracy, however, remains linked to that obtained by the eye of an operator (a specification required by the regulation).
Research and development in the use of ultrasound is enabling us to face the drawback due to the opacity of sodium — the distance measurements, imaging, thermometry and flowmetry, the detection of phenomena leading to the presence of gas bubbles in sodium and volume inspections —. Ultrasounds are poorly attenuated in sodium (less than in water), except when it is engased: the gas rate of the liquid must be monitored because it has very important effects on its acoustic properties.
The performances of the inspection techniques carried out while the reactor is in operation depend on the performances of the ultrasound techniques and on the constructive arrangements instigated during the design. Iterative work enables us to adapt the design towards the implementation of inspection and to develop inspection tools to meet the required controls.
These include:
– putting of sodium immersed targets in place that “grasp” the waves (or any other intrusive vector) well and reflect them towards the signal receivers;
– design favoring accessibility to include the sensors up until the zones to be controlled;
– recourse to inspectable witness zones;
– recourse, if necessary, to structures that can be taken apart (components such as heat exchangers, pumps, sodium purifiers);
– the development of repair techniques in situ (under an argon gas bell);
– the design of telemanipulation techniques (rigid rod, articulated arms, etc., whose trajectory must be defined and controlled); and
– the development of robots that embed the control and repair tools.
3.3.1. The case of acoustic measurements
Local or global acoustic measurements seem to be an interesting path to consider because of their non-intrusive nature, despite the need to locate acoustic sensors above the reactor slab (which avoids or removes the introduction through the upper slab that is already obstructed, or the risk of generating additional impurities in the liquid sodium) and because of the speed of information transmission.
The complexity of such a measurement will come from the multiplicity of noise sources (moving pumps, flows internal to the main tank, etc.) and their variation (according to the regime) and hence from the difficulty to measure and discriminate against the origin of an acoustic variation.
Knowledge of the source noise of the defect to be detected, gained through experiments carried out in experimental devices, for instance, will not be sufficient to establish the feasibility of the detection principle. For this, data about the background noise is also necessary: its intensity, spectral distribution, attenuation, etc. In the absence of a French reactor in operation that could be representative of the future SFR, these data will come from simulations produced by ad-hoc tools (possibly validated by data coming from the operation of Superphenix, Phenix, or even models if need be).
3.4. Conclusion
As you will have noticed, the acoustic processes are a major part of the means of surveillance and inspection of future SFR reactors: the complexity of the propagation modes of ultrasound in structures immersed in liquid sodium has led to the development of simulation tools and signal processing methods, an activity that fulfills the goals of the actions promoted by GIS-3SGS.
1 Chapter written by Jean-Philippe NABOT, Olivier GASTALDI, François BAQUÈ, Kévin PAUMEL and Jean-Philippe JEANNOT.