On March 11, 2011, an earthquake, followed shortly thereafter by a tsunami, severely damaged four reactors at Japan's Fukushima Daiichi nuclear power plant. The worst case scenario occurs: the reactor cores can no longer be cooled. The next day, a first hydrogen explosion occurred in reactor building 1 and then in the three other buildings, which significantly worsened the accident situation and the release of radioactive substances into the environment.
The risk associated with hydrogen is well known and has been studied since the 1980s after the first major nuclear accident in a reactor at the Three Mile Island Power Plant in the United States. The sequence of phenomena is as follows: during a prolonged failure of the cooling systems of the core of a nuclear reactor, the fuel rods heat up. This promotes a chemical reaction of oxidation of the zirconium (a metal) fuel cladding and then the remaining metal structures of the reactor core by water vapor. This reaction produces dihydrogen and is itself exothermic. It is likely to run away if the rate at which the heat is dissipated is much slower than that produced. In this case, the fuel rods can heat up to the melting point. This reaction also releases a large amount of hydrogen (H2), a flammable gas: “In a 900 megawatt reactor (the smallest in the current nuclear power plant), the complete oxidation of the metallic components would produce around 900 kilograms of hydrogen,” explains Ahmed Bentaib, researcher in the Major Accidents Division of the IRSN.
When adequate cooling options are not available, the molten materials form a magma called “corium” (see “Nuclear accident: How to contain a reactor's molten core?”, Pour la Science #514, August 2020). This then flows to the bottom of the reactor vessel until it penetrates it and then erodes the concrete floor of the containment whilst continuing to produce hydrogen and a significant amount of carbon monoxide. Carbon (CO), another flammable gas.
During these different phases there is a risk of explosion as the gases accumulate in flammable clouds in certain areas of the containment. The presence of a very weak energy source, such as contact with a 580°C surface, is enough to trigger ignition. This initially creates a slow flame (velocity of the order of one meter per second), which can accelerate very significantly under the effect of the turbulence generated by the presence of obstacles and the transition to the detonation zone. (Speed in the order of kilometers per second). ). The pressure peaks resulting from the accelerated flames can endanger the integrity of the reactor building's containment or damage safety-relevant systems.
Reduce risk
One way to reduce the risk associated with hydrogen is to install passive autocatalytic recombiners. Using catalytic plates, these systems recombine hydrogen with oxygen in the form of water vapor and carbon monoxide (CO) in the form of carbon dioxide (CO2). Since this chemical reaction is exothermic, it itself receives a passive heat supply from the surrounding gas masses. On behalf of the Nuclear Safety Authority (ASN), the IRSN assessed the performance of these devices installed by EDF in all French reactors between 2004 and 2007. Depending on the special features, each room contains twenty to one hundred and sixteen recombiners.
Principle of a passive autocatalytic recombiner. When it comes into contact with the catalyst plates (usually platinum or palladium, orange in the lower part of the housing), hydrogen (H2) combines with oxygen (O2) to form water vapor (H2O). Likewise, carbon monoxide (CO) is converted into carbon dioxide (CO2). These devices are designed as “chimneys”: the heat created by exothermic chemical recombination reactions thermally “attracts” the surrounding gases (chimney effect) and thus ensures passive operation.
After the Fukushima Daiichi accident, additional safety assessments conducted by IRSN in France showed that the effectiveness of this solution in certain configurations still needs to be proven. This is the case in various scenarios for which recombination kinetics may not be sufficient. Furthermore, an assessment coordinated by IRSN in collaboration with 27 European, Canadian and Japanese partners also highlighted the lack of knowledge about the operation of recombiners in the late stages of a major accident.
At this stage, the low oxygen content of the enclosure atmosphere and the presence of carbon monoxide affect its performance. “We are participating in research within the Amhyco project, funded by the European Commission and led by the Polytechnic University of Madrid, to remedy this,” confides Ahmed Bentaib. They will enable us to suggest improvements to the risk management guidance. »
Maintain state-of-the-art assessment tools
The risk must therefore be assessed taking into account the variety of serious accident situations, the measures planned by the operator to deal with the accident and the means used. To this end, IRSN has developed a multi-stage assessment methodology, including the selection of relevant scenarios with regard to hydrogen-related risks. These were previously simulated using the Astec software, which calculates the course and consequences of major accidents from the triggering event to the release of radionuclides into the environment (see “Understanding and preventing a nuclear reactor meltdown”, Pour la Science n ° 548, June 2023) .
This provides essential data such as the kinetics and location of the generation of various flammable gases during the accident. This information serves as input for further, more detailed calculations that focus on the phases where risk is greatest. These calculations, based on three-dimensional resolution of fluid mechanics, heat and gas transport equations, are used to evaluate the effects of safety systems such as recombiners and water atomization. Water, a system activated to lower the pressure and force the fission products to the bottom of the containment.
These calculations accurately describe the development of the containment atmosphere during the accident. This makes it possible to detect flammable clouds and the danger of rapid flames in the event of a fire. Under these circumstances, the calculations carried out make it possible to evaluate the pressure and temperature loads on the safety container and the safety devices and to derive the effects on their functions. To this end, IRSN conducts and participates in experimental programs, the results of which contribute to maintaining the state-of-the-art of its tools.
Measure the gas composition in the housing
“The simulations showed in particular that the commissioning of the water sprinkler system had to be delayed,” explains Ahmed Bentaib (see infographic opposite). This leads to the condensation of the water vapor present in the containment container, which contributes to the atmosphere becoming more flammable. »
In fact, it has been observed experimentally and numerically that the use of spray techniques increases the spread of flames under certain conditions, particularly at high hydrogen concentrations. This is due to the turbulence it causes. Therefore, it is important to know when to trigger it. “By delaying it by three to six hours, we give the recombiners time to significantly reduce the hydrogen concentration and thus avoid the conditions that promote flame acceleration. » This recommendation has been integrated into the management of major accidents in France.
When do you start sprinkling water? These three simulations are still images 5,300 seconds (approximately one hour and twenty minutes) after the start of a meltdown accident. They show the distribution of hydrogen in the reactor containment and the importance of the point at which the water spray system is put into operation. Dozens of recombiners (in red) limit the hydrogen content (here xH2, the H2 concentration increases from blue to red). In the first picture the sprinkler system was not triggered. In the following two images it was 4,800 seconds and 5,200 seconds after the accident began. The effect is significant and occurs in less than 7 minutes: very high local concentrations occur (as in the third picture). The risk is to promote the formation of a flammable cloud and, in the event of combustion, to cause so-called “flame acceleration”, which can compromise the integrity of the containment. IRSN modeling shows that delaying the activation of the sprinkler system by three to six hours after the start of the accident limits the risk of flame acceleration. About ten Raman spectrometers (in purple) would provide operators and emergency response teams with better information throughout the accident by informing them of the gas composition in the enclosure.
“In order to further reduce the risk associated with hydrogen and inform the operators and crisis teams responsible for managing the accident, it would be ideal to know precisely the composition of the atmosphere of the containment over time,” explains the researcher. We have reviewed the technologies available on the market. We have concluded that none of the existing measurement systems comply with French security requirements, which require that the measurement be carried out inside the containment so as not to create a possible route to bypass the containment. » The only solution: develop a sufficiently precise and robust in-situ measuring device that can withstand the conditions that prevail in the building during serious accident situations.
The developed system is based on Raman spectroscopy, an optical analysis technique. What is its principle? Knowing that a medium slightly changes the frequency of the light circulating there according to the vibration frequencies of the molecules present, the method consists of sending laser light into a medium and analyzing the frequency shift of the scattered light. This creates a spectrum (called the “Raman spectrum”) whose bands are characteristic of the molecules present and their concentration. “Unfortunately, in our case the measurement is not readable because the optical signals are noisy due to a phenomenon called the Cerenkov effect, which is related to radiation,” explains Ahmed Bentaib.
Unreadable Raman spectra. The irradiation (here 14 shades of gray per hour, left spectrum) prevents the differentiation of the bands that reveal the presence of molecules in the analyzed medium (right spectrum). This is called the “Cerenkov effect.” This problem could be solved by algorithmic signal processing.
“The challenge was to develop a device that could withstand the conditions prevailing in the containment in the event of a serious accident and in particular radiation, knowing that the dose rates in the event of a heart meltdown are in the order of 2,000 gray per hour.” The “The first prototype, developed within the framework of the Mithygène project in collaboration with the CEA, the CNRS, the Jülich Institute in Germany and the electronic device manufacturer Arcys, could not withstand more than 30 shades of gray per hour …,” he summarizes . Challenge overcome. “We have improved the performance of the Raman spectroscopy probe and developed signal processing algorithms capable of overcoming interference associated with the Cerenkov effect,” says the researcher.
The current pre-industrial prototype can therefore be operated on site in the event of a serious accident. And in addition to hydrogen, the device determines the concentration of five other gases in real time: oxygen, nitrogen, water vapor, carbon monoxide and carbon dioxide. “With the help of simulations, we have also proven that around ten well-placed probes provide a reasonably accurate picture of the composition of the gas mixture in the containment,” adds Ahmed Bentaib. This had never happened before. » This innovation is suitable for all reactor types. It could be used in new concepts of small modular reactors, or “SMRs.”
Download the PDF version of this article
(reserved for digital subscribers)