Summary and Conclusions from the International Seminar on In-Vessel Retention Strategy

F. FICHOTet al.

Summary and conclusions from the International Seminar on In-Vessel Retention Strategy

F.FICHOT

Institut de Radioprotection et de Sûreté Nucléaire

Cadarache, France

Email:

E. RAIMOND

Institut de Radioprotection et de Sûreté Nucléaire

Fontenay aux Roses, France

S. HERMSMEYER

European Commission, Joint Research Center

Petten, The Netherlands

B. CHAUMONT

Institut de Radioprotection et de Sûreté Nucléaire

Fontenay aux Roses, France

Abstract

On the 6-7 June 2016, IRSN hosted an international workshop about the « Strategy of In-Vessel Melt Retention: Knowledge and Perspectives ». The workshop was co-organized by JRC and IRSN, with the sponsorship of ETSON. With panel discussions and technical sessions, the workshop covered all the important issues related to in-vessel corium retention, from the physical understanding to regulatory frames. The major points of the safety demonstration were discussed. Some industrial aspects were also addressed. One of the objectives was to provide an orientation of R&D projects to strengthen IVR strategies, such as the H2020 IVMR project, coordinated by IRSN. The current approach followed by most experts for IVR is a compromise between a deterministic approach using the significant knowledge gained during the last two decades and a probabilistic approach to take into account large uncertainties due to lack of data for some phenomena and due to excessive simplifications of models. It was concluded that a harmonization of the positions of safety authorities on the IVR strategy is necessary to allow decision making based on scientific knowledge. For this, a consensus on several issues should be reached between R&D experts. This includes in particular the issues of the transient evolution of oxide and metal layers in the lower plenum and of the long term mechanical behavior of the thin “cold shell” resulting from vessel ablation..

1.INTRODUCTION

Gen III nuclear reactors are characterized by the consideration, at the design stage, of severe accidents with core melting and, therefore, the retention of corium in the containment. To reach this goal, the options selected for the design of those reactors may be divided into two different approaches. One approach is the stabilization of corium outside the vessel (EPR or some VVER cases, for example), the other aims at keeping the corium inside the vessel (IVR strategy). Thanks to panel discussion and technical sessions, the workshop covered all the important issues related to strategies of in-vessel corium retention, from the physical understanding to the safety demonstration and regulatory frames. Experts’ presentations provided an overview of the current knowledge and understanding about the physical processes associated with corium and its interaction with the reactor vessel. The major points of the safety demonstration were pointed out and discussed. Some industrial aspects were also addressed and discussed by reactor designers.

One of the objectives of the workshop was to provide a basis for the orientation of new research projects such as the European project IVMR. This project started in 2015 with duration of 4 years and is coordinated by IRSN. It aims at developing knowledge and tools to estimate the efficiency of the measures dedicated to stabilize corium in the reactor vessel in case of severe accident with core melting in reactors of 1000MWe or more.

The workshop gathered more than 130 participants from most of the main actors involved in issues related to severe accident management in order to share knowledge and positions:

‒Safety authorities: ASN - France, US NRC - USA, NSC – China, NRA - Japan, SEC-NRS – Russia, KINS – South Korea, STUK – Finland, ONR – United Kingdom, CNSC - Canada, SSM – Sweden, SNNI – Switzerland, SSTC-NRS – Ukraine, SUJB – Czech Republic, UJD - Slovakia

‒Technical Safety Organizations (TSO) and research institutes: IRSN - France, IBRAE – Russia, GRS – Germany, Bel V – Belgium

‒Utilities, Vendors: EDF, AREVA, Mitsubishi, SNERDI, Fortum

The presence of organizations who are not directly involved in IVR strategy (because ex-vessel corium stabilization is the main option in their country) showed that, in the current post-Fukushima context, all safety options deserve to be studied and that safety issues in- and ex- vessel are closely connected.

2.The IVR strategy as an option for SA mitigation

WENRA has recently initiated a work to harmonize the safety requirements at the European level. For new power plants, one of the safety objectives is to reduce potential releases, also in the long term which means that accidents leading to large or early releases must be practically eliminated and that other core melt accidents should cause only limited protective measures in area and time. The independence between all levels of defence-in-depth was also pointed out. In that respect, corium retention in the vessel should be examined, not as a unique option but as an option which would also increases the chances of success of the measures dedicated to protection of the containment. It was pointed out that deterministic analyses should cover core melt scenarios starting from all operational states. Postulated core melt accidents are typically considered with realistic assumptions and best estimate methodologies. The probabilistic safety assessment (PSA) is complementary to the deterministic analyses.

2.1.Panel discussion with Safety Authorities and TSO’s

A first panel with the participation of ASN, US NRC, NSC, ONR, SEC-NRS, KINS, SSM and SSTC-NRS was organized to let safety authorities and TSOs express their position about corium retention in general and more specifically about IVR strategy. The discussion focused on the three following questions:

1/ General position on IVR strategy: it was acknowledged by all participants that, for a safety authority, the IVR strategy remains an option that must be examined on a case by case basis. No general conclusions can be drawn on simple criteria such as reactor power or design. Each reactor design has to be analyzed independently of conclusions drawn for other designs, even similar ones. The implementation of the various operations necessary to flood the reactor pit and ensure the stabilization of corium in the vessel all have an important impact on the safety margins. Of course, the larger the reactor power, the more detailed the safety analysis must be because the safety margins are expected to be lower. One key conclusion was that it is important to harmonize the basis of knowledge at the R&D level first, in order to allow safety authorities to define safety criteria for IVR strategy which can be shared at the international level.

2/ Methodology to address IVR safety assessment (deterministic or probabilistic approaches) - Main issues for a robust assessment: there was no general position about the methodology to be used because, as it was pointed out in the previous point, a common basis of knowledge is still incomplete for the issues related to IVR (despite the large number of data already obtained in previous R&D programs). It was mentioned by several participants that safety assessment should not rely only on probabilistic approaches but should include also a deterministic point of view, in order to understand the main issues related to corium retention in the vessel and make a reliable estimate of the risks.

3/ Criteria of acceptability: short term and long term assessment of vessel integrity. At this point, it is clear that the criteria of acceptability may vary according to the reactor design. Among the most important criteria, the minimum vessel thickness appears as a critical one. It is obvious that the vessel would be highly ablated in case of IVR, with just a few centimeters of solid steel remaining at some places. In principle, it was demonstrated that such thickness would be enough to resist the static load corresponding to the weight of corium and small overpressure in LB LOCA but it was never really demonstrated if it would be enough to resist a significant spike of primary pressure, following for example, a reflooding, especially in case of SBO or SB LOCA scenarios. In addition, long term (i.e. several months) resistance of the ablated wall remains to be assessed. It was also mentioned that human factor might play an important role in IVR strategy because the decision of flooding rapidly the reactor pit is crucial for the success of IVR but is not reversible.

2.2.Examples of designs with implementation of IVR strategy

In spite of the difficulties to assess in a reliable manner the effectiveness of the IVR strategy, several vendors have designed reactors in which this strategy can be used, in case of severe accident with core melting. Some examples were given during the workshop for two large power PWRs (CAP-1400 and APR-1400), for one BWR (KERENA) and one PHWR (CANDU). Even though it is not the point of this paper to review all the elements which were given for the safety demonstration of those designs, it is interesting to identify the similarities between them and the design options which are considered as positive for IVR success.

1/ The passive reflooding of the reactor cavity, up to the level of the primary loops. This is even improved in the case of AP1000 and CAP-1400 where it is possible to flood a much larger volume covering most of the primary circuit, allowing a simultaneous in-vessel reflooding if the break is below the water level and the vessel depressurized.

2/ The design of vessel insulation in order to allow water circulation between the vessel wall and insulation and to allow steam venting in the top part.

3/ The simultaneous in-vessel water injection (direct such as in APR-1400 or not as in CAP-1400) in order to reduce the risk of focusing effect.

4/ The presence of a large mass of steel in the lower plenum (examples of KERENA and CAP-1400) which would most likely avoid the formation of a heavy metal layer and would increase the thickness of the top metal layer, thus reducing the risk of focusing effect.

5/ The absence of penetrations (as in several PWR and PHWR designs)

6/ The large water inventory (including safety injections and storage tanks) leading to delay the time of corium arrival in the lower plenum and therefore the residual power to extract (up to one day before corium arrival for the KERENA (see Figure 1Example of the KERENA design, from Fischer’s presentation) and CANDU designs, up to 3 days for VVER-TOI).

Figure 1Example of the KERENA design, from Fischer’s presentation

From those examples, it is worth noting that designers are aware that some uncertainties, such as complex physico-chemical phenomena in the corium still exist but they propose design options and measures which would, in any case, help to prevent large heat flux to the vessel. They also propose measures which should reduce the occurrence of fast transients which are critical for the safety assessment of IVR.In order to complete the international panorama given in that session, it is important to recall that In Russia, the IVR concept is not considered for high power VVERs, however for middle power reactors (VVER-600, 440) deterministic analyses demonstrated the effectiveness and reliability of decay heat removal by external reactor vessel cooling.

3.Overview of main results in the last decade and R&D needs

3.3.Corium pool

Several presentations were dedicated to the behavior of corium in the lower plenum and the resulting heat flux to the vessel, from the oxide pool and from the metal layer. Since MASCA results, it is well known that the thermochemical processes play an important role in the stratification of metal and oxide layers and hence on the heat transfers. The miscibility gap between molten fuel and molten steel may lead to various patterns of stratification (see Figure 2: Examples of MASCA MA-3 and MA-9 tests with a heavy metal layer and a light metal layer, respectively (from Bechta et al. presentation), with possible formation of a heavy metal layer for transient periods, which makes the thermal analysis more complex than in the standard analysis where the metal layer is always on top of the oxide pool. This is a major source of uncertainty. Further experiments are necessary to investigate situations which are more complex than in MASCA i.e.: situations with presence of an oxide crust separating metal and oxide, situations with oxidizing atmosphere, situations with solid debris in the molten pool.

Figure 2: Examples of MASCA MA-3 and MA-9 tests with a heavy metal layer and a light metal layer, respectively (from Bechta et al. presentation)

The knowledge about turbulent natural convection in the oxide and metal layers, and the correlations for heat transfers were also presented. Currently, the correlations used in codes are based on the same set of experimental data obtained in the past for the highest Rayleigh numbers that could be reached. Discrepancies between correlations exist at very high Rayleigh numbers but this uncertainty is known and can be taken into account in the analysis. However, very large uncertainties remain in the analysis of the top metal layer when it is thin (focusing effect). Several issues including scaling, 3D effects, radiative heat transfers and conduction in the vessel wall make the evaluation of the heat flux along a thin metal layer difficult, with large uncertainties. At present, this is a key-issue for the improvement of the safety demonstration of IVR. Experiments will be proposed in the IVMR and SAFEST projects to study some of these issues. In parallel, CFD appears as a promising way of tackling this problem.

3.4.Corrosion of vessel steel by corium and molten metal

At very high temperature, vessel steel reacts with both the oxide corium and the molten metal phase.

Corrosion of the vessel by the oxide corium was studied during the METCOR project (phases 1 and 2). It was observed that there are two steel corrosion mechanisms depending on the oxygen content of the system. Those mechanisms could be described by models. The rate of vessel corrosion was evaluated. It becomes significant at temperature above 1100°C. Therefore, corrosion does not affect the cold part of the vessel which is most resistant one.

The interaction of vessel steel with a molten metal containing Zirconium is exothermic and might be self-propagating under some conditions. However, this effect could not be observed clearly in the experiments performed up to now (possibly because of other effects causing heat losses). In the absence of further evidence, it is recommended to take into account the heat produced by intermetallic reactions in the models used for IVR.

3.5.External cooling

External reactor vessel cooling (ERVC) was studied rather extensively in the past leading to several correlations for the maximum heat flux that can be extracted from the vessel wall (CHF) as a function of the water flow rate, the local void fraction, the angle of the wall and the water subcooling. CHF increases with the mass flow rate, subcooling and inclination angle. Predictions are more complex for the complete natural circulation loop consisting of the channel between the vessel and the insulation and the volumes and pipes driving water back to the bottom of the vessel. The strategy of demonstration followed up to now relies on the design of a full scale loop representing a slice of the real geometry, with prescribed heat flux profiles applied along the internal vessel wall. This was done for the VVER-440 reactors in Europe, and results for CHF were obtained in full-scale experiments ULPU (in a semi-elliptical design) and CERES. This was also done for the AP-1000 design, with results obtained in the ULPU experiment (in a hemispherical design). Currently, a comparable experiment is under preparation at UJV (Czech Republic) for the VVER-1000 design, as a preliminary work for a possible implementation of IVR strategy in VVER-1000 reactors in Europe and Ukraine. Similar experiments were built in China (REPEC-2, 3D-IVR) to optimize the design of Chinese reactors. Those experiments have shown that values of the CHF between 1.2 and 2 MW/m2 can be reached, depending on the level of optimization of the loop. Apart from the design of the cooling loop, the effects of the surface characteristics and of the composition of water still need to be investigated in order to obtain clear conclusions. It is generally considered that water chemistry effects are less important than surface effects. It was demonstrated that the roughness and “micro-porosity” of the surface play a role to increase the CHF. Such natural surface roughness comes from the oxidation of the external surface of the vessel under normal operation. However, this oxidation is not controlled and it might be uncertain to rely on it to justify increased performances of ERVC. Alternative options such as industrial coatings are also studied, to guarantee an increase of the CHF. Up to 30% increase can be expected for the CHF. It is important to remind that the remaining thickness of vessel (non ablated) is inversely proportional to the heat flux through the vessel, which would be of the order of the CHF value (but lower, considering the safety margin). For a heat flux of 1MW/m2, the residual thickness is about 3cm. Therefore, it does not appear as a good option to find ways to increase the CHF above 3MW/m2 because the residual thickness would become lower than 1cm which would raise serious issues for the mechanical resistance. Another significant issue for the realization of the external vessel cooling is a transient heat transfer during vessel flooding. In normal operation conditions the vessel wall is heated to the temperature well above the water saturation temperature at the containment pressure (which is rather low), and the transient heat transfer during flooding may be different from the steady state or even fall into film boiling region. Thus this transient heat transfer problem should be carefully analyzed.