A preliminary study on concrete ablation during ex-vessel cooling in pre-flooded Westinghouse 3-loop NPP reactor cavity

Highlights

• An analytical model to evaluate the adequacy of ex-vessel cooling in pre-flooded condition is developed.

• The model considers the melting of dry debris bed and the pyrolysis of concrete.

• The results of the concrete ablation depth during ex-vessel cooling in Westinghouse 3-loop nuclear power plant are provided.

Abstract

We develop an analytical model about concrete ablation during ex-vessel cooling in a pre-flooded cavity. The model considers the melting of dry debris bed, the formation of crust, the growth of crust and the thermal ablation of concrete. This model can obtain concrete ablation depth during ex-vessel cooling. It is evaluated the adequacy of ex-vessel cooling in the pre-flooded reactor cavity of Westinghouse 3-loop nuclear power plants. If the debris bed is formed only in the zone directly below the reactor vessel narrowly and thickly, the characteristic of the debris bed should be like POMECO BED-5 and the pressure should be 2 bar or above to keep the integrity of the containment. However, if the debris bed is evenly distributed on the reactor cavity, it is assessed the pressure should be equal to or greater than 1 bar, 2 bar and 3 bar for POMECO BED-5, STYX BED and COOLOCED-8 BED, respectively.

Introduction

When thermal energy is produced by nuclear fission in nuclear power plants (NPPs), fission products are accumulated in nuclear fuels. Fission products are usually unstable and release decay heat as they continuously transformed into stable nuclides through radioactive decay. When the reactor shuts down, radioactive decay continues to occur even. Therefore, to release the decay heat to an ultimate heat sink in case of emergency, NPPs should be equipped with reliable safety features (e.g., safety injection system, auxiliary feedwater system and so on). However, when an initiating event such as Large-Break Loss-of-Coolant-Accident (LBLOCA) or Total Loss of Feed Water (TLOFW) occurs and the safety features fail to work properly, then the fuel would melt down due to the radioactive decay heat. This fuel melt accident called a severe accident.

At this case, the molten corium could damage the reactor vessel and eventually fall into a reactor cavity whenever it fails to cool down the molten fuel and discharge the decay heat to an ultimate heat sink outside the NPPs. Because the molten corium in the reactor cavity floor continuously emits thermal energy due to the radioactive decay, and it thermally attacks the structure of a containment. The leak-tightness integrity of the containment would be impaired if the molten corium is not still adequately cooled. Consequently, a large amount of radioactive material would be released to the environment. Thus, it is needed ex-vessel cooling to cool down the molten corium and discharge the accumulated heat to the outside in the case of in-vessel cooling failure.

There are four Westinghouse (WH) 3-loop NPPs (e.g., Kori unit 3 & 4 and Hanbit unit 1 & 2). Based on their Severe Accident Management Guidelines (SAMG), it is guided to inject the cooling water into the reactor cavity without considering whether the reactor vessel is breached or not at the moment of flooding the reactor cavity. Thus, in the WH (3-loop) NPPs, the reactor cavity may be in a dry state or a deep pool state at the moment of the reactor vessel failure.

The phenomena during the ex-vessel cooling in the pre-flooded cavity are quite different from ex-vessel cooling in the dry cavity condition. If the molten corium falls into the pre-flooded deep pool, the molten corium may be fragmented due to the Rayleigh-Taylor and Kelvin-Helmholtz instability (Matsuo et al., 2008). A steam explosion during the dropping process could occur, and the containment could be damaged by the dynamic loads from the steam explosion. Consequently, the containment may be damaged at the moment of the reactor vessel failure, thus, a large amount of radioactive material could be released to the outside at the early stage. Perhaps this is one of the worst cases caused by nuclear power plant accidents, because nearby residents may not have enough time to evacuate. However, in this study, we assume that there is no steam explosion during the molten corium falling. If there is no steam explosion, a heat-generating debris bed is formed in the reactor cavity. The major aim of this research is to make a model on concrete ablation in a pre-flooded cavity during the ex-vessel cooling and to do the safety analysis of this condition.

The major benefit of the ex-vessel cooling in the pre-flooded cavity is that as soon as the molten corium falls into the deep pool, the cooling is facilitated by boiling with the coolant. Moreover, operators do not need to take into consideration the moment of the reactor vessel failure; they just need to inject the coolant into the reactor cavity as soon as possible and then watch the ex-vessel cooling of the molten corium. However, as Greene et al. pointed out, the molten corium may not be well spread in the flooded reactor cavity floor (Greene et al., 1988). If this is the case, the radioactive decay heat is concentrated only on the area where the molten corium is spread. Therefore, if the cooling of the molten corium or the debris bed is not sufficient, it is possible thermally to attack the containment more severely in the pre-flooded cavity condition.

Several researches were found on the coolability of the debris bed in the pre-flooded condition. Firstly, FARO (Magallon and Huhtiniemi, 2001), MIRA (Haraldsson and Sehgal, 1999), and DEFOR (Kudinov and Davydov, 2013) and others were carried out on the particle size distribution of fragmented particles when molten materials fall into the water. Also, the COOLOCE (Takasuo, 2013), STYX (Lindholm et al., 2006), POMECO (Li et al., 2012) and others were carried out to obtain the dryout heat flux of the debris bed reflecting the particle distribution results. We conduct safety analysis assuming that the debris beds whose characteristics are the same as those of COOLOCE-8 BED, STYX BED, and POMECO BED-5 can be formed in the WH (3-loop) pre-flooded deep pool. There are several analytical models for obtaining the dryout heat flux of a heat-generating packed particle bed (Na et al., 2017b, Schulenberg and Müller, 1987, Tung and Dhir, 1988, Lipinski, 1982, Reed, 1982).

The reactor cavity floor shapes of the Korean NPPs (e.g., WH (3-loop), ARP1400 and OPR1000) are longish corridors rather than circular floors. Besides, the location of the reactor vessel is located at one end of the corridor. Therefore, more than half of the zone directly below the reactor vessel is surrounded by a wall and the other is open to the corridor. Therefore, it seems that more than half of the fragmented particles may be clustered on the narrow area forming a thick debris bed in the pre-flooded cavity floor even assuming the fragmented particles have good spreadability. Thus, the radioactive decay heat may be concentrated on the narrow area, which increases the possibility of damage in the leak-tightness of the containment.

However, if the reactor cavity is in a dry state at the moment of reactor vessel failure, then the molten corium may spread well and the thickness of the molten corium in the reactor cavity is uniform throughout the reactor cavity. Thus, the radioactive decay heat is evenly distributed on the reactor cavity floor. Then, the coolant is supplied into the reactor cavity, and the water promotes the cooling of the molten corium through boiling. The major benefit of this cooling method is that the thermal load is well distributed rather than concentrated on narrow area. However, if the water supply is delayed, the possibility of containment integrity failure increases. Although the radioactive decay heat is uniformly distributed, much of the generated heat is used to melt the concrete structures until the coolant is injected. To evaluate the adequacy of this cooling method, melt spreading experiments such as COMAS (Steinwarz et al., 2001), FARO L26S (Tromm et al., 2000), some experiments of Hokkaido university (Matsumoto et al., 2017, Ogura et al., 2018) and others were carried out. Furthermore, some researchers (Kawahara and Oka, 2012, Na et al., 2017a, Ye et al., 2013, Yeon et al., 2012) carried out numerical studies on the melt spreadability using three-dimensional calculation methods. There are SWISS (Blose et al., 1987), COMET-L3 (Miassoedov et al., 2010), MACE (Farmer et al., 2000) and other experiments on the molten corium-concrete interaction (MCCI).

In this paper, we investigate whether the integrity of the containment is maintained during the ex-vessel cooling process when the molten corium falls into the pre-flooded reactor cavity in the event of a severe accident of Kori Units 3 and 4 which are the WH (3-loop) NPPs. The framework of the analysis method is similar to the method used in APR1400 NPPs (Na et al., 2017b) and OPR1000 NPPs (Na et al., 2020) safety analysis. We use the one-dimensional dryout heat flux model of the heat-generating packed particles, and the assumption about the shape of the debris bed is used in the same way. However, we improve the analytical model in the study. According to the previous model, it is assumed that the integrity of all the deposited debris beds is maintained during the ex-vessel cooling. However, in this model, we consider the condition in which the dry debris bed melts, and the crust is formed and growing.

According to the Nuclear Safety Act of Republic of Korea, nuclear operators shall submit an Accident Management Program (AMP) by 2019. To review the adequacy of their AMP, according to the Safety Review Guideline for AMP (KINS, 2018), the containment is needed to keep its leak-tightness during the ex-vessel cooling. In this respect, the guideline requires that the integrity of the bottom containment liner plate of the containment should be maintained during the ex-vessel cooling.

The major aim of developing the current model is to use this analysis as one of references to assess the appropriateness of the submitted AMP. Therefore, when the model is derived, we use some conservative assumptions to obtain the conservative results in terms of the downward concrete ablation depth during the ex-vessel cooling. This paper also presents the safety analysis results of WH (3-loop) NPPs using the newly developed model.