Motivation

1. ERMSAR 2007 European Review Meeting on Severe Accident Research Karlsruhe, June 12-14, 2007Harmonisation of Level 2 PSA Issues: Feasibility study based on hydrogen combustion, consequences of RPV failure, iodine behaviour and molten core concrete interactionM.L. Ang1, B. Chaumont2, E. Grindon1, H. Loeffler3, C. Spengler31) AMEC NNC, Knutsford (GB) 2) IRSN, Fonteney Aux Roses (FR) 3) GRS, Köln (DE)

2. Motivation • PSA level 2 provides a structured framework for the evaluation of phenomena impacting severe accident progression. • The fundamental methodological steps have been consistently performed by SARNET partners. • Considerable variation in the detailed approaches and treatment of phenomena is observed in partner’s studies. • Greater consistency and transparency in practices is desirable. • This study is a first step to achieve such harmonisation in a level of detail which is beyond existing general guidance.

3. Introduction • In a specific task of SARNET PSA level 2 work the feasibility for the harmonisation of Level 2 practises in EC member states was examined for some important issues. • For the selected issues, a pragmatic methodology may include the following aspects: • Describe them in a qualitative way addressing the key features, • Identify what may be regarded as good practice from the review, • Derive general recommendations and, if possible, quantitative values on how to deal with the problem in a Level 2 PSA, • Identify how generic, or plant-type specific or plant specific the guidance is, • Agree on common areas where further research/development activities need to be performed.

4. Approach • The overall approach to this study, using the above framework, is as follows: • Step 1: Develop a questionnaire identifying the key issues for partners’ consultation. • Step 2: Partners to respond to the questionnaire • Step 3: Collate and extract recommendations/conclusions. • Step 4: Review of summary by partners. • These preliminary findings should provide an input to any future harmonisation initiative, including any development of guidelines.

5. Issues covered by present study • Hydrogen combustion • Consequences of RPV failure • Iodine behaviour • Molten core concrete interaction

6. Hydrogen combustion (1) • 11 partners contributed to the evaluation of 8 sub-issues • Hydrogen distribution: • All sources of hydrogen generation should be considered • MELCOR, MAAP or ASTEC (or lumped parameter codes for containment thermal hydraulics) are acceptable codes to assess hydrogen distribution. Adapted peaking factors should be considered. • Analysis of recombiner efficiency is essential (if applicable).

7. Hydrogen combustion (2) • Hydrogen ignition: • Different sources of ignition to be considered (random and continuous). Ignition probability is currently derived using expert judgment. • The ignition time may be considered as the worst physically possible time or as part of the assessment of uncertainties. • Combustion effects: • The different regimes of combustion should be considered, including possibility of deflagration and transition to detonation (DDT). • If there is no risk of DDT, the peak pressure may be assessed on the base of adiabatic isochoric complete combustion (AICC). • In the case of detailed calculations, a detailed nodalisation of the containment is necessary to correctly assess the flame speed.

8. Consequences of RPV failure • 8 partners contributed to the evaluation of 8 sub-issues • The present state of knowledge does not allow an accurate prediction of the failure size of the reactor vessel. • The corium composition and physical state at vessel rupture should be deduced from severe accident code calculations • Vessel lifting at vessel failure, including rocketing, at high or intermediate pressure should be assessed. • Geometrical details are essential to determine the corium entrainment in HPME scenarios. • Combined effects of DCH and hydrogen combustion should be considered. • Given the present state of knowledge (i.e. no coupled calculations), care must be taken regarding the dependencies between the different phenomena.

9. Iodine behaviour • 10 partners contributed to the evaluation. • The fundamental processes are still regarded as not fully understood, creating a high level of uncertainty. • Reluctance within the chemistry community to make explicit recommendations for PSA applications. • Some partners currently do not address chemistry effects. • Several partners use global modifying factors. • Difficult to draw any general conclusions about the completeness and verification of current approaches. • The use of stand alone iodine codes as an integral part of the PSA is not generally seen as practical.

10. Molten core concrete interaction • 8 partners contributed to the evaluation • The key information is the time dependent erosion rate. The present uncertainty is about 50% after approx. 2 days of interaction. • Check and improve the MCCI models concerning the 2D erosion behaviour based on qualified experiments performed most recently or in the near future. • A code benchmark for a generic reactor application is recommended. • The impact of water is still not sufficiently resolved in the current evaluation approaches. • The impact of pool temperature on fission product release may have to be examined. The uncertainty range is about 500 K.

11. Conclusions • Harmonisation of Level 2 PSA practices is difficult and challenging. • There are a number of limitations: • Harmonisation of detailed treatment of individual issues may be hindered due to limited information accessibility. • Severe accident code analysis is a significant and integral part of a Level 2 PSA. For their correct use, it is necessary to know where code limitations are relevant for the overall PSA. • A prescription on how expert judgment should be used and how limitations are accounted for is currently lacking. • For four issues, good practice or at least a summary of partner’s approaches has been presented. • This present SARNET approach should provide a template and input for any future initiatives on this harmonisation issues.