The PSA procedure allows practitioners to better understand the most causes prone to initiate nuclear accidents and to identify the most critical elements of the systems. However, despite the remarkable reliability of the PSA methodology, lessons learnt from the Fukushima Daiichi nuclear disaster point out the necessity of upgrading the current methodological framework related to areas such as cascading and/or conjunct events characterization, structural and equipment responses and uncertainties treatment. New developments in those areas would even enable the extension of their use in accident management.
The NARSIS project aimed at proposing progress in Probabilistic Safety Assessment (PSA) Fundamentals for NPP Safety exposed to multiple external natural hazards events, relying on a multidisciplinary approach and a threefold methodology:
1. Theoretical improvements in natural hazards assessment and their impacts, including the evaluation of the uncertainties and the reduction of subjectivity in expert judgments;
2. Verification of the findings in the frame of the safety assessment through adequate model reduction strategies for less time-demanding simulations and finally,
3. Application of the outcomes at the demonstration level by providing supporting decision tools for severe accident management.
Five key objectives were to be addressed:
1) Improving the characterization of natural external hazards, focusing on concomitant events (either simultaneous-yet-independent or cascading).
2) Improving the physical and functional vulnerability assessment of main critical SSCs of a NPP subjected to complex agressions, by introducing some vector-based fragility curves/surfaces, correlation effects and consequent damage scenarios.
3) Implementing an integrated multi-risk and safety assessment framework (e.g. using Bayesian Networks, BNs), for consequences assessment (large early release frequencies, core damage, ...) including sensitivity analyses.
4) Improving the evaluation and treatment of uncertainties, including uncertainties related to the integration of the expert judgment in the PSA.
5) Verifying and testing the proposed methodologies on simplified (generic) as well as on real NPPs representative of the European generation II & III fleet. Comparison of different safety assessment strategies (e.g. fully deterministic or probabilistic, and mixed deterministic-probabilistic).
