Report

Teaser, summary, work performed and final results

Periodic Reporting for period 1 - NARSIS (New Approach to Reactor Safety ImprovementS)

Teaser

The Probabilistic Safety Assessment (PSA) procedure allows practitioners to better understand and to estimate the likelihood of the most causes prone to initiate nuclear accidents and to identify the most critical elements of the systems. However, despite of the remarkable...

Summary

The Probabilistic Safety Assessment (PSA) procedure allows practitioners to better understand and to estimate the likelihood of the most causes prone to initiate nuclear accidents and to identify the most critical elements of the systems. However, despite of the remarkable reliability of the current PSA procedures, the accident that occurred at the Fukushima Daiichi Nuclear Power Plant (NPP) in March 2011 highlighted a number of challenging issues with respect to the current application of PSA and the validity of its results. The procedures in place at this time did not manage to properly forecast the cascading scenario and its impacts on the integrity of the structures. But, moreover, they did not succeed to anticipate the disastrous consequences on the functionality of the protecting systems. Therefore, in the current continuous improvement process of safety in nuclear installations, improvements of the safety assessment methodology should be considered.

The NARSIS project aims at making significant scientific step forward, towards addressing the update of some elements required for the safety assessment of nuclear installations. These improvements mainly concern:
• A better characterization of natural external hazards, considering scenarios with concomitant external events, either simultaneous-yet-independent hazards or cascading events;
• A better evaluation of the fragility of the main elements (structures, equipment, pools, etc.) of the NPPs to account for conjunct effects (including ageing effects) and interdependencies under single or multiple external natural aggressions;
• A better risk integration combined with uncertainty characterization and quantification: different approaches are proposed, such as Bayesian Belief Networks (BBN) or Extended Best Estimate Plus Uncertainty (E-BEPU), which aim at allowing efficient risks comparison and accounting for all the possible risk interactions and cascade effects (both low-frequency / high consequences events and combined events).
• A better processing and integration of expert-based information within PSA: NARSIS will investigate the applicability and the benefits of using modern uncertainty theories both to represent in flexible manner experts’ judgments and to aggregate them to be used in a comprehensive manner.

The project does not aim at performing a complete review of the PSA procedures, but to propose some elements of improvement to be integrated in the current PSA.

Work performed

NARSIS has successfully achieved 16 deliverables of the 18 foreseen and 9 over 10 milestones have been met for the First Periodic Period on time.

The main results of the work done in this reporting period are:
• As part of the WP1, an extensive review of existing multi-hazard assessments and procedures has been undertaken for natural hazard assessments with respect to nuclear safety.
• As part of the WP2, the most critical NPP elements, based on their importance in the mitigation of natural external events has been identified and the compilation of current practices for fragility assessment has been performed.
• As part of the WP3: a review and comparison of risk integration methods from high risk industries (e.g. aviation, chemical and nuclear industries), has been performed with particular emphasis on methods able to incorporate low probability events, multi-hazards, and integration of human, social/organizational and technical aspects.
• As part of the WP4, a generic model of NPP representative of a virtual PWR has been detailed, which will used for numerical simulations required for fragility assessment (structures, equipment) and model reduction strategies as well as for safety analyses (L2-PSA). The focus has been put on the reactor building and associated systems design properties for a selected number of safe shut down paths.
• As part of the WP5, the definition and characterization of a “referential” real NPP has been performed in coordination with task relating to objective 4 (task 4.1), extending to the SSCs which are potentially relevant for the implementation of EDMG/SAMG strategies. The focus has been put on critical safety functions and underlying systems, with their redundancy, diversity, independency and separation principles.

Final results

The project proposes to provide new tools and methodologies for the PSA of NPPs, in order to fill the gaps and shortcomings identified :

• The PSA process is commonly adopted for seismic or flood hazard but its application to all types of external events is not straightforward.

• A single-hazard PSA is not able to account for cascading or conjunct events, which have been proven to lead to unforeseen dramatic consequences.

• In a probabilistic context, the exploration of extreme events (i.e. distribution tails) remains an issue, to which extensive Monte-Carlo simulation processes only provide a partial solution.

• The methodologies for quantification and propagation of uncertainty sources (more specifically uncertainty related to expert-based information) have been the object of significant improvements in other fields (e.g. related to human-environmental interactions), but their application to the PSA of NPPs remains to be demonstrated.

• Most of the current fragility models are focused on the physical damage of components, while the emphasis should be put on the functionality loss of the equipment, in light of the Level 1 PSA.

• While a multi-hazard harmonization has been recently addressed by a few FP7 projects (e.g. MATRIX for urban contexts, INFRARISK for road infrastructures), its application to much more complex systems such as NPPs remains an untested challenge and could bring new perspectives in nuclear safety analyses.

Manifold impacts are expected from the NARSIS project:

• Scientific outcomes will include an integrated framework useful for combined hazard scenarios analysis as well as uncertainty quantification and comparison. Theoretical improvements are also foreseen in risk assessment and PSA methodologies.

• Operational outcomes will be achieved through the demonstration of a supporting Decision-Making (DM) tool for Severe Accident Management Guidelines (SAMG) implementation, which will be useful for actual NPPs and will potentially contribute to reducing the residual risk from NPPs operation.

• At European level:

- Knowledge basis for national and EU policies: some key scientific knowledge should be made available to the decision and policy makers.

- Better safety culture: A main foreseen achievement of NARSIS regarding NPP safety improvement or risks control, which will lead to a better safety culture, is related to the SAM DM tool (demonstration purposes) and guidelines planned in WP5. Moreover, the project contribution is not only focused on technical aspects (i.e. design of safer systems) or on the direct influence of human behaviour (i.e. operator error) but also on the impact of social and organizational factors (i.e. policies and procedures, see WP3) on the outcome of safety performance.

- Another achievement concerns the pedagogic (e.g. lectures) materials, which will be produced by NARSIS project (mostly taught to Master students and young professionals).

- Foster European cooperation: As NARSIS consortium is composed of a large number of European key actors in the nuclear fields, having different safety cultures and practices, it will contribute to enhance the mutual understanding at European level.

Website & more info

More info: http://www.narsis.eu/.