OECD/CSNI Workshop on Best Estimate Methods and Uncertainty Evaluations Workshop Proceedings

1.1 Background Best-Estimate Methods plus Uncertainty Evaluation are gaining increased interest in the licensing process. On the other hand, lessons learnt from the BEMUSE (NEA/CSNI/R(2011)3) and SM2A (NEA/CSNI/R(2011)3) benchmarks, progress of UAM benchmark, and answers to the WGAMA questionnaire on the Use of Best-Estimate Methodologies show that improvements of the present methods are necessary and new applications appear. One can quote: - For BEMUSE user effect for base case calculations, user effect for the uncertainty method application and the need for developing methods to estimate input parameter and model key parameter uncertainties. - For UAM, concerns are on how to subdivide the complex system of coupled multi-physics and multi-scale simulations into different steps, to derive multi-group microscopic cross-section libraries and how to proceed to calculate the uncertainties in further steps. - For SM2A, which addressed other scenarios than LBLOCA, other physical model uncertainty may be of concern; also, probability distribution to characterize time delay for system recovery or for a repair may be of high importance as the sequences including such time delays could be impacted by power up-rate; - Examples of new applications: uncertainty and sensitivity analysis of CFD codes, uncertainties for coupled codes, connection with PSA. It is recognised that the continuous improvement of classical BEPU approaches as mentioned above, will favour gaining confidence on their use instead of the conservative or mixed conservative-BE approaches. Looking into the future, the application of the BEPU methods to new codes (e.g. CFD) is compulsory to increase the level of confidence of these applications or tools in the nuclear safety field. As computing capacity is expanding, the extension of BEPU methods to coupled codes (e.g. TH-neutronics) and to PSA (refinement of success criteria verification and safety margin assessment) is subject of increased interest in the nuclear safety community. 1.2 Objectives of the workshop The CSNI Working Group on Analysis and Management of Accidents (WGAMA) objective is to assess and where necessary strengthen the technical basis needed for the prevention, mitigation and management of potential accidents in nuclear power plants, and to facilitate international convergence on safety issues and accident management analyses and strategies. The objective of this workshop is to provide a forum for a wide range of experts to exchange information in the area of best estimate analysis and uncertainty evaluation methods and address issues drawn-up from BEMUSE, UAM and SM2A activities. Both, improvement of existing methods and recent new developments are included. As a result of the workshop development, a set of recommendations, including lines for future activities were proposed. 1.3 Organization of the workshop The Workshop was focused on BEPU methods involving thermo-hydraulics for system codes. Other applications, involving CFD, neutronics coupling or extension to PSA, were also be subject of the workshop. As a common denominator of the Workshop, BEPU methods and applications were exposed and the capacity to determine the preservation of safety margins through a transparent and solid technical basis debated. The Workshop was open to wider participation especially participants to OECD/CSNI exercises like BEMUSE, UAM, SM2A, to other benchmarks, as well as industrial partners, Technical safety organizations and safety authorities and academia. The organisation of the Workshop was divided into three parts: Opening session including key notes from OECD and IAEA representatives, Technical sessions, and a Wrap-up session. All sessions included a debate with participation from the audience constituted by 71 attendees. The workshop consisted of four technical sessions: a) Development achievements of BEPU methods and State of the Art The objective of this session was to present the different approaches to deal with Best Estimate codes and uncertainties evaluations. A total of six papers were presented. One initial paper summarized the existing methods; the following open papers were focused on specific methods stressing their bases, peculiarities and advantages. As a result of the session a picture of the current State of the Art was obtained. b) International comparative activities This session reviewed the set of international activities around the subject of BEPU methods benchmarking and development. From each of the activities a description of the objectives, development, main results, conclusions and recommendations (in case it is finalized) was presented. This session comprised a total of four papers. c) Applications: licensing, safety analysis support, regulatory body views and industry activities This session put the focus on the application of current methods in safety analyses. Contribution from industry, technical safety organizations and regulatory bodies was present. As a result, a view of the penetration of BEPU methods in current safety analyses was obtained, as well as an indication of the evolution in the near future. Elements such as licensing practices, assessment process, etc. were considered. This session comprised nine papers. d) BEPU methods extension to new fields Since the early days of BEPU methods development it was realized that this approach could be extended and its benefits applied to other areas beyond the classic deterministic environment. This session was focused on the extension of BEPU methods beyond this point. The session developed the extension to CFD, quantification of global safety margins, use of PSA, etc. with a total of seven papers. The Technical programme of the workshop is included as appendix I. 1.4 Results The workshop provided an interesting and useful forum for the participants to share and discuss their respective practices regarding BEPU methods and state-of-the-art, BEPU applications in the licensing, safety analysis support, regulatory body views and industry activities, as well as BEPU extension to new fields such as safety margin evaluation and CFD. 4. It is recommended to promote the advance into the development of deterministic/probabilistic methods for safety verification purposes. The use of pilot applications should be considered in order to check the benefits and potentials of this approach. • The importance of validating BEPU methods and of reducing user effects was also an implicit conclusion of several presentations. Benchmarking exercises like BEMUSE have shown that large scatter among the different participants are present, casting doubts with regard the application of the method in licensing. Although BEPU methods have been used in increasing number for licensing, it was also observed that some of their aspects (especially evaluation of the safety margin) may present challenges to some regulators. 5. It is recommended to pursue into the validation of BEPU methods in order to answer regulatory concerns. • It was observed that under the so called IAEA/SSG-2 Option 4, design basis analysis and design bases extension analyses can benefit from full application of BEPU methods together with credit to systems availability. At the moment this approach is in first steps and needs further development. A potential for a new design bases standard was acknowledged. 6. It is recommended to advance into the development works and potential safety benefits outcome of BEPU methods application with consideration of systems availability (option 4) The recommendations of the workshop can be summarized by emphasizing that a great advance has been achieved since the beginning of BEPU activities, pointing to the consolidation of practices and nevertheless identifying still pending concerns. While the added value to safety with origin in the use of BEPU methods is recognized, the full potential of BEPU approach to a comprehensive safety assessment is still pending with regard to development. The CSNI and other international organizations like IAEA should play a leading role in this advancement.