Working Group on Integrity and Ageing of Components and Structures (IAGE WG)

Transcription

1 For Official Use NEA/SEN/SIN/IAGE(2008)1 NEA/SEN/SIN/IAGE(2008)1 For Official Use Organisation de Coopération et de Développement Économiques Organisation for Economic Co-operation and Development 26-Mar-2008 English - Or. English NUCLEAR ENERGY AGENCY COMMITTEE ON THE SAFETY OF NUCLEAR INSTALLATIONS Working Group on Integrity and Ageing of Components and Structures (IAGE WG) WORKSHOP ON THE RECENT FINDINGS AND DEVELOPMENTS IN PROBABILISTIC SEISMIC HAZARDS ANALYSIS (PSHA) METHODOLOGIES AND APPLICATIONS FINAL PROGRAMME AND ABSTRACTS Hosted by Electricité de France (EdF) in Co-operation with the IAEA Lyon Congress Centre, Lyon France 7-9 April 2008 English - Or. English JT Document complet disponible sur OLIS dans son format d'origine Complete document available on OLIS in its original format

2 2

3 ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT The OECD is a unique forum where the governments of 30 democracies work together to address the economic, social and environmental challenges of globalisation. The OECD is also at the forefront of efforts to understand and to help governments respond to new developments and concerns, such as corporate governance, the information economy and the challenges of an ageing population. The Organisation provides a setting where governments can compare policy experiences, seek answers to common problems, identify good practice and work to co-ordinate domestic and international policies. The OECD member countries are: Australia, Austria, Belgium, Canada, the Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Japan, Korea, Luxembourg, Mexico, the Netherlands, New Zealand, Norway, Poland, Portugal, the Slovak Republic, Spain, Sweden, Switzerland, Turkey, the United Kingdom and the United States. The Commission of the European Communities takes part in the work of the OECD. OECD Publishing disseminates widely the results of the Organisation s statistics gathering and research on economic, social and environmental issues, as well as the conventions, guidelines and standards agreed by its members. * * * This work is published on the responsibility of the Secretary-General of the OECD. The opinions expressed and arguments employed herein do not necessarily reflect the official views of the Organisation or of the governments of its member countries. NUCLEAR ENERGY AGENCY The OECD Nuclear Energy Agency (NEA) was established on 1 st February 1958 under the name of the OEEC European Nuclear Energy Agency. It received its present designation on 20 th April 1972, when Japan became its first non-european full member. NEA membership today consists of 28 OECD member countries: Australia, Austria, Belgium, Canada, the Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Japan, Luxembourg, Mexico, the Netherlands, Norway, Portugal, Republic of Korea, the Slovak Republic, Spain, Sweden, Switzerland, Turkey, the United Kingdom and the United States. The Commission of the European Communities also takes part in the work of the Agency. The mission of the NEA is: to assist its member countries in maintaining and further developing, through international co-operation, the scientific, technological and legal bases required for a safe, environmentally friendly and economical use of nuclear energy for peaceful purposes, as well as to provide authoritative assessments and to forge common understandings on key issues, as input to government decisions on nuclear energy policy and to broader OECD policy analyses in areas such as energy and sustainable development. Specific areas of competence of the NEA include safety and regulation of nuclear activities, radioactive waste management, radiological protection, nuclear science, economic and technical analyses of the nuclear fuel cycle, nuclear law and liability, and public information. The NEA Data Bank provides nuclear data and computer program services for participating countries. In these and related tasks, the NEA works in close collaboration with the International Atomic Energy Agency in Vienna, with which it has a Co-operation Agreement, as well as with other international organisations in the nuclear field. OECD 2008 No reproduction, copy, transmission or translation of this publication may be made without written permission. Applications should be sent to OECD Publishing: rights@oecd.org or by fax (+33-1) Permission to photocopy a portion of this work should be addressed to the Centre Français d exploitation du droit de Copie (CFC), 20 rue des Grands-Augustins, Paris, France, fax (+33-1) , (contact@cfcopies.com) or (for US only) to Copyright Clearance Center (CCC), 222 Rosewood Drive Danvers, MA 01923, USA, fax , info@copyright.com. 3

4 4

5 COMMITTEE ON THE SAFETY OF NUCLEAR INSTALLATIONS The NEA Committee on the Safety of Nuclear Installations (CSNI) is an international committee made up of senior scientists and engineers, with broad responsibilities for safety technology and research programmes, and representatives from regulatory authorities. It was set up in 1973 to develop and co-ordinate the activities of the NEA concerning the technical aspects of the design, construction and operation of nuclear installations insofar as they affect the safety of such installations. The committee s purpose is to foster international co-operation in nuclear safety amongst the OECD member countries. The CSNI s main tasks are to exchange technical information and to promote collaboration between research, development, engineering and regulatory organisations; to review operating experience and the state of knowledge on selected topics of nuclear safety technology and safety assessment; to initiate and conduct programmes to overcome discrepancies, develop improvements and research consensus on technical issues; to promote the coordination of work that serve maintaining competence in the nuclear safety matters, including the establishment of joint undertakings. The committee shall focus primarily on existing power reactors and other nuclear installations; it shall also consider the safety implications of scientific and technical developments of new reactor designs. In implementing its programme, the CSNI establishes co-operative mechanisms with NEA s Committee on Nuclear Regulatory Activities (CNRA) responsible for the program of the Agency concerning the regulation, licensing and inspection of nuclear installations with regard to safety. It also co-operates with NEA s Committee on Radiation Protection and Public Health (CRPPH), NEA s Radioactive Waste Management Committee (RWMC) and NEA s Nuclear Science Committee (NSC) on matters of common interest. 5

6 6

7 Workshop on the Recent Findings and Developments in Probabilistic Seismic Hazards Analysis (PSHA) Methodologies and Applications The Workshop on the Recent Findings and Developments in Probabilistic Seismic Hazards Analysis (PSHA) Methodologies and Applications is to be held on 7 to 9 April 2008 in Lyon France and is hosted by Electricité de France (EdF) in co-operation with the IAEA. Objectives The Objectives of this Workshop are to discuss recent research and the regulatory, utility and industry issues associated with PSHA. The results of PSHA studies undertaken to date in regions with low to moderate seismicity exhibit large areas of uncertainty. Participants in this Workshop will therefore discuss how to improve knowledge and reduce uncertainty associated with the lack of strong-motion data and regional-specific attenuation relations. Material presented at the November 2006 specialists' meeting on the seismic probabilistic safety assessment of nuclear facilities will be used to support these discussions. The deliverables and expected results from the workshop are as follows:- A description of state-of-the-art PSHA methodology including the management of epistemic and random uncertainties. A description of the issues associated with PSHA results and potential resulting difficulties, in particular differentiation between areas of low to moderate seismicity as contrasted to areas of high seismicity. A comparison between PSHA studies conducted in different areas, including low to moderate and high seismicity areas. A verification of the consistency of PSHA results versus available observations, especially for return periods where records are available, in order to get an objective comparison and to improve the confidence in the results. A description of regulatory practice and the technical basis for safety standards in this area. Information obtained as a result of this Workshop should provide the key to understanding and interpretating potential differences and discrepancies, from which methods to address them can be proposed. Workshop Proceedings The Workshop Proceedings will be published, including papers from the opening session, conclusions and recommendations from the discussion groups, a summary of all panel discussions and information papers submitted by Workshop participants. Organising Committee The Organising Committee will organise the sessions and the final programme for the workshop. The following people are serving on the Workshop Organising Committee:- Emmanuel VIALLET EdF France Chair Andrew J. MURPHY NRC USA Chair Robert J. BUDNITZ Lawrence Berkeley National Laboratory USA Pierre LABBE EdF France Shizuo NODA JNES Japan Antonio GODOY IAEA International Alejandro HUERTA NEA International Secretariat 7

8 Workshop Secretariat Mr. Alejandro Huerta, Nuclear Safety Division, OECD Nuclear Energy Agency, Le Seine Saint Germain, Bat. B, 12, boulevard des Iles, Issy-les-Moulineaux, France, Tel: , Fax: , alejandro.huerta@oecd.org Local contact Mr. Emanuel Viallet, EdF SEPTEN, Division GS, avenue Dutriévoz, Villeurbanne Cedex Tel: , Fax: , emmanuel.viallet@edf.fr 8

9 FINAL PROGRAMME (1/6) Sunday, 6 April Pre-registration Monday, 7 April :30 Registration 09:00 Opening Session Chairs: Emmanuel Viallet (EdF) Alejandro Huerta (OECD/NEA) 09:00 Welcome Address E. Viallet, EdF 09:15 Introduction & Objectives of the Workshop A. Huerta, OECD/NEA 09:40 Invited Lecture Relation of the Fact on the Plant Suffered from an Extremely Strong Local Earthquake to Seismic PSA Analysis Results Prof. Heki Shibata, University of Tokyo, Japan 10:10 Coffee Break 10:40 Session 1 - Recent Advances in Regulations and Guidelines Chairs: Prof. Julian Bommer (Imperial College London) Pierre Labbé (EdF) 10:40 Recent PSHA Activities at the USNRC, Including Continuing Research A. J. Murphy, A. M. Kammerer and J. Ake, NRC, USA 11:05 Confidence Levels in Deterministic Seismic Hazard Estimation and Confrontation with Probabilistic Estimates D. Baumont, O. Scotti, M.E. Cushing, and C. Berge-Thierry, IRSN, France 11:30 Swiss Regulatory Review of the PEGASOS Study P. Zwicky, R.T. Sewell, and D. Giardini, Consultants to the Swiss Nuclear Safety Inspectorate (HSK), Switzerland 12:00 Lunch Break 13:00 Seismic Issues in Siting New Nuclear Power Plants in the Central and Eastern United States Y. Li and N. Chokshi, NRC, USA 13:25 The New Revision of the IAEA Safety Standard on the Evaluation of Seismic Hazards A. Gürpinar, IAEA 9

10 FINAL PROGRAMME (2/6) Monday, 7 April 2008 (Cont d) 13:50 Session 1 Panel Discussion 14:20 Invited Lecture - Empirical Ground-Motion Models for PSHA at Sites of Critical Facilities J.J. Bommer, Professor of Earthquake Risk Assessment, Director of Undergraduate Studies Civil & Environmental Engineering, Imperial College London, UK 14:50 Session 2 - Current PSHA practices and challenges Chairs: Antonio Godoy (IAEA) Noda-Shizuo 14:50 PSHA in Korea with Low-to-Moderate Seismicity; 20 Years Experience and a New Challenge J.M. Seo (KAERI) and K.H. Yun (KEPRI), Korea 15:15 The Finnish Approach to Seismic Hazard Analysis Case Loviisa P. Varpuso (Fortum Nuclear Service), J. Sandberg (STUK) and P. Välikangas (STUK), Finland 15:40 Current Hungarian Practice of Seismic Hazard Assessment: Methodology and Examples L. Tóth (GeoRisk Earthquake Research Institute), E. Győri (Seismological Observatory), T.J. Katona (Paks NPP), Hungary 16:05 Coffee Break 16:25 Accounting for Epistemic Uncertainties in Site Effect Studies: Examples of a Logic Tree Approach P.-Y. Bard (University Joseph Fouier), D. Fäh (Swiss Seismological Service), A. Pecker (Géodynamique et Structure), J. Studer (Studer Engineering), N. Abrahamson (Geoscience Department, PG&E) 16:50 Use of the Hybrid Empirical Method for Developing Attenuation Relationship in Regions of Low Seismicity K. Campbell, (USA) 17:15 Session 2 Panel Discussion 17:45 End of Day 1 10

11 FINAL PROGRAMME (3/6) Tuesday, 8 April :30 Invited Lecture - PEGASOS Advanced Ground Motion Models N. Abrahamson, Geoscience Department, PG&E, USA 09:00 Session 3 - Checks of Consistency of PSHA Results Versus Available Observations Chairs: Aybars Gürpinar (IAEA) Andrew J. Murphy (USNRC) 09:00 PSHA Results Versus Historical Seismicity; Example of France P. Labbé, EdF-SEPTEN, France 09:25 A Method for Comparison of Recent PSHA on the French Territory with Experimental Feedback N. Humbert, E. Viallet, C. Durouchoux, EdF, France 09:50 Consistency of PSHA Models in Acceleration and Intensity by Confrontation of Predictive Models to Available Observations in France C. Martin (Geoter International), R. Secanell (Geoter International), E. Viallet (EdF), N. Humbert (EdF), France 10:15 Coffee Break 10:45 Report on the Kashiwasaki Kariwa NPP after the 16 July Earthquake An IAEA Perspective A. Godoy, IAEA 11:10 Constraining Probabilistic Seismic Hazard Estimates with Strong Motion Data: Possibilities and Limitations C. Beauval (Géosciences Azur-IRD), P.-Y. Bard (University Joseph Fouier), S. Hainzl (Institute of Geosciences, Germany) 11:35 Verification of the National Seismic Hazard Maps for Japan with Actual Observed Strong Motion by K-NET in the Last Decade Y. Fukushima, Japan 12:00 Session 3 Panel Discussion 12:30 Lunch Break 11

12 FINAL PROGRAMME (4/6) Tuesday, 8 April 2008 (Cont d) 14:00 Session 4 - Advanced and Alternative PSHA Methods Chair: Robert Budnitz (LBNL) 14:00 Invited Lecture - A Bayesian Updating Framework for Prediction of Site-Specific Seismic Ground Motion T. Takada (University of Tokyo, Japan) and M. Wang (University of Waterloo, Canada) 14:30 Invited Lecture PSHA Application of Bayesian Techniques G. Woo, UK 15:00 Uncertainties in PSHA for regions of Low to Moderate Seismic Potential: The Need for a Structured Approach J.P. Ake, NRC, USA 15:25 An Evaluation of Epistemic and Random Uncertainties Included in Attenuation Relationship Parameters N. Humbert, E. Viallet, EdF/SEPTEN, France 15:50 Coffee Break 16:10 Damage Relevant PSHA R. T. Sewell, R.T. Sewell Associates, Consultant to HSK 16:35 On the Use of a Bayesian Updating Technique to get Probabilistic Seismic Hazard Assessment More Rugged E. Viallet, N. Humbert, EDF/SEPTEN, France 17:00 Session 4 Panel Discussion 18:15 22:30 Workshop Special Event : Tour and Dinner in Beaujolais 12

13 FINAL PROGRAMME (5/6) Wednesday, 9 April :30 Invited Lecture - Methodologies for Capturing the Scientific Basis for Seismic Source Characteristics in PSHA: An Example from the PEGASOS Project, Switzerland K.J. Coppersmith (Coppersmith Consulting, Inc.) and R.R. Youngs (Geomatrix Consultants, Inc.) 09:00 Session 5 - Comparisons Between PSHAs Conducted in Different Areas Chairs: Nilesh Chokshi (USNRC) Gordon Woo (UK) 09:00 Lessons Learned from the Application of PSHA Approach and Methodologies IAEA EBP Conclusions A. Gürpinar (IAEA) 09:25 Comparison of PEGASOS Results with Other Modern PSHA Studies M. Richner and S. Tinic, Nordostschweizerische Kraftwerke, NPP Beznau, Switzerland 09:50 Some Lessons Learned and Recent Advances in PSHA Modelling for Eastern US Stable Continental Region C. Stepp (USA) 10:15 Coffee Break 10:40 Session 5 Panel Discussion 11:10 Final Session - Conclusions, Lessons Learnt and Recommendations from the Workshop Chairs: Emmanuel Viallet (EdF) Andrew J. Murphy (USNRC) 11:10 Synthesis of Topic 1 Conclusions (To be named later) 11:40 Synthesis of Topic 2 Conclusions (To be named later) 12:10 Lunch Break 13:10 Synthesis of Topic 3 Conclusions (To be named later) Synthesis of Topic 4 Conclusions (To be named later) Synthesis of Topic 5 Conclusions (To be named later) 14:40 General Discussion 13

14 FINAL PROGRAMME (6/6) Wednesday, 9 April 2008 (Cont d) 15:10 Coffee Break 15:30 Conclusions, Lessons Learnt, and Recommendations - Synthesis of the 5 Synthesis Presentations 16:15 Closing Remarks A. Huerta (OECD/NEA), E. Viallet (EdF) 16:30 End of Workshop 14

15 Abstracts of Presentations 15

16 Opening Session Chairs: Emmanuel Viallet (EdF) Alejandro Huerta (OECD/NEA) Invited Lecture Relation of the Fact on the Plant Suffered from an Extremely Strong Local Earthquake to Seismic PSA Analysis Results Professor Heki Shibata, Dr. Emeritus of University of Tokyo We experienced that damaging of Kashiwazaki-Kariwa Nuclear Power Plant, which consists of seven BWR units on 16 July Fortunately, no critical damage to their reactors and nuclear related items, that is, there was no damage of the major Class A items. However, there were various damages on Class B and C items in the plant, and roads and other conventional structures were damaged as roads in out-side area. The main office building was damaged inside, and we couldn t expect any usual function. Two out of four trunk lines of 500 kv were disconnected from other power units. The three units were fully operated and were sending approximately 4,000 Mew to the net, and one unit was in the starting process, and all four units were safely shut down by seismic trigger signals even the maximum recorded acceleration was 680 gal against the corresponding design basis acceleration S2 was 273 gal. Now they, Tepco almost completed the examinations of the details of their two units, and they could not find any critical damages of their reactor cores and primary piping and safety related systems. Whether or not we should expect some failures of safety related system expected under 2~3 times DBE, S2 loadings like this time, this is a point of discussion here. We have been checking the SMA values on safety related items, whose definition is more deterministic as to be in other countries than Japan. We know that their values are usually around 10, but some of them are near to two or less. Also, we did the shaking tests of various items on their functions. Based on our knowledge through such a process, we could say that some of them might exceed the failure limit by the loading in this time. The fact where there was no critical affair as a nuclear reactor, would it give us how much S-PSA value of it? and how much different is it to the evaluated value at the time when we completed its design?. That is, by the fact or event can we evaluate the S-PSA value, which was calculated before the event, or can we find some index number like likelihood? The author likes to discuss this possibility. In other words, the two values, those are the value obtained by the fact and the calculated value in advance, is like the two values in Bayesian Analysis. For the evaluation in advance, the fragility curve has been used. This curve is obtained by several ways, a stochastic curve decided by two or three theoretical parameters, data obtained by a design process and data obtained by test of one or a fewer number of samples. In this process, the detailed definition of failure, both for functional and structural has not been established, especially structural failure. Therefore, the fragility curve is not so exact, to compare to the actual failure in the field that is the uncertainty of the definition of margin. or fragility curve. To grasp the real conditional probability of occurrence of the critical event, it is key to define the margin or failure as definite value. The seismic hazard curve is another problem. Those in this area might be different from evaluated value in advance in usual procedure. The area like here, Noto, and Mihama, where NPPs exist along Japans sea-side has an unique seismic hazard curve. The concept defined destructive area has not been established well. 16

17 The estimation of peak-accelerations of the area has not been established well. Especially, in the site area, it is not clear how to estimate it practically. The PGAs in units #1 and #5, which are situated at both points near to the boundaries are almost two third, where #1 is the highest, and the town along the street, Arahama was severely damaged in all areas in this earthquake. An earthquake on the Coast of Japans Sea has a much localized area of high PGA in some cases as this case. We have never evaluated the hazard curve of such an earthquake. If we try to combine these two subjects, how do we conclude that the fact occurred in this plant would be expressed by the post-evaluated S-PSA value. To evaluate S-PSA value on other NPPs such as Mihama, Tsuruga, or Shimane, which is situated along Japans sea coast may the ordinary approach bring the exact values to us in this case?. 17

18 Session 1 - Recent Advances in Regulations and Guidelines Chairs: Professor Julian Bommer (Imperial College London) Pierre Labbé (EdF) Recent PSHA Activities at the USNRC, Including Continuing Research A.J. Murphy, A.M. Kammerer, J. Ake This paper examines three phases of the utilization of probabilistic seismic hazard assessment (PSHA) at the US Nuclear Regulatory Commission. These phases include:- 1. NRC's development and rational/justification 2. NRC's practical application in a regulatory environment, and 3. NRC's continuing enhancement and refinement. Under the first phase, the NRC worked with the innovators of the approach to develop an approach to address the fundamental issues associated with the "deterministic" approach to siting critical facilities, such as, nuclear power plants, dams, liquefied natural gas storage and handling facilities, etc. These fundamental issues include:- 1. What are the uncertainties in the deterministic parameters used to describe to hazards being planned for?. 2. Are there abrupt cliffs in the boundaries of the hazard estimates?. 3. Can the estimates of the hazards be better described to limit contention among experts and among users and regulators?. 4. Is there a better seismic hazard characterization methodology?. This phase led to the development and implementation of the LLNL and EPRI PSHAs for the Eastern and Central United States, to the preparation of the Senior Seismic Hazard Analysis Committee's (SSHAC) guidance, and to the writing of NRC Regulatory Guide A principal activity during the second phase, which made extensive use of PSHA to address seismic vulnerabilities, was the Individual Plant Examination for External Events (IPEEE) Program. The staff utilized PSHA results to develop the scope of the search of vulnerabilities and to bin the level of search to be conducted by individual plants. PSHAs, following the new RG 1.208, are being submitted for a new generation of Early Site Permit and Combined Operating License applications (This particular aspect of PSHA utilization will be covered by a complimentary paper at this workshop by Li and Chokshi). While the paper acknowledges the first two phases, the heart of the paper will review the Seismic Research Plan from the NRC's Office of Nuclear Regulatory Research. The plan, in part, starts from the premise that PSHA has achieved a level of maturity and utilization and lays the foundation for the next generation for PSHA. The elements in the Plan pertinent to this Workshop include:- 1. Development of a "Next-Generation" attenuation relationship for the CEUS. 2. Lessons learned from the practical application of the SSHAC guidance. 3. Analysis of the effects of the application of CAV filtering. 4. Seismic source characterization for the CEUS. The paper will also report on a preliminary study on a comparison of historic seismicity with site-specific PSHA and deterministic controlling earthquakes or SSEs. 18

19 Session 1 - Recent Advances in Regulations and Guidelines Chairs: Professor Julian Bommer (Imperial College London) Pierre Labbé (EdF) Confidence Levels in Deterministic Seismic Hazard Estimation and Confrontation with Probabilistic Estimates D. Baumont, O. Scotti, M. E. Cushing, and C. Berge-Thierry Seismic Hazard Working Group, IRSN/DEI/SARG/BERSSIN Accounting for the inherent uncertainty in the data and models that feed seismic hazard estimates has been an issue for more than 30 years. Probabilistic seismic hazard methods provide so far the only established tools that formally account for uncertainty. In this study, we propose an alternative approach that allows qualifying confidence levels based on a deterministic method using the French nuclear regulatory guide RFS as an example. The four main steps in the deterministic RFS guidelines entail (1) the definition of a seismotectonic zonation, (2) the selection of most damaging historical or instrumental earthquakes within each zone, and their displacement as close as possible to the site of interest, within their own zone, (3) the evaluation of magnitude-depth characteristics and (4) the estimation of intensity or spectral acceleration at the site. The propagation of inherent uncertainties associated to each step allows proposing different hazard spectrum depending on the confidence level of interest. Finally, deagregation of the hazard spectrum over the entire frequency range allows identifying the earthquake parameters that best fit the seismic hazard spectrum for the chosen confidence level. Results of this deterministic procedure will be discussed and confronted with the results obtained in a recent site-specific PSHA study. 19

20 Session 1 - Recent Advances in Regulations and Guidelines Chairs: Professor. Julian Bommer (Imperial College London) Pierre Labbé (EdF) Swiss Regulatory Review of the PEGASOS Study Peter Zwicky 1, Robert T. Sewell 2, Domenico Giardini 3 Consultants to the Swiss Nuclear Safety Inspectorate (HSK) The PEGASOS study was undertaken by licensees of Swiss nuclear power plants (NPPs), and performed during the years of 2001 to The scope of the study included the four NPPs (five reactor units) that are currently operated in Northern Switzerland, with a study area extending to 300 km radius from the plants, and including several European countries. It was a very advanced PSHA study that employed, and extended, the Level-4 methodology documented in the Senior Seismic Hazard Analysis Committee (SSHAC) guidelines. PEGASOS extended the SSHAC methods by adopting full elicitation of the site-response component of the PSHA. PEGASOS was the first study to introduce full elicitation on maximum ground-motion truncations, applying period-dependent truncations both on the number of standard deviations of ground motion and on the absolute values of the motions. In addition, PEGASOS developed and implemented new standards in rigorous quality assurance of PSHA and database management of project information. PEGASOS included participation of 21 European experts in the earth sciences and engineering, in addition to a US-Swiss assemblage of project management team (PMT), team of technical facilitators-integrators (TFI), calculation team, and numerous resource experts and supporting personnel. PEGASOS was managed by Nagra, the Swiss National Cooperative for the Disposal of Radioactive Waste, who organized the major project elements into the following sub-projects: SP1 for seismic source modelling; SP2 for ground-motion modelling; SP3 for site-response modelling; and SP4 for PSHA calculations and related software/code development. This paper describes the approach and key findings of the regulatory review of PEGASOS. The regulatory review was initiated by the Swiss Nuclear Safety Inspectorate (HSK), who assembled a review team (HSK-RT) consisting of the HSK project officer and four experts in the various major facets of PSHA. HSK-RT conducted its review activities in both participatory and late stages. The review was comprehensive, following every major facet and development within the study. HSK-RT accepted the study as meeting both the intended SSHAC Level-4 methodology and the Project Plan agreed between Swiss NPPs and HSK. However, during the course of the PEGASOS study, HSK-RT identified and documented various open items (from its participatory review activities), and at the end of the study (from its late-stage review of the final PEGASOS submittal), HSK-RT reported on areas for potential refinement Basler & Hofmann, Consulting Engineerings AG, Zürich, Switzerland R.T. Sewell Associates, Consulting, Louisville, Colorado USA Swiss Federal Institute of Technology (ETH), Department of Seismology and Geodynamics; and Director, Swiss Seismological Service, Zürich, Switzerland 20

21 In addition to addressing the approach and findings of HSK-RT s review of PEGASOS, this paper discusses existing status concerning regulatory disposition on use of the PEGASOS results and on potential refinement of the PEGASOS results. 21

22 Session 1 - Recent Advances in Regulations and Guidelines Chairs: Professor. Julian Bommer (Imperial College London) Pierre Labbé (EdF) Seismic Issues in Siting New Nuclear Power Plants in the Central and Eastern United States Yong Li and Nilesh Chokshi US Nuclear Regulatory Commission United States Nuclear Regulatory Commission (NRC) is currently reviewing several new reactor applications and is expected to receive more applications in the near future. Most of the proposed new reactor sites are located in the Central and Eastern United States (CEUS). In contrasting to the Western United States, including California, CEUS is less active in earthquake activities and can be categorized as low to moderate seismic area. Lack of surface expression of active faults and lack of seismic recordings made seismic siting analyses more difficult, relative to the Western US. Probabilistic Seismic Hazard Analysis (PSHA) was endorsed by NRC about 10 years ago in Regulatory Guide 1.165, but not applied until the first early site permit (ESP) application review in Because of the seismic setting and progressive seismic hazard evolvement the safety review staff face challenges in every component of PSHA process when reviewing site-specific seismic characterization of new plant applications. These challenges include issues in:- 1. Seismic source characterization and updates. 2. Ground motion attenuation. 3. Stability of the methodology. 4. Minimum magnitude truncation. This paper summaries those challenges and discusses potential solutions to those challenges and other potential issues. 22

23 Session 2 Current PSHA Practices and Challenges Chairs: Antonio Godoy (IAEA) Noda-Shizuo PSHA in Korea with Low-to-Moderate Seismicity; 20 Years Experience and a New Challenge J.M. Seo (KAERI) and K.H. Yun (KEPRI) Korea is located in a stable continental region with a low to moderate seismicity. The first study on the seismic risk was performed in A historical earthquake catalogue, which spans approximately 1900 years, was re-evaluated for events greater than MMI IV during the study. In that study, the intensity and epicentre of the major earthquakes were mostly determined by the MMI intensity scale. The attenuation equation of Trifunac and Brady (1975) was used. Eleven PSHAs have been carried out so far for the NPPs located at four different sites in Korea since 1987 Among these, six PSHAs were performed until 1998 by using the basically similar catalogues to that of 1982, and different attenuation equations from foreign countries. The results showed large uncertainties and conservatism in the hazard curves. Sensitivity analyses indicated that attenuation equations were the primary source. A series of comprehensive research programs have been performed since 1996 to reduce the uncertainties. For example, shaking table tests on the ancient houses and stone pagodas were performed to identify the correlation between damage state and the acceleration level. Also, many site investigations were performed to characterize the capable faults near the NPP sites. As with the increases in the instrumental data, a few attenuation equations was developed based on the Korea-specific source mechanism and path effects. These research results were partly incorporated in the PSHAs performed after 1999, contributing to reduce the hazard level. Even with the data accumulation, we are still facing several issues to be solved because of lack of region - specific strong ground-motion data and geoscientific knowledge. Examples are, 1) strong ground-motion evaluation especially for the capable faults located very near to the site, 2) seismic source characterization based on the result of paleoseismological studies, 3) determination of the design acceleration level taking into account the high frequency motion, 4) re-evaluation of the historical earthquake catalogue, 5) derivation of the attenuation equation in terms of the spectral acceleration suitable for Korea. Recently, moderate-sized inland earthquake (20 January 2005, M=4,8) provided us with a chance to evaluate the aleatory uncertainty of the previously developed ground-motion attenuation relation (KEPRI05 model) in Korea and gave us a new insight about what we have accomplished and what should be done more with the large accumulated dataset from the dense seismic network. Some of the findings in the related researches, which we believe is a one step further to reduce the hazard uncertainty, are described in this paper. 23

24 Session 2 Current PSHA Practices and Challenges Chairs: Antonio Godoy (IAEA) Noda-Shizuo The Finnish Approach to Seismic Hazard Analysis - Case Loviisa P. Varpasuo (Fortum Nuclear Services Ltd) J. Sandberg and P. Välikangas (Radiation and Nuclear Safety Authority-STUK) In Finland seismic design is currently required for new nuclear power plant units and for major modifications of operating units, even though seismic activity is low. The design ground response spectrum shall correspond to recurrence time of years on the median confidence level. The peak ground acceleration shall be at least 0.1 g. However, when the operating units were designed in the 1970s there were no seismic requirements. Seismic risk analysis (seismic PSA) has been done for the operating units and is underway for Olkiluoto 3 EPR which is under construction. The seismic hazard analysis to be used in the Loviisa seismic PSA was started in the late 1980 s. The Loviisa seismic PSA was submitted to STUK in In 1999 the seismic response spectrum and peak ground accelerations were determined for the Loviisa and Olkiluoto sites to be used for possible new units. In connection with the Loviisa operating licence renewal in 2007 STUK required that the seismic PSA, including the seismic hazard analysis, is updated and the need for risk reduction is assessed. The updating process has been started by Fortum Nuclear Services with reassessment of the hazard curves. The seismic hazard analysis for the Finnish NPP sites is based on the principles originally developed by Cornell. The input parameters, especially the seismicity parameters and attenuation functions, have been updated according to the available new information. The seismic conditions in Finland, the evolution of the seismic hazard analysis methods and the main results for the Loviisa site are described in the paper. The status of seismic design and seismic risk analysis in Finland, regulatory requirements and the plans for updating regulatory guides are also described. 24

25 Session 2 Current PSHA Practices and Challenges Chairs: Antonio Godoy (IAEA) Noda-Shizuo Current Hungarian Practice of Seismic Hazard Assessment: Methodology and Examples László TÓTH 1,2, Erzsébet GYŐR 1, Tamás J. KATONA 3 Hungary is located in the Pannonian basin between the seismically very active Mediterranean and the nearly aseismic East European platform. Although the Pannonian basin being characterized as a low to medium level seismicity region, the risk of seismic activity is still a major concern. The high population density in the region, and the number of vulnerable industrial facilities e.g. nuclear power plants are being among the most sensitive ones regarding public concern. In the paper we give a brief review of seismic hazard methodology, present the latest seismic hazard map of the region and highlight some chief contributors (e.g. site effect, liquefaction) to seismic risk. The seismic hazard is the probabilistic measure of ground shaking associated with the recurrence of earthquakes. Seismic hazard shows the level of chosen ground motion that likely will not be exceeded in specified exposure times. Hazard assessment for civil use commonly specifies a 90% probability of non exceedance of some ground motion parameter (most often ground acceleration) for an exposure time of 50 years, corresponding to a return period of 475 years. For nuclear power plants and other installations representing higher environmental risk, it is common practice to use much lower probability events in the order of 10-4 to 10-6 /year. The methodology we used to assess seismic hazard and the associated uncertainty at nuclear and non-nuclear fields provide technically sound results and incorporate the following attributes: experience-based, data-driven, proper treatment of uncertainties, flexible and facilitate sensitivity analysis. Six essential steps are involved in the assessment of seismic hazard; Step1: Evaluation of seismic sources; Step 2: Characterization of seismic sources; Step 3: Consideration of ground motion attenuation; Step 4: - Creation a mathematical model to calculate seismic hazard; Step 5: Presentation of the hazard results; Step 6: Evaluation of site effect and potential secondary seismic effects Seismological Observatory, Hungarian Academy of Sciences, Budapest, Hungary Georisk Earthquake Research Institute, Budapest, Hungary Paks Nuclear Power Plant Ltd., Hungary 25

26 Session 2 Current PSHA Practices and Challenges Chairs: Antonio Godoy (IAEA) Noda-Shizuo Accounting for Epistemic Uncertainties in Site Effect Studies: Examples of a Logic Tree Approach P.-Y. Bard 1, D. Fäh 2, A. Pecker 3, J. Studer 4 and N. Abrahamson 5 Probabilistic estimates for site specific studies are most often separated into "regional" PSHA to estimate the rock spectra, and local analysis to derive the site amplification factors: this latter part is mainly deterministic and is applied to the rock spectra corresponding to the design return period. This presentation will focus on several developments achieved recently within the framework of the PEGASOS project to account for site effects in a fully probabilistic way, including both epistemic uncertainties and aleatory variabilities. This approach was based on a wide spectrum of computations using different 1D soil columns to account for the uncertainty of soil parameters, different input spectra to account for non-linear effects, and various modelling approaches (1D/2D, various incident wave fields, without / with close sources). The final uncertainty is estimated through a logic tree approach: not only the weights for each branch do vary from one expert to another, but also the structure of the logic tree itself. Some examples will be given and briefly discussed. Some examples of resulting median amplification curves, together with their variability and sensitivity to the level to input ground motion, will be shown. The effects on the final hazard curves are significant: while obviously frequency dependent, they also strongly depend on the return period because of non-linear effects: significant amplification at low return periods may be replaced by attenuation at much longer return periods, so that the shape of the hazard curves may be significantly modified with respect to rock hazard curves. This experience brought several very valuable learnings. Despite very different approaches, the variability of final median estimates proves to be relatively small from one expert to the other. The overall variability could probably be significantly reduced with more site specific data (on site earthquake recordings, specific measurements of NL parameters for instance), and may be also by coupling the estimations of response spectra ordinates with estimations of other ground motion parameters LGIT/LCPC, University. Joseph Fouier, Grenoble, France Swiss Seismological Service, ETH Zürich, Switzerland Géodynamique et Structure, Bagneux, France Studer Engineering, Zürich, Switzerland Geoscience Departmentt, PG&E, San Francisco, USA. 26

27 Session 3 Checks of Consistency of PSHA Results Versus Available Observations Chairs: Aybars Gürpinar (IAEA) Andrew J. Murphy (USNRC) PSHA Results Versus Historical Seismicity; Example of France Pierre Labbé, EdF-SEPTEN This proposed contribution is based on the comparison of two methods of seismic risk assessment in France. The first method is based on historical data. It considers statistics of earthquakes of Epicentral Intensities I 0 = V to IX (in European Macroseismic Scale EMS98) and statistics on areas affected by Intensities I lower than (or equal to) I 0. The risk is estimated in terms of return periods of given Intensities and in annual probability of occurrence of given damage degrees (according to EMS98 scale definition). Typical outputs are as follows: In average on the Metropolitan French Territory, The Intensity V return period is estimated 120 years. Probability of a degree 2 damage is around 10-4 for masonry buildings (10-5 for a degree 3). Other values are also calculated, depending on the region under consideration. The second method is based on convolution of seismic hazard data and masonry building fragility curves. Seismic hazard is described in the form of 3 different maps of the metropolitan French territory. These maps are outputs of probabilistic seismic hazard assessments of France and provide PGA values for a 475 year return period. These data are extrapolated to other return periods on the basis of classical relationship between PGA and return periods. Fragility of masonries is described in the form of a classical log normal distribution of the probability of exceedance of a given damage versus the PGA (median value a and variability β). A sensitivity study is carried out on the median value corresponding to damage degrees 2 and 3. For realistic values of a and β, one of the maps leads to an tremendous overestimate of the risk, as compared to the historically observed risk, another one leads to a slight underestimate while the third one is in the middle. The author considers that this approach could be used for reducing uncertainties in probabilistic seismic hazard assessment in low-moderate areas where sufficient historical data are available. 27

28 Session 3 Checks of Consistency of PSHA Results Versus. Available Observations Chairs: Aybars Gürpinar (IAEA) Andrew J. Murphy (USNRC) A Method for Comparison of Recent PSHA on the French Territory with Experimental Feedback N. Humbert, E. Viallet(EdF/SEPTEN), C. Durouchoux(EdF/TEGG) Recent Probabilistic Seismic Hazards Assessments (PSHA) studies on the French metropolitan territory exhibit very different seismic levels. The two PSHAs considered in this paper are the study published by the MEDD (French ministry of environment and sustainable development) in 2002 and the study conducted by a working group of AFPS (French association of earthquake engineering) in These two studies are clearly not consistent in term of median values: we observe ratios of approximately 2 in term of PGA for a given return period and ratios of 10 between return periods for a given PGA. These discrepancies are a real serious problem for French Earthquake engineering. The purpose of this study is to compare PSHA of the French territory with experimental feedback and gives an objective point of view of the confidence on probabilistic models of seismic hazard based on observations. The PSHA considered is compared with seismic experience feedback provided by the survey system of EdF sites and by the RAP (French seismograph network). This paper develops a methodology based on a probabilistic approach. A statistic methodology of determination of soil-structure interaction (SSI), and specific soil effect is developed for each site where feedback is collected. These probabilistic models developed for each SSI and soil effect allows us to compare PSHA with feedback by a integration of both epistemic and random variability at each step of the methodology. This paper gives very interesting results: even with a limited feedback, this method is efficient. This method gives an objective evaluation of the confidence we can associate to a PSHA. Moreover, this method may become a first step of a Bayesian approach for updating PSHA using observation as conditional event. 28

29 Session 3 Checks of Consistency of PSHA Results vs. Available Observations Chairs: Aybars Gürpinar (IAEA) Andrew J. Murphy (USNRC) Consistency of PSHA Models in Acceleration and Intensity by Confrontation of Predictive Models to Available Observations in France Martin 1, Ch., Secanel 1, R., Viallet 2, E. and N. Humbert 2 During recent years, different attempts have been made in France, thanks to the working groups of the French Association for Earthquake Engineering, to calibrate the predicative PSHA models in terms of PGA at short return periods, using a comparative method between the predictive PGA and the observed PGA. Significant differences were noted between predictions and observations. Some of them can be related to the relationships used to define a homogeneous magnitude from macroseismic data or from various magnitude scales, to the attenuation models for which the principle of statistical ergodicity is assumed in France, to the way the uncertainties are propagated into the predictive model and to the short period of observations. To overcome the difficulties due to the use of macroseismic and historical seismicity to predict probabilities of exceedance of physical values like PGA, a PSHA in intensity is also carried out for the whole French territory. This approach makes it possible to be freed from the majority of the difficulties encountered in the acceleration s model, thanks to the use of specifically French data to derive regional attenuation laws, and to a larger time-period of observations (5 centuries). Following this approach an acceptable agreement is found between predictions and observations, for the various seismotectonic contexts of France, including the most stable and active areas. The paper presents the comparisons between predictive and empirical models, both in acceleration and in intensity, and concludes on the feasibility of the actualization method to calibrate the predictive models. KEYWORDS Probabilistic seismic hazard, Predictive and empirical models, Calibration of predictive model, France. 1 2 Geoter International, Espace 890, 890 Route Nationale 96, Roquevaire, France. EdF - SEPTEN, avenue Dutrievoz Villeurbanne Cedex, France. 29

30 Session 3 Checks of Consistency of PSHA Results vs. Available Observations Chairs: Aybars Gürpinar (IAEA) Andrew J. Murphy (USNRC) Constraining Probabilistic Seismic Hazard Estimates with Strong Motion Data: Possibilities and Limitations 1 C. Beauval 2, PY. Bard 3, S. Hainzl 4 The amount of probabilistic hazard studies has been increasing in the last years, due to the requirements by new regulations that hazard be estimated in probabilistic terms. The present study aims at defining the possibilities and limits for a comparison test between predictions and observations, with an emphasis on tests based directly on the rate of ground-motion occurrences are favoured, rather than on the rate of earthquake occurrences. Based on the properties of Poisson processes, the minimum time window insuring reliable occurrence rate estimates at a site is computed and evaluated. For example, for ground motion with a 475 years return period at a site, a minimal years observation time window is required for estimating the rate with a 30% uncertainty (coefficient of variation: standard deviation divided by the mean). These values do not depend on the seismic activity level of the regions under study. An analysis of recorded ground-motions at the stations of the permanent French accelerometer network shows that at best, the occurrence rates can be estimated with an accuracy of 30% for very low acceleration levels ( g g for the station STET). A similar analysis, carried out at two stations with longer time histories and located in much more active regions (Greece and California), provides ground motion levels up to 0.1 g. The issue whether the results of a comparison test at low acceleration levels can be generalized to higher acceleration levels remains in any case open, even if using a ground-motion prediction equation uniformly valid for a wide range of accelerations Slightly modified from a paper to appear in an early issue of BSSA 2008 Corresponding author Géosciences Azur, IRD, 250 av. A. Einstein, Valbonne, France LGIT, BP 53 X, Grenoble Cedex 9, France 4 Institute of Geosciences, Karl-Liebknecht Str. 24, Potsdam University, D Potsdam Golm, Germany 30

31 Session 4 Advanced and Alternative PSHA Methods Chair: Robert Budnitz (LBNL) Invited Lecture - A Bayesian Updating Framework for Prediction of Site-Specific Seismic Ground Motion Tsuyoshi Takada 1 and Min Wang 2 The ground motion attenuation relations have been proposed on an empirical basis and widely been used for ground motion prediction in a convenient and efficient way. The accuracy of the predictions, however, is not so high given an earthquake magnitude and distance, the uncertainty of the relations has, say, 0.5 or more in a natural logarithmic standard deviation, which significantly governs the tail of the seismic hazard curve in a probabilistic seismic hazard assessment (PSHA). After the 1995 Kobe earthquake in Japan, many earthquake ground motion observations have been deployed throughout the country. They are called K-NET or KiKnet, current seismograph observation networks in Japan. Since the ample records have been accumulated in many observatories in the past, these data should be made use of for higher accurate prediction. Furthermore, earthquake ground motion observations are often made in special locations such as the sites of nuclear power plants in Japan. These observed records are very useful for constructing a site-specific attenuation relation. To do this, a Bayesian approach can be effectively utilized to update an existing attenuation relation with observed records for a specific site. The paper will demonstrate how to apply the approach and comparison of the updated results. It is interesting to show how the accuracy of the relation changes. 1 2 Graduate School of Engineering, the University of Tokyo, Japan Department of Civil and Environmental Engineering, University of Waterloo, Canada 31